T h e Ox f o r d H a n d b o o k o f

P OL I T IC S
I N M U SL I M
SOCIETIES
The Oxford Handbook of

POLITICS
IN MUSLIM
SOCIETIES
Edited by
MELANI CAMMETT
and

PAULINE JONES

1
 3
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.

Published in the United States of America by Oxford University Press

198 Madison Avenue, New York, NY 10016, United States of America.

© Oxford University Press 2022

All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by license, or under terms agreed with the appropriate reproduction
rights organization. Inquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above.

You must not circulate this work in any other form

and you must impose this same condition on any acquirer.

Library of Congress Cataloging-in-Publication Data: 2021039932

ISBN 978-0-19-093105-6

DOI: 10.1093/oxfordhb/9780190931056.001.0001

1 3 5 7 9 8 6 4 2

Printed by Marquis, Canada

 Contents

About the Editors ix

Contributors xi
Acknowledgments xv

1. Politics in Muslim Societies: What’s Religion Got to Do with It? 1

Melani Cammett and Pauline Jones

PA RT I : R E G I M E S A N D R E G I M E C HA N G E
2. Islam and Political Structure in Historical Perspective 33
Eric Chaney
3. State-​Formation, Statist Islam, and Regime Instability:
Evidence from Turkey 53
Kristin E. Fabbe
4. States, Religion, and Democracy in Southeast Asia: Comparative
Religious Regime Formation 75
Kikue Hamayotsu
5. Repression of Islamists and Authoritarian Survival in the Arab
World: A Case Study of Egypt 101
Jean Lachapelle
6. Regime Types, Regime Transitions, and Religion in Pakistan 115
Matthew J. Nelson
7. Regime Change under the Party of Justice and Development (AKP)
in Turkey 143
Feryaz Ocaklı
8. Islam, Nationalism, and Democracy in Asia: Nations under Gods
or Gods under Nations? 167
Maya Tudor
9. Military Politics in Muslim Societies 193
Nicholas J. Lotito
vi   Contents

PA RT I I : E L E C TOR A L P OL I T IC S , PA RT I E S ,
A N D E L E C T ION S
10. Voting for Islamists: Mapping the Role of Religion 213
Ellen Lust, Kristen Kao, and Gibran Okar
11. Party Systems in Muslim Societies 239
Elizabeth R. Nugent
12. Ideologies, Brands, and Demographics in Muslim Southeast
Asia: “Voting for Islam” 251
Thomas Pepinsky
13. Religion and Party Politics in India and Pakistan 271
Steven I. Wilkinson
14. Religion and Electoral Competition in Senegal 295
Dominika Koter
15. Clientelism, Constituency Services, and Elections in Muslim
Societies 313
Daniel Corstange and Erin York

PA RT I I I : P OL I T IC A L AT T I T U DE S A N D
B E HAV IOR B E YON D VOT I N G
16. Religiosity and Political Attitudes in Turkey during the AKP Era 335
S. Erdem Aytaç
17. Religious Practice and Political Attitudes among Shiʿites in Iran
and Iraq 355
Fotini Christia, Elizabeth Dekeyser, and Dean Knox
18. Repressive Religious Regulation and Political Mobilization in
Central Asia: Why Muslims (Don’t) Rebel 369
Dustin Gamza and Pauline Jones
19. How Extraordinary Was the Arab Spring? Examining “Protest
Potential” in the Muslim World 391
Avital Livny
20. Illicit Economies and Political Violence in Central Asia 411
Lawrence P. Markowitz and Mariya Y. Omelicheva
 Contents   vii

21. Piety, Devotion, and Support for Shariʿa: Examining the Link
between Religiosity and Political Attitudes in Pakistan 429
Niloufer A. Siddiqui
22. Mapping and Explaining Arab Attitudes toward the Islamic
State: Findings from an Arab Barometer Survey and Embedded
Experiment 449
Mark Tessler, Michael Robbins, and Amaney Jamal

PA RT I V: S O C IA L M OB I L I Z AT ION
23. Social Movements, Parties, and Political Cleavages in Morocco: A
Religious Divide? 481
Adria Lawrence
24. The Rise and Impact of Muslim Women Preaching Online 501
Richard A. Nielsen
25. Religion and Mobilization in the Syrian Uprising and War 521
Wendy Pearlman
26. Christian-​Muslim Relations in the Shadow of Conflict: Insights from
Kaduna, Nigeria 539
Alexandra Scacco and Shana S. Warren
27. New Media and Islamist Mobilization in Egypt 561
Alexandra A. Siegel
28. Islamically Framed Mobilization in Tunisia: Ansar al-​Sharia in the
Aftermath of the Arab Uprisings 581
Frédéric Volpi
29. Islamist Mobilization during the Arab Uprisings 597
Chantal Berman

PA RT V: E C ON OM IC P E R F OR M A N C E A N D
DE V E L OP M E N T OU TC OM E S
30. Religious Legitimacy and Long-​Run Economic Growth in
the Middle East 619
Jared Rubin
31. Islam and Economic Development: The Case of Non-​Muslim
Minorities in the Middle East and North Africa 637
Mohamed Saleh
viii   Contents

32. State Institutions and Economic Performance in

Nineteenth-​Century Egypt 655
Lisa Blaydes and Safinaz El Tarouty
33. Islam and the Politics of Development: Shrines and Literacy
in Pakistan 675
Adeel Malik and Rinchan Mirza
34. Islam and Economic Development in Sub-​Saharan Africa 697
Melina R. Platas
35. Islamic Finance and Development in Malaysia 719
Fulya Apaydin

PA RT V I : S O C IA L W E L FA R E A N D
G OV E R NA N C E
36. Colonial Legacies and Welfare Provision in the Middle East
and North Africa 739
Melani Cammett, Allison Spencer Hartnett, and
Gabriel Koehler-​Derrick
37. Welfare States in the Middle East 769
Ferdinand Eibl
38. Islamist Organizations and the Provision of Social Services 789
Steven Thomas Brooke
39. Exploring the Role of Islam in Mali: Service Provision,
Citizenship, and Governance 807
Jaimie Bleck and Alex Thurston
40. Islamist Parties and Women’s Representation in Morocco: Taking
One for the Team 827
Lindsay J. Benstead
41. The Islamic State as a Revolutionary Rebel Group: IS’s Governance
and Violence in Historical Context 853
Megan A. Stewart

Index 871
 About the Editors

Melani Cammett is the Clarence Dillon Professor of International Affairs in the

Department of Government and Director of the Weatherhead Center for International
Affairs at Harvard University. She also holds a secondary faculty appointment at the
Harvard T. H. Chan School of Public Health. Her current research explores “toleration”
after ethnoreligious conflict across diverse global regions, the politics of social service
delivery and public goods provision, and the historical roots of economic and social
development in the Middle East. Cammett has published four books and multiple articles
in academic and policy journals, and has received various fellowships and awards. She
currently serves on the Lancet Commission on Syria.

Pauline Jones is Professor of Political Science and Director of the Digital Islamic
Studies Curriculum at the University of Michigan. Her past work contributes broadly
to the study of institutional origin, change, and impact with an empirical focus on the
former Soviet Union, primarily the five Central Asia states of Kazakhstan, Kyrgyzstan,
Tajikistan, Turkmenistan, and Uzbekistan. Her current work explores the influence of
religion on political attitudes and behavior in Muslim majority states, focusing on the
relationship between religious regulation, religiosity, and political mobilization, and
how trust in religious leaders affects voluntary compliance with pro-social policies. She
has published in both leading academic and policy journals and authored four books.
 Contributors

Fulya Apaydin
Institut Barcelona d’Estudis Internacionals
S. Erdem Aytaç
Koç University
Lindsay J. Benstead
Portland State University
Chantal Berman
Georgetown University
Lisa Blaydes
Stanford University
Jaimie Bleck
University of Notre Dame
Steven Thomas Brooke
University of Wisconsin–​Madison
Melani Cammett
Harvard University
Eric Chaney
University of Oxford
Fotini Christia
Massachusetts Institute of Technology
Daniel Corstange
Columbia University
Elizabeth Dekeyser
Massachusetts Institute of Technology
Ferdinand Eibl
King's College London
Safinaz El Tarouty
Independent Scholar
xii   Contributors

Kristin E. Fabbe
Harvard University
Dustin Gamza
Independent Scholar
Kikue Hamayotsu
Northern Illinois University
Allison Spencer Hartnett
University of Southern California
Amaney Jamal
Princeton University
Pauline Jones
University of Michigan
Kristen Kao
University of Gothenburg
Dean Knox
University of Pennsylvania
Gabriel Koehler-​Derrick
Brown University
Dominika Koter
Colgate University
Jean Lachapelle
University of Gothenburg
Adria Lawrence
Johns Hopkins University
Avital Livny
University of Illinois at Urbana-​Champaign
Nicholas J. Lotito
Yale University
Ellen Lust
University of Gothenburg
Adeel Malik
University of Oxford
Lawrence P. Markowitz
Rowan University
 Contributors   xiii

Rinchan Mirza
University of Kent
Matthew J. Nelson
SOAS University of London
Richard A. Nielsen
Massachusetts Institute of Technology
Elizabeth R. Nugent
Yale University
Feryaz Ocaklı
Skidmore College
Gibran Okar
Independent Scholar
Mariya Y. Omelicheva
University of Kansas
Wendy Pearlman
Northwestern University
Thomas Pepinsky
Cornell University
Melina R. Platas
New York University Abu Dhabi
Michael Robbins
Arab Barometer
Jared Rubin
Chapman University
Mohamed Saleh
Toulouse 1 Capitole University; Institute for Advanced Study in Toulouse
Alexandra Scacco
WZB Berlin Social Science Center
Niloufer A. Siddiqui
University at Albany–​State University of New York
Alexandra A. Siegel
University of Colorado–​Boulder
Megan A. Stewart
American University
xiv   Contributors

Mark Tessler
University of Michigan
Alex Thurston
University of Cincinnati
Maya Tudor
University of Oxford
Frédéric Volpi
University of Edinburgh
Shana S. Warren
Innovations for Poverty Action
Steven I. Wilkinson
Yale University
Erin York
Princeton University
 Acknowledgments

The inspiration for this book came from a coauthored piece in the Annual Review
of Political Science that we published in 2014 titled “Is There an Islamist Political
Advantage?” The prevailing assumption that we questioned in that article is very much
linked to the presumptions about the role of religion in Muslim societies that we inter-
rogate in this volume. Yet, because the scope here is much broader, we decided to bring
together a diverse group of outstanding social scientists with area expertise in various
parts of the Islamic world. We are grateful to these scholars for accepting our invita-
tion to contribute to this volume and for embracing our goal to investigate rather than
presume both whether and how religion influences social, political, and economic
outcomes in Muslim societies.
The project benefited greatly from two book workshops that were held in Ann Arbor,
Michigan, in May 2019 and Cambridge, Massachusetts, in December 2019. The former
was generously funded by the Weiser Center for Emerging Democracies (WCED) and
the Digital Islamic Studies Curriculum (DISC) at the University of Michigan. Derek
Groom (WCED) and Evan Murphy (DISC) provided excellent administrative sup-
port. For the latter, we received generous funding from the Weatherhead Center for
International Affairs (WCFIA) at Harvard University. At the WCFIA, we are immensely
grateful to Sarah Banse, who organized and coordinated the Cambridge workshop, and
Kathleen Hoover, who provided administrative support throughout the process of put-
ting together the book.
Finally, we owe a debt of gratitude to both Molly Balikov and Alyssa Callan at Oxford
University Press. They have been a crucial source of support and an absolute pleasure to
work with throughout the process. We especially appreciate their patience, as the timing
of this Handbook coincided with the COVID-​19 pandemic, which contributed to delays
for us and many of the contributors to this volume.
Melani Cammett, Cambridge, Massachusetts, USA
Pauline Jones, Ann Arbor, Michigan, USA
 Chapter 1

P olitics in Mu sl i m
So ciet i e s
What’s Religion Got to Do with It?

Melani Cammett and Pauline Jones

In Western scholarship, depictions of social and political life in the Muslim world os-
cillate between emphasizing the importance of religion and rejecting its relevance alto-
gether. Based on an essentialist definition of religion, an earlier generation of scholars
claimed that Islam plays an outsize role (Huntington 1996, Kedourie 1994;, Lewis 1995;
Patai 1973). Others contend that religion has a minimal influence at best, especially
when religion is defined as doctrine (Anderson 2001; Lust 2011; Tibi 2012). Although
social scientists with area expertise have long cautioned against a reductionist un-
derstanding of Islam and politics, at least until very recently, Western scholarship has
regarded Muslim-​majority states1 as places where religion dominates social and polit-
ical life, requiring distinct analytical approaches (Cammett and Kendall 2021; Mitchell
2004; Said 1981; Tessler et al. 1999). The question thus remains: What’s religion got to do
with it? How—​if at all—​does religion influence social, political, and economic outcomes in
the Muslim world? In this chapter, we explore this question based on a definition of re-
ligion that moves beyond doctrine. Importantly, our definition—​and many of the core
claims we make based on this definition—​is not unique to Muslim-​majority states and
societies, but can apply more generally to Western and non-​Western contexts with dis-
tinct dominant religious traditions.
Our broader aim in this volume is to integrate the study of politics in Muslim societies
into mainstream comparative analytical frameworks. Toward this goal, we focus on
some of the most salient outcomes of interest in the social sciences, including regime
type, political behavior, social mobilization, economic development, and govern-
ance. We take seriously the relevance of religion in Muslim societies, while seeking
both to clarify its role and probe its limitations in explaining these social, political, and
2    Melani Cammett and Pauline Jones

economic outcomes. Collectively, the chapters in this volume provide a partial but ro-
bust illustration of emerging Western scholarship that takes this approach.2 They pro-
vide insight not only into whether religion shapes these outcomes, but also under what
conditions and through which mechanisms it does so. The broad geographical scope of
the contributions to this volume, moreover, enables us to both situate and evaluate these
claims in diverse contexts across the Muslim world.
Within much of the scholarly community, the claim that religion is not the root cause
of various social, political, and economic outcomes in the Muslim world has become the
conventional wisdom. While we share this assessment, we contend that it is based on a
narrow definition of religion as doctrine—​that is, the set of beliefs and practices that de-
limit a particular faith. Indeed, perhaps the most common finding across the chapters in
this volume is that Islamic doctrine has little or no impact on a wide array of outcomes,
including political attitudes toward democracy, governance and economic develop-
ment, political participation, voting behavior, and even support for Islamist parties. This
is not to say that doctrine is unimportant, but rather that that there is no singular inter-
pretation of Islamic doctrine (hence, the small d) and that its agency depends on other
aspects of religion that are often overlooked. We define religion as both the beliefs and
practices that constitute doctrine and the infrastructure that sustains these beliefs and
practices. As we argue in this chapter, doctrine is critical to understanding religion’s po-
tential force, whereas infrastructure—​specifically, social organization (i.e., institutions
and actors) and social identity—​is the key to understanding why, when, and where reli-
gion has agency to enact that force.
Based on this definition of religion, we can offer general insights about how religion
influences a variety of important social, political, and economic outcomes (see Figure
1.1). Our goal is not to make universal claims, but rather to understand the effects of
religious institutions, actors, and identities in a given context. As the contributions in
this volume illustrate, by this definition, the influence of religion can operate through
multiple mechanisms and is elevated under a diverse set of conditions. It thus can also
have distinct effects, depending on the particular context. At the same time, given their
cross-​regional breath, the chapters point to some common mechanisms, conditions,
and effects across Muslim-​majority societies, despite their immense diversity.
First, religion as social organization can provide a technology of mobilization, par-
ticularly under conditions of state repression. Religious institutions, such as mosques

Facilitating Conditions
- State repression
- Religious actors empowered
- Violent conflict/civil war

(Selected) Outcomes
Social organization - Social & political mobilization
Mechanisms - Voting behavior
- Coordinates & sanctions behavior - Collective action
RELIGION = - Regime legitimacy
doctrine - Shapes attitudes - Political preferences
- Support for democracy
- Creates social bonds - Economic opportunities
- Attitudes toward violence
- Governance
Social identity - Economic development

Figure 1.1 How religion influences social and political outcomes.

 Politics in Muslim Societies     3

or affiliated agencies, as well as religious elites with standing in their communities, can
coordinate and sanction individuals, helping to overcome barriers to collective ac-
tion needed for social and political change. Likewise, social networks, which are often
fostered or reinforced by religious institutions and actors and may be rooted in shared
religious identity, can facilitate mobilization around social and political ends. Social
networks can also contribute to individual decisions to participate in collective action,
for example by propagating specific interpretations of doctrine that can be used to jus-
tify certain social and political goals.
Second, religion as social identity can impact social and political outcomes through
the communal bonds it fosters among religious adherents or members of the same re-
ligion. Consistent with psychological theories of identity and intergroup relations, re-
ligion as social identity can play a powerful role in shaping individual and collective
attitudes and behavior and transforming them into action. This is particularly the case
under conditions of conflict or instability. For example, threats against members of a
religious group make individuals identify more strongly with the group in question
(Huddy et al. 2007), and religion can address a vital need for security and community
(Tajfel and Turner 1979). The power of religion as social identity is in fact what makes
it so attractive to nonreligious actors, such as political and communal elites, who de-
liberately deploy this identity in public pronouncements to shore up support. The
sectarianization of social identities during conflicts waged in the name of religion also
lends support to these psychological theories.
Finally, religion is best situated to influence social and political outcomes when
religious institutions and actors are empowered—​for example, via state institutions
that elevate their status or via access to scarce resources. Across the Muslim world,
historical and more contemporary developments have endowed religious elites with
political authority and control over economic resources, giving them an advantaged
position to influence not only popular attitudes and political behavior but also ac-
cess to social and economic opportunities. This edge may arise because religious
elites utilize their control over formal state institutions to constrain attitudes and
behavior or to undermine the development of new institutions that might rival their
own. It can also arise because religious elites utilize the promise of access to scarce
resources to incentivize popular compliance with directives. These resources en-
hance the psychological appeal of religion as a social identity, which can bolster the
ability of religious institutions and actors to influence social and political attitudes
and behavior.
In the remainder of this chapter, we illustrate these arguments in more detail with
specific references to the chapters in this volume. We begin by developing our concep-
tualization of religion, which builds on a sociological approach, and by justifying our
emphasis on religion as infrastructure—​that is, social organization and identity. The
subsequent three sections focus on how religion as social organization and religion as
social identity shape these outcomes, and then point to the areas in which even a mul-
tifaceted understanding of religion appears not to illuminate the outcomes in question.
We conclude by summarizing our main analytical points.
4    Melani Cammett and Pauline Jones

Defining Religion

A religion is a unified system of beliefs and practices relative to sacred

things, that is to say set apart and forbidden, beliefs and practices which
unite into one single moral community, called a church, all those who ad-
here to them.
—​Emile Durkheim, The Elementary Forms of the Religious Life, 1982, 129

[I]‌t is intellectually imprudent and historically misguided to discuss the

relationships between Islam and politics as if there were one Islam, time-
less and eternal.
—​Olivier Roy, The Failure of Political Islam, 1996, v.

Defining religion is an excursion into well-​traveled intellectual and spiritual territory.

Admittedly, there are many competing yet equally valid understandings of what re-
ligion is. Our aim is not to review or critique the rich and expansive literature across
disciplines that seeks to conceptualize either religion in general or Islam in particular.
Nor is it to provide a definition of religion that is appropriate for all times, places, and
questions. Rather, our intent is to offer a definition that serves the two-​fold purpose of
this volume: (1) to explore the role that religion plays in shaping contemporary social,
political, and economic outcomes in Muslim-​majority countries while probing its lim-
itations as an explanation; and (2) to do so in a way that integrates the study of Muslim
societies into mainstream comparative analytical frameworks in the social sciences in
the West. We seek to define religion, therefore, in a way that is analytically clear, empiri-
cally verifiable, and broadly applicable as well as contextually appropriate.
In sum, building on a sociological approach (particularly that of Durkheim), we con-
ceptualize religion as multidimensional. It consists not only of doctrine—​that is, “a
unified system of beliefs and practices”—​but also the infrastructure that sustains these
beliefs and practices. Renewed emphasis on how doctrine can help to explain social and
political outcomes is a welcome development. However, we place greater emphasis on
infrastructure, which consists of both social organization and social identity. Doctrine
is critical to understanding religion’s potential force. Social organization and social
identity are the key to understanding why, when, and where religion has agency to enact
that force. Combined, doctrine, social organization, and social identity underscore why
religion should not be viewed as merely a “moral community.”

Religious Doctrine and Its Limits

About a decade ago, political scientists across subfields drew attention to the neglect of
religion in social scientific inquiry and called for greater emphasis on understanding
the various ways in which religion influences politics (Bellin 2008; Wald and Wilcox
 Politics in Muslim Societies     5

2006). Some scholars also called for renewed attention to the importance of doctrine
(Grzymala-​Busse 2012; Philpott 2009), which had been overshadowed by the domi-
nance of the political economy approach to the study of religion and politics (see Gill
2001 for an overview). This increased attention to religion and doctrine coincided with
a proliferation of work on the Muslim world that questioned the prevalence of both
religion and doctrine in explanations for social, political, and economic outcomes in
Muslim states and societies. Although these seem to be contradictory trends, they are
compatible in that they both acknowledge, implicitly or explicitly, the simple fact that
“[d]‌octrine is critical, but it does not act alone” (Grzymala-​Busse 2012, 126).
The recognition that doctrine lacks independent agency does not diminish its im-
portance. We share the view with other scholars that doctrine is a core dimension of
religion. It is not only what enables us to distinguish one religion from another, but also
to understand religion’s potential force. Defined as a unified set of beliefs and practices,
doctrine is essential to understanding why an appeal to religion might resonate at the
individual level and how it can be used to shape political attitudes and behavior. At the
same time, focusing solely—​or even primarily—​on doctrine limits our understanding
of both whether religion influences the outcomes we care about and how it exerts this
influence, for three main reasons.
First, treating religion as doctrine alone risks mistakenly attributing a causal role
to religion. Although doctrine is critical to understanding religion’s potential force, it
does not have the agency to enact that force. Assigning agency to doctrine is partic-
ularly problematic when it comes to Islam, given the widely held presumption that
Islamic doctrine is unique because it permeates every aspect of Muslims’ lives, and
thus is the primary driver of their political attitudes and behavior (Cammett and Jones
2014). Moreover, the “current obsessions with Islamic texts” in the case of Islam has
contributed to this misperception by serving to obscure the diversity within and across
Muslim societies (Aydin 2017, 231).
Second, and related, doctrine does not speak with one voice. As others have
recognized, across time and space, all the major world religions have expressed mul-
tiple iterations of doctrine (Black and Patton 2015; Stepan 2000). A large part of the
reason for the “multivocal” nature of doctrine (Stepan 2000, 44) is that it is subject to
interpretation. This is particularly the case for Islam, which does not vest authority in
any one source for interpreting the Quran and includes multiple schools of thought.
Yet it is also because doctrine is embedded in particular social and historical contexts
that both gives it meaning and affirms this meaning among a particular community of
adherents (Gerrish 1988, 91), making it impossible to identify a singular interpretation.
This is especially relevant for Muslim communities, where “religion as a set of rules” is
only intelligible “in the social context of practices” (Ismail 2004, 618; see also Eickelman
and Piscatori 2004, ­chapter 1). Thus, doctrine is an important source for distinguishing
among different faiths, but it should not be considered consistent across adherents
within the same faith.
Third, doctrine is also not uniformly observed. Just as doctrine is multivocal, it is
also multi-​praxis. Individuals do not necessarily conform to all the beliefs and practices
6    Melani Cammett and Pauline Jones

espoused by their religion. Nor do they do so to the same degree. Thus, we cannot as-
sume that individuals who are members of the same religion understand or adhere to
doctrine in the same way. Definitions of “religiosity” try to take this into account by
conceptualizing and measuring religiosity as beliefs, behavior, and belonging (Smidt
2019), and differentiating among types of beliefs as well as individual versus communal
practices (Beatty and Walter 1984; Layman 1997). Although this approach was devel-
oped to study the relationship between religion and politics in predominately Christian
contexts, it has gained traction in the study of Muslim-​majority countries as well (Livny
2020; Pepinskly et al. 2018). The increasing attention to not only disaggregating but also
depoliticizing piety is a welcome development in light of the prevailing tendencies to
reify Muslim piety and to presume that it is consistent with a singular social, political,
and economic agenda.

Religion as Multidimensional
By now it is clear that we conceptualize religion as multidimensional. Drawing on
Durkheim’s well-​known definition, we argue that religion consists not only of doc-
trine, but also includes the infrastructure that sustains the beliefs and practices that
make up doctrine.3 Infrastructure consists of two separate but integrally related
parts: social organization and social identity. Social organization includes both
institutions and actors. Combined with doctrine, it forms the third core element of
religion, which we argue is not merely a “moral community” but actually a social
identity. Given the aforementioned limitations of doctrine, we emphasize the role of
infrastructure as the levers of religion’s influence on a variety of social, political, and
economic outcomes. Whereas doctrine is critical to understanding religion’s potential
force, infrastructure is the key to understanding why, when, and where religion has
agency to enact that force.
In conceptualizing religion as social organization, we include both the institutions
and actors that separate religion from other aspects of social and political life. This, of
course, applies only to religion in its modern form, which makes “institutional differ-
entiation” possible and distinguishes modern religion, for example, from folk religion
(Yang 2011, 34). The institutions that comprise religion include formal and informal
rules and physical structures, whereas actors operate both because of and within
these institutions. Infrastructure, therefore, may be extensive. But it is not ubiquitous.
Rather, it consists of specific and mutually acknowledged institutions (e.g., churches or
mosques) and actors (e.g., priests or imams) within a moral community. Identifying the
relevant institutions and actors for any given religion is essential for two reasons. First,
like doctrine, infrastructure must be contextualized if we are to understand the role of
religion across and within denominations. Second, being precise enables us to ascertain
what religious infrastructure is doing and what it is not doing—​that is, whether reli-
gion per se is the driver of the social, political, and economic outcomes we care about.
This is particularly important in the case of Islam because it eschews the widely held
 Politics in Muslim Societies     7

presumption that Islam has a holistic influence on Muslims’ lives, and thus that Islam
itself has agency over their political attitudes and behavior.
Religion as social identity emerges from the combination of doctrine and social or-
ganization. Albeit discrete concepts, there has been a tendency in the context of Muslim
societies to conflate religion as social identity with religion as a culture and with religi-
osity. On the one hand, social identity is distinct from cultural definitions of religion
in that it is not all-​encompassing such that it “formulat[es] conceptions of a general
order of existence” (Geertz 1966, 4). On the other hand, religion as social identity is dis-
tinct from religiosity in two main ways. First, although related to doctrine, it does not
depend on uniformity of interpretation or commitment. Social identity goes beyond
Durkheim’s notion of “a moral community” in that it recognizes there can be diversity
within a community concerning adherence to doctrine, and yet the community can
still be unified by religion.4 In other words, social identity is not based solely on the de-
gree to which each individual member of the group is committed to doctrine.5 Second,
whereas religiosity concerns individual level piety, social identity can be manifested at
both the individual and group level, whether this corresponds to a village or the nation
(or even the global). Personal religiosity can thus only shed light on individual attitudes
and behaviors, whereas social identity can improve our understanding of individual and
collective attitudes and behaviors, including local and national (and even international)
forms of mobilization.

What Religion Is Not

Providing a clear definition of what religion is—​and thus whether it affects social and
political outcomes—​also means clarifying what it is not. First and foremost, we do not
share the view that religion in general, and Islam in particular, is all-​encompassing,
and thus influences every aspect of an individual’s life. Contra recent work by Shahab
Ahmed (2015, 259), the “mere fact of [Islam’s] massiveness” does not mean that it al-
ways “exercises agency and power.” Second, religion is not an ideology in the sense that it
corresponds to a particular set of beliefs or preferences about how the political and eco-
nomic system should operate. Our understanding of the influence of religion on politics
in Muslim societies has been hindered by the tendency to treat Islam itself as a program
for social and political change rather than as a religion. Religion can certainly become
the basis for an ideology, but it is not in and of itself a call for action.
Third, and directly related, religion is not the same as the movements and parties that
appeal to religion in order to mobilize supporters. Although these actors deploy reli-
gion to construct an ideology, they are not religious actors because they are not part of
religious infrastructure. This is an important point of departure, because Islamists are
sometimes considered to be major religious agents in the Muslim world. If we define
Islamists as actors who seek to make Islamic principles the basis for public and private
life and aim to use state powers of legislation and enforcement to do so, then this char-
acterization may seem apt (see, for example, Brooke’s chapter in this volume). However,
8    Melani Cammett and Pauline Jones

we view Islamist parties and politicians not as religious actors per se, but rather as actors
deploying religion toward political ends. Some of their policy preferences and goals may
be rooted in interpretations of Islamic doctrine, such as proposed legislation around
the role of religion in the constitution or family law, but in almost all Muslim-​majority
countries they vie for influence in political systems that are not theocratic or controlled
by religious authorities. Finally, we do not define religion as the obverse of “secular”—​a
concept that is too often presumed to be transparent in its meaning and limited in its
scope when in fact it is neither (Asad 2003).

Religion as Social Organization

Rather than doctrine, most contemporary scholarship on politics in Muslim societies

focuses on the social organization of religion, as manifested in institutions and actors.
In studying how religion shapes social and political outcomes, religious institutions and
actors are central because they adopt and advocate particular interpretations of Islam
with respect to questions of governance, frame the choices facing elites and non-​elites
according to these interpretations, and incite and facilitate collective action. When reli-
gion is defined in terms of these subdimensions of social organization, a strong case can
be made that religion “matters” in Muslim-​majority countries—​and in non-​Muslim-​
majority contexts as well, albeit via distinct types of institutions and actors.

Religious Institutions
Religious institutions can take varied forms, including formal and informal rules, and
in some instances are rooted in physical infrastructure. In social science research on
Muslim-​majority societies, all of these types of institutions have been shown to affect
key outcomes of interest such as political regime transitions, socioeconomic develop-
ment, political behavior, and social mobilization. Relevant formal institutions might in-
clude Islamic laws such as those related to finance or capital accumulation, informal
institutions such as uncodified norms, and brick and mortar institutions such as
mosques or community centers. To the extent that doctrine shapes social and political
outcomes, it is funneled through these channels. Formal and informal institutions can
privilege particular interpretations of doctrine, thereby setting the parameters for the
attitudes and behaviors of community leaders as well as individuals within the commu-
nity, but doctrine alone is not the catalyst for mobilization.
The impact of Islamic financial institutions on investment and growth illustrates how
formal institutions associated with Islam affect development, an important outcome in
social science research in general, and a central focus of the growing literature on the
“great divergence” between the West and Muslim societies. Timur Kuran’s (2011) ac-
count of how Islamic law “held back” the Middle East is illustrative. In his framework,
 Politics in Muslim Societies     9

aspects of Islamic law related to inheritance, business partnerships, and waqfs or

charities deterred capital accumulation as well as the flexibility to adapt to changing
circumstances, hindering economic growth and development. In this volume, Fulya
Apaydin also points to the role of formal institutions in her analysis of the effects of
Islamic banking on capital accumulation and productive growth. The case of Malaysia,
where Islamic banks have grown rapidly since the early 1980s, seems to be a “most likely”
success case in light of recent research on the conditions under which shariʿa-​compliant
banks foster growth, because Malaysian Islamic financial institutions operate alongside
the regular banking sector. Furthermore, Islamic banks have integrated previously un-
banked populations into the financial system. Yet even in this context, the Malaysian
experience shows that Islamic banks have not produced sustainable growth but rather
have extended loans to finance real estate and consumer durables purchases to the detri-
ment of productive investment, benefiting large corporations and governments instead
of small and medium enterprises. In line with Kuran (2011), Apaydin therefore indicates
that a key set of formal institutions in Muslim societies are not conducive to growth and
development.
In his chapter, Mohamed Saleh identifies taxation as a distinct formal institutional
feature of Islamic law that can explain the relative underdevelopment of Muslim
populations—​in this case, the inferior socioeconomic status of Muslims in compar-
ison with Copts in Egypt. Saleh shows how the imposition of a regressive poll tax on
non-​Muslims, in effect from the mid-​seventh through the mid-​nineteenth centuries,
incentivized poorer Egyptians (who were all Copts, initially) to convert to Islam, leading
nonconvert Copts to shrink into a better-​off minority. Group-​based restrictions into
certain occupations reinforced these divisions over time. In this way, Islamic taxation
acted as a screening mechanism of who stayed Copt and who became Muslim, rather
than a cause of the underdevelopment of Muslims once they converted.
Informal institutions, or unwritten rules that are generated, disseminated, and en-
forced outside of official or written channels (North 1990), also shape political and
social behavior, and often explain how formal institutions actually operate in distinct
contexts (Helmke and Levitsky 2004, 726). In this volume, several chapters touch on
the ways in which informal institutions, notably norms, shape political attitudes and
behaviors. Based on a survey of Iraqi and Iranian Shiʿa on the pilgrimage to Karbala,
Fotini Christia, Elizabeth Dekeyser, and Dean Knox argue that communal practice,
which transmits norms through religious socialization, affects attitudes about the role
of religion in politics, perceptions of regime legitimacy, and support for democracy.
Among other findings, however, they show that religious Iraqis and Iranians, who are
more likely to engage in communal practices such as the pilgrimage, harbor distinct
views about the appropriate relationship between the state and religion. This suggests
that different norms emerged among Shiʿa in these two different national contexts,
resulting in distinct political attitudes. More generally, if informal institutions condition
the effects of formal institutions, then ongoing research should investigate how the same
Islamic institutions have affected long-​term trajectories of political and economic devel-
opment in varied ways across time and space.
10    Melani Cammett and Pauline Jones

Tangible brick and mortar institutions associated with Islam, most notably mosques,
can facilitate various forms of political mobilization by serving as sites for coordinating
mass protests or by facilitating electoral turnout. In her analysis of the Syrian uprising,
Wendy Pearlman underscores the widely accepted argument that religion was unrelated
to the motivations of the protestors, who were instead driven by political and economic
grievances. Yet she also details how religious infrastructure played a critical role in the
process of mass mobilization. Popular demonstrations were often initiated at mosques
at the end of Friday prayers. This was not because of religious grievances, but rather be-
cause mosques were essentially the only spaces available for mass congregations, given a
long history of state repression of independent sites for civic engagement. In this sense,
her analysis echoes that of McAdam (1999), who pointed to the pivotal role of Black
churches in organizing dissent during the US civil right movement. With regard to elec-
toral politics, Islamist parties have benefited from their linkages to religious institutions
such as mosques to drive voter turnout (Brooke 2019, and this volume; Clark 2004;
Wickham 2002) or their embeddedness in social networks, which such institutions help
to reinforce (Masoud 2014, 168).

Religious Actors
Religious actors, including religious elites and scholars or ʿulamaʾ, figure prominently
in accounts of politics in Muslim-​majority societies. As in the case of institutions, the
power of religious actors is generally not derived from doctrine but rather from their so-
cial and political status. Furthermore, doctrine itself does not determine the status of re-
ligious actors; rather, factors commonly emphasized in comparative historical analysis,
such as legacies of state formation, mobilizational resources and capacity, and strategic
behavior, provide more convincing explanations for the influence of religious actors
on governance and public life in Muslim-​majority countries. Given status, religious
authorities can play a pivotal role in interpreting doctrine and thereby enjoy enormous
discretionary power to shape the effects of Islam on social and political life. Thus, reli-
gious elites often deploy religious appeals to suit their aims or vision for society and pol-
itics. This raises further questions about why such language resonates among segments
of the mass public, which we address in the next section on social identity.
Under some historical conditions, religious actors, notably the ʿulamaʾ, acquire
disproportionate influence over social and political outcomes. In particular, scholars
of the Muslim world point to the role of religious authorities in hindering political
opening and the development of an independent commercial class, which is often
linked to the development of the rule of law and checks on absolute authority (Kuru
2019; Rubin 2017).6 For the most part, these accounts do not stress doctrine but rather
point to the ways in which state-​society relations and societal power dynamics shape
regime outcomes, although Rubin (2017, 205–​206) contends that Islamic doctrine
serves to legitimize political authority, whereas Christian, and especially Protestant,
doctrine does not.
 Politics in Muslim Societies     11

Regardless of the precise historical factors that led to the empowerment of religious
elites, the evidence suggests that when clerics with anti-​democratic or illiberal views
enjoy great influence in politics and society, Muslim-​majority countries are more likely
to exhibit a democratic deficit—​a finding that undoubtedly extends beyond the Muslim
world. A number of contributions to this volume support this claim. In his chapter, Eric
Chaney argues that the authoritarian equilibrium in the Muslim world stems from a se-
quence of historical developments culminating in the emergence of an influential class
of religious elites who brought about the “religionization” of civil society by the elev-
enth century CE. The formal institutions devised under the authority of these conser-
vative elites cemented authoritarian rule, which Chaney posits endured through at least
the nineteenth century. Based on a more temporally proximate argument, Maya Tudor,
too, points to the role of elites committed to particular religious ideas in undercutting
democratic development. Contrasting the case of Pakistan to that of Indonesia and
India, she shows how the fusing of the majority religion, Islam, with national identity
at the country’s founding hindered the formation and durability of democracy because
religious leaders were able to undermine minority rights. In his chapter on Pakistan,
Matthew J. Nelson also argues that conservative political and religious elites have pushed
through constitutional provisions that privilege their interpretations, diminishing the
rights of non-​Muslims and resulting in an illiberal form of democracy, which significant
portions of the population appear to endorse.
Religious elites are not always anti-​democratic. In the case of Senegal, as Dominika
Koter shows, religious leaders from different Muslim brotherhoods have played a key
role in driving voter mobilization, brokering votes, and even issuing directives to their
followers to go to the polls. Furthermore, her analyses of electoral data indicate that re-
ligion is not a strong predictor of vote choice and that Muslim religious leaders back
politicians across political and religious groups, facilitating more robust political com-
petition. The contrast between her findings with those of others in this volume, and with
the broader scholarship on authoritarian durability in the Middle East, underscores the
need for more cross-​regional comparative historical analyses to untangle the conditions
under which religious authorities adopt conservative positions that undermine dem-
ocratic competition. In a complementary analysis of three Southeast Asian Muslim-​
majority countries, Kikue Hamayotsu argues that an elevated role for religious leaders
in state formation undermines democratic consolidation. In her account, when “reli-
gious regimes,” or institutionalized relations between political and clerical authorities,
crystallize prior to the formation of the modern state and the rise of popular democratic
movements, religious elites are more likely to hinder democratic consolidation.
The role of powerful religious elites is also linked to economic and social outcomes
(Bazzi, Koehler-​Derrick, and Marx 2020; Chaney 2013). In this volume, Jared Rubin
maintains that the relative power of religious leaders impeded economic development
in Muslim countries over the long run in comparison with the West. To bolster their
authority, Muslim rulers relied more heavily on religious legitimacy than their Western
counterparts. As a result, economic elites had less influence on policymaking, leading
to weaker institutionalization of property rights, which are often linked to growth and
12    Melani Cammett and Pauline Jones

development. Focusing on social outcomes, notably literacy, Adeel Malik and Rinchan
Mirza also highlight the ways in which the elevated status of religious actors impedes
development in certain contexts. In Pakistan, shrine elites, or caretakers of the tombs
of saints venerated by Islamic mystical orders, became empowered after the coup
of General Zia ul-​Haq in 1977 through his selective opening of the electoral arena.
Although Zia presided over a major expansion of educational infrastructure, shrine
elites, who were averse to mass schooling, leveraged their newfound political power
to suppress literacy expansion in the regions they dominated. More broadly, Malik
and Mirza emphasize the interplay between religious power and formal institutional
arrangements, showing how the effects of historically embedded religious power can re-
main latent for a long time and only become salient after conducive policy shocks.
In some contexts, religious elites play a key role in providing or arranging access to
social services, as Steven Thomas Brooke’s overview of the literature on Islamist ser-
vice provision in this volume documents. In Mali, for example, Jaimie Bleck and Alex
Thurston’s chapter demonstrates that religious actors provide key services, such as ed-
ucation and justice, to the population. Although religious parties and politicians do
not exist in Mali, clerics engage indirectly in politics by shoring up support for polit-
ical authorities and exerting influence over policymaking through their patronage
networks. In this way, clerics effectively work within the system. In recent years, how-
ever, jihadists, who have an antagonistic relationship with the state, are increasingly
involved in service provision, potentially shifting the long-​standing pattern of accom-
modation between the state and religious actors in Mali.
Finally, by virtue of their status, religious elites may have unique abilities to mobi-
lize people even in the absence of providing social benefits. Based on a study of female
Muslim preachers in online fora, Richard A. Nielsen explains the growing appeal of
female preachers with a social movement logic. Islamist movements benefit from the
fact that female preachers face fewer constraints to preach online than they do offline
in order to tap into new audiences of both women and men. In this interpretation, re-
ligious ideas serve as a normative constraint that compels female preachers to operate
online, not as a cause of this growing phenomenon, and religious ideas about gender
segregation do not explain the popularity of female preachers.
In short, extensive evidence from this volume and from preexisting research indicates
that religious institutions and actors shape important social, political, and economic
outcomes, but few contemporary scholars endorse the claim that their influence is de-
rived solely from Islamic doctrine.

Religion as Social Identity

As argued previously, religion as social identity is distinct both from religion, under-
stood as a culture, and from the concept of religiosity. Yet they have often been conflated
in the case of Islam. This tendency has fostered the view that religion both pervades
 Politics in Muslim Societies     13

Muslim societies and that Muslim societies are deeply religious. Perhaps the best ex-
ample is the substantial—​and quite influential—​literature that emerged in the 1990s,
which made claims about the inherent incompatibility of Muslim-​majority states
and democratic regimes based on an essentialist view of Islamic culture and society
(Huntington 1996; Kedourie 1994; Lewis 1995). In sum, Muslims were presumed both
to have negative attitudes toward democracy—​with implications for both their voting
preferences and mobilization potential (i.e., they would necessarily support Islamists
[Woltering 2002; Yavuz 1997])—​and positive attitudes toward violence (Kramer 1997).
In this volume, we not only eschew a cultural definition of religion but also distin-
guish the role of religiosity, including how it is measured, from that of religion as a social
identity. In other words, religiosity and religious identity are not just distinct concep-
tually, they also influence attitudes and behaviors in different ways. More specifically,
several chapters in this volume examine the influence of religiosity and religion as social
identity on four main outcomes identified as salient in the literature: (1) attitudes toward
democracy, (2) voting behavior, (3) social mobilization and political participation, and
(4) support for violence and terrorism.
When it comes to attitudes toward democracy, the findings across the chapters in this
volume suggest that neither religiosity nor religious identity have an effect. However,
they also suggest that both religiosity and religious identity can influence the type of
democracy that emerges. These findings are largely consistent with more recent schol-
arship based on survey research in the 2000s, which argues that Muslims are mostly
supportive of democracy as a form of government and do not view democracy as neces-
sarily incompatible with their religious principles (Esposito and Mogahed 2008; Norris
and Inglehart 2003; Rose 2002; Tessler et al. 2012). Also like recent scholarship, they are
largely consistent across Muslim-​majority states in different regions of the world.
The chapters in this volume that examine the relationship between religiosity
and attitudes toward democracy find that being more devout does not promote anti-​
democratic attitudes among Muslims in countries as distinct as Turkey and Pakistan.
For example, based on data from a series of nationally representative surveys, S. Erdem
Aytaç’s chapter on religiosity and political attitudes in Turkey finds that devout Muslims
have actually become more favorable toward democracy under the Party of Justice and
Development (AKP). He argues that this is because the AKP has allowed religion to play
both a more public and visible role, leading to “a rapprochement of devout Muslims
with the political regime,” whereas democracy under the previous (centrist) party lead-
ership was associated with restrictive policies toward religion in public life. Similarly,
based on the most recent Pew Research Center survey data (2015) for Pakistan, Niloufer
A. Siddiqui argues that religiosity has a positive effect on attitudes toward democracy
as a form of government. She finds that higher levels of religiosity, regardless of how the
concept is measured, are linked to greater support for democracy.
Yet these two chapters reach slightly different conclusions regarding whether reli-
giosity is linked to support for less inclusive (or illiberal) forms of democracy. In the
case of Turkey, Aytaç finds that more devout Muslims are less likely to have a pluralistic
understanding of democracy. Instead, they are more likely to hold a populist view of
14    Melani Cammett and Pauline Jones

politics, and thus to support policies that benefit the majority in the name of popular
sovereignty. Such policies are consistent with the AKP’s platform, suggesting that devout
Muslims in Turkey have become more pro-​democracy with the election of a party that
represents their own views. At the same time, Aytaç finds that devout Muslims under
the AKP have become less conservative in their political views; for example, they are less
likely to favor the establishment of a shariʿa-​based religious state in Turkey. Siddiqui’s
chapter introduces the important caveat that whether religiosity promotes more con-
servative political attitudes depends on how it is measured. Her three-​fold measure of
religiosity maps onto the aforementioned distinction between beliefs, behavior, and
belonging: (1) dogmatism (“support for Islamic law in governance”), (2) piety (“partic-
ipation in religious ritual”), and (3) devotion (“role of religion in one’s life”). She finds
that when religiosity is measured as either behavior or belonging, higher levels of reli-
giosity are associated with more conservative political views that can make democracy
less inclusive, such as support for restrictions on blasphemous speech. However, when it
is measured as belief, higher levels of religiosity are not associated with more conserva-
tive political views.
What about the relationship between religious identity and democracy? Here, other
relevant chapters in this volume point to similar findings. Focusing on religion as a so-
cial identity at the national level, both Maya Tudor and Matthew J. Nelson argue that
when religion and the state are intertwined, religious identity constrains both democ-
racy in general and liberal democracy in particular. Tudor compares three postcolo-
nial countries in South and Southeast Asia (Pakistan, India, and Indonesia), arguing
that the strategies of indigenous elites to mobilize the population in support of their
nation-​building projects had long-​term consequences for the viability of democracy.
More specifically, when these elites chose to make religious identity central to “founding
nationalisms,” stable democracies were less likely to emerge. Pakistan is a case in point.
Whereas the anti-​colonial movements in India and Indonesia actively sought to sepa-
rate nation and religion, Pakistan’s independence movement deliberately fused religious
and national identities. Islam has thus played an oversized role in regime legitimation,
and, as a result, Pakistan has been plagued by “an unstable [democratic] regime with
autocratic tendencies.” Nelson corroborates this argument by demonstrating that both
autocratic military regimes and democratically elected civilian regimes in Pakistan have
endorsed conservative constitutional provisions that are consistent with the country’s
dominant interpretations of Islam. These provisions, which establish different rights for
Muslims and non-​Muslims and restrict speech that might be considered disrespectful to
Islam or Muslims, have served to institutionalize regimes that are both exclusionary and
illiberal.
The findings across the chapters in this volume are also mostly consistent with re-
cent scholarship concerning the influence of religious identity and religiosity on voting
behavior in the Islamic world. Democratization and multiparty competition in sev-
eral Muslim-​majority societies, particularly over (roughly) the past two decades, has
enabled scholars to investigate this question empirically (Masoud 2014; Pepinsky et al.
2012, 2018). These studies largely conclude that Muslims are strategic voters who are
 Politics in Muslim Societies     15

motivated to cast their vote for a party for much the same reasons as voters in other,
less developed democracies, notably access to public and private goods, economic
policy and performance, and ideological affinity. In other words, contra earlier studies
(Esmer and Pettersson 2007), religion does not seem to be the driving factor behind
Muslim voters’ preferences. Recent studies also highlight the diversity of electoral
agendas among Islamist parties, which offer different choices to voters (Kurzman and
Naqvi 2010).
The chapters by Thomas Pepinsky and Steven I. Wilkinson in this volume contribute
to this growing literature by problematizing the common inference that voting for a re-
ligious party can be interpreted as support for a religious agenda. Pepinsky’s analysis
is based on two Muslim-​majority societies (Indonesia and Malaysia) that differ in sev-
eral important ways, including their ethnic and religious composition, but have expe-
rienced competition between parties that “implicitly or explicitly invoke Islam” (i.e.,
Islamic parties) and those that do not (i.e., non-​Islamic parties). He argues that one of
the reasons it is so difficult to determine whether religiosity is driving voter support
for an Islamic party is that these parties differ from non-​Islamic parties in ways other
than the religious aspects of their platforms. Inferring that a voter’s support for a non-​
Islamic party is not linked to religiosity is also problematic because these parties can
be—​and have been, in the context of Southeast Asia—​responsible for implementing re-
ligious policies. In sum, Pepinsky cautions that “voting for an Islamic party is not al-
ways a vote for Islam and voting for a non-​Islamic party sometimes is.” Wilkinson
makes similar claims in the context of South Asia. He argues that Indian voters—​both
the majority Hindu population and the substantial Muslim minority—​have consist-
ently been motivated by economic issues rather than religious and caste identities. Thus,
votes from “a particular group . . . for a particular party” are often misinterpreted as a
“group vote” when they are actually a vote for development. Conversely, he argues that
the mediocre electoral support for religious parties in Pakistan is not an indication that
voters do not support their religious agenda, but rather that this agenda has already been
institutionalized.
The insight that voters in Muslim-​majority societies are not swayed solely or even
primarily by religious appeals is echoed in Daniel Corstange and Erin York’s chapter
comparing linkage strategies across regime and opposition parties in Lebanon, Yemen,
and Morocco. They argue that parties in Muslim societies operate under the same
constraints as parties in non-​Muslim societies at similar levels of political and economic
development when it comes to developing strategies to capture vote share. In partic-
ular, they “find evidence that Muslim and Arab polities are not, in fact, more clientelist
than comparable non-​Muslim states.” The incentive to engage in clientelism, moreover,
applies to Islamic and non-​Islamic parties alike. In both Lebanon and Yemen, for ex-
ample, all major parties—​not just the Islamist Hizbullah and Islah—​focus on delivering
services and offering material payoffs to their constituents in order to win elections, al-
though regime parties often have an advantage given their access to state resources.
The chapter by Ellen Lust, Kristen Kao, and Gibran Okar presents a more nuanced
view of the role that religion plays in shaping voter preferences, particularly regarding
16    Melani Cammett and Pauline Jones

Islamic parties. Based on an analysis of elections during political transition in three

Muslim-​majority countries in the Middle East (Egypt, Libya, and Tunisia), they
argue that religious factors are more important determinants of electoral support for
Islamists than economic or organizational factors. However, an important caveat is
that their argument is based on an understanding of religion that differs from the one
we offer in this chapter. Consistent with the other chapters in this volume, they find
that religious identity (i.e., measured as choosing “Muslim” as one’s primary identity)
did not influence the decision to cast a vote for Islamists in any of these three coun-
tries. More inconsistent with the predominant approach in this volume, they find that
religiosity (measured as the frequency with which one engages in religious practice)
did appear to affect electoral support for Islamists, albeit only in Tunisia. Religion as
they define it seems to have the greatest influence on voter preferences when it comes
to attitudes toward the relationship between religion and politics. In Egypt and Libya,
they find that those who favored religious over secular parties were more likely to sup-
port Islamists, while in Tunisia and Libya they find that those who responded affirm-
atively to the idea that religious leaders should influence politics were more likely to
support Islamists.
Religion as social identity seems to exercise the greatest influence on attitudes
and behavior in Muslim societies when it comes to social and political mobilization.
Whereas some scholars have pointed to religion’s intrinsic value as a mobilizational tool
(Aminzade and Perry 2001; Tarrow 1998), the chapters in this volume emphasize the role
of collective practice in shaping religious identity and, consequently, attitudes toward
activism. Avital Livny argues that the gap in political participation between Muslim and
non-​Muslim societies cannot be attributed to the degree of religiosity, but rather to how
it is expressed. That is, the type of practices in which Muslim engage, and not their level
of piety, affects political participation. Specifically, she argues that collective religious
engagement, such as mosque attendance, bolsters protest potential among Muslims
because it promotes social relationships based on interpersonal trust that can help to
overcome barriers to collective action. Dustin Gamza and Pauline Jones also explore
the relationship between communal practice and political mobilization. Their focus,
however, is different in that they examine Muslim attitudes toward political action in
response to state repression of individual (wearing the hijab) versus communal (gath-
ering in Qurʾan study groups) practices. In some respects, their findings are distinct
from those of Livny. Contrary to their own expectations, they find that when communal
practices are repressed, Muslims are less willing to support mobilization against the
state than they are when individual practices are repressed. However, like Livny, Gamza
and Jones link these findings to the social bonds that are forged through communal
practices. Importantly, these chapters (as well as the chapter by Christia, Dekeyser, and
Knox) are consistent with the broader literature on religion and politics across various
faiths that emphasizes the role of communal practices in shaping political attitudes and
behavior (Harris 1998; Putnam 2000), particularly when it comes to group mobilization
(Harris 1998; Jamal 2005; Ludden 2005).
 Politics in Muslim Societies     17

In seeking to explain violent forms of mobilization, past work has consistently

invoked religion as intrinsically related to violence (see Brubaker 2015 for an overview).
For the Muslim world in particular, the dominant narrative has long been that Muslim
societies are more prone to engage in or support violent conflict because Islamic doc-
trine legitimates violence and its devoted adherents are motivated by calls to defend
their religion (Jackson 2007). Two chapters in this volume provide a much different
view. They go beyond the existing literature finding no association between religi-
osity and support for violence (Clingingsmith, Khwaja, and Kremer 2009; Tessler and
Robbins 2007), by actually detecting a negative relationship. Based on an embedded
experiment in the 2016–​2017 Arab Barometer surveys, Mark Tessler, Michael Robbins,
and Amaney Jamal focus specifically on attitudes toward the Islamic State. Across seven
different Arab countries, they find widespread disagreement with both the goals and
tactics of the Islamic State among ordinary Sunni Muslims. Moreover, the greatest dis-
agreement seems to be among the most religious, who are more likely to endorse the
statement that “the actions of the Islamic State are not compatible with the teachings of
Islam.” In the case of Pakistan, Siddiqui finds that, whether measured as belief, behavior,
or belonging, greater religiosity has a negative effect on attitudes toward religious vio-
lence. Specifically, respondents who are either more devout or more dogmatic are also
more likely to reject the claim that violence can be justified on religious grounds.
More recent research has argued that the relationship between religion and violent
conflict is more complex and, like nonviolent forms of mobilization, may be mediated by
the salience of social identities forged through communal religious practice (Hoffman
and Nugent 2017). The chapter by Alexandra Scacco and Shana S. Warren corroborates
these claims, albeit in the opposite direction. Based on a study of Christian and Muslim
residents in Kaduna, Nigeria, who directly experienced large-​scale deadly religious vi-
olence, they argue that exposure to conflict can reinforce strong intragroup ties that
create obstacles to intergroup reconciliation after conflict.

Beyond Religion

Research on the Muslim world links social and political outcomes to a host of factors
beyond what we refer to as religious infrastructure. Indeed, because they have so con-
vincingly demonstrated that nonreligious factors drive political behavior and outcomes
in the Muslim world, scholars with expertise in Muslim-​majority countries often view
arguments rooted in religion with skepticism. Instead, they link a variety of analytically
distinct institutions and actors, which are not religious but are nonetheless sometimes
conflated with religion, to political and economic development, political values, col-
lective action and mass mobilization, and other social and political phenomena in the
Muslim world. Here we highlight some of the existing scholarship and chapters in this
volume that emphasize nonreligious institutions and actors.
18    Melani Cammett and Pauline Jones

Nonreligious Institutions
Research on the Muslim world points to a variety of institutional forms that arise from
nonreligious influences and range from those that originated in the deep past to those
from more contemporary periods. More temporally distant institutions date back to the
pre-​Islamic period and even earlier, the Muslim conquest, the subsequent rise of various
Islamic empires or, more proximately, the colonial era. Alternative accounts trace the
roots of the social and political phenomena in Muslim societies to the 20th century, when
many Muslim-​majority societies emerged from colonial occupations or were established
as independent states. Even when there is nothing particularly “Islamic” about these
institutions, because they are not part of the social infrastructure of Islam itself, they are
sometimes conflated with Islam because they interacted with religious institutions or ad-
vantaged religious actors in shaping economic and political development.
The deep history of state-​building is a prominent explanation for the distinctiveness
of politics and economic development in the Muslim world. More specifically, the de-
velopment of state institutions in the pre-​Islamic period, which resulted from the top-
ographical conditions in the regions where Islam first spread, has been linked to the
absence of democracy in Muslim-​majority polities (Hariri 2015; Kennedy 1981). In places
where agriculture called for irrigation techniques, more centralized and top-​down po-
litical forms developed to coordinate the prevailing mode of production and persisted
over time. During the colonial era, European powers were less likely to establish direct
rule and to attract European settlers in these contexts, inhibiting the transmission of
democratic institutions and practices (Hariri 2015). Beyond the troubling normative
implications of this argument—​that the suppression and occupation of indigenous
people is a major route to democratization in non-​Western settings—​historical evi-
dence shows that local political actors sought to construct representative institutions
in the 19th and 20th centuries, but that the French and British actively dismantled them
(Atallah 2010; Thompson 2020).
Another genre of long-​run historical argument points to distinct types of institutions
unique to Islamic societies but, again, that were not derived from Islam per se. Lisa
Blaydes and Eric Chaney (2013) argue that the development of military slavery
(mamluks) and a system of agricultural taxation (iqta), which developed after the Arab
conquests, hindered the establishment of checks on political authority. Rather than con-
scription from the local population, importing slave soldiers from afar ensured that
these coercive forces were dependent on the ruler and were not socially embedded.
Similarly, the iqta system, which granted tax rights to the landholder that could not be
inherited or sold in return for military service, prevented the rise of an independent class
of landholders that could counterbalance the ruler’s power, a pattern of state-​society re-
lations that is linked to the rise of representative institutions in the West. Blaydes and
Chaney emphasize that these institutions do not stem from Islamic doctrine, but, given
their prevalence across the Arab, Persian, and Turkic areas, mamlukism and iqta effec-
tively evolved to become properties of many Muslim societies with enduring effects.
 Politics in Muslim Societies     19

In this volume, Lisa Blaydes and Safinaz El Tarouty focus on nineteenth-​century

Egypt to show how this historical legacy of an unconstrained executive permitted exces-
sive borrowing, which paved the way for European interference. Although Egypt under
Muhammad Ali had embarked on state centralization efforts that were similar to those
occurring in Europe at the time, no domestic actor held sufficient power to rein in ex-
cessive spending. Their argument spotlights institutions that developed in a Muslim-​
majority society rather than features of Islam or Islamic institutions per se. It also raises
questions for further research about the conditions under which elites oppose ruler bor-
rowing, especially if they do not want to be taxed and external sources of capital are
available.
It is important to note that such long-​run historical approaches discount the impor-
tance of subsequent historical developments, such as colonialism, which they depict
as endogenous to the weakness of Muslim societies as a result of institutions dating to
the pre-​Islamic or classical Islamic periods (Blaydes 2017, 488). Even if the decline of
Muslim empires over centuries made them more vulnerable to foreign occupation and
control, however, colonialism could still have independent effects on social and polit-
ical outcomes in the colonized territories. As economic historians have shown, colo-
nial practices in the Ottoman Empire, such as the Capitulations or the establishment
of distinct commercial regulations for minority communities, had deleterious eco-
nomic effects on the nascent economies of the post-​independence Middle Eastern states
(Owen 1993; Owen and Pamuk 1999).
In this volume, several chapters argue that European colonialism led to variation in
social and economic development in Muslim societies. Melina R. Platas documents a
human development gap between Muslims and Christians in sub-​Saharan Africa, where
Muslim-​majority countries are on average poorer and exhibit inferior social outcomes.
She argues that norms and beliefs about gender equality, fertility, and schooling as well
as differential investment in education help to explain this gap. Yet the religious con-
tent of institutions or beliefs do not explain these differences; rather, distinct patterns
of colonial governance may have undercut the accountability of local leaders in Muslim
areas in comparison with those controlled by non-​Muslim authorities in the precolonial
and colonial periods. Melani Cammett, Allison Spencer Hartnett, and Gabriel Koehler-​
Derrick also focus on the variation in colonial institutions, in this case to explain distinct
patterns of welfare infrastructure development in post-​independence Middle Eastern
countries. They argue that the relative reliance on public versus private institutions
during the colonial period shaped the postcolonial extension of public health and ed-
ucation institutions. However, the effects of colonialism are also conditional on the na-
ture of the governing coalition at independence: when new ruling elites brought about
a revolutionary break from the past, new patterns of welfare regime performance arose.
Malik and Mirza’s chapter also stresses the impact of colonial influences. The British co-
lonial administration bolstered the power of historic religious elites by introducing pro-
perty rights that made religious succession more hereditary in shrine institutions and
structurally aligned the interests of shrine elites with other landholders. Furthermore,
20    Melani Cammett and Pauline Jones

by classifying shrine elites as “agrarian castes,” colonial authorities made them eligible to
receive landed gentry status.
Are Muslim societies stuck in a negative economic and political equilibrium as a
result of long-​run historical factors? A skeptic could reasonably argue that such his-
torical approaches rob post-​independence elites and citizens of their agency.7 Post-​
independence political and economic institutions, the result of elite choices and
bargaining among state and societal actors, constitute more proximate influences on so-
cial and political outcomes in the Muslim world and may even establish new “paths” that
erode the influence of historical legacies (Cammett 2017).
Some authors in this volume argue that political institutions established in the post-​
independence period shape current political and economic development. To explain
“executive aggrandizement” in Turkey, Feryaz Ocaklı emphasizes the deficiencies of
state institutions, which have permitted President Erdogan and his allies in the ruling
Justice and Development Party (AKP) to tilt elections in their favor and to staff gov-
ernment agencies with loyalists. Although the AKP is an Islamist party by many
definitions, its behavior is not rooted in religion but instead employs strategies that
were present during long stretches of secular rule in Turkish politics. Islamists are no
more or less trustworthy than their secular counterparts; rather, institutional weak-
ness enables incumbents—​in this case, Islamists—​to strengthen their grip on power.8
Elizabeth R. Nugent’s chapter, too, centers on more contemporary political institutions,
documenting that party systems in Muslim-​majority countries are less competitive
and institutionalized than their counterparts in non-​Muslim-​majority contexts. But
to understand the origins of these institutions, she argues, we should look to historical
developments dating back to colonialism and persisting over time. In particular, the
colonial powers established unrepresentative legislative institutions and favored some
groups over others in divide-​and-​rule strategies, practices that influenced the creation
of electoral rules in post-​independence regimes.

Nonreligious Actors
Many accounts of social and political phenomena highlight elite and non-​elite actors
beyond religious authorities. Relevant actors include political and economic elites, the
military, opposition figures and parties—​most notably Islamist parties, which we con-
sider to be nonreligious actors, as noted previously.
In their role as policymakers, political elites have the potential to shape political
and economic outcomes most directly. To explain post-​independence economic de-
velopment trajectories in Syria and Turkey, David Waldner (1999) argues that conflict
among elites compelled them to establish broad coalitions, which led to the creation
of institutional arrangements that undermined capital formation and investment. In a
similar vein, Ferdinand Eibl’s chapter in this volume explains the emergence of Middle
Eastern welfare regimes, which differ in their accessibility and generosity, in part as a
result of cross-​national variation in social coalitions at independence. Unlike Cammett,
 Politics in Muslim Societies     21

Hartnett, and Koehler-​Derrick, as well as Platas, Eibl focuses on the postcolonial period,
but like these authors he does not locate the roots of varied development trajectories in
religion or in religious institutions and practices.
Political elites across the Muslim world also affect outcomes by explicitly deploy reli-
gious appeals and using religious infrastructure to their advantage (Abboud 2015; Baram
1997)—​a phenomenon that extends to politicians in Western societies, for example, who
make explicit references to Christianity in their public statements. Pearlman’s chapter
provides a recent example in her account of the Syrian war. She argues that the sectarian
dimensions of the conflict escalated not because the opposition was motivated by re-
ligion or religiosity, but rather, in part, because the regime deliberately labeled polit-
ical activists as sectarian extremists. This contributed to perceptions of sectarian threat
among members of minority groups in Syria (Pinto 2017) and facilitated the entry of
armed Islamist groups into the conflict, some of whom were purposefully released from
Syrian jails to bolster the regime’s claim that the opposition was composed of religious
extremists.
What about the role of political elites when it comes to violence in the name of re-
ligion? Lawrence P. Markowitz and Mariya Y. Omelicheva’s chapter offers a unique
perspective by exploring what some have characterized as surprisingly low levels of
violence and terrorism in the Muslim-​majority countries of Central Asia. They argue
that state actors can reduce the incidence of terrorist violence by essentially partnering
with groups that would be most likely to engage in such behavior. Specifically, when the
state’s political and security elites are intertwined with informal and illicit economies
like the drug trade, they can actually decrease incentives and opportunities for violence.
The military as a corporate group is another key actor in social science research on the
Muslim world. This is ironic given the general presumption that religion and religious
mobilization is difficult to suppress in Muslim-​majority countries, and yet there are co-
ercive forces that routinely repress religion across the Muslim world. Nicholas J. Lotito’s
contribution to this volume focuses squarely on the military, demonstrating that it plays
a key role in perpetuating authoritarianism by quashing protests in Muslim societies,
a finding that does not hold in non-​Muslim contexts (Bellin 2004). In probing poten-
tial explanations, he finds no support for the claim that culture or religion shapes the
behavior of the military. Instead, chronic insecurity, foreign assistance, and features of
the organizational cultures of the armed forces may at least proximately explain high
levels of military repression in Muslim countries. Other chapters in this volume also
highlight the role of repression in sustaining authoritarianism—​or at least tempering
democratization—​and countering mass mobilization. For example, Nelson emphasizes
that the military retains special status in Pakistan, limiting the scope for democra-
tization; Jean Lachapelle argues that the repression of Islamists helps to cement au-
thoritarianism in Egypt by garnering support from non-​Islamist publics, who regard
the incumbent authoritarian ruler as the least worst option (see also Lust 2011); and
Pearlman contends that state repression drove the sectarianization of the Syrian war.
The choices, strategies, and ideologies of political parties also figure prominently
in some accounts of politics in Muslim-​majority countries. Much research focuses on
22    Melani Cammett and Pauline Jones

Islamist parties, which have become increasingly prominent actors since at least the
1980s in many Muslim countries and performed well in founding elections in Egypt and
Tunisia after the Arab uprisings. A growing research program explores their behavior in
and out of office and the determinants of individual support for Islamists.
Multiple chapters in this volume assess the mobilizational strategies and appeal of
Islamist and other religious parties. Wilkinson’s chapter analyzes the shifting influence
of religion on politics in India, which is home to the world’s largest Muslim-​minority
population, and Pakistan. In recent decades, religious parties have gained strength in
both countries, albeit for different reasons in these distinct contexts. In India, the BJP
has steadily gained power, culminating in a national electoral victory in 2014. In order
to broaden its electoral appeal, however, the party downplayed its hardline Hindu na-
tionalist agenda—​at least until recently—​and instead played up its message of good gov-
ernance and development. For political elites in Pakistan, where religion was central
to the founding national myth (see Tudor’s chapter in this volume), activating religion
is a strategic choice aimed at cementing winning coalitions and marginalizing rival,
nonreligious parties and politicians.
In the aftermath of the Arab uprisings, Islamists were the major beneficiaries in
some countries, yet they played distinct and often marginal roles in organizing and
participating in the uprisings themselves. In her chapter, Chantal Berman asks why
Islamist mobilization was more visible in protests in some countries than in others, and
whether Islamist beliefs shaped the likelihood that supporters of Islamists would partic-
ipate in the protests. Comparing Islamists in Egypt, Morocco, and Tunisia, she argues
that the decision to urge supporters to protest depended on their experiences of state
repression. In Egypt, where the Muslim Brothers operated in a “middle ground” be-
tween repressed outsider and regime insider, party leaders called upon their supporters
to participate. Conversely, in Tunisia, where Ennahda was totally repressed under Ben
Ali, party leaders could not mobilize their followers. Finally, the Moroccan Party of
Justice and Development (PJD), a member of the loyal opposition, urged its supporters
not to protest. Berman then analyzes how their followers responded, finding that some
followed their leaders’ edicts to either protest or abstain, while others protested without
formal Islamist leadership, galvanized by economic grievances. Thus, strategic rather
than religious factors shape the behavior of Islamist parties, and religion may have
shaped the motivations to protest of only a portion of participants in the Arab uprisings.
In her chapter, Alexandra A. Siegel explicitly argues that religion does not explain the
success of Islamist mobilization in Egypt during the Arab uprisings. Rather, the su-
perior capacity of Islamists to project a coherent narrative, in this case through social
media, enabled them to more successfully advance their goals than other opposition
groups. At the same time, she cautions, the very technologies that facilitated the Muslim
Brotherhood’s mobilization efforts have also undermined it by accelerating organiza-
tional fragmentation, amplifying extremists, and providing the regime a new tool to re-
press the Brotherhood.
Once elected, how do Islamists perform in office? In particular, given their reputation
as conservative actors, how do they represent female constituents? Lindsay J. Benstead’s
 Politics in Muslim Societies     23

chapter analyzes these questions in the case of Morocco, finding that Islamist parties
address the concerns of women more effectively than their “secular” counterparts. Her
data show that women were more likely to contact both female and Islamist deputies
concerning personal and community problems. She argues that an “Islamic mandate
effect,” or a commitment to serve marginalized communities while respecting piety
by using female members to mobilize female voters, helps to explain these findings.
Corroborating existing research (Blaydes 2014; Meyersson 2014), Benstead therefore
finds that Islamists exhibit responsiveness to female constituents, at least in some policy
domains.
Beyond Islamist parties that contest elections, the Arab uprisings and other regional
and global developments led to a renewed focus on ideologically or tactically extremist
Islamist actors, such as Salafis and armed jihadist groups. Focusing on the aftermath of
the 2011 uprisings in Tunisia, Frédéric Volpi’s chapter traces the rise of Salafi networks,
specifically Ansar al-​Sharia, which gained appeal through its “Islamically framed”
messages and activism. Ultimately, however, the organization failed for reasons that had
little to do with religion, notably its inability to rein in internal disagreements and to im-
pose party discipline on its supporters. Even the Islamic State, which some argue deploys
a traditionalist interpretation of Islam to justify its mode of governance and goals, does
not deploy religious repertoires to govern and mobilize. As Megan A. Stewart argues in
this volume, there is nothing particularly Islamic about its governance and warmaking
strategies, which mirror standard approaches implemented by leftist rebel groups
during the Cold War.
Despite the preponderant focus in Western social science research on Islamists and
on their apparent ability to outmaneuver other opposition groups, Adria Lawrence
warns that the Islamist-​secularist “divide” is overplayed. Instead, she contends in
her chapter, religious frames are not the unique purview of Islamists. Ostensibly sec-
ular parties sometimes justify their policies on religious grounds, while Islamists ex-
press support for nonreligious or even secular goals. Similarly, in his chapter on Islamic
parties in Southeast Asia, Pepinsky argues that support for religious parties should not
be interpreted as support for a religious agenda, nor should voting for nonreligious
parties be viewed as opposition to religion in politics.
As we noted at the outset, most social scientists who specialize on Muslim-​majority
countries argue that religion does not explain party-​voter linkages in the Muslim
world—​even when explaining the electoral successes and mobilizational appeal of
Islamist parties (Blaydes 2010; Cammett 2014; Cammett and Jones 2014; Lust-​Okar
2005; Masoud 2014; Pepinsky, Liddle, and Mujani 2012). Similarly, in their analyses of
party strategies in Muslim societies, Corstange and York find that Arab and Muslim
polities are not characterized by more clientelism than others, and political and eco-
nomic factors rather than cultural attributes such as religion are associated with the
use of clientelist outreach. Although ruling parties have an advantage in using this
linkage strategy given their greater access to the resources at the heart of most clientelist
exchanges, other parties can successfully employ constituency service to garner support
as an alternative approach.
24    Melani Cammett and Pauline Jones

Conclusion

In this chapter, we explore how—​if at all—​religion influences social, political, and eco-
nomic outcomes in the Muslim world. When religion is defined narrowly as doctrine—​
that is, the set of beliefs and practices that delimit a particular faith—​we argue that
religion is not the root cause of various social, political, and economic outcomes in the
Muslim world. We offer a broader conceptualization of religion as both the beliefs and
practices that constitute doctrine and the infrastructure that sustains these beliefs and
practices. We argue that doctrine is critical to understanding religion’s potential force,
whereas infrastructure—​specifically, social organization and social identity—​is the key
to understanding why, when, and where religion has agency to enact that force. Based
on this definition of religion, we can offer general insights about how religion influences
a variety of important social, political, and economic outcomes that can be applied be-
yond Muslim-​majority states. First, religion as social organization (i.e., institutions and
actors) can provide a technology of mobilization, particularly under conditions of state
repression. Second, religion as social identity can impact social and political outcomes
through the communal bonds it fosters among religious adherents or members of the
same religion. Finally, religion is best situated to influence social and political outcomes
when religious institutions and actors are empowered.
Western scholarship has often depicted social and political life in the Muslim world
as disproportionately influenced by religion. For at least some scholars based in Western
institutions, however, the picture is more nuanced, especially in more contemporary re-
search. The contributions to this volume are representative of this emerging and impor-
tant scholarship. Collectively, they ask not only whether religion matters, but also how
it matters. The authors represent a diverse set of scholars in terms of their methodo-
logical approaches and geographical specialization. The breadth of their area expertise
enables us to gain insight on the role of religion in Muslim-​majority countries across re-
gions and to move beyond the predominant focus on the Middle East and North Africa.
Because these insights concern some of the most salient outcomes in the social sciences,
they also help to integrate the study of Muslim societies into mainstream comparative
analytical frameworks.
At the same time, we acknowledge that there are some important limitations to a
volume that focuses solely on Western scholarship. A crucial next step is to bring these
perspectives in dialogue with the work of scholars based in Muslim-​majority coun-
tries. Knowledge production from the diverse countries of the Muslim world is vital to
broaden the scope of what is deemed important to study, what factors are most impor-
tant to explaining these outcomes, and how they should be studied. On both empirical
and normative grounds, it is urgent for scholars in Western institutions to learn more
from their colleagues in the places they study. Only though this engagement will we be
able to fully explore the role that religion and other factors play in shaping contempo-
rary social, political, and economic outcomes in Muslim-​majority countries.
 Politics in Muslim Societies     25

Acknowledgments
We are grateful for the input of all contributors, especially Chantal Berman, Steven Brooke,
Ellen Lust, Adeel Malik, Matthew Nelson, and Megan Stewart for their detailed feedback on
this essay. All errors are our own.

Notes
1. Our focus is primarily (though not exclusively) on Muslim-​majority societies, which we
define in demographic terms. We include countries in which the majority of the population
are nominally categorized as “Muslim,” whether Sunni, Shiʿa, or another subgroup therein.
Globally, this amounts to approximately fifty countries with broad geographic scope be-
yond the usual suspects in the Middle East, notably the Balkans, Central Asia, South
and Southeast Asia, and sub-​Saharan Africa. (For a full list of countries, see footnote 1
in Corstange and York’s chapter.) This more inclusive scope of Islamic countries not only
fulfills a demographic definition of Muslim-​majority, but also expands the scope of our
analysis. Some scholars explicitly differentiate between how Islam is practiced and institu-
tionalized in the Middle East versus other regions. Here, we seek to investigate whether the
influence of religion on social and political outcomes is also distinct.
2. By “Western scholarship,” we mean academic research produced in institutions located in
the West, notably Europe and North America. As we emphasize in the conclusion, Western
scholarship should engage more extensively with scholarship generated by researchers in
Muslim-​majority countries.
3. Like Durkheim’s, our conceptualization combines a substantive and a functionalist defini-
tion of religion. It both “informs us of what the sacred is” (substantive) and “what religion
does” (functionalist) (Droogers 2018, 6).
4. As others have argued, the applicability of Durkheim’s definition to modern and non-​
Christian contexts is limited because it emphasizes “unity” based on an “image of one society
with one religion” (Yang 2011, 32).
5. This is consistent with Avital Livny’s (2020, 125) novel conceptualization of group identity in
the Muslim context as “primarily about an attachment to other members of one’s religious
community . . . rather than a sense of individualized attachment to a set of beliefs or duty
to God.”
6. Another genre of long-​term historical argument credits a distinct set of actors—​slave
soldiers—​with enabling the relative lack of checks on absolute authority in the Muslim
world. Rulers’ reliance on a military class that was not conscripted from the local population
allegedly reduced the checks on their authority (Blaydes and Chaney 2013). For discussions
of these types of long-​term historical arguments, see Cammett (2017, 14–​15) and Kuru (2019,
197–​202).
7. Personal communication with Tarek Masoud, Harvard Kennedy School, January 25, 2021.
8. Masoud (2008) makes a similar argument.
26    Melani Cammett and Pauline Jones

References
Abboud, Samer N. 2015. Syria. Cambridge: Polity.
Ahmed, Shahab. 2015. What is Islam?: The Importance of Being Islamic. Princeton, NJ: Princeton
University Press.
Aminzade, Ronald R., and Elizabeth J. Perry. 2001. “The Sacred, Religious, and Secular in
Contentious Politics: Blurring Boundaries.” In Silence and Voice in the Study of Contentious
Politics, edited by Doug McAdam, Elizabeth J. Perry, William H. Sewell, Jr., Sidney Tarrow,
Charles Tilly, Ronald R. Aminzade, and Jack A. Goldstone, 155–​179. Cambridge: Cambridge
University Press.
Anderson, Lisa. 2001. “Arab Democracy: Dismal Prospects.” World Policy Journal 18, no.
3: 53–​60.
Asad, Talal. 2003. Formations of the Secular: Christianity, Islam, and Modernity. Stanford,
CA: Stanford University Press.
Atallah, Sami. “The Gulf Region: Beyond Oil and Wars.” In Democracy in the Arab
world: Explaining the Deficit, edited by Samir Makdisi and Ibrahim Elbadawi, 166–​195.
London: Routledge, 2010, ch. 6.
Aydin, Cemil. 2017. The Idea of the Muslim World: A Global Intellectual History. Cambridge,
MA: Harvard University Press.
Baram, Amatzia. 1997. “Neo-​Tribalism in Iraq: Saddam Hussein’s Tribal Policies, 1991–​96.”
International Journal of Middle East Studies 29, no. 1: 1–​31.
Bazzi, Samuel, Gabriel Koehler-​ Derrick, and Benjamin Marx. 2020. “The Institutional
Foundations of Religious Politics: Evidence from Indonesia.” Quarterly Journal of Economics
135, no. 2: 845–​911.
Beatty Kathleen M., and Oliver Walter. 1984. “Religious Preference and Practice: Reevaluating
Their Impact on Political Tolerance.” Public Opinion Quarterly 48, no. 1: 318–​329.
Bellin, Eva R. 2004. “The Robustness of Authoritarianism in the Middle East: Exceptionalism
in Comparative Perspective.” Comparative Politics 36, no. 2: 139–​157.
Bellin, Eva. 2008. “Faith in Politics: New Trends in the Study of Religion and Politics.” World
Politics 60, no. 2: 315–​47.
Black, B., and L. Patton. 2015. “Introduction.” In Dialogue in Early South Asian Religions: Hindu,
Buddhist, and Jain Traditions, edited by B. Black and L. Patton, 1–​ 24. Abingdon,
UK: Routledge.
Blaydes, Lisa. 2010. Elections and Distributive Politics in Mubarak’s Egypt. New York: Cambridge
University Press.
Blaydes, Lisa. 2014. “How Does Islamist Local Governance Affect the Lives of Women?”
Governance 27, no. 3: 489–​509.
Blaydes, Lisa. 2017. “State Building in the Middle East.” Annual Review of Political Science
20: 487–​504.
Blaydes, Lisa, and Eric Chaney. 2013. “The Feudal Revolution and Europe’s Rise: Political
Divergence of the Christian West and the Muslim World before 1500 CE.” American Political
Science Review 107, no. 1: 16–​34.
Brooke, Steven. 2019. Winning Hearts and Votes: Social Services and the Islamist Political
Advantage. Ithaca, NY: Cornell University Press.
Brubaker, Rogers. 2015. “Religious Dimensions of Political Conflict and Violence.” Sociological
Theory 33, no. 1: 1–​19.
 Politics in Muslim Societies     27

Cammett, Melani. 2014. Compassionate Communalism: Welfare and Sectarianism in Lebanon.

Ithaca, NY: Cornell University Press.
Cammett, Melani. 2017. “Development and Underdevelopment in the Middle East.” In
Handbook of the Politics of Development, edited by Carol Lancaster and Nicolas van de Walle,
1–​36. New York: Oxford University Press.
Cammett, Melani, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17: 187–​206.
Cammett, Melani, and Isabel Kendall. 2021. “Political Science Scholarship on the Middle
East: A View from the Journals.” PS: Political Science & Politics: 1–​8. https://​doi.org/​10.1017/​
S1049096520002061
Chaney, Eric. 2013. “Revolt on the Nile: Economic Shocks, Religion, and Political Power.”
Econometrica 81, no. 5: 2033–​2053.
Clark, Janine A. 2004. Islam, Charity, and Activism: Middle Class Networks and Social Welfare
in Egypt, Jordan, and Yemen. Bloomington, IN: Indiana University Press.
Clingingsmith, David, Asim Khwaja, and Michael Kremer. 2009. “Estimating the Impact of the
Hajj: Religion and Tolerance in Islam’s Global Gathering.” Quarterly Journal of Economics
124: 1133–​1170.
Droogers, André. 2018. “Defining Religion: A Social Science Approach.” In The Oxford
Handbook of the Sociology of Religion, edited by Peter B. Clarke, 263–​279. Oxford: Oxford
University Press.
Eickelman, Dale F., and James Piscatori. 2004. Muslim Politics. Princeton, NJ: Princeton
University Press.
Esmer, Yilmaz and Thorleif Pettersson. 2007. “The Effects of Religion and Religiosity on Voting
Behavior.” In The Oxford Handbook of Political Behavior, edited by Russell J. Dalton and
Hans-​Dieter Klingemann, 187–​193. New York: Oxford University Press.
Esposito, John, and Dalia Mogahed. 2008. Who Speaks for Islam? What a Billion Muslims Really
Think. New York: Gullop Press.
Geertz, Clifford. 1966. “Religion as a Cultural System.” In Anthropological Approaches to the
Study of Religion, edited by Michael Banton, 1–​46. London: Tavistock.
Gerrish, B. A. 1988. “”The Nature of Doctrine.” Journal of Religion 68, no. 1: 87–​92.
Gill, Anthony. 2001. “Religion and Comparative Politics.” Annual Review of Political Science
4: 117–​138.
Grzymala-​Busse, Anna. 2012. “Why Comparative Politics Should Take Religion (More)
Seriously.” Annual Review of Political Science 15: 421–​442.
Hariri, Jacob Gerner. 2015. “A Contribution to the Understanding of Middle Eastern and
Muslim Exceptionalism.” Journal of Politics 77, no. 2: 477–​490.
Harris, Margaret. 1998. “A Special Case of Voluntary Associations? Towards a Theory of
Congregational Organization.” British Journal of Sociology 49, no. 4: 602–​618.
Helmke, Gretchen and Steven Levitsky. 2004. “Informal Institutions and Comparative
Politics: A Research Agenda.” Perspectives on Politics 2, no. 4: 725–​740.
Hoffman, Michael T., and Elizabeth Nugent. 2017. “Communal Religious Practice and
Support for Armed Parties: Evidence from Lebanon.” Journal of Conflict Resolution 61, no.
4: 869–​902.
Huddy, Leonie, Stanley Feldman, and Christopher Weber. 2007. “The Political Consequences
of Perceived Threat and Felt Insecurity.” ANNALS of the American Academy of Political and
Social Science 614: 131–​153.
28    Melani Cammett and Pauline Jones

Huntington, Samuel P. 1996. The Clash of Civilizations and the Remaking of World Order.
New York: Simon & Schuster.
Ismail, Salwa. 2004. “Being Muslim: Islam, Islamism and Identity Politics.” Government and
Opposition 39, no. 4 (Autumn): 614–​631.
Jackson, Richard. 2007. “Constructing Enemies: Islamic Terrorism in Political and Academic
Discourse.” Government and Opposition 42, no. 3: 394–​426.
Jamal, Amaney. 2005. “The Political Participation and Engagement of Muslim Americans: Mosque
Involvement and Group Consciousness.” American Politics Research 33, no. 4: 521–​544.
Kedourie, Elie. 1994. Democracy and Arab Political Culture. London: Frank Cass.
Kennedy, Hugh. 1981. The Early Abbasid Caliphate: A Political History. London: Croom Helm.
Kramer, Martin. 1997. “Ballots and Bullets: Islamists and the Relentless Drive for Power.”
Harvard International Review 19, no. 2: 16–​19.
Kuran, Timur. 2011. The Long Divergence: How Islamic Law Held Back the Middle East.
Princeton, NJ: Princeton University Press.
Kuru, Ahmet T. 2019. Islam, Authoritarianism, and Underdevelopment: A Global and Historical
Comparison. Cambridge: Cambridge University Press.
Kurzman, Charles, and Ijlal Naqvi. 2010. “Do Muslims Vote Islamic?” Journal of Democracy 21,
no. 2: 50–​64.
Layman G. 1997. “Religion and Political Behavior in the United States.” Public Opinion
Quarterly 61, no. 2: 288–​316.
Lewis, Bernard. 1995. Cultures in Conflict: Christians, Muslims, and Jews in the Age of Discovery.
New York: Oxford University Press.
Livny, Avital. 2020. Trust and the Islamic Advantage: Religious-​Based Movements in Turkey and
the Muslim World. Cambridge: Cambridge University Press.
Ludden, David. 2005. Making India Hindu: Religion, Community, and the Politics of Democracy
in India. Oxford: Oxford University Press.
Lust, Ellen. 2011. “Missing the Third Wave: Islam, Institutions, and Democracy in the Middle
East.” Studies in Comparative International Development 46, no. 2: 163–​190.
Lust, Ellen. 2005. Structuring Conflict in the Arab World: Incumbents, Opponents, and
Institutions. New York: Cambridge University Press
Masoud, Tarek. 2008. “Islamic Parties and Democracy: Are They Democrats? Does it Matter?”
Journal of Democracy 19, no. 3: 19–​24.
Masoud, Tarek. 2014. Counting Islam: Religion, Class, and Elections in Egypt.
New York: Cambridge University Press.
McAdam, Doug. 1999. Political Process and the Development of Black Insurgency, 1930-​1970.
Chicago: University of Chicago Press.
Meyersson, Erik. 2014. “Islamic Rule and the Empowerment of the Poor and Pious.”
Econometrica 82, no. 1: 229–​269.
Mitchell, Timothy. 2004. “The Middle East in the Past and Future of Social Science.” The Politics
of Knowledge: Area Studies and the Disciplines, edited by David L. Szanton, 74-​118. Berkeley,
CA: University of California Press.
Norris, Pippa, and Ronald Inglehart. 2003. “Islamic Culture and Democracy: Testing the ‘Clash
of Civilizations’ Thesis.” Comparative Sociology 1, no. 3–​4: 235–​265.
North, Douglass. 1990. Institutions, Institutional Change and Economic Performance.
New York: Cambridge University Press.
Owen, Roger. 1993. The Middle East in the World Economy 1800–​1914. London: I.B. Tauris.
 Politics in Muslim Societies     29

Owen, Roger, and Şevket Pamuk. 1999. A History of Middle East Economies in the Twentieth
Century. Cambridge, MA: Harvard University Press.
Patai, Raphael. 1973. The Arab Mind. New York: Scribner.
Pepinsky, Thomas B., R. William Liddle, and Saiful Mujani. 2012. “Testing Islam’s Political
Advantage: Evidence from Indonesia.” American Journal of Political Science 56, no.
3: 584–​600.
Pepinsky, Thomas B., R. William Liddle, and Saiful Mujani. 2018. Piety and Public
Opinion: Understanding Indonesian Islam. New York: Oxford University Press.
Philpott, Daniel. 2009. “Has the Study of Global Politics Found Religion?” Annual Review of
Political Science 12: 183–​202.
Pinto, Paulo Gabriel Hilu. 2017. “The Shattered Nation: The Sectarianization of the Syrian
Conflict.” In Sectarianization: Mapping the New Politics of the Middle East, edited by Nader
Hashemi and Danny Postel, 123–​142. New York: Oxford University Press.
Putnam, Robert. 2000. Bowling Along: The Collapse and Revival of American Community.
New York: Simon & Schuster.
Rose, Richard. 2002. “How Muslims View Democracy: Evidence from Central Asia.” Journal of
Democracy 13: 102–​111.
Roy, Olivier. 1996. The Failure of Political Islam. Cambridge, MA: Harvard University Press.
Rubin, Jared. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the Middle East Did
Not. New York: Cambridge University Press.
Said, Edward. 1981. Covering Islam: How the Media and Experts Determine How We See the Rest
of the World. New York: Vintage.
Smidt, Corwin E. 2019. “Measuring Religion in Terms of Belonging, Beliefs, and Behavior.” In
Oxford Research Encyclopedia of Politics. Oxford: Oxford University Press.
Stepan, Alfred C. 2000. “Religion, Democracy, and the ‘Twin Tolerations.’” Journal of
Democracy 11, no. 4: 37–​57.
Tajfel, Henri, and John C. Turner. 1979. “An Integrative Theory of Inter-​group Conflict.” In
The Social Psychology of Inter-​group Relations, edited by W. G. Austin and S. Worchel, 33–​47.
Monterey, CA: Brooks/​Cole.
Tarrow, Sidney. 1998. “Framing Contention.” In Power in Movement: Social Movements and
Contentious Politics, 106–​122. Cambridge: Cambridge University Press.
Tessler, Mark, Amaney Jamal, and Michael Robbins. 2012. “New Findings on Arabs and
Democracy.” Journal of Democracy 23, no. 4: 89–​103.
Tessler, Mark, Jodi Nachtwey, and Anne Dressel. 1999. Area Studies and Social Science: Strategies
for Understanding Middle East Politics. Bloomington, IN: Indiana University Press.
Tessler, Mark, and Michael H. Robbins. 2007. “What Leads Some Ordinary Arab Men and
Women to Approve of Terrorist Acts against the United States.” Journal of Conflict Resolution
51: 305–​328.
Thompson, Elizabeth F. 2020. How the West Stole Democracy from the Arabs: The Syrian Arab
Congress of 1920 and the Destruction of Its Historic Liberal-​Islamic Alliance. New York: Grove
Atlantic.
Tibi, Bassam. 2012. Islamism and Islam. New Haven: Yale University Press.
Wald, Kenneth D., and Clyde Wilcox. 2006. “Getting Religion: Has Political Science
Rediscovered the Faith Factor?” American Political Science Review 100: 523–​529.
Waldner, David. 1999. State Building and Late Development. Ithaca, NY: Cornell
University Press.
30    Melani Cammett and Pauline Jones

Wickham, Carrie Wickham. 2002. Mobilizing Islam: Religion, Activism and Political Change in
Egypt. New York: Columbia University Press.
Woltering, Robbert A. F. L. 2002. “The Roots of Islamist Popularity.” Third World Quarterly
23:1133–​1143.
Yang, Fenggang. 2011. “A Definition of Religion for the Social-​Scientific Study of Religion.”
In Religion in China: Survival and Revival under Communist Rule, 25–​43. Oxford: Oxford
University Press.
Yavuz M. H. 1997. “Political Islam and the Welfare (Refah) Party in Turkey.” Comparative
Politics 30: 63–​82.
 Pa rt I

REGIMES AND
R E G I M E C HA N G E
 Chapter 2

Isl am and P ol i t i c a l
Structure in H i stori c a l
Perspec t i v e

Eric Chaney

For centuries, scholars have argued that religion and politics are uniquely linked in
the Islamic world. Lapidus (1996, 3–​4) sums up this scholarship, noting how it claims
that “binding religious norms” dictate that “Islam [encompass] all domains including
law and the state” impeding the “secularization and development” of Islamic societies.
According to this line of thought, the union of religion and politics traces its roots to the
actions of the Prophet Muhammad himself and can explain the region’s historical insti-
tutional path and present-​day democratic deficit.
This chapter combines historical evidence with recent research to propose an alterna-
tive narrative for the institutional evolution of the Islamic world. In particular, it roots
the region’s democratic deficit in a sequence of critical junctures in history and not in
Islamic doctrine. Thus, I hypothesize that the Arab conquests led to a spread of tribal
modes of governance and the rise of military slavery. These institutional changes, in turn,
contributed to the decay of secular bureaucratic structures, the emergence of an influential
class of religious scholars, and the “religionization” of civil society. I conclude by arguing
that the region’s present-​day democratic deficit is rooted in the crystallization of this
­institutional framework more than four centuries after the Prophet Muhammad’s death.
Figure 2.1 illustrates the chapter’s main thesis. Figure 2.1a provides the partial rela-
tionship between democracy and adherence to Islam in 2010, conditional on conti-
nent indicator variables.1 Figure 2.1b shows this same relationship, adding a dummy
for the prevalence of parallel cousin marriage to the continent indicators, which I claim
proxies for the spread of Arabian tribal forms of societal organization.2 As Figure 2.1
makes clear, the addition of the parallel cousin indicator variable completely explains
the “Islam democratic deficit” from a statistical standpoint, suggesting that the spread
of tribal forms of organization may enhance our understanding of the region’s political
difficulties.
34   Eric Chaney

I hypothesize that following the Prophet Muhammad’s death, tribal elites (many of
whom were late converts to Islam) seized power and sought to undermine the positions
of Muslims from outside pre-​Islamic elite circles. To undercut growing attempts by
this disempowered group to constrain their rule, rulers harnessed the fiscal systems of
the conquered non-​Muslim populations to develop the “slave soldier” military system
(Blaydes and Chaney 2013). Under this system, foreigners played an increasingly cen-
tral role in the military, eventually destroying the landed elites on which pre-​Islamic
bureaucracies depended. Tribal networks centered around religious elites replaced these
structures, leading to a collapse in secular state capacity and the religionization of civil
society. The concomitant militarization of society and the resulting surge in the political
power of religious leaders was the ultimate product of the victory of tribal forms of so-
cietal organization over previous modes of governance centered around landed elites.
These institutional developments culminated in the eleventh century CE, leading to in-
stitutional forms that persisted in many areas through the nineteenth century. This insti-
tutional persistence, in turn, explains the democratic deficit documented in Figure 2.1.
The evidence presented in this chapter weighs against the traditional view that the
region’s autocratic institutions are rooted in Islam. Rather, the historical record and em-
pirical evidence seems more consistent with the view espoused by the Enlightenment
thinkers credited with laying the ideological underpinnings of modern democratic
institutions in Europe (Acemoglu and Robinson 2006, 69; Israel 2011). These thinkers,
on the whole, viewed Islam as a “pure monotheism of high moral calibre” and tended
to attribute the negative aspects of some interpretations of Islam to the actions of rent-​
seeking elites in the political and ideological arenas (Israel 2006, 615, 616, 620).
The remainder of this chapter proceeds as follows. The first section provides a brief
overview of the traditional view of the political impact of the intertwining of religious
and political authority throughout Islamic history. The second section provides quali-
tative evidence casting doubt on this view. The third and fourth sections propose an al-
ternative narrative. The fifth section provides preliminary empirical support consistent
with this alternative view and briefly explores the ideological impact of these historical
institutional arrangements. A final section concludes the discussion.

Religious Origins?

Traditionally, Western scholarship highlighted the religious origins of the Islamic state
(e.g., Lewis 1993).3 While there are variations on this theme, its logical thrust rests on two
premises: (1) the Prophet Muhammad combined both religious and political functions,
and (2) this union of religion and politics became fixed in religious norms from an early
date. These two premises are used to argue that Islamic doctrine legitimates institutions
that link politics and religion.4
Religion is believed to have affected the Islamic world’s institutional development
from the start. After the death of Muhammad in 632 CE, the caliphate continued to
 (a) .5 JPN
TWN MNG
IND
ISR CYP
KOR PHL CPV MUS IDN
ZAF LBN KGZ TUR COM
NPL BWA
LSOGEO KEN GHA MYS PAK
ARM ZMB BEN SLE BGD MLI SEN
NAMBDI MWI
LBR AUS GNB
NZL
THA LKA
BTN COD MOZ CHL
URYCRICAN
USA TTO GIN IRQ
SLB
JAM
NIC
PAN
PER NGA
KHM HUN
IRL
IT ASWE
LTU
FIN
NOR
ESP
SVK
PRT
POL GRC
DNK
SVN
GBRAUT
DEU
CHENLD
GAB GTM
BRA
DOM
SLV
PRY ARG
MEX NER
Residual Normalized Polity Score

EST
HRV
ROU FRA
HND
BOLCOL BGR MNE MKD DJIDZA ALB
0

ZWE LVA BEL

CZE
MDA
ETH SRB GUY
MDG SGP CIVPNG
ECU BFA YEM SOM
CAF UGA TZA UKR JOR
TJK
AGO TGO TCD SDN MRT
RUS EGY
MMR COG
RWA CMR KAZ TUN
LAO
VNMCHN GNQ HTI SYR KWT AZE
IRN GMB
ARE BHR OMN MAR
PRK ERI TKM UZB LBY
FJI
−.5

VEN QAT SAU

SWZ

CUB
BLR
−1

−.5 0 .5 1
Residual Proportion Muslim
[N = 160; b = −0.21; se = 0.07]

(b) LBN
.5

IRQ
CPV JPNTWN MNG MUS ISR CYP DZA
ZAF IND COM
LSO
BWA KOR KEN PHL GHA IDN
ZMB BEN NZL
AUS SLE KGZ MLI SEN TUR
NAMBDI NPL
SDN LBR
MWIGEO URY
USA
CHLCRICAN TTO
MRT YEMGNB MYS PAK
COD ARM MOZ TJK SLB GIN BGD
Residual Normalized Polity Score

NIC
PER
PANEGY
JAM JOR
THA LKA PRT
NOR
POL
SVK
LTU
IRL
ESP
IT ASWE
HUN
FIN GBR
SVN
GRC
DNK CHENLD
DEU
AUT
PRY
GTM
DOM
SLV
BRAMEX
ARG NGA
BTN GAB ROU
EST
HRV FRA TUN
BGR MNE MKD NER ALB
HND
BOLCOL
0

KHM MDA
LVA BEL
CZE SRB GUY DJI
ZWE PNG ETH MAR
ECU SYR KWT
LBY CIV IRN
MDG ARE UKR OMN BFA SOM
UGA BHR
CAF TZA TKM RUSUZB
AGO TGO SGP
QAT SAU TCD
RWA
COG CMR HTI
GNQ GMB
MMR KAZ
VNM
LAOCHN VEN FJIERI AZE
−.5

SWZ PRK
CUB
BLR
−1

−.5 0 .5 1
Residual Proportion Muslim
[N = 160; b = 0.03; se = 0.08]

Figure 2.1: Islam, democracy, and tribal structures.

Note: Both graphs use ISO country codes to denote countries. For a link between country codes and country names, see https://​www.iban.com/​
country-​codes.
36   Eric Chaney

combine religious and political functions. However, as the Islamic state expanded, po-
litical institutions and Islamic ideals increasingly diverged.5 This led religious circles to
view the caliphate as fundamentally illegitimate, forcing the caliphs to import slaves to
staff their armies from the ninth century CE onward. Pipes (1981) directly links military
slavery to this legitimacy crisis, arguing that the introduction of slave soldiers resulted
from the failure of Islamic polities to successfully integrate religion and politics.
This hypothesis has the virtue of neatly explaining many features of the Islamic
world’s institutional history. Military slavery characterized much of the Islamic world
for roughly a millennium and has no equivalent outside the region. Moreover, it seems
to concisely account for the eventual collapse of the institutional framework of Islam’s
“golden age.”
Gibb (1955a, 1955b) provides one overview of the classical interpretation of this pro-
cess. He begins by explaining how Muslim rulers rapidly developed political structures
built on Roman and Persian foundations. These institutions were “peculiarly inorganic,”
due in part to Islam’s inability to legitimize secular state structures. The ensuing legiti-
macy crisis led to Muslim indifference and withdrawal from the political arena, and to
the concomitant militarization of government power through the use of slave armies.
This militarization gradually undermined early Islamic state structures, resulting in the
emergence of a more legitimate—​but less intellectually dynamic—​institutional frame-
work in the eleventh century CE (Gibb 1955b, 124–​128).
Much of this account has withstood the test of time. For example, scholars continue to
believe that the institutional changes of the eleventh century CE ushered in a novel and
remarkably enduring institutional framework that characterized much of the Islamic
world until the nineteenth century (e.g., Lambton 1991, 53).6

The Traditional Narrative, Tribes, and

Military Slavery

Despite possessing many strengths, this narrative suffers from some difficulties. First,
the fact that military slavery was absent from much of the Islamic world (e.g., much of
Southeast Asia; see Crone 1980, 80) casts doubt on the idea that this institution is rooted
in Islamic doctrine. In addition, Lapidus (1975, 1996) notes that while religion and state
were united in some Islamic societies, those in which state and religion were clearly
differentiated represented the historical norm (Lapidus 1996, 24).
On the theoretical side, although Muhammad indisputably combined both religious
and political functions, the enduring institutional consequences of this union are less
clear. The Qurʾan precluded another prophet, and Muhammad did not name a successor
prior to his death. Muhammad’s death consequently necessitated institutional change
 Islam and Political Structure in Historical Perspective    37

via his replacement, but denied unequivocal divine sanction to any successor, since no
one could claim his prophetic status doctrinally.
Muslim elites sought to solve the resulting power vacuum through the creation
of the caliphate, which as an institution emerged from political struggles following
Muhammad’s death (e.g., Madelung 1997). The tribe of Quraysh, which had dominated
much of the Arabian Peninsula prior to the advent of Islam, ended up controlling the
caliphate. This resulted in the exclusion (and subsequent alienation) of many of the most
devout and prominent Muslims from other tribes (e.g., the Ansar) from positions of
influence. The subsequent struggle for power eventually led to the emergence of the
Umayyad Caliphate in 661 CE, in which the clan of Umayya from within the tribe of
Quraysh converted the caliphate into a hereditary dynasty.
The following century was dominated by a search for reliable nontribal sources of mil-
itary power (Crone 1980).7 The Umayyads first relied on Syrian tribal forces, but quickly
transitioned to professional armies staffed by both Arabs and non-​Arabs. These profes-
sional armies suffered from factional rivalries little better than the tribal ones, which
led to the downfall of the Umayyads and the rise of the Abbasids. The Abbasids faced
similar difficulties, which they attempted to resolve through the use of military slavery.
Blaydes and Chaney (2013, 2016) build on this literature, arguing that the introduction
of military slavery was a critical juncture in which rulers sought to lessen constraints on
their rule by establishing loyal armies that were not susceptible to local power struggles.
This view is consistent with Ibn Khaldun (1989, 141), who argued that the introduction
of slave soldiers enabled the ruler to claim “royal authority all for himself ” in a tribal
milieu.
How did tribal social structures lead to military slavery? I hypothesize that by
facilitating collective action in a tribal setting, Islam endowed tribes with greater
political power (see, e.g., Chaney 2013).8 These tribal units used their political influ-
ence to resist a reversion to pre-​Islamic institutional forms, yet were not a suitable
base for autocratic state structures, since tribal societies tend to be “egalitarian, con-
sensus based, and fractious” (Fukuyama 2011, 303). Military slavery provided one solu-
tion to this tension, by allowing rulers to maintain a cohesive military without relying
on domestic social structures that tribal elites deemed unjust. In other words, in the
absence of military slavery, the tribalization of society combined with Islam would
have led to the disintegration of the military, as tribal units undermined the hierar-
chical structures on which pre-​Islamic military structures depended. Military slavery
prevented this from happening and represented a “brilliant adaptation designed to
create a strong state-​level institution against the backdrop of one of the most power-
fully tribal societies on earth” (Fukuyama 2011, 303).9 This institutional innovation, in
turn, allowed the Arabs to become “the only tribal conquerors to have caused the cul-
tural traditions of highly civilized peoples to be reshaped around their tribal heritage”
(Crone 1994, 459).
38   Eric Chaney

Toward an Alternative Narrative

Ibn Khaldun (1989) famously argued that the first millennium of Islamic political his-
tory could not be disentangled from tribal forms of societal organization. In this and
the following section, I explain how these social structures both impacted institutional
developments and spread across areas that were conquered by Arab forces.
Following the conquests by Arab-​ Islamic tribes, society was divided as the
conquerors segregated themselves from the conquered. In an attempt to consoli-
date power, caliphs began to disempower conquering tribal elites by relying on the
pre-​conquest institutional forms that existed in the conquered areas. Tribal elites
responded by rallying around religious ideals. This led to the emergence of two
competing forms of social organization. The first—​henceforth tribal—​descended
from the tribal forms of organization brought by the conquering Arab-​Muslims,
whereas the second—​henceforth landed—​traced its roots to the more autocratic
traditions of the pre-​Islamic empires of the newly conquered lands.10 Although the
caliphs emerged from the tribal conquerors, they adopted the more centralized gov-
ernment of the conquered, seeing that as a more secure (and autocratic) form of power
for themselves. Other tribal leaders resented this and used existing tribal and religious
traditions to constrain caliphal power.
The tribal society used parallel cousin (bint ʿamm) marriage to maintain and ex-
pand clanic structures, emphasizing the “personal liberty and dignity which was ex-
pected by the Arabian tribesmen,” and the idea that “all men were on the same level
before God.”11 The landed society was more “aristocratic,” being based to a large degree
on “agrarian traditions such as those which had been kept alive from Sasanian times
among the landed gentry of the Iranian highlands” (Hodgson 1974a, 253, 280, 281). These
two movements developed competing interpretations of Islam. The tribal group largely
adhered to traditionalist Islam and argued that following the death of Muhammad, re-
ligious authority passed to those who remembered his teachings.12 Thus, while polit-
ical power might reside with the caliph, religious authority was dispersed among the
scholars who “owed their authority entirely to their learning.” As the caliphate became
increasingly autocratic, this group refused to “recognize the legality of the [resulting] re-
gime or to cooperate with it” (Athamina 1992, 153–​154).13 Largely excluded from political
power, they began the process of developing what would eventually become Islamic law.
Some scholars have referred to this group as the “constitutionalists,” in the sense that
they argued that political power was subject to the constraints imposed by this group’s
interpretation of Islamic law.14
The landed group “tended to cooperate with the state” and adhered to ration-
alist interpretations of Islam (Lapidus 2014, 130).15 Adherents of these rationalist
interpretations benefitted from the expansion of secular bureaucratic institutions under
the Abbasids. They were often highly educated, “trained to think on the basis of human
rather than revealed information,” and “disinclined to defer to religious scholars” (Crone
 Islam and Political Structure in Historical Perspective    39

2006, 23–​24). These individuals sought to retain pre-​Islamic social structures, while also
absorbing certain Arabic and Islamic elements (Gibb 1982, 66).
This institutional and ideological divide came to a head during the Mihna (833–​849),
when caliphs attempted to force religious scholars to adopt rationalist interpretations of
Islam. The subsequent ideological confrontation has been viewed as an attempt to sub-
sume the religious scholars and their institutional framework into the growing bureau-
cratic machinery of the Abbasid Caliphate (Watt 1963, 44). The traditionalist religious
scholars—​buttressed by popular support—​proved stronger in what has been described
as the “victory of the constitutionalist[s]‌” (Watt 1963, 52). This victory thwarted the
growing autocratic trajectory of the caliphate, allowing Islamic law to continue its de-
velopment outside state structures, and perpetuating the tension between tribal and
landed forms of organization.
This victory demonstrated the growing strength of the tribal type of societal organi-
zation, which, given its characteristics, was an ill-​suited military base for caliphal autoc-
racy (Lapidus 1982, 718). Yet the strength of this organization also precluded a transition
to pre-​conquest social and institutional structures. As noted above, I hypothesize that
the resulting tension explains the ultimate emergence and rapid spread of military
slavery. In other words, caliphs could not co-​opt this tribal/​traditional element, but they
also could not rely on it to support autocratic institutional arrangements, and so they
turned to military slavery.

The Origins of a New

Political Equilibrium

Institutionalized military slavery has been called a “major innovation in Middle Eastern
history,” introduced in response to a “pressing military and political need for loyal
troops” (Lapidus 2014, 86). While caliphs initially sided with landed elites and the ex-
isting forms of pre-​conquest government, this need ultimately resulted in the destruc-
tion of pre-​conquest bureaucratic institutions and the emergence of a new political
framework. For centuries, however, the stalemate between landed forms of societal or-
ganization (on which secular bureaucratic structures rested) and more egalitarian tribal
arrangements helped power development.
Traditionalist religious leaders—​ as representatives of tribal forms of social
organization—​crafted Islamic law to “neutralize caliphal power in favour of retaining
power diffused in society at large” (Hodgson 1974b, 119). These constraints are believed
to have undermined institutional elitism, encouraged “equality among the Muslim
faithful,” and resulted in “a high degree of social mobility” (Hodgson 1974a, 303–​304).
This meritocratic environment was arguably pivotal in the cultural, technological,
and scientific advances collectively known as the Islamic Golden Age.16 However,
the secular bureaucracy that rested on landed elites also played an important role in
40   Eric Chaney

these developments by constraining the populist and anti-​science tendencies of the

traditionalists, providing an environment in which Hellenistic, philosophical, and
secular views flourished.17 In other words, these two competing forms of societal or-
ganization checked each other, creating a supportive environment for intellectual
discovery.
Military slavery eroded this balance of power by undercutting the landed elites on
which secular bureaucratic institutions rested. Rulers assigned military slaves the right
to collects taxes in lieu of salary (these administrative grants were known as iqtas), and
these soldiers were rotated from iqta to iqta at the will of the ruler (Lapidus 2014, 198–​
199). Given their short time horizons, military slaves had “no permanent interest in the
land and were concerned in the main with squeezing in the shortest possible time as
much as they could out of the land in their temporary possession” (Lambton 1991, 53).
Predictably, this predatory behavior decreased revenue flows to the bureaucracy. In ad-
dition to undermining bureaucratic structures through this channel, military slaves also
destroyed the power of landed elites more broadly (Lapidus 2014, 199). The historian
Miskawayh (who died in 1030 CE) provides one contemporary description of this pro-
cess, noting how after the introduction of slave soldiers, “the bureaucracy was shut down
and the bureaucratic secretaries and tax collectors disappeared [ . . . instead the land was
managed by] slave soldiers and agents who did not take care of what was in their power
or manage the land in a fruitful manner” (Miskawayh 1915, 97–​98).
Traditionalist religious elites benefitted from these developments, filling the void
created by weakening bureaucratic structures. As the bureaucracy increasingly
struggled to provide basic public goods and forms of civic organization, the religious
scholars “took charge of judicial administration, local police, irrigation, public works
and taxation” (Lapidus 1996, 13). These developments encouraged the spread of tribal
forms of social organization and Islamization.18 The ultimate result was that tradition-
alist religious leaders took charge of civil society as “old landowning and officeholding
notables lost their power to new military regimes dominated by slave soldiers” (Lapidus
2014, 224). Since military slaves did not have a vested interest in local prosperity or sta-
bility, they did not fill the void of the bureaucracies they replaced, and they failed to con-
strain and balance the tribal traditionalists the way the landed bureaucrats had.
These societal changes culminated in what is known as the “Sunni revival.” This term,
which goes back at least to Gibb (1955b), reflects the historical emphasis on the ideolog-
ical shift away from rationalist Islam favored by secular bureaucrats and toward more
traditionalist interpretations. Gibb (1955b, 126) emphasized the top-​down aspects of the
shift, arguing that the Seljuks supported institutions such as madrasas to engineer the
spread of what he called “orthodox” Islam, as well as to train a new “orthodox bureauc-
racy” to replace its secular precedent. Scholars employed in these institutions provided
the intellectual underpinning for this new orthodoxy.
In a seminal article, Makdisi (1973, 168) questioned this view, suggesting that the
Sunni revival was instead a bottom-​up process in which traditionalist Islam triumphed
over rationalist alternatives. This alternative view is consistent with historical evi-
dence that traditionalist Islam had gained societal influence prior to the advent of the
 Islam and Political Structure in Historical Perspective    41

Seljuks in the middle of the eleventh century CE. In addition, it sees Seljuk politicians
as maximizing agents more interested in co-​opting influential religious leaders than in
championing one particular ideology over another (Makdisi 1981, 237).
The preceding discussion helps explain the Seljuks’ actions during the revival.
When the Seljuks conquered much of the Islamic world in the eleventh century, they
encountered an environment in which tribal forms of societal organization centered
around traditionalist religious elites had become dominant.19 Consequently, they
devised a new set of formal institutions to reflect this reality. In other words, while Seljuk
policy generated a discrete change in formal institutions, these changes should ulti-
mately be viewed as the reflection of the final victory of traditionalist-​tribal forms of
societal organization over rationalist-​landed alternatives, a victory that had taken place
before the Seljuks arrived on the scene.

Islamic Law, Traditionalism, and the Institutions of the

Sunni Revival
An influential literature has highlighted the long-​run impact of Islamic law on a host
of political and economic outcomes (e.g., Kuran 2011). The previous discussion, how-
ever, suggests that this legal system may have been endogenous to the political system
in which it was embedded. Indeed, in recent decades, scholars have increasingly linked
the institutionalization of Islamic legal structures to the broader changes that occurred
during the Sunni revival. For example, Ephrat (2000, 1) notes how during the Sunni re-
vival, “the four Sunni schools of legal interpretation (madhahib) were consolidated as
scholarly establishments, the nuclei of the Sufi fraternities were formed, and the ‘law
college’ (madrasa) and Sufi hostel (khanqah or ribat) were founded, based on substan-
tial pious endowments (awqaf).” This new institutional framework—​which marked the
emergence of a “pattern of governance that was to be emulated and reinforced until the
nineteenth century”—​was characterized, in part, by the dissociation of legal and state
structures (Hallaq 2009, 77, 146).
Such evidence raises the possibility that Islamic law was ultimately developed to or-
ganize a tribal society outside of traditional bureaucratic structures. Indeed, what even-
tually became traditionalist Islamic law began to develop among Arab tribes in garrison
cities and drew heavily on pre-​Islamic tribal law (Hallaq 2009, 42–​43). Non-​Muslim
populations and many recent converts continued to be subject to more bureaucratic
pre-​Islamic legal structures of the conquered areas (Glick 1995, 53). In other words, legal
institutions appear to have paralleled societal structures as there arose a divide between
those living under “tribal” law and “landed” law.
The historical record suggests that tribal forms of organization may have encouraged
traditionalist interpretations of Islam. Initially, tribal legal structures differed mark-
edly by geographic location, as the Arab Muslim populations drew on different tribal
precedents. Traditional Islam worked to counter these centrifugal tendencies through
the construction of religious authority (Hallaq 2009, 49, 58) by drawing heavily on
42   Eric Chaney

“traditions” that claimed to be the words or deeds of the Prophet Muhammad as handed
down by a line of trustworthy persons (Schacht 1993, 34). These traditions once linked
to Muhammad were used to buttress religious authority, stifle dissent, and induce con-
formity against an otherwise diffuse and fractious tribal backdrop. In other words, in a
tribal society resistant to authority, religious authority became a central component in
ordering tribal communities. Rationalist Islam made it more difficult to organize tribal
units by undermining such authority, suggesting one reason why Islamic law developed
in a traditionalist direction in tribal communities.
As was the case with the broader political framework, the transition to tribal law was
not immediate: for centuries, secular bureaucratic elites competed with traditionalist
religious leaders “for recognition of their capacity to establish rules for effective govern-
ment” (Bligh-​Abramski 1992, 42).20 The Sunni revival marked the final victory of tra-
ditionalist Islamic law over competing interpretations. Thus, the Seljuks crafted a new
institutional arrangement under which rulers relied on traditionalist religious leaders
to administer the tribally organized populations that now constituted their domains
(Hallaq 2009, 150). In this new equilibrium, “the ruling elite received the coopera-
tion of the scholars and their promotion of its legitimacy, while the scholars received
a salary, protection, and the full right to apply the law as they saw fit” (Hallaq 2009,
152). Military leaders henceforth used religious endowments organized under the law
of waqf—​particularly madrasas—​to purchase the allegiance of many of the most prom-
inent religious leaders, who performed many of the functions previously carried out by
the bureaucracy. In so doing, these new elites encouraged the spread of traditionalist
legal structures, which represented a “watershed in the history of Muslim institutions”
(Makdisi, 1981, 237).
In sum, the evidence suggests that the victory of traditionalist interpretations of
Islamic law was rooted in the attempt to organize a tribal society outside bureaucratic
structures. This evidence weighs against the traditional view that the dissociation of the
religious legal system from state structures was the deterministic consequence of the
original union of religion and politics in Islamic doctrine. Instead, the preceding dis-
cussion has traced this and other institutional developments to the evolution of the bal-
ance of political power in a monotheistic tribal milieu. In particular, I have hypothesized
that the emergence of religious authority and the ultimate victory of traditionalist
interpretations of Islam were closely related to the victory of tribal forms of societal or-
ganization over landed alternatives.

Diffusion, Persistence, and

Ideological Impact

The Sunni revival’s institutional framework spread rapidly across the Islamic world
at the time, becoming “the norm almost everywhere” in the areas of initial Muslim
 Islam and Political Structure in Historical Perspective    43

conquest (Hodgson 1974b, 57). As the Islamic world expanded, Muslim dynasties at
times attempted to export this institutional framework to conquered regions (Lapidus
2014, 331). Yet more recently incorporated regions often enjoyed significant autonomy
and did not experience many of the societal changes that occurred following the Arab
conquests. This may be due, in part, to political changes during the early modern pe-
riod stemming from technological change (such as the spread of the use of gunpowder)
and/​or other causes.21 Whatever the reasons, an institutional divide arose in the Islamic
world between areas that had been incorporated early into the Islamic world and those
incorporated later.
As a consequence, the institutions of the Sunni revival did not flourish outside of
tribal areas and regions that were conquered in the centuries following the original
Arab conquests. Thus, in areas such as Malaysia and Indonesia, pre-​Islamic societal
structures largely survived conversion to Islam (Geertz 1968, 9). Rulers did not use mil-
itary slaves (Crone 1980, 263), Islamic law seems to have been confined largely to family
matters in much of this area (Hadi 2004, 238), and local populations did not adopt par-
allel cousin marriage (Korotayev 2000). These observations are consistent with the view
that the revival’s institutions were rooted in tribal social structures rather than in Islamic
doctrine.
Subsequent dynasties found tribal areas difficult to govern due to the “weakness of
the state apparatus,” and continued to rely on military slaves to help administer these
regions (Lapidus 2014, 367, 374). When European colonial powers arrived in the nine-
teenth and twentieth centuries, they found weak civil societies and exploited this sit-
uation to their advantage. Consequently, when these areas gained independence after
World War II, the new states were weakly institutionalized, with little history of civil so-
ciety or checks and balances to draw on. Chaney (2012) explains how this environment
contributed to the emergence of autocratic institutions, providing empirical evidence in
support of this narrative.
Table 2.1 replicates the analysis in this paper, while adding a variable for the incidence
of parallel cousin marriage to the analysis to further investigate the historical narrative
developed in the previous sections.22
Column 1 confirms the conventional wisdom that across the world larger Muslim
populations are associated with lower levels of democracy. Column 2 shows that
this relationship persists even after including an indicator for whether a country was
fuel-​endowed or not. In column 3, I include a dummy variable equal to 1 if parallel-​
cousin marriage is preferred, as coded in Shulz (2019, 50). When I do this, the coeffi-
cient on Muslim majority drops in absolute value and is only statistically significant
at the 10 percent level. The negative link between Islam and democracy completely
disappears when continent dummies are included, as shown in Table 2.1, column 4. In
column 5, I demonstrate that this coefficient and that on a variable measuring the ex-
tent to which a present-​day country was conquered by Arabs are almost identical. In
other words, today the only areas in the Muslim world that demonstrate a democracy
deficit are those that were conquered by Arabs and those where parallel cousin mar-
riage is preferred.
44   Eric Chaney

Table 2.1: Arab Conquest, Tribal Structures, and the Democracy Deficit in 2010
Dependent Variable:

Normalized Polity Score 2010 Cousin

(1) (2) (3) (4) (5) (6) (7)

Muslim Majority -​0.31*** -​0.26*** -​0.11*** 0.02 0.05 0.39*** 0.01

(0.05) (0.06) (0.06) (0.06) (0.07) (0.08) (0.06)
Fuel-​Endowed -​0.25*** -​0.08 -​0.11 -​0.15* 0.45*** 0.18**
(0.09) (0.09) (0.09) (0.09) (0.10) (0.08)
Parallel Cousin -​0.37*** -​0.34***
(0.08) (0.08)
Arab Conquest -​0.32*** 0.76***
(0.09) (0.11)
N 160 160 160 160 160 160 160
Continent Dummies? No No No Yes Yes Yes Yes
R2 0.19 0.23 0.31 0.44 0.43 0.56 0.77

Notes: robust standard errors in parentheses. ***, **, and * indicate significance at the 1 percent,
5 percent, and 10 percent levels, respectively. See text for details.

Columns 6 and 7 are consistent with the claim that these two variables measure the
same treatment (up to measurement error). Column 6 shows that, in general, Muslim
countries are more likely to prefer cousin marriage. Column 7, however, shows that
this effect is entirely driven by countries that were conquered by Arab armies. In other
words, the results are consistent with the hypothesis developed in this chapter that
Arab Conquest → Tribalization → Weak Civil Society → Autocracy today. More broadly,
these results cast doubt on the hypothesis that something inherent in Islamic doctrine
impedes democratization.

Ideological Impact
Did the tribalization of society have an ideological impact? The narrative proposed in
this chapter suggests that it should have led to the spread of traditionalist interpretations
of Islam. Analysis in Makdisi suggests that this indeed occurred and that traditionalism
“in the history of Muslim religious thought has been minimized, and its importance
overlooked” (Makdisi 1963, 37).23
While the place of traditionalism within Islamic religious history remains a topic of
ongoing research, many have argued that the ascendancy of traditionalist views was
detrimental for the region’s intellectual development. For example, scholars have often
highlighted the negative effects of the doctrine of occasionalism, which denied the
 Islam and Political Structure in Historical Perspective    45

necessary relationship between causes and effects (Huff 2003, 110, 113). Some have gone
so far as to attribute the decline of scientific activity in the Islamic world to these ide-
ological developments, arguing that if not for this ideological shift, “the Arabs might
have been a nation of Galileos, Keplers, and Newtons” (Sachau 1879, x). Of course, such
ideological explanations face the difficulty that, as previously noted, traditionalism de-
veloped centuries prior to the Sunni revival, when the Islamic world was indeed a leader
in scientific production.
In a recent working paper, Chaney (2016) provides empirical evidence linking the
shift in the region’s intellectual development to the institutional changes described in this
paper. In particular, he shows that the decline in scientific output coincides with a rise in
the proportion of authors affiliated with madrasas. Madrasas, in turn, are believed to have
been controlled by traditionalist religious leaders who used them as a “weapon” to under-
mine rationalism (Makdisi 1984, 21). Such evidence suggests that the ideological changes
that occurred during the revival were a product of the societal changes documented in
the chapter, rather than a cause of them. Consistent with this view, the following centuries
were characterized by relative stability in the dominant interpretations of Islam, which
contrasted with the pre-​revival period’s “wide range of movements and intellectual
currents featuring a varied mixture of reason and revelation” (Hallaq 2009, 502).

Conclusion

This chapter has provided an overview of the Islamic world’s institutional develop-
ment from the death of the Prophet Muhammad to the present. It has provided evi-
dence weighing against the traditional view attributing the region’s political structures
to Islamic doctrine. Rather, the historical evidence raises the possibility that influential
interpretations of Islamic law and doctrine are rooted in the tenacious persistence of
tribal forms of social organization following the Arab conquests. I have hypothesized
that these forms of societal organization encouraged the introduction of military
slavery, leading to a drop in secular state capacity and the tribalization of civil society.
This institutional legacy, not Islam, lies at the root of much of the region’s ongoing
difficulties today.
Somewhat paradoxically, given the anti-​tribal bent of much of Muhammad’s mes-
sage, the historical evidence suggests that Islamic law developed to help organize tribal
populations outside of traditional bureaucratic state structures.24 I have hypothesized
that this environment encouraged interpretations of Islam emphasizing conformity and
obedience and discouraging free thought in order to maintain cohesion across tribal
units and maintain order for military regimes. In this sense, the analysis is consistent
with claims by Muslim intellectuals that some influential interpretations of Islam—​
rooted in part in the teaching of the religious elites of the revival and after—​can best be
characterized as distortions of Muhammad’s message generated by rent-​seeking elites
rather than as the essence of Islamic doctrine.25
46   Eric Chaney

The analysis in this paper shows that the relationship between Islam and political
structure is not static, and raises the possibility that the dominant interpretation of Islam
in a given time and place is related to the political equilibrium. In particular, the results
suggest that rationalist interpretations of Islam are more likely to flourish in strongly in-
stitutionalized societies.

Acknowledgments
I thank Christian Sahner, Muhammad Saleh, Jesus Fernández-​Villaverde, and the editors for
helpful comments and conversations. All remaining errors are mine.

Notes
1. Data used to generate this graph are drawn from Chaney (2012), except for the dummy for
parallel cousin marriage, which is taken from Shulz (2019, 50, Figure A.5) (I slightly modify
his data by setting modern-​day Israel to zero). See Chaney (2012) for evidence that these
patterns have remained roughly constant over the past decades. For more information on
the data see this source and the analysis below.
2. A parallel cousin is descended from a parent’s same-​sex sibling. The hypothesis that this
form of marriage essentially proxies for the spread of Arabian social structures goes back at
least to Korotayev (2000).
3. Given the many variations of state structures in the Islamic world, the definition of the
“Islamic state” is by necessity arbitrary. For the purposes of this paper, I will use this term
to refer to the political structures that developed in the core areas (i.e., from the Nile to the
Oxus Rivers) of the Islamic world. This political evolution, which crystallized in the middle
of the eleventh century, led to the development of the “most common state type of Islamic
society” (Lapidus 2014, 492).
4. The concept of legitimacy is vague. Given its widespread use in the secondary literature,
however, I use it when referring to this literature. Lipset (1959, 86) defines legitimacy as
“the capacity of a political system to engender and maintain the belief that existing political
institutions are the most appropriate or proper ones for society.” In other words, legitimacy
is related to individual perceptions of optimality. As a concept, it only is relevant to the ex-
tent that people’s perception of what is legitimate deviates from what is optimal to believe
in a given environment. In a religious context, legitimacy is usually tied to the pronounce-
ment of religious elites (Rubin 2017, 32). If this is correct, legitimacy (for the purposes of this
paper) is closely related to the ability of religious leaders to convince the population that
suboptimal institutions (from the point of view of the population) are optimal. In other
words, the concept of religious legitimacy is tightly linked to rent-​extraction by religious
elites through their control over beliefs.
5. Between 632 and 750 CE, Arab-​Muslim armies conquered a vast geographic region reaching
from modern-​day France to Pakistan.
6. This framework was characterized by, among other things, the spread of the use of religious
endowments such as waqfs and a decline in formal governmental structures. See the section
titled “The Origins of a New Political Equilibrium” below for additional details.
 Islam and Political Structure in Historical Perspective    47

7. The existing literature tends to stress the instability of tribal military forces, which are
“egalitarian by nature [and] resistant to the establishment of political structures outside
of the tribe” (Guichard 1976, 517). Similarly, Athamina (1998, 374) argues that the loyalty
of Arab tribesmen to rulers was “doubtful at best because their true loyalty lay solely with
their tribes.” See Moscona, Nunn, and Robinson (2019) for empirical evidence linking
tribal environments and conflict.
8. This idea goes back at least to Ibn Khaldun, who argued that “Arab peoples in particular
need religion to achieve the levels of [cooperation] essential to conquest and royal au-
thority because their natural savagery ordinarily inhibits cooperation. Religion enables
them to restrain themselves and cooperate in a common cause. It supplements family
loyalties to create wider, more encompassing solidarities” (quoted in Lapidus 1990,
18). Similarly, Lapidus (1990, 31) argues that “Islam could be used for the unification of
fragmented populations.” This view is also consistent with a large literature arguing for
the importance of ideology in solving the collective action problem (e.g., Acemoglu and
Robinson 2006, 126).
9. This naturally raises the question of why the Arabian Peninsula generated this type of so-
ciety to begin with. Crone (1994, 449) suggests that it may be related to the physical envi-
ronment of the Arabian Peninsula and the historical reliance on camel pastoralism.
10. The terms tribal and landed are admittedly imprecise. Throughout this chapter, I follow
Crone (1994, 447) and define tribe as a “a descent group [ . . . ] which constitutes a polit-
ical community.” Landed is best identified with the Persian aristocracy, which is “generally
believed to have existed in pre-​Islamic Persia” (Lambton 1967, 41), although similar ar-
rangement likely existed in other regions prior to the Arab conquest). This societal divide
resulted in the emergence of “two institutional structures with different elites and a dif-
ferent ethos” (Lapidus 1996, 11). While there may be differences in how these two groups
are defined and debates regarding the extent to which these societies were feudal (Zakeri
1995, 13), the claim that what I call landed pre-​Islamic arrangements were not tribal is
uncontroversial.
11. Thus, this tribal society grew, incorporating new converts to Islam from non-​Arabian
backgrounds through the institution of wala (see Bernards and Nawas [2005] for an
overview of this institution). Given present-​day anthropological evidence discussed
throughout this paper, it seems likely that these new arrivals adopted some Arabian tribal
practices, and that the “full-​fledged Islamic civilization” that seems to have emerged by the
tenth century CE (Nawas 2006, 75, 88) contained many of these tribal characteristics.
12. Makdisi (1963, 22) concisely differentiates between the two groups as follows: “The
traditionalists made use of reason in order to understand what they considered as the le-
gitimate sources of theology: scripture and tradition. What they could not understand
they left it as it stood in the sources; they did not make use of reason to interpret the
sources metaphorically. On the other hand, the rationalists advocated the use of reason on
scripture and tradition; and all that they deemed to contradict the dictates of reason they
interpreted metaphorically in order to bring it into harmony with reason.”
13. For additional evidence linking religious elites with tribal elites excluded from power, see
Bligh-​Abramski (1992, 54).
14. This divide emerged very early, and by 700 CE it was clear that a “wedge existed between
the ruling elite and the emerging religio-​legal class” (Hallaq 2009, 128).
15. Crone (1980, 64) argues that “from the start” the Abbasids supported rationalists
who “as religious disputants and propagandists played an integral part in the Abbasid
48   Eric Chaney

establishment” and “soon inherited the bureaucracy.” Crone (2006, 24) argues that the
preference of bureaucratic officials for such rationalism continued long after the Mihna.
See Zakeri (1995, 11) for evidence that rationalist Islam was linked to the landed aristocracy.
16. In addition, it helps explain why the pre-​conquest bureaucracies did not generate similar
results.
17. See Chaney (2019) for evidence that the bureaucracy played a central role in encouraging
scientific development.
18. Islamization and tribal forms of organization did not necessarily go hand in hand
(Guichard 1976, 223), but the ultimate collapse of landed state structures encouraged
tribalization (in some ways in a similar manner to the process through which peasants
adopted a lord to protect their property and lives during the feudalization of Europe). In
a recent article, Rapoport and Shahar (2012, 29) suggest that this tribalization of society
in regions conquered by the Arabs may have been widespread. This hypothesis is also
consistent with modern-​day evidence that the form of cousin marriage practiced in the
Arabian Peninsula is primarily practiced today in areas that were conquered by Arab tribal
forces (e.g., Korotayev 2000), and with the analysis below.
19. For evidence that the demise of the landed elites preceded or at the very least coincided
with the Sunni revival, see Bulliet (1972, 22, 25).
20. Lambton (1967, 42) notes the link between the landed aristocracy and the bureaucracy by
noting that after the Arab conquest the landed elites “retained their importance and con-
tinued to act as a link between the government and the people. Their functions, however,
were bureaucratic.”
21. For example, from the late sixteenth through eighteenth centuries, the breakdown of mil-
itary slavery combined with a host of additional factors to generate “profound changes in
the Ottoman political system” (Lapidus 2014, 366).
22. As Chaney (2012) notes, the econometric framework is kept simple, running a regression of
the form Democracyic = β1 Muslimic + β2 ArabConquesticic + γ′x + εic where Democracyic
is the 2010 polity score in country i on continent c normalized to lie on the interval [0,1]
(higher scores more democratic), Muslimic is an indicator equal to 1 if at least half the
country’s population is Muslim, ArabConquesticic is a measure of the extent to which a
country was conquered by Arab forces, and x is a vector of covariates, including continent
dummies, and an indicator equal to 1 if a country is fuel-​endowed. See Chaney (2012) for
further details on these variables and their construction.
23. Although the idea that tribal forms of societal organization encourage more conservative
interpretations of Islam awaits systematic empirical investigation, there is some quali-
tative evidence supporting this view. For example, the cash waqf—​which was viewed by
conservative Muslims as violating Islam’s usury ban, but nevertheless used in much of
the Ottoman Empire—​was adopted very late if at all in areas conquered by Arab forces
(Mandaville 1979, 308).
24. Ibn Khaldun (1989) recognized that the Prophet Muhammad taught against the very trib-
alism that he argued was the motor of Islamic political history. Thus, he notes that “we find
that Muhammad censured group feeling [i.e., tribal asabiyya] and urged us to reject it and
leave it alone. He said: ‘God removed from you the arrogance of pre-​Islamic times and its
pride in ancestry. You are the children of Adam, and Adam was made of dust.’”
25. For example, al-​Afghani criticized post-​revival Islam’s “passivity and resignation,” which
he viewed as a corruption of true Islam, which was “a religion of reason and the free use of
 Islam and Political Structure in Historical Perspective    49

the mind” (Lapidus 2014, 519). For a more recent take on this theme, see https://​www.inde-
pendent.co.uk/​voices/​commentators/​islamic-​history-​is-​full-​of-​free-​thinkers-​but-​recent-​
attempts-​to-​suppress-​critical-​thought-​are-​9993777.html.

References
Acemoglu, Daron, and James Alan Robinson. 2006. Economic Origins of Dictatorship and
Democracy. Cambridge: Cambridge University Press.
Athamina, Khalil. 1992. “The Ulama in the Opposition: the Stick and Carrot Policy in Early
Islam.” Islamic Quarterly 36, no. 3: 153–​178.
Athamina, Khalil. 1998. “Non-​Arab Regiments and Private Militias during the Umayyad
Period.” Arabica 45, no. 4: 347–​378.
Bernards, Monique, and John Nawas. 2005. Patronate and Patronage in Early and Classical
Islam. Leiden: Brill.
Blaydes, Lisa, and Eric Chaney. 2013. “The Feudal Revolution and Europe’s Rise: Political
Divergence of the Christian West and the Muslim World before 1500 CE.” American Political
Science Review 107, no. 1: 16–​34.
Blaydes, Lisa, and Eric Chaney. 2016. “The Political Economy Legacy of Institutions from the
Classical Period of Islam.” In The New Palgrave Dictionary of Economics, edited by Steven
Durlauf and Lawrence Blume. London: Palgrave Macmillan.
Bligh-​Abramski, Irit. 1992. “The Judiciary (Qadis) as a Governmental-​Administrative Tool in
Early Islam.” Journal of the Economic and Social History of the Orient 35, no. 1: 40–​7 1.
Bulliet, Richard. 1972. The Patricians of Nishapur: A Study in Medieval Islamic Social History.
Cambridge, MA: Harvard University Press.
Chaney, Eric. 2012. “Democratic Change in the Arab World, Past and Present.” Brookings
Papers on Economic Activity 42, no. 1: 363–​414.
Chaney, Eric. 2013 “Revolt on the Nile: Economic Shocks, Religion and Political Power.”
Econometrica 81, no. 5: 2033–​2053.
Chaney, Eric. 2016. “Religion and the Rise and Fall of Islamic Science.” Working paper.
Cambridge, MA: Harvard University.
Chaney, Eric. 2019. “Religion, Political Power and Human Capital Formation.” In Advances in
the Economics of Religion, edited by Jean-​Paul Carvalho, Sriya Iyer, and Jared Rubin, 437–​
448. Cham, Switzerland: Palgrave.
Crone, Patricia. 1980. Slaves on Horses. Cambridge: Cambridge University Press.
Crone, Patricia. 1994. “The Tribe and the State.” In The State: Critical Concepts, edited by John.
A. Hall, 446–​476. London: Routledge.
Crone, Patricia. 2006. “Post-​Colonialism in Tenth-​Century Islam.” Der Islam 83, no. 1: 2–​38.
Ephrat, Daphna. 2000. A Learned Society in a Period of Transition: The Sunni ʿUlamaʾ of
Eleventh-​Century Baghdad. Albany: State University of New York Press.
Fukuyama, Francis. 2011. The Origins of Political Order. London: Profile Books.
Geertz, Clifford. 1968. Islam Observed. Chicago: University of Chicago Press.
Gibb, Hamilton Alexander Rosskeen. 1955a. “An Interpretation of Islamic History I.” The
Muslim World 45, no. 1: 4–​15.
Gibb, Hamilton Alexander Rosskeen. 1955b. “An Interpretation of Islamic History II.” The
Muslim World 45, no. 2: 121–​133.
50   Eric Chaney

Gibb, Hamilton Alexander Rosskeen. 1982. “The Social Significance of the Shuubiya.” In
Studies on the Civilization of Islam, edited by Stanford J. Shaw and William R. Polk, 62–​73.
Princeton, NJ: Princeton University Press.
Glick, Thomas. 1995. From Muslim Fortress to Christian Castle. Manchester, UK: Manchester
University Press.
Guichard, Pierre. 1976. Al-​Andalus: Estructura antropológica de una sociedad islámica en
Occidente. Barcelona: Barral Editores.
Hadi, Amirul. 2004. Islam and State in Sumatra. Leiden: Brill.
Hallaq, Wael. 2009. Shari’a. Cambridge: Cambridge University Press.
Hodgson, Marshall. 1974a. The Venture of Islam. Vol. 1. Chicago: University of Chicago Press.
Hodgson, Marshall. 1974b. The Venture of Islam. Vol. 2. Chicago: University of Chicago Press.
Huff, Toby. 2003. The Rise of Early Modern Science: Islam, China and the West.
Cambridge: Cambridge University Press.
Ibn Khaldun. 1989. The Muqaddimah. Princeton, NJ: Princeton University Press.
Israel, Jonathan. 2006. Enlightenment Contested. New York: Oxford University Press.
Israel, Jonathan. 2011. Democratic Enlightenment: Philosophy, Revolution, and Human Rights
1750–​1790. New York: Oxford University Press.
Korotayev, Andrey. 2000. “Parallel-​Cousin (FBD) Marriage, Islamization, and Arabization.”
Ethnology 39, no. 4: 395–​407.
Kuran, Timur. 2011. The Long Divergence: How Islamic Law Held Back the Middle East.
Princeton, NJ: Princeton University Press.
Lambton, Ann. 1967. “The Evolution of the Iqta in Medieval Iran.” Iran 5: 41–​50.
Lambton, Ann. 1991. Landlord and Peasant in Persia. London: I.B. Tauris.
Lapidus, Ira. 1975. “The Separation of State and Religion in the Development of Early Islamic
Society.” International Journal of Middle East Studies 6, no. 4: 363–​385.
Lapidus, Ira. 1982. “Review of Slave Soldiers and Islam: Genesis of the Military System by
Daniel Pipes.” Journal of Interdisciplinary History 12, no. 4: 716–​7 18.
Lapidus, Ira. 1990. “Tribes and State Formation in Islamic History.” In Tribes and State
Formation in the Middle East, edited by Philip S. Khoury and Joseph Kostiner, 25–​47.
Berkeley: University of California Press.
Lapidus, Ira. 1996. “State and Religion in Islamic Societies.” Past & Present 151: 3–​27.
Lapidus, Ira. 2014. A History of Islamic Societies. New York: Cambridge University Press.
Lewis, Bernard. 1993. Islam and the West. New York: Oxford University Press.
Lipset, Seymour. 1959. “Some Social Requisites of Democracy: Economic Development and
Political Legitimacy.” American Political Science Review 53, no. 1: 69–​105.
Madelung, Wilferd. 1997. The Succession to Muhammad. Cambridge: Cambridge
University Press.
Makdisi, George. 1963. “Ash’ari and the Ash’arites in Islamic Religious History II.” Studia
Islamica 18: 19–​39.
Makdisi, George. 1973. “The Sunni Revival.” In Islamic Civilization, 950–​1150, edited by D.
H. Richards, 155–​168. Oxford: Bruno Cassirer.
Makdisi, George. 1981. The Rise of Colleges. Edinburgh: Edinburgh University Press.
Makdisi, George. 1984. “The Juridical Theology of Shafiʿi: Origins and Significance of Usul al-​
Fiqh.” Studia Islamica 59: 5–​47.
Mandaville, Jon. 1979. “Usurious Piety: The Cash Waqf Controversy in the Ottoman Empire.”
International Journal of Middle East Studies 10, no. 3: 289–​308.
Miskawayh. 1915. Tajārib al-​Umam. Cairo: al-​Tamadun al-​Sina’iya.
 Islam and Political Structure in Historical Perspective    51

Moscona, Jacob, Nathan Nunn, and James Alan Robinson. 2019. “Kinship and
Conflict: Evidence from Segmentary Lineage Societies in Sub-​Saharan Africa.” NBER
Working Paper no. 24209. Cambridge, MA: National Bureau of Economic Research.
Nawas, John. 2006. “The Birth of an Elite: Mawali and Arab Ulama.” Jerusalem Studies in Arabic
and Islam 31: 74–​91.
Pipes, Daniel. 1981. Slave Soldiers and Islam: The Genesis of a Military System. New Haven,
CT: Yale University Press.
Rapoport, Yossef, and Ido Shahar. 2012. “Irrigation in the Medieval Islamic Fayyum: Local
Control in a Large-​Scale Hydraulic System.” Journal of the Economic and Social History of the
Orient 55, no. 1: 1–​31.
Rubin, Jared. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the Middle East Did
Not. Cambridge: Cambridge University Press.
Sachau, Edward. 1879. The Chronology on Ancient Nations: An English Version of the Arabic Text
of the Atharul-​Bakiya of Albiruni. London: W. H. Allen.
Schacht, Joseph. 1993. An Introduction to Islamic Law. Oxford: Clarendon Press.
Shulz, Jonathan. 2019. “Kin Networks and Institutional Development.” Working paper.
Rochester, NY: SSRN.
Watt, William Montgomery. 1963. “The Political Attitudes of the Mu‘tazilah.” Journal of the
Royal Asiatic Society of Great Britain and Ireland 1, no. 2: 38–​57.
Zakeri, Mohsen. 1995. Sasanid Soldiers in Early Muslim Society. Wiesbaden: Harrasowitz
Verlag.
 Chapter 3

Stat e-​F orm ati on, Stat i st

Isl am , and Re g i me
Instabi l i t y
Evidence from Turkey

Kristin E. Fabbe

Religion, and particularly the forces of political Islam and state secularism, have been
central to discussions of regime stability in the Turkish case.1 After its founding as a
single-​party and eventually self-​declaredly secular state under the leadership of Mustafa
Kemal Atatürk in the 1920s, the country effectively transitioned to a multiparty system
in 1950. Yet intense polarization, political instability, and military interventions have
propelled the Turkish Republic into crisis about once every decade since. Successful
military coups in 1960, 1971, and 1980, the “postmodern coup” of 1997, the 2007 mili-
tary “E-​memorandum,” and most recently the failed coup of 2016 have all interrupted
the process of democratic deepening. In light of Turkey’s most recent bouts of political
upheaval, scholars have struggled to definitively characterize Turkey’s regime “type,”
alternatively labeling it “tutelary democracy” (Höjelid 2009), “elusive democracy”
(Tezcür 2010), “secular democracy” (Secor 2011), “delegative democracy” (Taş 2015),
“illiberal democracy” (Göl 2017), “competitive authoritarianism” (Çalışkan 2018; Esen
and Gumuscu 2016), “neo-​fascism” (Tuğal 2016), “spindle autocracy”(White 2017),
“fragile or weak authoritarianism”(Akkoyunlu and Öktem 2016), and finally, building
on Guillermo O’Donnel’s concept, a “brown country” that functionally and territorially
mixes key democratic and authoritarian characteristics (Kurban 2020).
These difficulties in classifying Turkey’s regime reflect the broader absence of regime
stability in Turkey’s modern history. That is to say, they reflect a lack of strong demo-
cratic or authoritarian consolidation. Scholars have attributed this instability, in part, to
the cycling of one of the country’s master narratives of conflict, which allegedly pits the
forces of political Islam against Kemalist secularism. Guardians of the Kemalist tradi-
tion have been depicted as “working from inside the state system” to limit expressions of
54   Kristin E. Fabbe

religion in the public sphere and repress pious elements in society via assertive forms of
state secularism (laïcité—​lâiklik) (Kuru 2009). Political Islamists, by contrast, gained at-
tention for “working from outside the system” to mobilize religious identity and contest
the state’s repression of religious freedom via local politics and the grass roots (White
2002). More recently, after nearly two decades of rule by the Justice and Development
Party (AKP), which was first democratically elected in 2002, it is argued that the tables
have been turned. Now in control of the system, political Islamists are said to be using it
to flout the very democratic institutions that brought them to power, to desecularize the
state, and to impose Islamic values (Cornell 2017).
To make sense of these upheavals, and to shed light on some of the reasons why both
democracy and authoritarianism have “failed to stick” in the Turkish case, this chapter
advocates for a more long-​term historical assessment of the relationship between reli-
gion and regime. The chapter makes two interlocking arguments. First, using evidence
from the late Ottoman Empire and early Republican Turkey, I argue that processes of
state formation shaped the subsequent trajectory of Islamist politics in modern Turkey,
which came to be dominated by statist or state-​centric political Islamist currents. Power
arrangements fortified during state formation imbedded elements of the dominant re-
ligion (Sunni Islam) within the state and subsequently structured “the deep rules of the
game” in Turkish politics. As such, far from being a reactionary force working primarily
from outside of the state system, political Islamists in Turkey learned to operate largely
from within state organs and with statist prerogatives. They continue to do so.
Second, and relatedly, although Turkey’s political Islamists have indeed used grass-​
roots strategies to inspire and mobilize the masses (Delibas 2015; White 2002), legacies
of late Ottoman and Kemalist state-​building have contributed to another set of strategies
at the elite level that are less visible and still poorly understood. State-​centric Islamists
in Turkey have wielded their moral authority to homogenize and nationalize society,
as well as to build and reorient the state in their own image. They have used coalition-​
building, coalition-​balancing, and patronage politics to expand their foothold in state
institutions. They have consistently and steadily gained influence through a patient
strategy of temporary bargains with the anti-​democratic forces of Kemalist secularism
against mutual enemies (leftists, minority groups, etc.). Finally, state-​centric Islamists
have aspired for institutional capture rather than protracted power sharing—​much like
their Kemalist counterparts. The emergence of state-​centric political Islam has thus
contributed to regime instability because many big political battles are fought within the
critical institutional corridors of the state—​security, judiciary, education, etc.—​and are
thereby destabilizing to it, whether in democratic or autocratic form.
Before turning toward empirics and further elaborating the arguments presented
thus far, a few definitional notes are in order. First, in its approach to religion, the his-
torical analysis in this chapter focuses on how the emerging Turkish state engaged three
discrete aspects of the dominant majority religion, Sunni Islam: religious elites, religious
institutions, and religious attachments. Disaggregating religion in this way shows appre-
ciation for the fact that religious authority—​which can equally be a potential challenge
or supplement to state authority—​is embodied in the form of individual power brokers
 State-Formation, Statist Islam, and Regime Instability    55

(elites), their organizational structures (institutions), and norms regarding identity and
belonging (attachments). Such an approach also restores agency to Sunni Muslim reli-
gious elites in the history of state-​building, secularization, and modernization, which is
key to understanding how subsequent power relations between the emergent machine
of the centralized state and the traditional loci of religious authority evolved.2
My use of the term “political Islam” corresponds to the Turkish terms İslâmcılık
(Islamism) or siyasal İslâmcılık (political Islamism), which are used to describe the
agenda and activities of individuals and groups who claim to defend the religious sphere
from various forms of encroachment and to protect the interests of the pious and reli-
gious elites (Fabbe and Balikçioglu 2019). “Secularization” is a more difficult term to pin
down. As I have noted elsewhere, I harbor deep skepticism about the usefulness of sec-
ularism and secularization as face-​values concepts, while also admitting that the terms
are hard to avoid (Fabbe 2019). Here I contend that state formation is a kind of seculari-
zation, insomuch as it involves the expansion of state-​centric norms of sovereignty and
institutional hegemony over specific “disciplinary” domains (in the Foucauldian sense)
of human interaction. Hence, when I write of secularization or secularizing strategies,
I refer to their Jacobin orientation; that is, to a belief in the possibility of transforming
society through totalistic political action designed to reorient loyalties toward the state
by controlling traditional influences and generating modern/​civic identities (Eisenstadt
1999; Fabbe 2019).
Finally, the labels of “Kemalist” and “Islamist” are not without their problems,
paradoxes, or ambiguities—​but both terms are used widely and self-​referentially in
Turkey (White 2002), and thus they are also retained here.

State Formation and the Historical

Precursors of Statist Islam

In the nineteenth and early twentieth centuries, the emerging Turkish state made du-
rable advances in sovereignty and hegemony by deploying the symbolic resources of
Sunni Muslim identity, by leveraging religious institutions, and by imbedding indi-
vidual religious elites into nascent, state-​centric structures of education, law, and bu-
reaucracy as part of the state-​building process. Specifically, institutional changes in
education and law that were set into motion during the reign of Sultan ʿAbdulḥamīd II
(r. 1876–​1909), and were continued subsequently under the rule of the Committee for
Union and Progress (CUP) and in the early Kemalist period, demonstrate how modern
Turkish state formation led religious elites to develop state-​centric prerogatives and
orientations. Using evidence from the periods of Hamidian, CUP, and Kemalist rule,
I show how a lack of manpower for modernizing reforms led the state to expand into ed-
ucation and law by creating new institutional layers on the edge of traditional religious
structures. Institutional layering was further undergirded by the incremental, piecemeal
56   Kristin E. Fabbe

co-​optation of a sizable number of religious elites, a process that realigned their interests
and rendered them largely state-​centric. As a result of this state formation process, the
traditional religious institutions of Sunni Islam gradually eroded in the late Ottoman
period, and the interests of Islamists increasing assumed state-​centric orientations.

The Late Ottoman Legal and

Educational Landscape under Sultan
ʿAbdulḥamīd II

Sultan ʿAbdulḥamīd II, who ascended to the throne in 1876 at the age of thirty-​four, came
to power in the midst of a gathering political storm. Foreign (British, French, Russian)
interventions had whittled away Ottoman influence. The internationally imposed au-
tonomous provinces of the empire—​Serbia, Romania, Bulgaria, Crete, Samos, Mt.
Lebanon, Egypt—​began to slip further from Ottoman control, largely as a result of
European meddling (Findley 2010). By the 1870s and 1880s, Ottoman political elites and
the empire’s Muslim population became increasingly unwilling to accept any kind of
European interference in Ottoman affairs, with the French conquest of Tunisia (1881)
and the British occupation of Egypt (1882) particularly offensive to them. Furthermore,
for the empire’s Christian elite, the idea of Ottoman citizenship was beginning to lose its
appeal, making way for the politics of self-​determination (Rodogno 2012).
ʿAbdulḥamīd II thus inherited a host of political and fiscal problems, while also
under pressure to centralize, modernize, and consolidate his power base. Despite these
impediments, ʿAbdulḥamīd II ruled as an absolute monarch for some thirty years and
managed to make great strides toward developing modern infrastructure, even amid
the impending chaos. In large part this was because, from approximately the mid-​1870s
onward, he supported modernizing reforms in the fields of education and law based on
a strategy of institutional layering, which continuously blurred the distinction between
“state” and “religious” institutional structures and helped consolidate state authority.
For example, in the field of education, the term ibtidā‘ī mektebi, which is often used
interchangeably with the terms “modern,” “secular,” or “state” schools in the secondary
literature, was actually also the standard term for traditional Qurʾan schools in many
official government documents during this period. The “usage of the term ibtidā‘ī
mektebi both for Qurʾan schools and for government schools in official documents,”
argues Akşin Somel (2001, 109), is “related to the ‘historical relict’ of considering pri-
mary education as an integral part of religious instruction.” Work by Benjamin Fortna
also illustrates that during the reign of ʿAbdulḥamīd II the character and content of
late Ottoman “state” education became more Islamic than it had been immediately
after the first wave of Tanzimat reforms. Fortna writes that “Western subjects were de-​
emphasized and a considerable amount of classically Islamic content such as Quran,
 State-Formation, Statist Islam, and Regime Instability    57

ḥadīth and jurisprudence were introduced into an otherwise largely secular lesson plan”
(Fortna 2011, 81).
This blurring of lines between religious and state institutional structures in the field
of education was born from the imperatives of state-​building under constraints—​
especially severe shortages in civilian-​trained (in this case, nonreligiously trained)
teachers. This can be illustrated by figures compiled by the Ottoman statistical office for
the year 1897–​1898 (1313). The total number of schools is listed as 29,081, consisting of
55 “high schools” (mekātib-​i ʿidādiyye), 412 “middle schools” (mekātib-​i rüşdiyye), and
28,614 “primary/​elementary schools” (mekātib-​i ibtidāiyye). However, the same records
do not indicate there being anywhere near an adequate amount of nonreligiously trained
teachers to fill these schools—​the total number of state-​employed teaching staff and
other personnel was listed as only 3,613! Furthermore, the number of students enrolled
at the empire’s 14 state-​created training programs for elementary school teachers (dāru’l-​
muʿallimīn) was only 277 that same year, meaning that a new crop of available teachers
was not on the horizon (Güran 1997, 106–​109). Elites at the helm of the Ottoman state
were aware of this manpower predicament, as evidenced by an 1888 document issued by
the Ottoman Commission for Education, which stated that, “regarding the appointment
of individuals capable of teaching in the new method, because of serious concerns about
empty teaching slots and the absence of teachers, recourse has to be taken by continuing
[to appoint] those with traditional training in existing [religious] methods” (Kodman
1980, 240).
In the legal realm, too, centralizing reforms had not and did not ostracize the reli-
gious establishment. The Provincial Law of 1864 marked the beginning of a separation
of administrative and judicial powers and paved the way for the consolidation of the
Niẓāmiye3 court system as a set of institutions that was theoretically distinct from the
shariʿa courts. At this stage, however, the entire judicial system (including commercial
and criminal elements) was still subordinate to the Şeyḫülislām, and thus the religious
establishment (Rubin 2011, 21–​29). What is more, the beginning of the new Niẓāmiye
system did not immediately herald the end of the traditional shariʿa courts, the ḳāḍī
system, and/​or religious legal order. For one, Aḥmed Cevdet Paşa, a shariʿa jurist, who
can by no means be labeled a “secularist,” spearheaded the Niẓāmiye system. Between
1870 and 1877, Cevdet Paşa led an effort to codify the shariʿa based on the corpus juris
of the Hanafite school. These developments culminated in the creation of the Mecelle
(Mecelle-​yi aḥkām-​ı ʿadliye), a civil code that was a hybrid legal artifact with European
structure but distinct Islamic features and content (Hallaq 2009, 409–​411). The Mecelle
was applied in both traditional shariʿa courts and in the new Niẓāmiye system, showing
the close relationship between the two.
Furthermore, the creation of the Niẓāmiye courts did not occur at the expense of re-
ligious elites’ career advancement. Shariʿa judges instead just assumed a new duty as the
chief judges of the new Niẓāmiye courts, now under the new title of nāʾib. The tradi-
tional kadıship thus evolved into something akin to “fellowship” for medrese students.
Research into the personnel records of late Ottoman nāʾibs reveals that many entered
58   Kristin E. Fabbe

the religious hierarchy of kadıship in their student years and then were promoted to
the position of nāʾib in the new Niẓāmiye system, marking a high degree of fluidity be-
tween the traditional shariʿa system and the new state-​centric Niẓāmiye courts, as well
as new pathways for upward mobility (Akiba 2005). Similarly, analysis of the judicial
personnel in the Niẓāmiye courts shows that initially many hailed from the medreses,
the new Niẓāmiye law school (which itself drew from the medreses), and the School for
Shariʿa Judges (Akiba 2018; Rubin 2011), established in 1855.

The Rise of the Committee for Union

and Progress (CUP) and its Approach
to the Religious Establishment

Sultan ʿAbdulḥamīd II’s approach to religion had, in many ways, bolstered his authority
and centralization. Nonetheless, iron-​fist tendencies and a penchant for the brutal sup-
pression of potential opponents earned him a number of enemies. As discontent with
the seemingly endless military campaigns and economic crises brewed among the lower
ranks of the bureaucracy and military, organized opposition took shape (Karal 1962;
Pekdemir 2008). Namely, the Young Turk movement (the intellectual movement from
which the Committee for Union and Progress [CUP] eventually emerged) appeared on
the scene in the late nineteenth century.
The movement began as a group of exiles, intellectuals, military officers, and civil
servants connected by virtue of their opposition to ʿAbdulḥamīd II’s regime. The
organization’s self-​proclaimed goal was a return to the political structures of the First
Constitutional Period—​ namely the reinstatement of the constitutional monarchy
and revival of the Ottoman Parliament, which had been enacted by the Basic Law/​
Constitution (Ḳānūn-​ı Esāsī) of 1876. This First Constitutional Period was short-​lived,
lasting only two years before being abolished by ʿAbdulḥamīd II, who then reinstated
an absolute monarchy with himself at the helm in 1878. The CUP won a major victory in
July 1908, when the group’s “Young Turk Revolution” succeeded in forcing ʿAbdulḥamīd
II to acquiesce to their demands that the 1876 constitution and Ottoman Parliament
be reinstated. Although the event was a major victory for the Young Turks, it did not
provide the CUP with a free hand to systematically disenfranchise religious influence.
Nonetheless, as Şükrü Hanioğlu has observed, there was also growing sentiment within
CUP circles that the social position of the ulema and its role in state affairs should be
substantially curtailed (Hanioğlu 1995, 2005) in the interests of state centralization and
modernization.
The religious elite were more than a mere nuisance for the reform-​minded CUP.
As available data reveal, by virtue of size and position, the ulema was both a potential
breeding ground for collective resistance to the forces of further state centralization and
 State-Formation, Statist Islam, and Regime Instability    59

a reservoir of expertise and manpower that had to be taken very seriously. Approximately
3,278 upper-​level ulema members (at the rank of ḳāḍī, müftü, and nāʾib) were, at least in
theory, operating across the empire (Albayrak 1996, vol. 1). To this figure needs to be
added the number of medrese instructors, ṭalebeler/​softā (students), imamlar (prayer
leaders), müezzinler (call to prayer reciters), and ḥāfıẓlar (Qurʾan reciters). Precise
numbers are elusive, but some basic estimates are possible: in 1892 Ottoman authorities
estimated the number of medrese students in the capital alone to be near to 12,000, a siz-
able number indeed (Bein 2006, 289), while another estimate put the combined number
of imamlar, müezzinler, and ḥāfıẓlar empire-​wide at 188,000 (Karpat 1985, 218). Finally,
detailed qualitative analysis of the career trajectories of 140 randomly selected ulema
members working during the final decades of the empire reveals that over two-​thirds
of them transferred between (or simultaneously worked in) “traditional” religious
institutions and new “state-​centric” institutions (Fabbe 2019). Put simply, the Sunni
Muslim religious establishment was sizable and embedded.
As relations between the CUP and the Sunni Muslim religious establishment
unfolded, two points became increasingly clear. First, the primary preferences of these
religious elites lay not in some fundamental, religiously dictated distaste for moderniza-
tion or even centralization, but in a more basic desire to preserve their own position in
society, their institutions, and their livelihood. This meant that the Sunni Muslim reli-
gious establishment was largely divided regarding its attitudes toward the CUP (Ardıc̦
2012; Bein 2011). Second, the CUP recognized the importance of religious attachments
and moral authority. The organization extensively used religious language and rhet-
oric to broaden its mass appeal and to placate men of faith as well as the pious. In CUP
publications, evidence to this effect abounds, and CUP leaders often claimed to be better
Muslims than the Sultan-​Caliph himself (Emil 1979; Gündüz 2008; Hanioğlu 2001;
Mardin 1964).
While the CUP struggled to formalize its political influence with ʿAbdulḥamīd II still
on the throne, opposition to the CUP began to rear its head. Political factions opposed
to a CUP takeover consolidated. On September 14, 1908, the Liberal Union (ʿOsmānlı
Aḥrār Fırḳası) was founded, a liberal group popular with minorities and in favor of
decentralization—​as opposed to the CUP’s centralization policies. This group was
headed by Prince Ṣabāḥaddīn, nephew and fierce opponent of ʿAbdulḥamīd II (Kuran
1948, 229).4 Furthermore, the İttiḥād-​ı Muḥammedī Fırḳası (Society of Muhammadan/​
Moslem Union), a religiously conservative Islamic faction demanding the full imple-
mentation of the shariʿa, also appeared on April 5, 1909 (Tunaya 1984, 1, 182–​200). Less
than a week prior, this faction’s leader had led a large uprising in the capital, demanding
a full implementation of the shariʿa while the Liberal Union stood by in tacit support
(Albayrak 1987; Atilhan 1956; İrtem 2003; Kutay 1977). Despite the ostensibly Islamic
aims of the uprising, it seems that, at its core, the real subject of dispute was whether
state-​centric expansion should proceed or not. Otherwise, it is difficult to explain the
strange merging of political bedfellows that led the Liberal Union—​a vociferous oppo-
nent of Sultan ʿAbdulḥamīd II that was backed by Christian and Jewish minorities—​to
60   Kristin E. Fabbe

support a group of pious, Sunni Muslim, anti-​CUP demonstrators protesting for the full
implementation of the shariʿa.
More interestingly, for our purposes here, despite the emergence of an “Islamic party”
and the existence of a “religious uprising” at this juncture in 1909, the Sunni Muslim re-
ligious establishment itself did not fully coalesce behind either; rather, it fractured fur-
ther still. And, while exceptions certainly did exist, the lines of these fractures began
to take a clearer form. Upper-​level ulema members with a secure position in the reli-
gious establishment’s hierarchy often tended to view themselves as better off defending
the CUP’s state centralization efforts and the constitution (Ṭanīn 1908; İ. Kara 1994,
2005; Kuran 1948; Tunaya 2003; Unat 1960). They wanted to maintain their social status
and position, which was still possible in the contemporary political climate as a result
of the incentives being offered by the CUP leadership in return for compliance. Low-​
level religious functionaries and students of religion, on the other hand, had much
less to gain by siding with the CUP and tended to support the protestors (Farhi 1971).
Religious students and low-​level functionaries saw the very institutions that were sup-
posed to afford them upward mobility being neglected, as upper-​level ulema joined new
state-​centric institutions of education and law, not to mention entered the political fray.
Lower-​level members of the Sunni Muslim religious establishment had no place in the
new centralized state system from the outset, and they thus attempted to resist as the
CUP tightened its grip on power.
The CUP’s response to the uprising was as swift as it was brutal, replete with arrests,
public hangings, and widespread general repression. It also, however, included def-
erential gestures toward those members of the Sunni Muslim religious establish-
ment that had supported it. First, the CUP institutionalized an ulema branch of its
organization in order to demonstrate the seriousness of its commitment to religion
(Hanioğlu 2001, 306). Second, it maneuvered to acquire a fatwa providing an official
religious justification for ʿAbdulḥamīd II’s dismissal, thus sealing his fate (Hür 2008).
Finally, even though the 1876 constitution had been reinstated, the reforms of 1909
made the document substantially more “Islamic” in both tone and content than the
1876 original. An alteration to Article 10, for example, added a reference to the shariʿa
in connection with the lawful reasons for arrest, and a change in Article 118 made
Ḥanafī jurisprudence (fiqh) a major source for new legislation (Hanioğlu 2008, 160).
The constitution thus earned the praise of many influential Sunni Muslim religious
elites (Berkes 1964, 368–​370).
What is shocking is that certain members of the Sunni Muslim religious establish-
ment continued to back and legitimate the CUP’s next set of moves, which made it clear
that the CUP sought to decrease the administrative and spiritual influence of a key re-
ligious institution, the Meşīḫat (the office of the Şeyḫülislām). The CUP let its motives
toward the office of the Şeyḫülislām be known early in 1909, when one of its key activists
began publicly advocating that all judicial and educational institutions should be
transferred to the Ministry of Justice and Ministry of Education, respectively (Bein 2011,
20). Key ulema came to the CUP’s defense, providing religious justifications for treating
 State-Formation, Statist Islam, and Regime Instability    61

the Meşīḫat as a subordinate arm of the state completely devoid of autonomous spiritual
authority (Ardıc̦ 2012, 40). The CUP won an additional victory in 1910 when an ulema
member who was sympathetic to reorganizing the administrative, judicial, and educa-
tional functions of the religious establishment was appointed to the role of Şeyḫülislām
(Bein 2011, 40). By 1912 political maneuvering and rhetorical battling within the reli-
gious establishment’s ranks, centering on control of the office of the Şeyḫülislām, had
fractured the Sunni Muslim religious establishment into three main groups: those who
supported the CUP and bolstered its religious credentials, those who mounted defen-
sive opposition to CUP endeavors, and those who kept their heads down in an attempt
to remain apolitical while maintaining their careers (Bein 2011, 101). From 1912 through
the First World War, these groups jockeyed for influence, undermining one another in
the process.

Co-​optation and
Professionalization: Islamists in an
Era of State Secularism

The religious establishment, for its part, was certainly not looking to forfeit authority
and autonomy to state-​centric control. Nor were members of the religious establishment
entirely ignorant of the fact that reformers were being strategic in their appropriation of
religious rhetoric. At this point, however, the divisions within the religious establish-
ment were so acute that its collective defenses had been greatly weakened. Furthermore,
the continued availability of new professional avenues for Sunni Muslim religious elites
in this crucial reform period aligned a significant portion of their interests with those
of the emerging state, especially as religious institutions themselves languished. This
gradual shifting of allegiances paved the way for the Kemalists’ final push to completely
subordinate Sunni Muslim religious institutions to state-​centric purposes. Later in the
chapter I provide examples of this phenomenon in education, law, and the creation of
the bureaucratic institution that eventually became Turkey’s Presidency of Religious
Affairs (Diyānet İşleri Re’īsliği, later to become the Diyânet İşleri Başkanlığı, and from
now on referred to herein simply as “the Diyanet.”

Education

From the aforementioned qualitative analysis of 140 ulema biographies (Fabbe 2019),
I find that approximately fifteen ulema members taught in the new primary educa-
tion system at some point, and forty additional ulema members fulfilled some sort of
62   Kristin E. Fabbe

educational role in connection with the state as the medrese system was reformed and
brought under the state mantle prior to being “abolished” in 1924. Local records paint a
similar picture. For example, records from the Annual Education Yearbooks (Maʿārif
Sālnāmesi) from the Diyarbekir vilayet from 1898 to 1903 show that the vast majority
of teachers at all levels came from the learned religious class (Şimşek 2006). Beyond
attempts to expand modern teacher-​training programs, which ulema members were
encouraged to join, I have found no evidence to suggest that the CUP made any effort to
rid the state educational system of its religiously trained personnel.
Notably, even as religious teachers were co-​opted and became a mainstay in the pri-
mary schools for the time being, religious instruction itself was on the decline. More
specifically, the religious content of official primary-​school curricula was not imme-
diately eradicated—​even with the Kemalist revolution—​but rather slowly phased out
over the course of thirty to forty years. Tracing staffing figures, changes in textbook
content and class scheduling, and the observations of primary-​school students all
emphasize that religious messages and morality actually began to be stripped away
from educational texts long before schools were stripped of their “religious” teachers
(Doğan 1994; Somel 2001; Ünal 2008). In other words, the Second Constitutional
Period is characterized by a cadre of religious functionaries teaching on the basis of
materials that were slowly moving away from traditional Islamic doctrine and toward
a modern synthesis that fused a nominal version of Sunni Islam with an emergent
state-​centric nationalism.
Finally, the gradual decline and concomitant nationalization of the late Ottoman
medrese education system shows that, although the medreses were indeed formally
abolished by the Kemalist government in 1924, the gradual processes of change leading
up to that point had weakened these religious institutions to such a degree that aboli-
tion and usurpation of the system’s resources was a viable policy. Two newly fashioned
“medreses” that focused on the professional preparation of religious functionaries
were opened in late 1912, the Medresetü‘l-​vā‘izīn and the Medresetü‘l-​eimme ve’l-​ḫuṭebā
(Akiba 2003). This trend toward professionalization culminated in the “Reform of the
Medreses” bill (Iṣlāḥāt-​ı Medāris Niẓāmnāmesi) in 1914. This sweeping reform program
centralized, regularized, and effectively “nationalized” medrese education in Istanbul
(similar changes were later made outside the capital), organizing it into a three-​tiered
system (primary, secondary, higher) modeled on the structure of the state school system.
New curricular changes were also slowly introduced in an effort to blend the content of
the medrese and state systems closer together (Kütükoğlu 1978; Ürgüplü 2015).5 Finally,
and most significantly, with the 1914 reforms, caps were placed on medrese enrollment.
In 1914 enrollment in the capital was limited to 2,880, and in 1917 it was further reduced
to 1,350 (Bein 2011, 66). Interestingly, senior representatives from the medrese system
fully endorsed these caps, arguing that it helped to maintain student quality and weed
out free riders.
Thus, when the republican regime passed the Unification of Education Law (Tevḥīd-​i
Tedrīsāt Ḳānūnu) on March 3, 1924, which effectively transferred all medreses to the
 State-Formation, Statist Islam, and Regime Instability    63

control of the state Ministry of Education (Ayhan 1999), the medrese system itself had
been substantially weakened. Even as the medreses were forced to shutter their doors,
most of the ulema members who worked in them were not left out in the cold. The
Kemalists softened the blow by adding Article 4 to the Unification of Education Law,
which provided for the opening of a network of so-​called Imam Hatip schools, which
were designated with training prayer leaders and preachers who would comply with
principles of the new republic. These schools continue to exist to this very day and en-
roll a substantial portion of secondary and high school students in Turkey (Öcal 2007;
Ünsür 2005).

Law

In the legal realm, similar dynamics played out as reformers slowly phased out shariʿa
courts and grew the system of centralized Niẓāmiye courts described earlier in this
chapter by co-​opting Sunni Muslim religious elites. Data derived from the catalog of
Ottoman court records (Akgündüz and Türk Dünyası Araştırmaları Vakfı 1988) show
that in 1908, the year of the Young Turk Revolution, 121 of the 166 established judicial
centers for shariʿa law (73 percent) were still producing bound volumes of witness tes-
timony and judicial verdicts, meaning, of course, that they were still active and hearing
cases. By 1922, however, that number appears to have dropped to 32, meaning that only
19 percent of these centers were still active. The interesting question, then, is not how
Mustafa Kemal closed the shariʿa courts in 1924, but rather how these courts were suc-
cessfully phased out in the period before he rose to power. Evidence suggests that the
layering of modern and traditional legal institutions (combined with the gradual co-​
optation of judicial personnel from the religious establishment) created a situation in
which many ulema became tightly enmeshed with the emerging state’s legal apparatus
in the late Ottoman period, even after the Young Turk Revolution. Although this lay-
ering took numerous forms, particularly important in this respect was the fluidity with
which personnel flowed between the two systems for decades.
In the 1870s the Ministry of Justice had assumed the role as the administrative
headquarters for the Niẓāmiye court system, while the shariʿa courts remained under
the direction of the Meşīḫat (office of the Şeyḫülislām). As historians have observed,
however, the structure of the Niẓāmiye court system actually created a sort of “spe-
cial partnership” between the Ministry of Justice and the office of the Şeyḫülislām in
the domain of civil law that endured through the Second Constitutional Period. This
was because the civil sections increasingly operated in accordance with Niẓāmiye
procedure, but were presided over by judges from the religious establishment (Akiba
2018; Rubin 2011). For example, Rubin (2011, 78) finds that the majority of judges who
presided over the Niẓāmiye courts of first instance—​or the civil sections thereof, to
the extent that any distinction was maintained—​were nāʾibs. That is, they were shariʿa
64   Kristin E. Fabbe

judges from the ranks of the ulema and employed by the office of the Şeyḫülislām.
Rubin also finds that, according to Ottoman statistical yearbooks (sālnāmeler), these
were the same individuals who served as ḳāḍīs in the local shariʿa courts (147). There
is no evidence to suggest that the Ministry of Justice made any effort to expel nāʾibs
from the civil section of the Niẓāmiye system or rid the court of members of the re-
ligious elite, even after the Young Turk Revolution, unless they resisted reform very
directly (Akiba 2018, 234).
Qualitative analysis from the aforementioned 140 randomly selected ulema
biographies corroborates these findings. Fifty-​nine of these ulema members worked as
nāʾibs in the Niẓāmiye court system. Of these fifty-​nine individuals, twenty-​seven si-
multaneously served as both a nāʾib in the Niẓāmiye system and a ḳāḍī in a shariʿa court.
Furthermore, an additional eleven individuals did some sort of clerkship or scribal work
in the Niẓāmiye courts over the course of their lifetime (Fabbe 2019).
Given the high degree of fluidity and overlap between the Niẓāmiye and shariʿa sys-
tems for decades, it is perhaps not surprising that the religious establishment offered
little protest in 1917 when two important laws were passed that fused the institutions fur-
ther. The first law, which was actually even signed by the Şeyḫülislām, subordinated the
shariʿa courts to the Ministry of Justice. This law stripped the office of the Şeyḫülislām of
some of its most critical administrative assets without actually threatening the existence
of the shariʿa courts or the jobs of those who worked in them. The second law was the
Law of the Shariʿa Court Procedure, which further formalized the unification of judicial
procedure across the various courts and thereby made a significant step toward formally
centralizing the judiciary under an ordered set of regulations devised by the state (Rubin
2011, 26–​50).
One final piece of evidence from the Turkish ulema biographies and career trajectories
helps to summarize the data presented here about religious co-​optation into new state-​
centric institutions for law and schooling. Of a sample of seventy-​eight biographies of
ulema members (Miller 2005, 103) that extend directly into the Republican period, we
see the following trajectory for members of the religious establishment:

• 22 went on to work in the Ministry of Islamic Affairs and Religious Endowments/​

Diyanet, and 24 went on to work in the Justice Ministry
• 14 went on to secure teaching positions
• 11 remained unemployed
• 6 were imprisoned or fled the country
• 1 committed suicide

In other words, these biographies depict a situation in which a large segment of active
ulema members were given the opportunity to incorporate themselves into the nascent
institutional framework of the centralized state and, as a result, opted for cooperation
with, rather than active resistance to, policies to completely subordinate and dismantle
religious institutions once the Kemalists seized power.
 State-Formation, Statist Islam, and Regime Instability    65

Forging the Religious Bureaucrat

As the above examples demonstrate, work in the state’s judiciary and education systems
was a popular alternative to obsolescence or resistance for religious elites. So too was
serving as a bureaucrat in the predecessor of Turkey’s ministry for religious affairs. The
creation of this bureaucracy, its objectives, its exclusion of non-​Sunni Muslims, and its
subsequent growth and entrenchment stand as perhaps the best examples that the “sec-
ular” Turkish state created institutional conduits for religious elites and future Islamists
that flowed within and through the arteries of the state.
The immediate roots of Turkey’s Diyanet and its religious bureaucrats can be traced
to 1920, when the Şerʿiye ve Evḳāf Vekāleti (Ministry of Islamic Affairs and Religious
Endowments) was established after the opening of the Turkish Grand National
Assembly by Mustafa Kemal’s newly established shadow government in Ankara. Amit
Bein’s excellent history of the late Ottoman ulema shows that this short-​lived ministry
“served as a de facto alternative to the Şeyḫülislām and was primarily staffed by reform-​
minded ulema” (Bein 2011, 71–​72). In 1922 the Grand National Assembly also moved
to strip Sultan Vaḥīdeddīn of his role as caliph on the basis of treachery, transferring
the position instead to the more pliable ʿAbdülmecīd Efendi and sending out a circular
to inform all local imams and ḫatībs about this change.6 Soon thereafter, close on the
heels of the Treaty of Lausanne, in August 1923 the Kemalist assembly approved a reso-
lution to make Ankara the capital of the new state while retaining Istanbul as the seat of
the actual Caliphate (Ahmad 2003, 87). With these moves, the Ankara assembly man-
aged to temporarily preserve the symbolic authority of Islamic leaders and the Caliphate
while isolating them ever further from the inner workings of the new government in the
Republic’s first crucial months.
Finally, in March 1924, Kemal abolished the Caliphate completely in a demonstration
of national authority and state sovereignty. Interestingly, the actual proposal to abolish
the Caliphate was presented by the religious scholar and Sufi leader Sheikh Ṣafvet Efendi,
who rationalized that, “because the Caliphate is essentially intrinsic to the concept and
purview of the government and republic, the office of the Caliphate is abolished” (Ardıc̦
2012, 296–​297). A telegraph was sent to müftülüḳs (mufti offices) announcing the aboli-
tion and further instructing them that Friday sermons should emphasize prayers for the
well-​being (selāmet) and prosperity (saʿādet) of the religious community, whereas any
other sermon or message that would instigate chaos should be avoided.7 Simultaneously,
the Ministry of Islamic Affairs and Religious Endowments was shuttered and replaced
by the even more “domesticated” Diyānet İşleri Re’īsliği (Presidency of Religious Affairs,
later to become the Diyânet İşleri Başkanlığı). From now on, all cadres of preachers—​
vā‘iz, mü‘ezzin, imam, and/​or ḫatīb—​would be civil servants working for the Diyanet.
They would be required to send in their documents, including the information on their
identity cards, so as to monitor their compliance with the state. Local müftülüḳs were
turned into representatives of the Diyanet in their respective jurisdictions, sometimes
66   Kristin E. Fabbe

making decisions on behalf of the Diyanet as needed. 8 Simultaneously, the resources

and property of the religious establishment—​and especially the former Ministry of
Islamic Affairs and Religious Endowments—​were dismantled and conferred to other
departments such as the Ministry of Education.9
Strikingly, between 1920 and 1925, Mustafa Kemal and his allies did not take many
hostile actions against the Sufi orders or tekkes. With the abolition of the Caliphate
and the creation of the Diyanet, however, a Rubicon had been crossed. Soon there-
after the Kemalist government abolished the shariʿa courts,10 forced medreses to
close their doors, and imposed a number of other decrees that demonstrated the he-
gemony of state sovereignty over the religious realm. With respect to the Sufi orders,
the Sheikh Saʿīd Rebellion in the predominantly Kurdish southeast in February 1925
marked a critical turning point, for it provided direct evidence that tekkes and tarikats
(especially in non-​ethnically Turkish areas) could still potentially serve as reservoirs
of resistance to state consolidation and modernization. The next fall, with the pas-
sage of law 677v, the religious orders were formally abolished and their tekkes closed.
In the debate over the passage of the law, almost no one, not even those with me-
drese educations and Sufi backgrounds, defended the tekkes (M. Kara 2001, 19–​34;
Silverstein 2011, 84–​88). By the 1930s the term ulema itself was even banished from the
official lexicon (Bein 2011, 103).
The Diyanet, however, remained. With an official budget and state protection, it
created something akin to a bureaucratic shelter for those religious elites who were, if
not entirely comfortable, then at least somewhat secure working in the service of the
state. The Diyanet also quickly went about a general process of granting those members
of the religious establishment who were sympathetic to the Kemalist cause with official
papers certifying their affiliation with the Diyanet, so that they could circulate freely.11
The Diyanet also continued to pay out salaries to a number of religious teachers and
officials whose status was questionable.12 That said, the Diyanet was only viable as a
stable bureaucratic institution because the religious establishment that staffed it had
grown collectively weak as a result of educational and legal reforms based on strategies
of co-​optation, internal divisions within the establishment, and reformers’ ability
to instrumentalize religious rhetoric and attachments for nationalist purposes. The
Diyanet’s position relative to the government was made even weaker over the course
of the next decade. At its absolute nadir in 1931, the Diyanet even lost effective admin-
istrative authority over mosque officials and employees for a significant period of time
(Aytür, Çelik, and Şahinaslan 1989; Bein 2011, 143).
Although largely powerless back then, what remained of the Diyanet was not without
purpose. The Kemalists (and subsequent governments) used the Diyanet in a “pro-
ductive” capacity to help forge a bridge between traditional Sunni Muslim religious
attachments and modern nationalist loyalties. Now at the almost complete behest of the
Republic, religious administrators and civil servants were given a clear new mission: to
help finish the job of transforming the masses into loyal citizens of the state through
state-​sponsored and religiously grounded Turkification and nationalization initiatives
(Fabbe 2019; Lord 2018). Immediately after the Caliphate’s demise in March 1924, the
 State-Formation, Statist Islam, and Regime Instability    67

Kemalist government released a circular requiring all religious elites to use the occa-
sion as an opportunity to proselytize on the virtues of the nation-​state in public prayers
and sermons.13 In the 1920s and 1930s, the government tasked religious elites in the
Diyanet with producing modern Turkish translations of the Qurʾan and other sacred
texts, as well as publishing sermons for use in mosques across the country (Bein 2011,
121). Remarkably, in 1925 Ḥamdī Aḳseki, who would later serve as the Diyanet president
between 1947 and 1951, even wrote and published an instructional guide to teach Islam
to privates in the military, titled ʿAskere dīn kitābı (The book of religion for soldiers). The
book was reprinted several times and depicts military service as both a civic and a reli-
gious duty (Gürpinar and Kenar 2016, 66–​67).
It was not until 1928 that the Turkish Constitution substituted the line declaring Islam
as the religion of the state with the pronouncement that Turkey was a secular (lâik) re-
public. And even then, it did not abolish the Diyanet. The triumph of the Kemalists was
therefore not the separation of religion and state, but the usurpation and domestica-
tion of the Sunni Muslim religious establishment by the state. This was achieved grad-
ually and largely before the Kemalists came to power by the co-​optation of religious
elites, the gradual undermining of religious institutions, and the repurposing of reli-
gious attachments. Weak and diminished though they were as a group, many members
of the ulema had not been ostracized by the state but retained a steady foothold as civil
servants. This form of institutionalization not only cemented the peculiar synergies be-
tween Kemalism and Islam (and thus religion and state), but set the stage for Islamists’
eventual renaissance and re-​emergence as a state-​centric force for change in subsequent
decades.

Statist Islam into the

Post-​Kemalist Era

Turkey’s religious bureaucracy was at its weakest in the 1930s. In the Republican Era,
Islamist critics who remained outside the institutional framework of the Diyanet
or the state were stigmatized (by the state and its Islamist cadres alike) as pro-​shariʿa
(şeriâtçılık), backward, and/​or reactionaries (mürteci or irticâ yanlısı). Some were
sanctioned by the Kemalist state accordingly, whereas others broke rank with the
Kemalists in favor of other emerging political factions or apoliticism (Fabbe 2019, 123–​
125; Fabbe and Balikçioglu 2019). Regardless, by the 1940s Islamists were exerting in-
fluence on the state and on politics. Members of parliament began debating the role of
religion in primary education, with many wishing for a greater inclusion of religious
curriculum. Optional religion courses were placed in public schools in 1949, Qurʾan
seminaries/​courses were given permission to function, and a Faculty of Divinity was
established at Ankara University that same year. By 1950 the abolition on using Arabic
instead of Turkish for the called to prayer (ezân) was also lifted (Gözaydın 2008, 222).
68   Kristin E. Fabbe

All of these moves signified a growing space for religious activity, albeit mostly under
the mantle of the state. Space limitations do not allow for a detailed analysis of subse-
quent state-​centric patterns in Islamist behavior in the post-​Kemalist era, which could
easily warrant a chapter (or two) on their own. Nonetheless, I offer here some very brief
pieces of suggestive evidence revolving around key Islamist actors in contemporary
Turkish politics to illustrate this point. Key Islamist figures emerged from within the
state and have worked through it—​despite their allegiances to informal (and ostensibly
illegal) cemaats (religious brotherhoods) and tarikats (religious orders). Although many
have been vehement critics of Kemalist policies, they have also partnered with Kemalists
at various points over the last seven decades.
One example is Mehmet Zahid Kotku, who had been initiated into the influential
Khālidī branch of the Naqshbandī Order in 1918, took over leadership of the cemaat
in 1952, and became an influential, state-​sponsored prayer leader at the İskender Paşa
Mosque. Kotku became increasingly popular among student circles, growing a com-
munity of followers. In his sohbets (religious conversations) and teachings, Kotku
encouraged his disciples to be active in worldly and state affairs, and he and his cemaat
quickly emerged as one of the most influential leaders of Turkish political Islam. Kotku
and his followers did not shun the state, nor did Kotku “hold esteem for more radical
Islamists” who viewed state sovereignty with scorn (Mardin 2005, 158). Rather, he and
his movement embraced a synthesis of Turkish nationalism and Islam, positioning his
cemaat to help establish, support, and promote political, economic, and press activi-
ties (153–​155). By the 1960s, Naqshbandī-​Khālidī Sheikh Kotku’s sohbets began to take
place in important state institutions, including the State Planning Organization (Devlet
Planlama Teşkilatı), thereby signaling this cemaat’s willingness to participate in insti-
tutionalized politics. Preeminent figures that later came to dominate Islamists politics
in Turkey—​and contributed to major regime ruptures in the 1970s, 1980s, 1990s, and
2000s—​including Necmettin Erbakan and Recep Tayyip Erdoğan, both participated in
Kotku’s sohbets in the 1970s (Silverstein 2011, 102–​103) and subsequently followed his
statist approach.
Similarly, an offshoot of Said Nursî’s Nurcu cemaat call Hizmet (Service) emerged
around the popular preacher Fethullah Gülen, who also worked for the Diyanet and
followed a statist agenda—​although not always openly. Gülen worked as a Diyanet-​
appointed imam from the late 1950s to the 1980s, during which time he learned how
to position himself with other state actors and avoid restrictions on his community of
followers (Lord 2018, 221). His movement grew in popularity as a result of his oratory
skills and his writings, and recordings of his preaching began circulating widely in the
1980s. To further build its base, the movement typically established individual charitable
foundations with no official link to one another, although they were all tacitly and ideo-
logically connected to Gülen’s teachings. In this manner, the group developed a formi-
dable network of subsidized private educational institutions, preparatory schools, and
dormitories, which filled widening gaps in the state’s developmental capabilities (White
2002, 207). Gülenists also used charitable trusts, nongovernmental organizations,
firms, business associations, media outlets, and certain arms of the state bureaucracy to
 State-Formation, Statist Islam, and Regime Instability    69

expand their influence and support (Lord 2018, 219–​225). The police schools, the mili-
tary schools, the judiciary, the Ministry of Education, the Diyanet, and the human re-
source arms of all of these state institutions were all targeted by movement members to
increase their influence in the state. In the wake of the 1980 coup, especially, Fethullah
Gülen worked directly with the Turkish junta government, demanding that the religious
community comply with military directives and even providing a religio-​legal opinion
in favor of the coup d’état (Çakır 1994; Yavuz 2005).
Many additional examples—​the long string of Islamist parties influenced by Erbakan
and Erdogan and their unlikely coalition partners over the last six decades, the role of
Islamists in the recent Ergenekon and KCK trials, and the intra-​Islamist splits within
the Turkish state and Diyanet that recently came to a head with the failed coup of 2016—​
could be provided, but space does not allow. In the interests of brevity, suffice it to say
that modern Turkish politics has been characterized by existential power struggles for
the control of state institutions that involve Islamists and are largely invisible to the ge-
neral public as they unfold. These often paranoid and obsessive elite efforts to maintain
hegemony within state institutions have destabilized both democracy and autocracy.
They birth acute periods of crisis in which elite actors (Kemalist and Islamist alike) turn
against the perceived “enemies within” the state to shift balances of power—​causing re-
peated democratic and autocratic ruptures.

Conclusions

Instead of centering the story of religion-​state relations in Turkey on the Kemalist phe-
nomenon and its accoutrements, I have shown that it can be illuminating to focus on the
critical historical antecedents that structured the nascent state’s expansion into educa-
tion and law from roughly 1870 to 1928. When set in this broader historical perspective,
the supposedly miraculous story of Turkey’s Kemalist secularization, as well as its more
recent (and, to some, disappointing) “reversal,” begins to seem much less surprising.
The Kemalists could relegate religion by decree because they were confronting a gradu-
ally but greatly weakened religious establishment, many members of which already had
a vested interest in the triumph of the state and state-​centric institutions. Furthermore,
state formation in Turkey forged institutional and ideational links between the pious
and the national, and it drew on Sunni Muslim religious resources, networks, and priv-
ilege to help create state-​centric loyalties. The eventual counterpart to the Kemalist’s
secularizing achievements was the emergence of Islamists with statist characteristics
and prerogatives. Thus, this chapter contributes to a growing body of work that has in-
creasingly, and in my view rightfully, questioned the validity of viewing political Islam
as diametrically opposed to the forces of state secularism (Baskan 2014; Fabbe 2019;
Lord 2018; Tepe 2008; Tuğal 2009; Turam 2007; White 2013, 2017).
Some of the reasons behind Turkey’s failures of democratic maintenance are re-
lated to religion, but not in the way political science scholars deductively theorize
70   Kristin E. Fabbe

(Islamic values in conflict with democracy, competing overarching doctrines of shariʿa

versus sovereignty, etc.). From the outset, the Turkish state relied on religion as much
as it repressed it, and the nature of this reliance has contributed to regime instability.
Democracy in Turkey was not derailed by a group of outsider Islamists seeking to over-
throw the state and impose a new religious order or an Islamic state. Turkey’s political
Islamists have long been a fixture in the state apparatus, they tend to value the state as
a valid social ordering principle, and they have gradually worked through state arteries
and institutions to promote their agenda. Many contemporary observers of Turkish pol-
itics have therefore inadvertently gotten things backward: it is not the rise of Islamist
party politics that is reshaping the state from the outside; rather, the gradual and ever-​
closer intertwining of Islamist movements and the state has recently culminated in its
ultimate political expression with the AKP—​though the failed 2016 coup and the sub-
sequent purge of Gülenists from the state led to yet another outbreak of instability.
Scholars that seek to understand if and how religion impacts regime type in Muslim ma-
jority societies are thus urged to shine a light on how the historical foundations of state
secularism have shaped the evolution of political Islam.

Notes
1. This chapter has drawn heavily upon Kristin E. Fabbe, Disciples of the State? Religion
and State-​building in the Former Ottoman World (Cambridge: Cambridge University
Press, 2019), © Kristin Fabbe 2019, reproduced with permission of Cambridge
University Press.
2. In addition to housing Turkish-​speaking Sunni Muslims, the Ottoman imperial terri-
tories that became Turkey, and even the modern Turkish state, included a substantial
number of religious, ethnic, and linguistic minorities, including (but not limited to) Alevis,
Arab Christians, Jacobites and Assyrians, Armenians, Circassians, Karamanlides, Greek
Orthodox Christians, Kurds (both Kirmanji and Zaza), Jews, Dönme, Laz, Pontics, and
Roma. Religious heterogeneity—​like other forms of identity-​based diversity—​was neither
entirely static nor exogenous to the politics of state-​building in this period, though it is be-
yond the scope of this chapter. In other work I demonstrate that fusions between Sunni
Islam and the Turkish state pushed states toward an exclusionary style of religious nation-
alism that was further consolidated through state-​sponsored efforts to eradicate religious
diversity via forced migration and ethnic cleansing. See Fabbe (2019), chap. 8.
3. The term Niẓāmiye is derived from the word niẓām, meaning order and regularity.
4. In addition to religious minorities, the LU also drew some of its support from CUP defectors
who had become disillusioned by the organization’s aims and/​or frustrated by their own
inability to secure a more prominent position. Because the party was weak compared to
the CUP, it occasionally joined forces with smaller left-​wing parties such as the center-​left
Democratic Party (Fırḳa-​yı ʿİbād). In essence, then, the organization was a strange hodge-
podge of political factions united in little more than their opposition to the CUP’s aggres-
sive drive toward state centralization.
5. Several years later, however, the curriculum was again readjusted to ensure that the ulema
members held most of the faculty positions and did not lose their jobs.
 State-Formation, Statist Islam, and Regime Instability    71

6. Cumhuriyet Arşivi/​Tarih 14/​12/​1922/​Fon 51.0.0.0/​Yer 13.113.66. Another copy of this an-

nouncement can be found in Cumhuriyet Arşivi/​Tarih 1/​1/​1923/​Fon 51.0.0.0/​Yer 2.4.3.
7. Cumhuriyet Arşivi/​Tarih 7/​3/​1924/​Fon 51..0.0.0/​Yer 2.1.30; Tarih 6/​3/​1924/​Fon 51..0.0.0/​
Yer 2.12.8.
8. See Cumhuriyet Arşivi/​Tarih 7/​3/​1924/​Fon 51..0.0.0/​Yer 2.134; Cumhuriyet Arşivi/​Tarih 17/​
6/​1924/​Fon 51..0.0.0/​Yer 3.25.6; Cumhuriyet Arşivi/​Tarih 22/​9/​1925/​Fon 51..0.0.0/​Yer 3.16.2.
9. Several telegraphs discuss the future of the medrese buildings after the Ministry of Islamic
Affairs and Religious Endowments was reduced to a directorate. See Cumhuriyet Arşivi/​
Tarih: 1/​2/​1925/​Fon 30..10.0.0/​Yer 192.313..9/​Documents 1–​6. Documents 4–​6 announce
that some of the dilapidated and abandoned medrese buildings should be transferred to
the Ministry of Education for repurposing.
10. Cumhuriyet Arşivi/​Tarih 29/​3/​1924/​Fon Kodu 30..0.0.018.1.1/​Yer 9.19..1.
11. Cumhuriyet Arşivi/​Tarih 17/​6/​1924/​Fon Kodu 51..0.0.0/​Yer 3.25..6; Tarih: 18/​10/​1924/​Fon
Kodu 51..0.0.0/​Yer 5.43..11. In some cases there were delays; see Cumhuriyet Arşivi/​Tarih
22/​9/​1925/​Fon Kodu 51..0.0.0/​Yer 3.16..2.
12. Cumhuriyet Arşivi/​Tarih 5/​11/​1925/​Fon Kodu 51..0.0.0/​Yer 12.99..20; Tarih: 16/​1/​1926/​Fon
Kodu 51..0.0.0/​Yer 12.100..11.
13. Cumhuriyet Arşivi/​Tarih 7/​3/​1924/​Fon Kodu 51..0.0.0/​Yer 2.1..30.

References
Ahmad, Feroz. 2003. Turkey: The Quest for Identity. Oxford: Oneworld.
Akgündüz, Ahmet, and Türk Dünyası Araştırmaları Vakfı, eds. 1988. Şer’iye Sicilleri: Mahiyeti,
Toplu Kataloğu, ve Seçme Hükümler. Türk Dünyası Araştırmaları Vakfı Yayınları 52, 63.
Istanbul: Türk Dünyası Araştırmaları Vakfı.
Akiba, Jun. 2003. “A New School for Qadis: Education of Sharia Judges in the Late Ottoman
Empire.” Turcica 35: 125–​163.
Akiba, Jun. 2005. “From Kadi to Naib: Reorganization of the Ottoman Sharia Judiciary in the
Tanzimat Period.” In Frontiers of Ottoman Studies: State, Province, and the West, Vol. 1, ed-
ited by Colin Imber and Keiko Kiyotaki. London: I.B. Tauris.
Akiba, Jun. 2018. “Sharīʿa Judges in the Ottoman Nizāmiye Courts, 1864–​1908.” Osmanlı
Araştırmalar/​Journal of Ottoman Studies 51: 209–​237.
Akkoyunlu, Karabekir, and Kerem Öktem. 2016. “Existential Insecurity and the Making of a
Weak Authoritarian Regime in Turkey.” Southeast European and Black Sea Studies 16, no.
4: 505–​527. https://​doi.org/​10.1080/​14683857.2016.1253225.
Albayrak, Sadık. 1987. 31 Mart Vak’ası Gerici Bir Hareket Mi? Istanbul: Bilim–​Araştırma
Yayınları.
Albayrak, Sadık. 1996. Son Devir Osmanlı Uleması (İlmiye Ricalinin Teracim-​i Ahvâli). Vol. 1, A
-​ F. Istanbul: Büyükşehir Belediyes Kültür İşleri Daire Başkanlığı Yayınları.
Ardıc̦, Nurullah. 2012. Islam and the Politics of Secularism: The Caliphate and Middle Eastern
Modernization in the Early 20th Century. SOAS/​Routledge Studies on the Middle East 16.
London and New York: Routledge.
Atilhan, Cevat Rifat. 1956. 31 Mart Faciası: Ilim Işığında ve Tarih Önünde. Istanbul: Aykurt
Neşriyat.
Ayhan, Halis. 1999. Türkiye’de Din Eğitimi, 1920–​1998. Marmara Üniversitesi İlâhiyat Fakültesi
Vakfı Yayınları, no. 173. Istanbul: Marmara Üniversitesi İlahiyat Fakültesi Vakfı.
72   Kristin E. Fabbe

Aytür, Nihat, Yaşar Çelik, and Enver Şahinaslan. 1989. “Diyanet İşleri Başkanlığı Teşkilat
Tarihçesi.” Diyanet Dergisi, 25.1 January, February, March 1989, 31–​66.
Baskan, Birol. 2014. “State Secularization and Religious Resurgence: Diverging Fates of
Secularism in Turkey and Iran.” Politics and Religion 7, no. 1: 28–​50. https://​doi.org/​10.1017/​
S1755048313000278.
Bein, Amit. 2006. “Politics, Military Conscription, and Religious Education in the Late
Ottoman Empire.” International Journal of Middle East Studies 38: 283–​301.
Bein, Amit. 2011. Ottoman Ulema, Turkish Republic: Agents of Change and Guardians of
Tradition. https://​doi.org/​10.1515/​9780804777766.
Berkes, Niyazi. 1964. The Development of Secularism in Turkey. Montreal: McGill University
Press. http://​www.deslibris.ca/​ID/​448332.
Çakır, Ruşen. 1994. Ayet Ve Slogan: Türkiye’de İslami Oluşumlar. Edited by Yaşadığımız Dünya
Dizisi. Istanbul: Metis Yayınları.
Çalışkan, Koray. 2018. “Toward a New Political Regime in Turkey: From Competitive toward
Full Authoritarianism.” New Perspectives on Turkey 58 (May): 5–​33. https://​doi.org/​10.1017/​
npt.2018.10.
Cornell, Svante. 2017. “A Religious Party Takes Hold: Turkey.” SAIS Review of International
Affairs 37, no. 1S: S-​21–​S-​38. https://​doi.org/​10.1353/​sais.2017.0013.
Delibas, Kayhan. 2015. The Rise of Political Islam in Turkey: Urban Poverty, Grassroots Activism
and Islamic Fundamentalism. Library of Modern Turkey, vol. 10. London: I.B. Tauris.
Doğan, Nuri. 1994. İlk ve Orta Dereceli Okul Ders Kitapları ve Sosyalleşme, 1876–​1918. 1. basım.
İnceleme-​Araştırma 34. Istanbul: Bağlam.
Eisenstadt, S. N. 1999. Fundamentalism, Sectarianism, and Revolution: The Jacobin Dimension
of Modernity. Cambridge Cultural Social Studies. Cambridge: Cambridge University Press.
Emil, Birol. 1979. Mizancı Murad Bey : Hayatı Ve Eserleri. Edebiyat Fakültesi Yayınları, no. 2417.
Istanbul: Edebiyat Fakültesi Basımev.
Esen, Berk, and Sebnem Gumuscu. 2016. “Rising Competitive Authoritarianism in Turkey.”
Third World Quarterly 37, no. 9: 1581–​1606. https://​doi.org/​10.1080/​01436597.2015.1135732.
Fabbe, Kristin. 2019. Disciples of the State?: Religion and State-​Building in the Former Ottoman
World. Cambridge: Cambridge University Press. https://​doi.org/​10.1017/​9781108296878.
Fabbe, Kristin, and Efe Balikçioglu. 2019. “Political Islam.” In The Routledge Handbook of
Turkish Politics. London: Routledge.
Farhi, David. 1971. “The Şeriat as a Political Slogan—​Or the ‘Incident of the 31st Mart.’” Middle
Eastern Studies 7, no. 3: 275–​299. https://​doi.org/​10.1080/​00263207108700182.
Findley, Carter V. 2010. Turkey, Islam, Nationalism, and Modernity: A History, 1789–​2007. New
Haven, CT: Yale University Press.
Fortna, B. 2011. Learning to Read in the Late Ottoman Empire and the Early Turkish Republic.
Göl, Ayla. 2017. “The Paradoxes of ‘New’ Turkey: Islam, Illiberal Democracy and
Republicanism.” International Affairs 93, no. 4: 957–​966. https://​doi.org/​10.1093/​ia/​iix135.
Gözaydın, İstar B. 2008. “Diyanet and Politics.” The Muslim World 98, no. 2–​3: 216–​227. https://​
doi.org/​10.1111/​j.1478-​1913.2008.00220.x.
Gündüz, Mustafa, ed. 2008. İçtihad’ın İçtihadı: Abdullah Cevdet’den Seçme Yazılar.
Ankara: Lotus Yayınevi.
Güran, Tevfik, ed. 1997. Osmanlı Devleti’nin ilk istatistik yıllığı 1897 [The first statistical year-
book of the Ottoman Empire]. Tarihi Istatistikler Dizisi 5. Ankara: T.C. Başbakanlık Devlet
İstatistik Enstitüsü.
 State-Formation, Statist Islam, and Regime Instability    73

Gürpinar, Doğan, and Ceren Kenar. 2016. “The Nation and Its Sermons: Islam, Kemalism
and the Presidency of Religious Affairs in Turkey.” Middle Eastern Studies 52, no. 1: 60–​78.
https://​doi.org/​10.1080/​00263206.2015.1076797.
Hallaq, Wael B. 2009. Sharīʿa: Theory, Practice, Transformations. Cambridge: Cambridge
University Press.
Hanioğlu, M. Şükrü. 1995. The Young Turks in Opposition. Studies in Middle Eastern History.
New York: Oxford University Press.
Hanioğlu, M. Şükrü. 2001. Preparation for a Revolution: The Young Turks, 1902–​1908. Studies in
Middle Eastern History. Oxford: Oxford University Press.
Hanioğlu, M. Şükrü. 2005. “Blueprints for a Future Society: Late Ottoman Materialists on
Science, Religion and Art.” In Late Ottoman Society: The Intellectual Legacy, edited by
Elisabeth Özdalga, 28–​116. New York: Routledge.
Hanioğlu, M. Şükrü. 2008. A Brief History of the Late Ottoman Empire. Princeton, NJ: Princeton
University Press.
Höjelid, Stefan, ed. 2009. Turkey: From Tutelary to Liberal Democracy? Lund: Sekel.
Hür, Ayşe. 2008. “31 Mart ‘Ihtilal-​i Askeriyesi.’” Taraf, April 6, 2008.
İrtem, Süleyman Kani. 2003. 31 Mart İsyanı ve Hareket Ordusu: Abdülhamid’in Selanik Sürgünü,
edited by Osman Selim Kocahanoğlu. Istanbul: Temel Yayınları.
Kara, İsmail. 1994. İslâmcıların Siyasi Görüşleri. Istanbul: İz Yayıncılık.
Kara, İsmail. 2005. Din Ile Modernleşme Arasında Çağdaş Türk Düşüncesinin Meseleleri.
Istanbul: Dergah Yayınları.
Kara, Mustafa. 2001. “Bir Şeyh Efendinin Meşrutiyet ve Cumhuriyete Bakışı.” Tasavvuf İlmi ve
Akademik Araştırma Dergisi 6: 19–​34.
Karal, Enver Ziya. 1962. Osmanlı Tarihi, VIII Cilt. Birinci Meşrutiyet ve İstibdat Devirleri.
Ankara: Türk Tarih Kurumu Basımev.
Karpat, Kemal H. 1985. Ottoman Population, 1830–​1914: Demographic and Social Characteristics.
Turkish and Ottoman Studies. Madison: University of Wisconsin Press.
Kodman, Bayram. 1980. Abdülhamid Devri Eğitim Sistemi. Istanbul: Ötüken.
Kuran, Ahmed Bedevî. 1948. İnkılâp Tarihimiz Ve İttihad Ve Terakki. Istanbul: Tan Matbaası.
Kurban, Dilek. 2020. Limits of Supranational Justice: The European Court of Human Rights and
Turkey’s Kurdish Conflict. Cambridge: Cambridge University Press.
Kuru, Ahmet T. 2009. Secularism and State Policies toward Religion: The United States,
France, and Turkey. Cambridge Studies in Social Theory, Religion, and Politics.
Cambridge: Cambridge University Press.
Kutay, Cemal. 1977. 31 Mart Ihtilalinde Abdülhamid. Istanbul: Kalem Yayınevi.
Kütükoğlu, Mübahat S. 1978. “Dârü‘l-​Hilâfeti’l-​Aliyye Medresesi ve Kuruluşu Arifesinde
İstanbul Medreseleri.” İslâm Tetkikleri Enstitüsü Dergisi 7, no. 1–​2: 1–​212.
Lord, Ceren. 2018. Religious Politics in Turkey: From the Birth of the Republic to the AKP.
Cambridge Middle East Studies. New York: Cambridge University Press.
Mardin, Şerif. 1964. Jön Türklerin Siyası Fikirleri 1895-​1908. Ankara: Türkiye İş Banası Kültür
Yayınları.
Mardin, Şerif. 2005. “Turkish Islamic Exceptionalism Yesterday and Today: Continuity,
Rupture and Reconstruction in Operational Codes.” Turkish Studies 6, no. 2: 145–​165. https://​
doi.org/​10.1080/​14683840500119478.
Miller, Ruth Austin. 2005. Legislating Authority: Sin and Crime in the Ottoman Empire and
Turkey. Middle East Studies. New York: Routledge.
74   Kristin E. Fabbe

Öcal, Mustafa. 2007. “From the Past to the Present: Imam and Preacher Schools in Turkey—​
An Ongoing Quarrel.” Religious Education 102, no. 2: 191–​205. https://​doi.org/​10.1080/​
00344080701285477.
Pekdemir, M. Kemal. 2008. Tarihin En Tartışmalı Padişahı Abdülhamid. 1. baskı.
Istanbul: Neden.
Rodogno, Davide. 2012. Against Massacre: Humanitarian Interventions in the Ottoman Empire,
1815–​1914. Princeton, NJ: Princeton University Press.
Rubin, Avi. 2011. Ottoman Nizamiye Courts: Law and Modernity. New York: Palgrave
Macmillan.
Secor, Anna. 2011. “Turkey’s Democracy: A Model for the Troubled Middle East?” Eurasian
Geography and Economics 52, no. 2: 157–​172. https://​doi.org/​10.2747/​1539-​7216.52.2.157.
Silverstein, Brian. 2011. Islam and Modernity in Turkey. New York: Palgrave Macmillan.
Şimşek, Mehmet. 2006. Amid’den Diyarbekir’e Eğitim Tarihi. 1. baskı. Istanbul: Kent Yayınları.
Somel, Selçuk Akşin. 2001. The Modernization of Public Education in the Ottoman Empire,
1839–​1908: Islamization, Autocracy, and Discipline. The Ottoman Empire and Its Heritage,
vol. 22. Leiden and Boston: Brill.
Ṭanīn. 1908. “Cemʿiyet-​i İttiḥādiye-​Yi ʿİlmiye.” Ṭanīn, August 14, 1908.
Taş, Hakkı. 2015. “Turkey—​From Tutelary to Delegative Democracy.” Third World Quarterly
36, no. 4: 776–​791. https://​doi.org/​10.1080/​01436597.2015.1024450.
Tepe, Sultan. 2008. Beyond Sacred and Secular: Politics of Religion in Israel and Turkey. Stanford,
CA: Stanford University Press.
Tezcür, Güneş Murat. 2010. Muslim Reformers in Iran and Turkey: The Paradox of Moderation.
Austin: University of Texas Press. http://​site.ebrary.com/​id/​10392320.
Tuğal, Cihan. 2009. Passive Revolution: Absorbing the Islamic Challenge to Capitalism. Stanford,
CA: Stanford University Press.
Tuğal, Cihan. 2016. The Fall of the Turkish Model: How the Arab Uprisings Brought down Islamic
Liberalism. London and New York: Verso.
Tunaya, Tarık Zafer. 1984. Türkiye’de Siyasal Partiler. Vol. 1. Istanbul: Hürriyet Vakfı Yayınlar.
Tunaya, Tarık Zafer. 2003. İslamcılık Akımı. Istanbul: Bilgi Üniversitesi Yayınları.
Turam, Berna. 2007. Between Islam and the State: The Politics of Engagement. Stanford,
CA: Stanford University Press.
Ünal, Uğur. 2008. II. Meşrutiyet Öncesi Osmanlı Rüşdiyeleri, 1897–​1907: Programlar, Ders
Içerikleri, Istatistikler. Ankara: Gazi Kitabevi.
Unat, Faik Reşit, ed. 1960. İkinci Meşrutiyetin İlanı ve Otuzbir Mart Hadisesi (II. Abdülhamidʾin
Son Mabeyn Başkâtibi Ali Cevat Bey’in Fezlekeʾsi). Ankara: Türk Tarih Kurumu Basımev.
Ünsür, Ahmet. 2005. Kuruluşundan Günümüze İmam Hatip Liseleri. Istanbul: Ensar Neşriyat.
Ürgüplü, Ali Suat, ed. 2015. Şeyhülislam Ürgüplü Mustafa Hayri Efendi’nin Meşrutiyet, Büyük
Harp ve Mütareke Günlükleri. Istanbul: İş Bankası.
White, Jenny. 2002. Islamist Mobilization in Turkey: A Study in Vernacular Politics.
Seattle: University of Washington Press.
White, Jenny. 2013. Muslim Nationalism and the New Turks. Princeton Studies in Muslim
Politics. Princeton, NJ: Princeton University Press.
White, Jenny. 2017. “Spindle Autocracy in the New Turkey.” Brown Journal of World Affairs,
Fall/​Winter, 2017.
Yavuz, M. H. 2005. “Neo-​Nurcular: Gülen Hareketi.” In Modern Türkiye’de Siyasi Düşünce /​Cilt
6: İslamcılık, edited by Tanıl Bora and Murat Gültekin, 295–​307. Istanbul: İletişim.
 Chapter 4

States, Re l i g i on,
and Demo c rac y i n
Sou theast Asia
Comparative Religious Regime Formation

Kikue Hamayotsu

The long-​established “secular” premise of modern liberal democracy—​separation be-

tween church and state—​is under critical scrutiny. Newer scholarship in the fields of
religion and politics and comparative democratization demonstrates that complete sep-
aration of church and state or active marginalization of religion in the realm of politics
and policymaking is not the norm across various political regimes and various religious
traditions. Furthermore, comparative studies of Christian-​dominant democracies sug-
gest that religious institutions could play a positive role in facilitating a transition to
and consolidation of democracy (Driessen 2014; Kalyvas 2000; Kselman and Buttigieg
2003). However, we are still unsure how and under what conditions particular state-​
religion relations are formed and consolidated in Muslim-​majority societies. Moreover,
whether newly emerging Muslim-​dominant democracies are similarly equipped to
guarantee state protection of basic democratic rights and freedom of various religious
(and nonreligious) communities and populations is still contested.
This chapter seeks to account for the political origins of what I call “religious regimes”
through a comparative historical analysis of three Muslim-​majority Southeast Asian
nations: Indonesia, Malaysia, and Brunei. Despite relatively similar historical, geo-
political, and sociocultural conditions—​and against their original intentions—​these
three nations have established distinctive types of religious regimes since indepen-
dence: secular-​dominant (Indonesia), established-​religion (Malaysia), and religious
monarchy (Brunei). The chapter also asks what the effects of the type of religious
76   Kikue Hamayotsu

regime on state protection of democratic civil rights of religious (and nonreligious)

communities and members are.
I advance two theoretical arguments. First, I argue that the varying pathways adopted
by these three nations are the result of state formation—​specifically, the ways in which
secular political elites pursue and consolidate their state powers and domination fol-
lowing independence. Second, my comparative case studies demonstrate that the type
of religious regimes formed in Muslim-​majority societies—​as well as the timing of reli-
gious regime formation—​will likely impact the ability of late-​developing democracies
to protect basic civil rights and freedom of religious and secular communities. This
addresses an unresolved question pertinent to the prospect of modern constitutional
democracy in the age of thriving religious fundamentalism and what Stathis Kalyvas
(2003) calls “unsecular politics.” In contrast to Christian-​dominant Europe, where
mostly secular regimes were concurrently formed to facilitate democratic rule and con-
solidation, postcolonial Muslim societies saw the institutional foundations of religious
regimes built earlier, and prior to the rise of democratic movements and institutions,
allowing established religious authorities greater influence to complicate the trajectories
of democratic consolidation.

Comparative Religious Regime

Formation: State, Organized Religion,
and Democracy

Institutional separation of church and state, or what some scholars term “political sec-
ularism,” has long been an established premise in the formation of modern democracy
across Christian-​dominant Europe and America (Toft, Philpott, and Shah 2011, chap. 3).
Scholars argue that such an institutional condition is a prerequisite for establishing
modern democracy and expanding constitutional civil rights and freedom (Calhoun,
Juergensmeyer, and VanAntwerpen 2011, Introduction; Kselman and Buttigieg 2003).
Culturalists highlight cultural and/​ or theological differences between religious
traditions to argue that religious and state powers should be closely intertwined in
Muslim societies according to Islamic theology and traditions. An ideological antago-
nism by clerical elites and Islamists to political secularism and Western liberal democ-
racy is often said to be a justification for blurred relations between the two institutional
powers and a collective aspiration to build an Islamic state and society guided by shariʿa
(Islamic laws). According to this position, a lack of political secularism is also a source
of autocratic rule and illiberal characteristics and attitudes in predominantly Muslim
societies (Tessler 2000, 2002; Tibi 2008, 2012). Furthermore, many scholars suggest
that flourishing religious fundamentalism, radical Islamism, and theocratic rule across
Muslim societies in the past few decades are the result of Muslim grievances and re-
sistance against the penetration of the Western powers and cultural traditions that
 States, Religion, and Democracy in Southeast Asia     77

Islamists perceive as “infidel” and an insult against Islam (Barber 1995; Dalacoura 2005;
Huntington 1993; Inglehart 2000; Kepel 1997; Mamdani 2004; Roy 2007).
However, such theoretical premises based on religious differences are under critical
scrutiny among newer scholarship of religion and politics, comparative regimes, and
democratization. Comparative studies have contributed to new classifications of, and
explanations for, a variety of forms of political secularism and relations between polit-
ical and religious authority—​not only across various religious traditions, but also within
them (Calhoun, Juergensmeyer, and VanAntwerpen 2011; Fox 2006, 2008; Fox and
Sandler 2005; Künkler, Madeley, and Shankar 2018; Kuru 2009; Shah, Stepan, and Toft
2012; Toft, Philpott, and Shah 2011). Ahmet Kuru (2007, 2009), for example, advocates
for the typologies of “passive secularism” and “assertive secularism,” not only to cate-
gorize state-​religion regime types, but also to account for state approaches to religious
institutions and authority both across and within religious traditions. His comparative
case studies of Christian-​dominant France and the United States and Muslim-​dominant
Turkey demonstrate that types of state-​religion relations are attributed to the presence
(or lack) of dominant religious authority in imperial ancient regimes prior to the forma-
tion of a modern state and national identity.
Furthermore, some scholars focus on the institutional characteristics of dominant re-
ligious authority and its relations with the state to account for democratic transition and
consolidation—​or enduring autocracy—​across various religious traditions (Kalyvas
2000; Koesel 2014; Künkler and Leininger 2009; Menchik 2018; Stepan 2001). Stathis
Kalyvas, for example, argues that the rise of Christian democracy in Catholic Belgium
and the failure of Muslim Algeria to attain democratic transition results from the highly
centralized ecclesial order of Catholicism and decentralized Islamic authority (Kalyvas
2000). The concept of “twin tolerations” Alfred Stepan advanced is another shining ex-
ample of this scholarly trend. He argues that it is mutual respect for autonomous rela-
tions and interactions between the secular state and religious authority that facilitate the
consolidation of modern liberal democracy (Stepan 2001). Overall, there is increasing
consensus that it is not simply separation between church and state that facilitates dem-
ocratic governance, but possibly “friendly relations” and collaboration between the two,
which allow religious authority to play a positive role in the expansion of democracy
(Driessen 2009, 2010).
Despite this theoretical progress, the extent to which, and the manner in which, reli-
gious authority is organized and institutionalized, and how clerical authority and elites
relate to secular political authority among Muslim societies, is relatively understudied
and unknown.1 Moreover, the ability of emerging Muslim-​dominant democracies to
ensure equal membership and protection for not only religious, but also nonreligious,
populations, a basic condition for a modern democratic state, is still questionable. As
Mirjam Künkler and Shylashri Shankar (2018) rightly point out, Muslim societies,
whether democratic or not, tend to be insensitive, or often unambiguously violent to
nonbelievers, not allowing their citizens “not to believe.” A number of Muslim-​majority
states tend to be aggressive in requiring and coercing religious membership, identity,
and piety among citizens. The conditions for religious freedom and minority rights
78   Kikue Hamayotsu

to believe (and not believe) are not necessarily protected by democratic regimes, as
espoused by theorists of modern democracy, Robert Dahl and Alfred Stepan in partic-
ular (Dahl 1986, 1991; Stepan 2001, 2011).
Even though the culturalist premise that particular religious traditions—​especially
Islam—​are unfit for the establishment of liberal democracy is largely discredited in
theory, we are still uncertain whether emerging predominantly Muslim democracies
are willing—​and able—​to protect various religious and nonreligious communities and
populations in reality, when they are concurrently experiencing thriving religious fun-
damentalism. Islamist organizations, leaders, and ideologies continue to command
considerable influence and appeal to mobilize a large segment of Muslim populations,
posing a major threat to the democratic rights of non-​ Muslim and minority
communities. According to the ILGA-​RIWI Global Attitudes Survey on Sexual, Gender
and Sex Minorities, Muslim-​majority nations undeniably lag far behind other nations
and religious traditions in terms of gender rights and equality, especially women and
LGBTQ communities.2 Homosexuality is typically illegal and criminalized according to
shariʿa and/​or civil laws. The question then becomes, what are the conditions that could
facilitate state protection and promotion of fundamental civil rights for believers and
nonbelievers alike in Muslim-​majority nations?
My comparative historical analysis of religious regime formation in Muslim-​
dominant nations in Southeast Asia seeks to fill these theoretical gaps. I conceptualize
the religious regime as institutional relations between political authority (a secular
state) and clerical authority (organized religion). Along the lines of the state-​religion
regimes analyzed by Ahmad Kuru (2009), religious regimes in Muslim-​majority na-
tions are broadly categorized into three ideal types: secular, established-​religion, and
religious (as indicated in Table 4.1).
My regime typology focuses on the relative powers permitted to religious institutions
and clerical elites (ʿulamaʾ) in four areas: (1) executive, (2) legislative, (3) judiciary,
and (4) official constitutional position of Islam. This is done in order to assess the re-
lations and interactions between secular political and religious authorities in public
policymaking. These types are Weberian ideal types and continuum. Some cases fit per-
fectly within these ideal types, while many others sit on a continuum between them.
The secular regime, for example, according to my typology, does not mean that reli-
gious authority has little to no role to play in state policymaking or politics. Religious
institutions and clerical elites can legitimately participate in political activities and
pursue state offices to influence state policymaking and gain access to state resources, as
seen in Indonesia or Turkey. However, religious authority is not officially equipped with
veto power to overturn or supersede laws, rulings, or policies made by secular political
and legal authority, in the way that they are in established-​religion or theocratic regimes,
such as Malaysia and Brunei.
Likewise, the religious regime allows secular elites and institutions to influence and
participate in policymaking and government. The key distinctive feature of the religious
regime is that religious authority and laws (shariʿa) are paramount and predominant.
Secular elites and institutions are not equipped to veto decisions and policies made by
 States, Religion, and Democracy in Southeast Asia     79

Table 4.1: Types of Religious Regimes in Muslim-​Majority Nations

Secular Established-​Religion Religious

Executive Secular Secular Religious (monarchic or

theocratic)
Legislative Secular Secular Shariʿa/​fatwa
Judiciary Secular Mixed Shariʿa/​fatwa
State religion Islam not privileged/​ Islam Islam
nominal
Examples Indonesia, Senegal, Malaysia, Pakistan, Brunei, Saudi Arabia
Turkey Bangladesh (monarchic)
Iran, Taliban’s
Afghanistan
(theocratic)

Source: Kuru 2009, 6–​10, 247–​254.

ʿulamaʾ according to their theological rationales and interpretations. Moreover, my ty-

pology makes an important distinction between two subcategories of the religious
regime—​monarchic and theocratic rule—​according to this veto power principle. In
monarchic religious regimes, a hereditary ruler is usually the head of the state religion.
In contrast to theocratic rule, ultimately his power or authority are not defined by theo-
cratic expertise, or by divine will. He reserves discretion to make, remake, and unmake
shariʿa or theologically certified opinions (fatwa) and policies according to his worldly
will and needs, in almost the same manners that Christian kings or sultans ruled their
realms in premodern Europe or empires in the name of God.
The religious regime is distinct and independent from the political regime as a regime
typology. It is analytically beneficial for comparative analysis, because the latter tends
to ignore the institutional and ideological powers of organized religion or religious
authority. Moreover, it is the timing of religious regime formation, and the patterns of
interactions between the two institutional powers—​secular and religious—​that this
chapter suggests will impact the ability of a democratic regime to protect constitu-
tional and civil rights of religious (and nonreligious) citizens and minorities in the con-
text of thriving religious fundamentalism in Muslim-​majority nations. What factors or
conditions have contributed to the diverging patterns of state-​religion relations across
Muslim-​majority nations? What are the effects of the type of religious regime on the
democratic rights and freedom of religious (and nonreligious) communities, majority
as well as minority alike?
This chapter focuses on three Muslim-​majority nations in Southeast Asia—​Indonesia,
Malaysia, and Brunei—​to answer these questions. Despite relatively similar religious
and cultural traditions, historical legacies of colonial rule by European powers, and ge-
opolitical conditions, these nations have established three distinctive types of religious
80   Kikue Hamayotsu

regime that fit the three ideal types mentioned above, offering valuable comparative case
studies: secular (Indonesia), established-​religion (Malaysia), and religious monarchy
(Brunei). They all adhere to Sunni Islamic traditions and Muslim societies are com-
monly considered to be more moderate, tolerant, and peaceful than their counterparts
elsewhere, in the Middle East, Africa, or South Asia (Hefner 2000; Menchik 2016;
Peletz 2002; Sloane-​White 2017). Moreover, Indonesia and Malaysia achieved a rel-
atively peaceful transition to democratic rule—​in 1998 and 2018, respectively—​after a
few decades of autocratic rule, whereas Brunei has established one of the most stable
absolute monarchies, allowing empirical observation of interactions between various
political regimes and religious authority. Further, in the run-​up to independence from
colonial rule, the place of Islam, the majority faith, was commonly a major source of
constitutional debates and political conflict among Muslim ruling elites. The sec-
ular Muslim rulers had no intention to build a state primarily based on, or guided by,
shariʿa and clerical authority. Arguably, these Southeast Asian nations could have taken
converging paths, resulting in a mostly similar religious regime. In contrast to the orig-
inal intentions of the founding fathers to build a largely secular modern state with cere-
monial religious characters, however, the three nations have taken diverging pathways
to establish three distinctive types of religious regime.
I advance two theoretical arguments. First, the religious regimes formed in Muslim-​
dominant Southeast Asian nations are the result of state formation, specifically the ways
in which secular political elites pursue and consolidate their state powers and domina-
tion during post-​independence institution building. When secular rulers compete with
traditional clerical elites (ʿulamaʾ) over the ideological foundations of the nascent state
and distributions of state powers and resources, thereby not needing the latter’s socio-
cultural assets or legitimization for their regime survival, the state tends to adopt a more
secular and pluralistic position, as attested in a largely secular-​dominant Indonesia. On
the other hand, when secular rulers need the cultural assets and legitimizing powers of
religious authority to achieve their domination and survival, the state is likely to grant a
privileged position and exclusive patronage to the majority religion in the formal con-
stitutional and bureaucratic structures. Malaysia’s established-​religion regime, where
Islam is the state religion, and to a much greater extent Brunei’s religious monarchy are
testament to this theoretical model.
Second, my comparative analysis suggests that the type—​and the timing—​of reli-
gious regime formation accounts for the ability—​and willingness—​of late-​developing
democracies such as Indonesia and Malaysia to protect basic civil rights and freedom
of religious (and nonreligious) communities and people. In contrast to Christian-​
dominant Europe, where mostly secular regimes were formed concomitantly to facil-
itate democratic transition and consolidation, late-​developing Muslim democracies
have seen the consolidation of religious regimes prior to the formation of modern state
and popular democratic movements and consolidation. As a result, late-​developing
Muslim democracies tend to see more aggressive religious authorities contesting
and complicating the trajectories of democratic consolidation. The secular regime of
Indonesia is better equipped than the established-​religion regime of Malaysia to tame
 States, Religion, and Democracy in Southeast Asia     81

assaults against religious minorities, thanks to the institutionalized religious subordi-

nation to the secular state authority. However, in practice, even a consolidated secular
democracy such as Indonesia is constrained by thriving religious fundamentalism and
the latter’s ideological and political appeals to mobilize a large segment of pious Muslim
populations, because of the timing of religious regime formation.

Indonesia: The Making of a Secular-​

Dominant Regime in the Face
of Powerful Islamism

Indonesia has established a largely secular religious regime as a result of secular-​

sponsored state-​building since independence, under the autocratic regimes of the first
president, Sukarno (1945–​1965) and the second, Suharto (1966–​1998). The secular dom-
ination of the Indonesian state—​constitutional domination of secular political elites,
and subordination of religious authority and elites to the state —​has taken place against
all odds, as a comparatively large Muslim population (close to 90 percent), active and
competitive religious civil society, and historically powerful Islamist movements have
contested the secular state. The largely secular nature of the religious regime has been
consolidated even further after a relatively peaceful consolidation of democratic rule
since 1998, primarily to reinforce the secular nationalists’ hold on to state powers, while
marginalizing radical forms of Islamism.

Making of the Unitary State of the Republic and

the Pluralist National Identity, Pancasila
In the process of preparation for independence of the republic and crafting of the con-
stitution, the place of Islam, the majority faith, in the making of a new state and unifi-
cation of a fragile archipelagic nation was much more contentious in Indonesia than in
its Muslim neighbors. In contrast to colonial rule in British Malaya, Dutch colonial rule
in the East Indies was generally much more hostile to Islam and suspicious of religious
authority as a potential source of political disturbances and opposition against colonial
rule. The Dutch promotion of secular indigenous elites and legal traditions in the colo-
nial state, and the exclusion of ʿulamaʾ and Islamic traditions from that state—​as well
as toleration of Christian missionaries in indigenous communities—​gave rise to deep
anticolonial and anti-​Christian sentiments, as well as powerful nationalist movements
led by ʿulamaʾ (Laffan 2003), a phenomenon not witnessed in British Malaya. It was an
ad-​hoc alliance of these nationalist movements and political parties, Islamist and sec-
ular, that fought and won the revolutionary war against the Dutch to gain full territorial
independence in 1949.
82   Kikue Hamayotsu

To what extent the state integrates and elevates shariʿa and privileges Islam and the
majority Muslim community was a main source of constitutional debates, further
deteriorating political rivalry between secular and religious nationalist leaders (Boland
1982; Cammack 1989; Cammack and Feener 2012; van Dijk 1996; Elson 2009; 2010; Elson
2012; Elson 2013). The secular Muslim elites, especially President Sukarno, a secular na-
tionalist, had no intention to build a new republic based on Islam or guided by shariʿa.
In fact, Sukarno was determined to make a secular modern state—​wherein political au-
thority is emancipated from, and superior to, religious authority, which Sukarno and
other secularists considered feudalistic and unfit for building a modern republic. For
the secularists, Atatürk’s Tukey was a model for state-​building. On the other hand, re-
ligious elites from major Islamic movements—​traditionalist Nahdlatul Ulama (NU)
and modernist Muhammadiyah—​as well as more radical Islamist groups, were eager
to build the new republic based on Islam and guided by religious authority and shariʿa.
They insisted that the constitution should assure the majority faith a superior position
as the ideological and legal foundation of the state (Cribb 1999; Feith and Castles 1970;
Kahin 2013).
In the face of the still extremely frail foundations of political sovereignty, territorial
and national unification, and imminent threats of secessionist movements in predom-
inantly Christian regions, Sukarno’s Indonesia could not afford to lose the allegiance of
non-​Muslim regions and populations. Without royal or aristocratic lineage or affinity
in the feudalistic Indonesian society, his legitimacy and survival as the founding presi-
dent were largely dependent on his charismatic personal characters and political skills
to integrate the culturally and ideologically divided nation together and make it a more
egalitarian society (Benda 1965; Cribb 1999; Feith and Castles 1970). In order to pre-
vent the disintegration of the nascent state and cultivate pluralistic national identity, he
invented a novel national doctrine called Pancasila (meaning “five precepts” in Sanskrit)
as the constitutional foundation of the state-​building.3 The state’s elevation of Pancasila
was much to the chagrin of the religious leaders who were determined to see Islam priv-
ileged as the state religion. Having their Islamist ambitions defeated constitutionally,
some of the regions dominated by Islamist leaders rebelled against the central govern-
ment to create an Islamic state (Boland 1982; Formichi 2012; Harvey 1977; Horikoshi
1975). However, Islamist rebellions were militarily pacified, while Islamist parties and
leaders were banned and marginalized thereafter.

Consolidation of Military Rule and Subordination

of Religious Authority to the Secular State
The political ascent and consolidation of military-​dominant autocratic rule under
General Suharto largely relied on his pledges to restore political order and stability to
attain economic growth and improve the basic welfare of impoverished populations.
Secular political and business elites generally perceived Islamism or organized
 States, Religion, and Democracy in Southeast Asia     83

religion as a major threat not only to political and social stability, but also, and more
importantly, to their own domination after the state-​sponsored mass termination of
the organized Left, Communism. Secular dominant state-​building after the fall of
Sukarno in 1966 primarily countered traditionally strong Islamism to cement insti-
tutional and ideological allegiance of religious authority and elites to the secular state
(McVey 1983).
First, the state requires popular allegiance to the state-​sanctioned official religions
(originally four, increased to six in 2001): Islam, Roman Catholic, Protestant, Hindu,
Buddhism, and Confucianism. The Ministry of Religious Affairs was created in 1946
to placate the religious leaders, but does not privilege Islam even ceremonially, in stark
contrast to Malaysia and Brunei.4 State jurisdiction over religion (Islam in particular)
was, and remains, limited primarily to family and ritualistic matters. Second, failures
of religious elites to establish Islam as the state religion meant that religious organiza-
tions and clergies were largely left autonomous to run their religious missions, activi-
ties, and programs, and to meet the spiritual and welfare needs of Muslim populations.
The two largest Islamic associations, NU and Muhammadiyah, traditionally monop-
olize the domestic religious market, and compete with one another over membership,
resources, ownership, authority, and power within the Muslim community, thereby
retaining considerable social influence (Bush 2009; Federspiel 1970; Feillard 1996,
1997; Jung 2009).
However, the regime adopted various measures to limit the political and ideological
powers of religious civil society, thus subjugating it to the secular state. Ideologically,
the government enforced a policy called asas tunggal (sole ideological foundation) to
require all religious and social organizations to pledge absolute allegiance to Pancasila,
the state ideology widely viewed by religious elites as a secular state’s political tool to un-
dermine religious authority. Politically, the regime reorganized all the political parties
into the three state-​sanctioned parties—​the regime party, a Muslim-​based party, and
a non-​Muslim (largely Christian and secular) party—​to further curtail the political
powers of organized religion. The secular domination of the state has changed little,
even after Suharto’s efforts to woo the modernist Muslim communities in the final years
of his rule, primarily to ensure his regime’s survival and domination in the state (Hefner
1993; Liddle 1996).

Democratic Rule and Consolidation of the Secular

Pluralist State
Successful transition to consolidated democracy since 1998 has not altered the subordi-
nate position of Islam and religious authority to the secular state at all levels (legislative,
executive, and judicial). Democratic legislative elections introduced in 1999 have led to
a mushrooming of political parties, including Islamist parties. In contrast to post–​Arab
Spring Egypt and Tunisia (or Turkey), however, Islamist parties proved to be rather
84   Kikue Hamayotsu

weak from the onset (Fealy 2009; Hamayotsu 2011a, 2011b, 2012a). In general, Islam does
not seem to sell among Muslim voters, despite flourishing Islamist and radical (and even
terrorist) organizations and religious conservatism under democratic rule (Kuipers
2019; Pepinsky, Liddle, and Mujani 2012, 2018).
At the executive level, the ascension to the presidency of the exceptionally liberal
cleric and former chairman of NU Abdurrahman Wahid in 1999 brought about un-
precedented institutional changes favorable to religious and ethnic minorities. The
country’s fifth president, Megawati Sukarnoputri (2001–​2004), chairwoman of the
secularist PDIP, and her successor, the former military general Sushilo Bambang
Yudhoyono (2004–​2014) had little interest in altering the essentially secular domi-
nation of the state, although Yudhoyono acknowledged the political benefit of cler-
ical elites and organizations to strengthen his support base in civil society (Bush 2015;
Hamayotsu 2013).
The decentralization of administrative and fiscal powers to more than five hundred
districts and cities, and the introduction of direct elections of the local executive offices
in 2005, also contributed to the increasingly competitive religious market (Hoesterey
2015). A number of districts and cities in traditionally conservative regions issued local
by-​laws and executive orders to introduce religiously inspired regulations, such as
the banning of alcoholic beverages, fueled further anxieties about sweeping Islamism
(Buehler 2016; Bush 2008). In practice, however, there is no adequate evidence to in-
dicate that religious authority has gained more legal or administrative powers to su-
persede the secular dominant state and legal institutions and political authority under
democratic rule. In stark contrast to the strengthened religious legal authority and co-
ercive powers in Malaysia and Brunei, the secular regime of Indonesia, in fact, has been
further consolidated to keep religious authority and institutions firmly subordinate to
the secular state.
First, the religious bureaucratic and legal reforms over the past few decades have
limited state jurisdiction over shariʿa strictly to matrimonial matters (marriage and
divorce), while placing it under the secular Supreme Court (Cammack 1989, 1997;
Feener and Cammack 2007). Except for the special autonomous region of Aceh, there
is no equivalent of the powerful independent shariʿa courts or religious apparatus de-
veloped in Malaysia or Brunei that could constitutionally supersede and veto the sec-
ular courts and laws (Feener 2013). Second, the government passed laws to restrict
civil society’s ideological deviation from Pancasila, thereby tightening state control
over religious organizations. President Joko Widodo, who took office in 2014, is a sec-
ular nationalist PDIP politician who is overwhelmingly popular among minorities,
and he has gone even further to tighten the state’s coercive power over Islamism to
defend the secular multireligious foundation of the state. He enacted the Executive
Regulation in Lieu of Law (Perppu) in 2017, widely seen as targeting radical Islamism,
authorizing the government to ban any organizations deemed to contradict Pancasila,
or “national unity.”5
 States, Religion, and Democracy in Southeast Asia     85

Consolidated Democracy and Religious Conservatism

In short, the democratically elected state has consolidated the secular regime even fur-
ther, keeping religious authority tightly subordinate to secular political authority in
the state. However, the rising influence of radical Islamism and religious conservatism
during the successful consolidation of democracy has posed a major challenge to the
democratic regime. Even though the secular regime is better equipped with state and
legal apparatuses to protect the pluralistic foundations of democratic rule and tame reli-
gious radicalism than the other religious regimes, secular elites, be they executive or leg-
islative, are still on the defensive when it comes to religiously sensitive issues and policy
areas. They are usually unwilling to interfere overtly in sensitive areas such as LGBTQ
rights or anti-​Christian movements—​not because the state privileges Islam, but, iron-
ically, because of democratic elections. Muslim politicians are afraid of losing Muslim
votes by offending socially influential ʿulamaʾ and their religious decrees, now that their
political survival and legitimacy are dependent on the ballots. Against the backdrop of
deep-​seated, and politicized, religious identity, Muslim politicians find themselves vul-
nerable to popular religious sentiments.

Malaysia: Marriage of Ethnic

and Religious Identity in the
Established-​R eligion Regime

In contrast to a secular Indonesia and a number of other Muslim-​majority states,

Malaysia has formed the established-​religion regime, wherein the dominant religion
and its ecclesiastic order are constitutionally integrated into the secular state and legal
structures. The majority faith—​Islam—​has constitutionally been the official religion of
the Federation of Malaysia since independence in 1957. The constitution confers Islam
and the majority community privileges, a superior position that Indonesian clerics
and Islamists have fought hard legally and illegally, but failed or compromised on, as
discussed above. Although secular authority largely dominates the executive, legislative,
and judicial branches, the party-​dominant regime has actively incorporated and ele-
vated religious authority in the state—​especially the powers of shariʿa courts and laws,
and provision of welfare services and education.
The formative years of state-​building under British colonial rule and after indepen-
dence is essential to account for the way in which the established-​religion regime was
formed. Under the constitution, nine sultans or hereditary Malay rulers (constitutional
monarchs) serve as the head of the state religion in a respective state, a unique institu-
tional innovation and legacy of British colonial rule. The institutional power of religious
86   Kikue Hamayotsu

authority and institutions, however, was extremely small in practice and largely ceremo-
nial at the onset. Malay political elites who formed the ethnic-​based nationalist party,
the United Malays National Organization (UMNO), to prepare for British-​sponsored
independence had no intention of building an independent state guided by Islam or
shariʿa. Nor did they have an interest or need to share powers with ʿulamaʾ, whom they
generally considered irrelevant in the building of a modern state. However, the Malay
elites depended on and needed the symbolic powers of Malay hereditary rulers to ne-
gotiate the constitutional superiority of the Malay community and their own political
domination in the new independent state (Fernando 2006). The official religion and
established-​religion regime were originally introduced, and gradually consolidated, not
because of strong Islamism or powerful clerical authority, but rather to strengthen the
state dominated by the ethnic Malay elites in an ethnically divided plural society and
economy.

Formation of the National Faith and Religious Authority

under the Secular State
The institutional foundations of the established-​religion regime—​reminiscent of the
Church of England under the British Crown—​were originally laid under British rule. In
stark contrast to the Dutch East Indies, the “divide and rule” policies in British Malaya
institutionalized and reinforced the cultural and religious authority of the Malay hered-
itary rulers within the indigenous Malay communities. Unlike the Dutch rulers, who
were either suspicious or hostile to traditional religious authority, and to Islam in ge-
neral, British officials considered the latter an integral source of political legitimacy and
stability, and of social order in the midst of a booming colonial economy and the rapid
social and demographic transformations that were underway.
Under the auspices of royal and imperial powers, the emerging secular state assumed
and sponsored religious functions and activities, thereby sanctioning an official inter-
pretation and jurisdiction of religion in the areas that secular officials considered the
purview of the colonial state. Each state government codified and legislated “Islamic
laws,” created shariʿa courts, and appointed mufti (Islamic supreme jurists) and shariʿa
judges to be integrated into the lower echelons of the colonial state and administrative
structures. Wealthy states and sultans such as Johor also founded and sponsored tradi-
tional religious schools for Malay children, who were generally very poor and lacked
access to modern education.6 Moreover, all the Malay rulers established a Council of
Religious Affairs to take care of general affairs of religion. One of the most important
functions assigned to state religious officials was to collect and manage zakat (Islamic
tithe) and waqf (religious endowment), essential financial sources to fund their modest
offices, salaries, and activities (Roff 1974; Yegar 1979).
The establishment of the official faith under colonial rule meant that the secular
state attained legitimacy and a near monopoly over the interpretation, codification,
 States, Religion, and Democracy in Southeast Asia     87

and regulation of religious matters, including law, education, and proselytization.

Importantly, the religious regime formed in British Malaya effectively hindered the
emergence of vibrant religious civil associations and movements, potentially chal-
lenging the monopoly of the religious market and civil society by state-​sanctioned re-
ligious authorities (Jesudason 1995). However, clerical authorities integrated into the
colonial state were typically underqualified loyal servants for the Malay monarchs. They
lacked the administrative and legal expertise and the financial base to wield influence,
in contrast to secular, English-​educated westernized Malay administrative and legal
elites who came to dominate the state after the peaceful departure of colonial rulers.
Overall, religious elites and officials were poorly funded and largely neglected in the
secular-​dominant state formation in the formative decades after independence. The
Malay nationalist elites, especially the first prime minister, Tunku Abdul Rahman, were
determined that an independent Malaysia be secular, and that state and religion be sep-
arate. However, they needed the symbolic and legitimizing powers of Malay hereditary
rulers to claim their political domination and entitlement to state powers and resources.
Having shared the same royal and aristocratic lineage with the hereditary rulers, Tunku
and other Malay elites had no intention to demolish or destabilize the feudalistic social
and political order centered on royal powers, of which they were an integral part. Thus,
they negotiated and granted a constitutionally privileged—​but merely ceremonial—​
position to Islam and Malay monarchs only to ensure the state was controlled by the
ethnic majority, with mostly immigrant and economically affluent ethnic minorities,
Chinese and Indian, as subordinate partners.

Expansion and Centralization of Religious Authority

under the Ethnic Party Regime
Party-​dominant autocratic rule under UMNO in the post-​racial riots of 1969 fur-
ther consolidated the established-​ religion regime, especially after Mahathir
Mohamad became prime minister (1981–​2003). The federal government—​the pow-
erful Prime Ministers’ Office in particular—​has built and centralized the religious
bureaucracies and functions to gain unprecedented authority and power to enforce
a range of “Islamization” policies in order to strengthen the ethnic-​based party re-
gime. Against the backdrop of growing religious fundamentalism in civil society at
home and abroad, the party-​dominant regime co-​opted one of the most influential
and charismatic Islamist leaders in the country, Anwar Ibrahim, into the ruling party
and government. Anwar and his men associated with the Malaysian Muslim Youth
Movement (ABIM) quickly rose up in the dominant party to achieve strategic and
powerful positions in government. Mahathir’s action was widely regarded as a re-
gime strategy to counter the electoral challenge of the Malay-​based Islamist party, the
Malaysian Islamic Party (PAS), within the increasingly self-​consciously Islamic Malay
populations (Jomo and Cheek 1988).
88   Kikue Hamayotsu

Unlike Suharto’s Indonesia, however, Malaysia’s party regime had to work within
the institutional framework of the established-​religion regime, wherein religious au-
thority was integrated into constitutional monarchic powers. Nine sultans, heads of
the state religion in a respective state, were easily defensive of their already dimin-
ished turf of influence and authority. They perceived that Malay politicians were ea-
gerly attempting to strip them of religious authority and power against the backdrop
of constitutional reforms intended to weaken monarchic power and authority under
Mahathir. In order to work around this constraint, the federal government invested
abundant political, administrative, and economic capital to centralize and expand re-
ligious bureaucracies and authority under the prime minister’s office. It amended the
constitution and enacted legislatures not only to upgrade the jurisdiction of shariʿa
courts and laws independently from—​and on a par with—​the secular civil courts, but
also to build an extensive network of modern Islamic schools and shariʿa-​based finan-
cial, corporate, welfare, and charity institutions across the country (Hamayotsu 2004,
2018; Sloane-​White 2017).
As a result, the party regime further strengthened and consolidated the established-​
religion regime inherited from the British Crown, turning the majority faith and reli-
gious authority into extremely powerful functionaries of, and ideological capital for,
the secular modern state. As the religious and ethnic identities are meshed to cement
the deep-​seated ethnic cleavage and politically salient communal identity, the legal and
living conditions of ethnic and religious minorities, and of nonbelievers, suffered ne-
glect and deterioration. Furthermore, powerful state-​sanctioned religious authorities
were equipped with extensive formal coercive powers to monitor, regulate, and crim-
inalize behaviors and movements they classified as “unorthodox,” “heretical,” or “for-
bidden” to certify legal and physical assaults, accusations, and discriminations against
gender and sexual minorities.

Transition to Democracy and Growing Religious

Nationalism and Conservatism
The dominant party regime collapsed unexpectedly through competitive and peaceful
elections in 2018, giving way to democratic rule, but the powerful established-​religion
regime remains intact. As discussed in the case of democratic Indonesia, the religious
regime continues to condition significant debates, laws, and policies that pertain to
democratic and civil rights, communal and minority rights, and religious freedom for
believers and nonbelievers alike (Hamayotsu 2012b; Moustafa 2018; Neo 2015; Shah and
Sani 2010). In Malaysia, it is the state-​sponsored religious authority, subordinate to sec-
ular elites, that has been among the most eager to elevate the position of Islam and inte-
grate and expand shariʿa within the state and legal structures. Their authority and powers
have been strengthened considerably so as to sponsor state-​formation dominated by the
Malay ruling elites and to consolidate the ethnic-​based party regime.
 States, Religion, and Democracy in Southeast Asia     89

Nascent democratic rule, brought about by an anti-​regime coalition of urban-​

based ethnic minorities and progressive Muslims, is perceived as a threat to the ide-
ological and institutional foundations of the ethnic-​based majoritarian rule and the
established-​religion regime. As a result, the political regime change has given rise
to interreligious tensions and aggressive pro-​shariʿa and anti-​minority nationalist
movements, not only in civil and political societies, but also within the hearts of state
and judicial bureaucracies dominated by politically and culturally conservative Muslim
elites. Newly elected Malay elites are not equipped or ready to tame radical and con-
servative Islamism, because their political survival and legitimacy is still dependent on
their ability and willingness to preserve the constitutional privileges granted to Islam
and the majority Muslim community. Promotion of minority rights—​especially sexual
minorities—​is especially damaging and costly for their political and regime survival.

Brunei: A Marriage of Political and

Religious Authorities in the Making
of a Religious Monarchy

Religious monarchic rule in Brunei is a relatively new institutional innovation, a result

of state-​formation almost exclusively sponsored by the absolute monarch (the house of
Bolkiah) to ensure sovereignty, legitimacy, and survival of his kingdom. The sultan is
the head of state and state religion (Islam), similar to the Malaysian sultans. However,
the regime’s religious characteristics have grown much greater under the current sultan,
Hassanal Bolkiah, who came to power in 1967 and has actively cultivated and centralized
religious as well as political authority. Under his personalistic rule, the state has built an
extensive religious apparatus to legislate and implement laws and policies not only to
run the government, but also to regulate the society and economy according to shariʿa.
In contrast to Indonesia’s secular-​dominant and Malaysia’s established-​religion regimes,
political and religious authorities here have been increasingly fused to form one nearly
undistinguishable and mutually reinforcing authority to rule this tiny oil-​rich autocratic
state. The regime is couched in the uniquely Islamic concept of sultanate. Brunei looked
as Malaysia’s established-​religion regime at independence, and in fact planned to join the
Malaysian federation. However, after three decades of royal-​sponsored state-​building, it
is almost identical to Saudi Arabia and other small old-​rich Arab monarchies. It is a re-
ligious regime, but not theocratic, differing from post-​revolution Iran or the Vatican,
where religious authority and rule are supreme and absolute. ʿUlamaʾ are extremely
powerful functionaries of the state, but their official positions and powers are ultimately
subordinate to the sultan.
The sultan is unconditionally above the law, with no other institutions or individuals
equipped to check or veto his absolute powers. The sultan monopolizes absolute
90   Kikue Hamayotsu

executive and legislative powers—​being prime minister; minister of finance, defense,

foreign affairs, and trade; chief commander of the army and police; and head of the
state religion (Islam) concurrently. His rule is not dictatorial, however. The government
emphasizes the rule of law and provides exceptionally generous welfare to his subjects.
He is popular, despite international condemnation against his initiatives to expand en-
forcement of shariʿa, including the stringent Islamic criminal codes known as hudud
(Müller 2016, 2017). Under this religious regime, dominant secular and clerical powers
are married and fused to run a powerful authoritarian state monopolizing the com-
manding heights of the economy—​oil and gas. The formation of the religious monar-
chic regime is primarily intended to achieve legitimacy and survival of hereditary rule
of the house of Bolkiah before (and after) Brunei gained full independence in 1984 from
British imperial rule.

Searching for Regime Survival: Political Origins

of Religious Monarchy
Brunei is ruled by one of the oldest dynasties and is one of the few surviving absolute
monarchies in the world today (Fanselow 2014, 90; Saunders 2013). It is a mistake to
assume that the sultanate of Brunei—​officially named Brunei Darussalam—​was rich,
powerful, and religious from its inception. On the contrary, the previous 28th sultan,
Omar Ali Saifuddien III (1950–​1967), who negotiated the country’s independence with
the British and Malaysian governments, did not have the intention or the ability to
build an independent modern state purely guided by shariʿa or Islam. His primary con-
cern was the survival of his tiny kingdom and vulnerable royal household, surrounded
by much larger and powerful neighbors, Malaysia, Indonesia, and the Philippines.
The sultan originally intended to join the Federation of Malaysia, then ruled by like-​
minded ethnic-​Malay Muslim aristocratic elites, to enter the Council of Malay Rulers
(or sultans) as an equal royal member and constitutional monarch of the federation. He
and a small assembly of Malay aristocrats in the Bruneian court were typically educated
in Britain or English-​medium elite boarding schools in Malaysia or Singapore. Just like
Malaysian royals and aristocratic elites, they were usually Anglophone and westernized
to behave just like British aristocrats (Milner 1981, 1994). Sultan Omar had the only one
wife, which is very rare for powerful Muslim kings and royals elsewhere. He was strongly
opposed to his son, then Crown Prince Hassanal Bolkiah, attempting to marry a second
wife when his son fell in love with a much younger Malay commoner, so as to safeguard
their pure royal blueblood and hereditary rule. Furthermore, the sultan was knighted to
officially carry the British nobility title conferred by the British Crown, exhibiting his af-
finity with, and allegiance to, the latter (Saunders 2013, chaps. 8–​10).
The sultan and his court were born royal, with a lineage said to be at least six centuries
old. In this regard, his claim to the throne and state powers was relatively unquestioned
among his Malay subjects. In practice, however, the sultanate was very poor and not
 States, Religion, and Democracy in Southeast Asia     91

even financially self-​sufficient to maintain his household, let alone to protect his tiny
territory and sovereignty, until a British petroleum company finally discovered oil and
natural gas in the 1970s. Because of the lack of the fiscal, administrative, military, and
security powers to maintain his household and government, and to protect his kingdom
and subjects, the sultan and his advisors were on the payroll of the British government
and heavily dependent on the British Crown as a protectorate for their immediate sur-
vival (Saunders 2013).
The sultan cared for the general well-​being of the Muslim community and Islam, sim-
ilar to the Malaysian sultans. The kingdom had traditionally been defined as an Islamic
state on paper, making Islam the central pillar of the ethnic-​and religious-​based na-
tional identity and state ideology; it was later propagated more officially as Melayu Islam
Beraja (Malay Islamic Monarchy, or MIB), a concept introduced under British imperial
rule.7 Originally, however, the national ideology was merely a vague and symbolic con-
cept to bolster royal sovereignty and legitimacy to rule his Malay subjects. It was prima-
rily meant to enshrine superordinate and absolute powers of the sultan as the guardian
of his Malay subjects, Malay culture, and Muslim faith. In contrast to Malaysia, other
ethnic and religious minorities, including the largest minority (Chinese, around 10 per-
cent), were not granted equal citizenship, let alone constitutional rights, because of this
iron rule of ethnic-​and religious-​based birthright to be legally recognized as citizens of
the religious monarchic state.

Bureaucratization and Expansion of the National Religion

and Ideology
The absence of an Islamist threat to the state and active religious civil society in the run-​
up to full independence, the state (sultan), subsidized by the British Crown, had built
the slowly emerging “national church,” or ecclesiastic bureaucracies, in order to legiti-
mize and strengthen the absolute powers of hereditary rule. The sultan is the head of the
state religion, but unlike his largely ceremonial Malaysian constitutional monarchs, the
current sultan, Hassanal Bolkiah, had actively expanded and sponsored the state appa-
ratus, which ran religious (Islamic) affairs, to turn religious institutions and clergies into
powerful functionaries of the absolute state.
The Ministry of Religious Affairs established in 1959 is exclusively responsible for all
matters considered to be the traditional purview of religious authority in Islam: building
and maintenance of mosques, managing pilgrimage, compulsory Islamic education,
propagation, alms, charities and social welfares, halal certification, and banks and finan-
cial institutions, among others. The ministry also manages the shariʿa judiciary, which
runs in parallel with the civil judiciary introduced by the British. The religious ministry
is among the largest government agencies (Müller 2018, 153–​155).
All Muslim clergies are loyal and favorite servants of the sultan, authorized
and sponsored by the state to carry out their religious missions and activities. The
92   Kikue Hamayotsu

Department of State Mufti (Islamic supreme jurist), created independently under the
prime minister’s (sultan’s) office, is the chief interpreter of Islam, holding the exclusive
authority to issue fatwas, which have legal effect (Müller 2018, 154).8 As the sultanate
became wealthier and self-​sufficient to build and manage its state apparatus, the cler-
ical officials employed by the state have become extremely powerful politically to in-
fluence public policies, and they enjoy generous budgets to expand massive religious
bureaucracies and operations, primarily because of their mutually beneficial relations
with the throne. For example, the Ministry of Religious Affairs has seen an increase of
its already generous budget when other strategic ministries, such as Defense, saw theirs
drastically cut when oil and gas prices plummeted to an alarmingly low level (Müller
2016, 163). Any competing autonomous clerical organizations, let alone anti-​clerical or
anti-​monarchic movements, almost disappeared due to the religious as well as political
markets solely monopolized and controlled by the absolute state.

Consolidation of Religious Monarchy under Sultan

Hassanal Bolkiah
The sultanate is generally committed to adherence to, and promotion of, moderate
Sunni Islamic traditions and watchful of infiltration of radical Salafi and Wahhabi
Islamism, as well as secular liberal traditions seen as alien and infidel by the religious
regime. For the sultan, it is essential to maintain the doctrinal and ideological cohe-
sion and purity of the Malay population and popular obedience to the state religious
authority and monarch. Doctrinal and ideological deviation from the state-​sanctioned
religious dogmas and codes is not simply a religious offense, but is considered to be sub-
versive to the national ideology, MIB, and the state (i.e., the sultan) and closely watched
by the state religious authorities (Müller 2018).
The current sultan has further strengthened religious authority and power of the state
to cultivate his popularity, legitimacy, and, ultimately, hereditary rule in the kingdom.
In 2014, Brunei enforced the Syariah Penal Code Order 2013, a legal reform intended
to bolster and reinforce shariʿa laws even further, to the detriment of secular civil laws.
The initiative stirred much debate and condemnation internationally, since it includes
the severest punishments and penalties for shariʿa-​derived criminal offenses while
tightening regulatory apparatus and procedures to enforce these criminal codes (Müller
2015, 2016).
This proposal to strengthen shariʿa rule and the religious regime and apparatus was
not a result of a penetration of puritanical or radical Islamism or increasing popular
Islamic activism and movements in civil society. Such activism in Brunei is close to non-
existent. But rather, it was the result of the mutually beneficial enterprise by the two pil-
lars of the religious monarchic regime—​monarch and ʿulamaʾ—​to consolidate further
cohesion and domination of the national identity, and the institutional and ideological
foundations of their absolute rule and survival.
 States, Religion, and Democracy in Southeast Asia     93

Conclusion

This chapter has accounted for the manners in which the three Muslim-​majority na-
tions in Southeast Asia have adopted and consolidated distinctive patterns of institu-
tional relations between political and religious authority to form diverging religious
regimes: the secular-​dominant regime of Indonesia, the established-​religion regime of
Malaysia, and the religious monarchic rule of Brunei. My comparative historical anal-
ysis highlights the analytical utility of the religious regime to demonstrate that the reli-
gious regime formation is distinctive from the formation and consolidation of political
regimes. It argues that the distinctive patterns of religious regime formation are the re-
sult of state-​formation, specifically the ways in which secular political elites pursued and
consolidated their state powers and domination following independence.
The religious regime formation is equally useful in accounting for the role of religious
authority—​and the ability and willingness of nascent Muslim-​dominant democracies—​
to protect constitutional rights and freedom of religious (and nonreligious) communities
and citizens. Denial or negligence of fundamental constitutional freedom may not be so
surprising in an absolute kingdom such as Brunei or Saudi Arabia. However, the de-
clining constitutional rights of religious (and nonreligious) communities and citizens
in a relatively successful consolidated democracy such as Indonesia are unexpected and
as yet unexplained (Hamayotsu 2014). Declining religious freedom and tolerance—​and
rising communal tension—​among various religious communities in a democratizing
Malaysia is equally perplexing, given its sizable non-​ Muslim urban middle-​ class
populations, modern economy, and absence of large-​scale communal violence since
1969. The discussion above shows that the declining constitutional rights and freedoms,
and the collective desires to elevate shariʿa and the majority faith, are not primarily
attributed to “Islam,” as the culturalists may like to insist, nor to political or electoral
pressures from Islamism or radicalism. Rather, it is the result of the timing of religious
regime formation, especially the manner in which the institutional relations between
political and religious authority are formed and consolidated.
In contrast to Christian-​dominant Europe, as well North America, where secular
regimes were formed concomitantly to facilitate democratic rule and rights and pro-
tect religious freedom, the religious regimes in Muslim-​dominant societies were formed
earlier and prior to the rise of modern democratic movements and rule. As a result,
more aggressive religious authorities and a religious fundamentalism tend to com-
plicate and compromise the state’s ability—​and willingness—​to protect democratic
rights and freedom of citizens, both religious and secular alike, even after democratic
consolidation.
Finally, but not least, the chapter echoes other comparative scholars and studies that
champion the theoretical and analytical value of institutional and ideological powers
of organized religion in political transformations and policymaking (e.g., Gill 2007;
Grzymala-​Busse 2012; Grzymała-​Busse 2015; Kalyvas 1996; Koesel 2014; Künkler and
94   Kikue Hamayotsu

Shankar 2018; Kuru 2009; Levine and Mainwaring 1989; Norris and Inglehart 2011;
Philpott 2007; Stepan 2001; Toft, Philpott, and Shah 2011). In Muslim societies, organ-
ized religion and its relations with the state will continue to shape and condition the po-
litical regimes and well-​being of citizens for many decades to come.

Notes
1. There are some country-​specific case studies, such as on Egypt, Tamir Moustafa, “Conflict
and Cooperation between the State and Religious Instituitons in Contemporary Egypt,”
International Journal of Middle East Studies 32, no. 1 (2000); Malika Zeghal, “Religion
and Politics in Egypt: The Ulema of Al-​Azhar, Radical Islam, and the State (1952–​94),”
International Journal of Middle East Studies 31, no. 3 (1999); or Pakistan and Malaysia, Seyyed
Vali Reza Nasr, Islamic Leviathan: Islam and the Making of State Power (Oxford: Oxford
University Press, 2001).
2. https://​ilga.org/​ilga-​riwi-​global-​attitudes-​survey.
3. The five precepts are: belief in the One and Only God; just and civilized humanity; unified
Indonesia; democracy guided by the wisdom of the representatives of the people; and social
justice for all Indonesians.
4. For a brief history of official religions, see https://​www.insideindonesia.org/​
the-​sixth-​religion.
5. The president readily invoked the law to outlaw one of the most prominent Islamist organi-
zations, Hizbut Tahrir Indonesia (HTI).
6. In predominantly Malay rural states such as Kedah, Kelantan, and Terengganu, local re-
ligious teachers ran their traditional religious schools supported by respective Malay-​
Muslim community.
7. MIB privileges Malay (Melayu) ethnic supremacy, Islam (as interpreted by the state),
and the monarchy (Beraja). For the subsequent institutionalization and indoctrina-
tion of the concept, see Dominik M. Müller, “Hybrid Pathways to Orthodoxy in Brunei
Darussalam: Bureaucratised Exorcism, Scientisation and the Mainstreaming of Deviant-​
Declared Practices,” Journal of Current Southeast Asian Affairs 37, no. 1 (2018): 150–​152.
8. A fatwa is a theological opinion and usually does not have a legal effect according to the
Islamic tradition.

References
Barber, Benjamin R. 1995. Jihad vs. MacWorld: How the Planet is Both Falling Apart and Coming
Together and What This Means for Democracy. New York: Times Book.
Benda, Harry J. 1965. “Political Elites in Colonial Southeast Asia: An Historical Analysis.”
Comparative Studies in Society and History 7, no. 3: 233–​251. doi:10.2307/​177791.
Boland, B. J. 1982. The Struggle of Islam in Modern Indonesia. The Hague: Martinus Nijhoff.
Buehler, Michael. 2016. The Politics of Shari’a Law: Islamist Activists and the State in
Democratizing Indonesia. Cambridge: Cambridge University Press.
Bush, Robin 2008. “Regional Sharia Regulations in Indonesia: Anomaly or Symptom?” In
Expressing Islam: Religious Life and Politics in Indonesia, edited by Greg Fealy and Sally
White, 174–​191. Singapore: Institute of Southeast Asian Studies.
 States, Religion, and Democracy in Southeast Asia     95

Bush, Robin. 2009. Nahdlatul Ulama and the Struggle for Power within Islam and Politics in
Indonesia. Singapore: Institute of Southeast Asian Studies.
Bush, Robin. 2015. “Religious Politics and Minority Rights during the Yudhoyono Presidency.”
In The Yudhoyono Presidency: Indonesia’s Decade of Stability and Stagnation, edited by
Edward Aspinall, Marcus Mietzner, and Dirk Tomsa, 239–​257. Singapore: Institute of
Southeast Asian Studies.
Calhoun, Craig, Mark Juergensmeyer, and Jonathan VanAntwerpen. 2011. Rethinking
Secularism. Oxford: Oxford University Press.
Cammack, Mark. 1989. “Islamic Law in Indonesia’s New Order.” International and Comparative
Law Quarterly 38, no. 1: 53–​73. doi: 10.2307/​760426.
Cammack, Mark. 1997. “Indonesia’s 1989 Religious Judicature Act: Islamization of Indonesia or
Indonesianization of Islam?” Indonesia 63: 143–​168. doi:10.2307/​3351514.
Cammack, Mark E., and R. Michael Feener. 2012. “The Islamic Legal System in Indonesia.”
Pacific Rim Law and Policy Journal 21, no. 1: 13.
Cribb, Robert. 1999. “Nation: Making Indonesia.” In Indonesia beyond Suharto: Polity, Economy,
Society, Transition, edited by Donald K. Emmerson, 3–​38. Armonk, NY: M.E. Sharpe.
Dahl, Robert A. 1986. Democracy, Liberty, and Equality Oslo: Norwegian University Press.
Dahl, Robert A. 1991. Democracy and Its Critics. New Haven, CT: Yale University Press.
Dalacoura, Katerina. 2005. “US Democracy Promotion in the Arab Middle East since 11
September 2001: A Critique.” International Affairs 81, no. 5: 963–​980.
Driessen, Michael D. 2009. “Religion, State, and Democracy: Analyzing Two Dimensions
of Church-​ State Arrangements.” Politics and Religion 3, no. 1: 55–​ 80. doi:10.1017/​
S1755048309990435.
Driessen, Michael D. 2010. “Religion, State, and Democracy: Analyzing Two Dimensions
of Church-​ State Arrangements.” Politics and Religion 3, no. 1: 55–​ 80. doi:10.1017/​
S1755048309990435.
Driessen, Michael D. 2014. Religion and Democratization: Framing Religious and Political
Identities in Muslim and Catholic Societies. Oxford: Oxford University Press.
Elson, R. E. 2009. “Another Look at the Jakarta Charter Controversy of 1945.” Indonesia
88: 105–​130.
Elson, R. E. 2010. “Nationalism, Islam, ‘Secularism’ and the State in Contemporary Indonesia.”
Australian Journal of International Affairs 64, no. 3: 328–​343.
Elson, R. E. 2012. “Absent at the Creation: Islamism’s Belated, Troubled Engagement with Early
Indonesian Nationalism.” In Anthony Reid and the Study of the Southeast Asian Past, 303–​
335. Singapore: Institute for Southeast Asian Studies.
Elson, R. E. 2013. “Two Failed Attempts to Islamize the Indonesian Constitution.”
Sojourn: Journal of Social Issues in Southeast Asia 28, no. 3: 379–​437.
Fanselow, Frank. 2014. “The Anthropology of the State and the State of Anthropology in
Brunei.” Journal of Southeast Asian Studies 45, no. 1: 90–​112. doi:10.1017/​S0022463413000611.
Fealy, Greg. 2009. “Indonesia’s Islamic Parties in Decline.” Inside Story, May 11. https://​
insidestory.org.au/​indonesias-​islamic-​parties-​in-​decline/​.
Federspiel, Howard M. 1970. “The Muhammadijah: A study of an Orthodox Islamic Movement
in Indonesia.” Indonesia 10: 57–​79.
Feener, R. Michael. 2013. Shari’a and Social Engineering: The Implementation of Islamic Law in
Contemporary Aceh, Indonesia : Oxford: Oxford University Press.
Feener, R. Michael, and Mark E. Cammack, eds. 2007. Islamic Law in Contemporary
Indonesia: Ideas and Institutions. Cambridge, MA: Harvard University Press.
96   Kikue Hamayotsu

Feillard, Andrée. 1996. “Traditionalist Islam and the Army in Indonesia’s New Order: The
Awkward Relationship.” In Nahdlatul Ulama, Traditional Islam and Modernity, edited by
Greg Barton and Greg Fealy, 42–​67. Clayton, Australia: Monash Asia Institute.
Feillard, Andrée. 1997. “Traditionalist Islam and the State in Indonesia: The Road to
Legitimacy and Renewal.” In Islam in an Era of Nation-​States: Politics and Religious Renewal
in Muslim Southeast Asia, edited by Robert W. Hefner and Patricia Horvatich, 129–​154.
Honolulu: University of Hawai‘i Press.
Feith, Herbert, and Lance Castles, eds. 1970. Indonesian Political Thinking 1945–​1965. Ithaca,
NY: Cornell University Press.
Fernando, Joseph M. 2006. “The Position of Islam in the Constitution of Malaysia.” Journal of
Southeast Asian Studies 37, no. 2: 249–​266. doi:10.2307/​20072709.
Formichi, Chiara. 2012. Islam and the Making of the Nation: Kartosuwiryo and Political Islam in
20th Century Indonesia. Leiden: KITLV Press.
Fox, Jonathan. 2006. “World Separation of Religion and State into the 21st Century.”
Comparative Political Studies 39, no. 5: 537–​569. doi:10.1177/​0010414005276310.
Fox, Jonathan. 2008. A World Survey of Religion and the State. Cambridge: Cambridge
University Press.
Fox, Jonathan, and Shmuel Sandler. 2005. “Separation of Religion and State in the Twenty-​First
Century: Comparing the Middle East and Western Democracies.” Comparative Politics 37,
no. 3: 317–​335. doi:10.2307/​20072892.
Gill, Anthony. 2007. The Political Origins of Religious Liberty. Cambridge: Cambridge
University Press.
Grzymala-​ Busse, Anna. 2012. “Why Comparative Politics Should Take Religion
(More) Seriously.” Annual Review of Political Science 15, no. 1: 421–​ 442. doi:10.1146/​
annurev-​polisci-​033110-​130442.
Grzymała-​Busse, Anna. 2015. Nations under God: How Churches Use Moral Authority to
Influence Policy: Princeton, NJ: Princeton University Press.
Hamayotsu, Kikue. 2004. “Islamization, Patronage and Political Ascendancy: The Politics and
Business of Islam in Malaysia.” In The State of Malaysia: Ethnicity, Equity and Reform, edited
by Edmund Terence Gomez, 229–​252. London: Routledge-​Curzon.
Hamayotsu, Kikue. 2011a. “The End of Political Islam? A Comparative Analysis of Religious
Parties in the Muslim Democracy of Indonesia.” Journal of Current Southeast Asian Affairs
30, no. 3: 133–​159.
Hamayotsu, Kikue. 2011b. “Political Rise of the Prosperous Justice Party in Post-​Authoritarian
Indonesia: Examining the Political Economy of Islamist Mobilization in a Muslim
Democracy.” Asian Survey 51, no. 5: 971–​992.
Hamayotsu, Kikue. 2012a. “A House for All Indonesian Muslims? The United Development
Party is Determined to Survive as a Political Force.” Inside Indonesia 108. https://​www.
insideindonesia.org/​a-​house-​for-​all-​muslims?highlight=WyJoYW1heW90c3UiXQ%3D%3D
Hamayotsu, Kikue. 2012b. “Once a Muslim, Always a Muslim: The Politics of State Enforcement
of Syariah in Contemporary Malaysia.” Southeast Asia Research 20, no. 3: 399–​421.
Hamayotsu, Kikue. 2013. “The Limits of Civil Society in Democratic Indonesia: Media Freedom
and Religious Intolerance.” Journal of Contemporary Asia 43, no. 4: 658–​677. doi:10.1080/​
00472336.2013.780471.
Hamayotsu, Kikue. 2014. “Conservative Turn? Religion, State and Conflict in Indonesia.”
Pacific Affairs 87, no. 4: 815–​825.
 States, Religion, and Democracy in Southeast Asia     97

Hamayotsu, Kikue. 2018. “The Political Origins of Islamic Courts in Divided Societies: The
Case of Malaysia.” Journal of Law and Religion 33, no. 2: 248–​270.
Harvey, Barbara Sillars. 1977. Permesta: Half a Rebellion: Ithaca, NY: Cornell Modern Indonesia
Project, Southeast Asia Program, Cornell University.
Hefner, Robert W. 1993. “Islam, State and Civil Society: ICMI and the Struggle for Indonesian
Middle Class.” Indonesia 56: 1–​36.
Hefner, Robert W. 2000. Civil Islam: Muslims and Democratization in Indonesia. Princeton,
NJ: Princeton University Press.
Hoesterey, James Bourk. 2015. Rebranding Islam: Piety, Prosperity, and a Self-​Help Guru.
Stanford, CA: Stanford University Press.
Horikoshi, Hiroko. 1975. “The Darul-​Islam Movement in West Java (1948–​62): An Experience
in the Historical Process.” Indonesia 20: 59–​86. doi:10.2307/​3350996.
Huntington, Samuel P. 1993. “The Clash of Civilizations?” Foreign Affairs 72, no. 3: 22–​48.
Inglehart, Ronald. 2000. “Culture and Democracy.” In Culture Matters: How Values Shape
Human Progress, edited by Lawrence E. Harrison and Samuel P. Huntington, 80–​97.
New York: Basic Books.
Jesudason, James V. 1995. “Statist Democracy and the Limits of Civil Society in Malaysia.”
Journal of Commonwealth and Comparative Politics 33, no. 3: 335–​356.
Jomo, Kwame Sundaram, and Ahmed Shabery Cheek. 1988. “The Politics of Malaysia’s Islamic
Resurgence.” Third World Quarterly 10, no. 2: 843–​868.
Jung, Eunsook. 2009 Taking Care of the Faithful: Islamic Organizations and Partisan
Engagement in Indonesia. PhD diss. Department of Political Science, University of
Wisconsin-​Madison.
Kahin, Audrey. 2013. “Natsir and Sukarno: Their Clash over Nationalism, Religion and
Democracy, 1928–​1958.” In Encountering Islam: The Politics of Religious Identities in Southeast
Asia, edited by Yew-​Foong Hui, 191–​217. Singapore: Institute of Southeast Asian Studies.
Kalyvas, Stathis N. 1996. The Rise of Christian Democracy in Europe. Ithaca, NY: Cornell
University Press.
Kalyvas, Stathis N. 2000. “Commitment Problems in Emerging Democracies: The Case of
Religious Parties.” Comparative Politics 32, no. 4: 379–​398.
Kalyvas, Stathis N. 2003. “Unsecular Politics and Religious Mobilization: Beyond Christian
Democracy.” In European Christian Democracy: Historical Legacies and Comparative
Perspectives, edited by Thomas Kselman and Joseph A. Buttigieg, 293–​320. Notre Dame,
IN: University of Notre Dame Press.
Kepel, Gilles. 1997. Allah in the West: Islamic Movements in America and Europe. Translated by
Susan Milner. Stanford, CA: Stanford University Press.
Koesel, Karrie J. 2014. Religion and Authoritarianism: Cooperation, Conflict, and the
Consequences. Cambridge: Cambridge University Press.
Kselman, Thomas, and Joseph A. Buttigieg, eds. 2003. European Christian Democracy: Historical
Legacies and Comparative Perspectives. Notre Dame, IN: University of Notre Dame Press.
Kuipers, Nick. 2019. Who’s Running on Islam in Indonesia? New Mandala, March 26. https://​
www.newmandala.org/​whos-​running-​on-​islam-​in-​indonesia/​.
Künkler, Mirjam, and Julia Leininger. 2009. “The Multi-​Faceted Role of Religious Actors
in Democratization Processes: Empirical Evidence from Five Young Democracies.”
Democratization 16, no. 6: 1058–​1092.
98   Kikue Hamayotsu

Künkler, Mirjam, John Madeley, and Shylashri Shankar, eds. 2018. A Secular Age be-
yond the West: Religion, Law and the State in Asia, the Middle East and North Africa.
Cambridge Studies in Social Theory, Religion and Politics. Cambridge: Cambridge
University Press.
Künkler, Mirjam, and Shylashri Shankar. 2018. “Introduction.” In A Secular Age beyond
the West: Religion, Law and the State in Asia, the Middle East and North Africa, edited by
John Madeley, Mirjam Künkler and Shylashri Shankar, 1–​ 32. Cambridge: Cambridge
University Press.
Kuru, Ahmet T. 2007. “Passive and Assertive Secularism: Historical Conditions, Ideological
Struggles, and State Policies toward Religion.” World Politics 59, no. 4: 568–​594.
Kuru, Ahmet T. 2009. Secularism and State Policies toward Religion: The United States, France
and Turkey. New York: Cambridge University Press.
Laffan, Michael Francis. 2003. Islamic Nationhood and Colonial Indonesia: The Umma below
the Winds. London and New York: RoutledgeCurzon.
Levine, Daniel H., and Scott Mainwaring. 1989. “Religion and Popular Protest in Latin
America: Contrasting Experiences.” In Power and Popular Protest: Latin American Social
Movements, edited by Susan Eckstein, 203–​240. Berkeley: University of California Press.
Liddle, William R. 1996. “The Islamic Turn in Indonesia: A Political Explanation.” Journal of
Asian Studies 55, no. 3: 613–​634.
Mamdani, Mahmood. 2004. Good Muslim, Bad Muslim: America, the Cold War, and the Roots
of Terror. New York: Pantheon Books.
McVey, Ruth. 1983. “Faith as the Outsider: Islam in Indonesian Politics.” In Islam in the Political
Process, edited by James P. Piscatori, 199–​225. Cambridge: Cambridge University Press.
Menchik, Jeremy. 2016. Islam and Democracy in Indonesia: Tolerance without Liberalism.
Cambridge: Cambridge University Press.
Menchik, Jeremy. 2018. “Soft Separation Democracy.” Politics and Religion 11, no. 4: 863–​883.
doi:10.1017/​S1755048318000329.
Milner, Anthony. 1981. “Islam and Malay Kingship.” Journal of the Royal Asiatic Society of Great
Britain and Ireland 1: 46–​70.
Milner, Anthony. 1994. The Invention of Politics in Colonial Malaya: Contesting Nationalism and
the Expansion of the Public Sphere. Cambridge: Cambridge University Press.
Moustafa, Tamir. 2018. Constituting Religion: Islam, Liberal Rights, and the Malaysian State.
Cambridge Studies in Law and Society. Cambridge: Cambridge University Press.
Moustafa, Tamir. 2000. “Conflict and Cooperation between the State and Religious Institutions
in Contemporary Egypt.” International Journal of Middle East Studies 32, no. 1: 3–​22.
Müller, Dominik M. 2015. “Brunei.” In Keeping the Faith: A Study of Freedom of Thought,
Conscience, and Religion in ASEAN, edited by David Cohen and Kevin Tan, 54–​98.
Depok: West Java Human Rights Resource Centre.
Müller, Dominik M. 2016. “Brunei in 2015: Oil Revenues Down, Sharia on the Rise.” Asian
Survey 56, no. 1: 162–​167. doi:10.1525/​as.2016.56.1.162.
Müller, Dominik M. 2017. “Brunei Darussalam in 2016: The Sultan Is Not Amused.” Asian
Survey 57, no. 1: 199–​205. doi:10.1525/​as.2017.57.1.199.
Müller, Dominik M. 2018. “Hybrid Pathways to Orthodoxy in Brunei
Darussalam: Bureaucratised Exorcism, Scientisation and the Mainstreaming of Deviant-​
Declared Practices.” Journal of Current Southeast Asian Affairs 37, no. 1: 141–​183. doi:10.1177/​
186810341803700106.
 States, Religion, and Democracy in Southeast Asia     99

Nasr, Seyyed Vali Reza. 2001. Islamic Leviathan: Islam and the Making of State Power
Oxford: Oxford University Press.
Neo, Jaclyn L. 2015. “Competing Imperatives: Conflicts and Convergences in State and Islam in
Pluralist Malaysia.” Oxford Journal of Law and Religion 4, no. 1: 1–​25.
Norris, Pippa, and Ronald Inglehart. 2011. Sacred and Secular: Religion and Politics Worldwide.
Cambridge: Cambridge University Press.
Peletz, Michael G. 2002. Islamic Modern: Religious Courts and Cultural Politics in Malaysia.
Princeton, NJ: Princeton University Press.
Pepinsky, Thomas B., William R. Liddle, and Saiful Mujani. 2012. “Testing Political Islam’s
Economic Advantage: The Case of Indonesia.” American Journal of Political Science 56, no.
3: 584–​600.
Pepinsky, Thomas B., R. William Liddle, and Saiful Mujani. 2018. Piety and Public
Opinion: Understanding Indonesian Islam. Oxford: Oxford University Press.
Philpott, Daniel. 2007. “Explaining the Political Ambivalence of Religion.” American Political
Science Review 101, no. 3: 505–​525. doi:10.2307/​27644463.
Roff, William R. 1974. “The Origin and Early Years of the Majlis Ugama.” In Kelantan: Religion,
Society and Politics in Malay State, edited by William R. Roff, 101–​52. Kuala Lumpur: Oxford
University Press.
Roy, Olivier. 2007. Secularism Confronts Islam. Translated by George Holoch.
New York: Columbia University Press.
Saunders, Graham. 2013. A History of Brunei. London and New York: RoutledgeCruzon.
Shah, Dian Abdul Hamed, and Mohd Azizuddin Mohd Sani. 2010. “Freedom of Religion in
Malaysia: A Tangled Web of Legal, Political, and Social Issues.” North Carolina Journal of
International Law and Commercial Regulation 36, no. 3: 647–​688.
Shah, Timothy Samuel, Alfred Stepan, and Monica Duffy Toft. 2012. Rethinking Religion and
World Affairs. Oxford: Oxford University Press.
Sloane-​White, Patricia. 2017. Corporate Islam. Cambridge: Cambridge University Press.
Stepan, Alfred. 2001. “The World’s Religious Systems and Democracy: Crafting the ‘Twin
Tolerations.’” In Arguing Comparative Politics, 213–​253. Oxford: Oxford University Press.
Stepan, Alfred. 2011. “The Multiple Secularisms of Modern Democratic and Non-​democratic
Regimes.” In Rethinking Secularism, edited by Craig Calhoun, Mark Juergensmeyer, and
Jonathan VanAntwerpen, 114–​144. Oxford: Oxford University Press.
Tessler, Mark. 2000. “Democratic Concern and Islamic Resurgence: Converging Dimensions
of the Arab World’s Political Agenda.” In Democracy and Its Limits: Lessons from Asia, Latin
America, and the Middle East, edited by Howard Handelman and Mark Tessler, 262–​289.
Notre Dame, IN: University of Notre Dame Press.
Tessler, Mark. 2002. “Islam and Democracy in the Middle East: The Impact of Religious
Orientation on Attitudes toward Democracy in Four Arab Countries.” Comparative Politics
34, no. 3: 337–​354.
Tibi, Bassam 2008. “Islamist Parties: Why They Can’t Be Democratic.” Journal of Democracy 19,
no. 3: 43–​48.
Tibi, Bassam. 2012. Islamism and Islam. New Haven, CT: Yale University Press.
Toft, Monica Duffy, Daniel Philpott, and Timothy Samuel Shah. 2011. God’s Century: Resurgent
Religion and Global Politics. New York: W. W. Norton.
van Dijk, C. 1996. “Ulama and Politics.” Bijdragen tot de Taal-​, Land-​en Volkenkunde 152, no.
1: 109–​143. doi:10.2307/​27864720.
100   Kikue Hamayotsu

Yegar, Moshe. 1979. Islam and Islamic Institutions in British Malaya: Politics and Implementation.
Jerusalem: Magnes Press.
Zeghal, Malika 1999. “Religion and Politics in Egypt: the Ulema of al-​Azhar, Radical Islam, and
the State (1952–​94).” International Journal of Middle East Studies 31, no. 3: 373–​375.
 Chapter 5

Re pression of I sl a mi sts
and Au thori ta ria n
Survival in t h e A ra b
Worl d
A Case Study of Egypt

Jean Lachapelle

Anger at police brutality played a central role in ousting the Egyptian dictator Hosni
Mubarak in 2011. A few months before the uprising that toppled Mubarak, a picture
of the mangled face of Khaled Said, a young Alexandrian murdered by police officers,
had been circulated on social media, triggering deep outrage among middle-​class
Egyptians. Such a vivid illustration of the regime’s brutality sparked a wave of protest
among young urbanites in Cairo and Alexandria beginning in June 2010. “We are all
Khaled Said” was the title of the Facebook page created to commemorate the victim, a
digital space that became a forum for citizens concerned with the repressive excesses of
the Mubarak regime. The online forum would become a crucial coordinating tool for
the historic January 25, 2011, protests that launched a country-​wide uprising (Clarke and
Kocak 2020).
Tragically, a revolt primarily motivated by indignation against arbitrary repression
led to the establishment of a regime more repressive than its predecessor. Egypt’s brief
democratic experiment after Mubarak’s ouster was interrupted by a military coup in
July 2013 led by then-​Defense Minister Abdel Fattah al-​Sisi. The coup was followed by
high levels of lethal violence against protesters, accompanied by the Egyptian judiciary’s
sentencing of hundreds of people to death for a range of political crimes.
Since 2013 this new authoritarian regime has engaged in intense repression just as
brutal—​if not more so—​than under Mubarak. For instance, between July 2013 and
April 2014, over 40,000 alleged members of the Muslim Brotherhood were detained
or arrested. The Egyptian judiciary sentenced hundreds to death, and security forces
102   Jean Lachapelle

killed several hundred in street clashes (Human Rights Watch 2014; Rutherford 2018,
195–​196). Yet the violence in the early years after the coup did not trigger the same waves
of anti-​regime protest and indignation as before.
This chapter answers two interrelated questions regarding these events in Egypt,
which shed broader light on the nature of Islamist repression across the Arab world.
After the 2013 coup, why did state violence increase so much in Egypt? Why, despite its
scale, did this repression not backfire as it had in the past? One possible answer is that
the violence of the new regime acted as a deterrent for protests. Influential scholarship
explains low levels of popular mobilization under authoritarianism by pointing to the
risks inherent in political activism in undemocratic spaces (Kuran 1991). According to
this view, the absence of significant anti-​regime protests in the early years of the Sisi re-
gime might be due to the regime’s repressiveness and the real risks of violence or death
that citizens would incur for public demonstrations. However, we also know that repres-
sion often promotes further mobilization rather than diminishing it (Davenport 2007;
Francisco 2004; Sullivan and Davenport 2017), and that repression may decrease dissent
in some situations but increase it in others (Davenport 2007, 8). Given how repression
can have opposite effects, an argument focusing mainly on deterrence seems insufficient
for explaining low levels of mobilization.
This chapter argues that fear alone cannot explain low mobilization among non-​
Islamists against the new regime after Egypt’s 2013 coup. On the contrary, less mobili-
zation occurred because the post-​coup regime was popular among some segments of
the population, who saw its repressive nature as necessary for thwarting the perceived
Islamist threat. Non-​Islamists were willing to accept repression if it meant wiping the
Muslim Brotherhood off the political map. Such popular support for repression shaped
the political calculation of the new regime, and drastically decreased the costs of repres-
sion it would face.
This chapter offers a new perspective for understanding the repression of Islamists
that extends beyond Egypt to the Arab world more generally. It emphasizes the role of
non-​Islamist audiences, defined as civilian groups who do not subscribe to the Islamists’
political project, in shaping the costs and benefits of repression for Arab autocrats.
Taking seriously the role that non-​Islamist audiences play enriches how existing schol-
arship addresses the causes of repression against Islamists, which has primarily focused
on either the level of direct threat that Islamists pose to autocratic rulers—​a threat that
arises from Islamists’ presumed ideological appeal or organizational capacity—​or the
coercive capacity of Arab states. By contrast, I demonstrate that explaining when and
why Arab rulers repress Islamists requires paying greater attention to the attitudes of
non-​Islamists, who constitute civilian audiences whose support authoritarian regimes
can seek to cultivate.
This chapter is organized as follows. First, I review existing scholarship that explains
repression in authoritarian regimes through three lenses: threat perception, state
capacity, and elite competition. Second, I explain the significance of non-​Islamist
audiences and show how they conditioned repression against the Muslim Brotherhood
in Egypt and helped the post-​2013 regime gain popularity. I provide a case study of the
 Repression of Islamists and Authoritarian Survival    103

non-​Islamist media discourse and provide evidence for civilian support for repression
against the Muslim Brotherhood based on an analysis of opinion pieces published in
Al-​Masry al-​Youm, Egypt’s main privately owned newspaper. Focusing on the pe-
riod between July 3, 2013—​the day of the coup—​and August 14, 2013—​the day of a
large-​scale massacre of supporters of the Muslim Brotherhood in Cairo—​this anal-
ysis demonstrates how prominent non-​Islamist voices expressed consistent support
for a hard line against the protesters who mobilized in support of ousted president
and Muslim Brotherhood member Mohamed Morsi. I further suggest that such sup-
port from non-​Islamists boosted the new regime’s confidence that repression would
not backfire, thus contributing to the government’s choice to use massive repression.
Finally, I briefly examine other episodes of repression in the Arab world and clarify how
my argument travels beyond Egypt.

Existing Literature

When and why do authoritarian regimes use repression against Islamist groups?
Existing scholarship in political science points to three sets of explanations.
The first set of arguments focuses on the level of threat that an opposition group poses
to a regime (Davenport 2007). Islamist groups are viewed as posing a unique threat to
regimes because such groups are understood as cohesive, disciplined, and capable of
mobilizing large numbers of followers. Strong organizational strength has roots in so-
cialization and recruitment processes. Recruits to prominent Islamist organizations
such as the Egyptian Muslim Brotherhood often go through a lengthy socialization,
which ensures “solidarity and commitment among its members” (al-​Anani 2016, 68).
The organization tightly controls who moves up its ranks (Kandil 2015, 48–​53) and
shapes a member’s entire social life. Given these characteristics, regimes may repress
Islamist organizations to eliminate the threat they pose. For instance, Carrie Wickham
(2013, 112) explains the Mubarak regime’s repression of the Muslim Brotherhood in the
mid-​2000s as follows: “Although the Brotherhood was not the first or boldest group to
challenge the Mubarak regime, it was the largest and best organized, making it both a
greater threat and an easier target.”
Another threat-​based argument focuses on the content of Islamists’ political mes-
sage. Islamists are often considered to have a unique ability to craft and disseminate
messages that resonate powerfully among Muslim populations because they tap into
extant civilizational repertoires of religious obligations and political conduct that can
be harnessed for mass mobilization (Lewis 1988). As Mark Tessler (2011, 47) argues,
“Islam has shaped the Arabs’ history, helps to define their collective national identity,
and gives spiritual meaning to the lives of millions, even many who are not personally
devout.” From this perspective, an ability to evoke shared belief systems gives Islamists
a presumed advantage in mobilizing supporters relative to non-​Islamists, which makes
the former potentially more threatening.1
104   Jean Lachapelle

Islamists’ threats to Arab autocrats may derive from the former’s proximity and privi-
leged connection to ordinary citizens and local communities, which makes them strong
contenders, especially during elections (Brooke 2017; Cammett and Issar 2010; Cammett
and Jones 2014). For instance, Tarek Masoud’s (2014) study of Egypt shows that Islamists
operating under Mubarak’s regime had access to a dense network of mosques that
allowed them to better convey electoral messages to voters than competing leftist parties
which lacked access to such networks. The delivery of healthcare services through ro-
bust infrastructures, such as hospitals and clinics operated by Islamists, also sustained
an image of “competence and probity” that helped Islamists at the polls (Brooke 2017,
42).2 The fact that states cannot easily dismantle such networks gives Islamists a pow-
erful advantage in recruiting members capable of challenging the regime (Cammett and
Jones 2014, Clark 2004; Wickham 2002). It follows that regimes repress Islamists to pre-
vent them from winning elections. Lisa Blaydes (2011) has shown how, in Egypt under
Mubarak, security forces ramped up repression against the Muslim Brotherhood in the
lead-​up to elections to reduce their electoral gains.
According to this view, authoritarian regimes in the Arab world have different abilities
to neutralize Islamists’ political messages. Rulers that explicitly claim religious legiti-
macy, such as the king of Morocco, might be better equipped than regimes founded on
secular ideologies at thwarting threats that Islamists pose. By contrast, for the Arab re-
publics such as Egypt, Syria, and Iraq, Islamists might pose a more severe threat because
they offer a fundamental alternative to the regime’s foundational ideologies of Arab na-
tionalism. Scholars have argued that Islamists began posing an acute challenge to these
regimes once their foundational ideologies failed to deliver on initial promises of ec-
onomic and political emancipation (Ajami 1981; Burgat 1995; Kepel 2000; Roy 1994).
Secular authoritarian rulers may use repression against Islamists to prevent them from
deploying powerful religious symbols that they are ill-​equipped to counter in other ways
(Gerges 2018; Sassoon 2016).
The second type of argument for explaining authoritarian repression focuses on the
state. Prominent literature in political science shows that state capacity shapes patterns
of repression in myriad ways. For instance, states with low coercive capacity are unable
to collect accurate intelligence about dissidents, which encourages indiscriminate re-
pression (Greitens 2016; Policzer 2009). Given the secrecy of Islamist organizations,
their tight-​knit nature, and their embeddedness in local communities, Arab regimes
have faced great difficulty monitoring them. For example, Saddam Hussein’s regime
in Iraq struggled with “[g]‌athering accurate intelligence about the activities of reli-
gious activists” belonging to the Daʿwa Party because the party was “insular, made of
individuals with family ties and marriage connections” (Blaydes 2018, 264). Lacking
capacity to monitor and collect information about Islamist organizations and their ac-
tivities, the regime resorted to collective punishment that included mass killings and
incarceration. Such indiscriminate repression in turn strengthened group identities in
ways that further deepened the problem of monitoring dissidents.
Finally, a third approach to explaining repression centers on competition among
coercive agencies. According to this argument, repression reflects coercive agencies’
 Repression of Islamists and Authoritarian Survival    105

efforts to demonstrate their political indispensability to the autocrat by uncovering and

thwarting dangerous plots. The result is a form of “coercive outbidding” that results in
indiscriminate repression (Greitens 2016). An especially vivid illustration of this phe-
nomenon is a large-​scale violent crackdown against the Egyptian Muslim Brotherhood
during the mid-​1960s. In 1965, then-​president Gamal Abdel Nasser ordered mass arrests
of the Islamist organization. The Military Intelligence Department (MID) claimed
it had uncovered a plot by the Muslim Brotherhood to assassinate Nasser. Research
drawing on biographies suggests that the MID’s primary motivation for leveling these
accusations against the Brotherhood was to undermine its bureaucratic rival, the
General Investigations Directorate (GID), Egypt’s secret police under the direction
of the minister of interior. By exposing a dangerous plot to assassinate the president
that the secret police had allegedly missed, the MID sought to expose the Ministry of
Interior’s incompetence and show that the GID had been asleep at the wheel (Kandil
2014, 75). Large-​scale repression of the Muslim Brotherhood ensued.
In sum, existing scholarship addresses state repression against Islamists through
three lenses, focusing on threats, state capacity, and institutional competition. While
such studies provide valuable insight into the role of Islamist organizations’ strength
and the state, they tend to overlook the influence of social forces and especially non-​
Islamist groups that may share an authoritarian regime’s worry about Islamists. In the
next section, I discuss how such non-​Islamist audiences affect repression.

The Role of Non-​Islamist Audiences

The term “non-​Islamist audiences” refers to civilian groups that do not subscribe to
the Islamists’ political project. I argue that these groups’ attitudes influence an author-
itarian ruler’s calculation of the costs and benefits of using repression against Islamist
organizations. When non-​Islamists support the repression of Islamists, the negative
consequences to regimes’ legitimacy are weak. By neutralizing the perceived threat that
the Islamists pose, regimes present themselves as defenders of stability and can signal to
non-​Islamist audiences a willingness to protect them. Indeed, repression may increase
an authoritarian regime’s popularity when non-​Islamists perceive Islamists as posing an
acute threat.
Non-​Islamist attitudes toward Islamists are not constant. Therefore, regimes’ costs
and benefits of repressing vary over time. When non-​Islamists have felt threatened
by Islamists, they have been more likely to support repression against them. Consider
Egypt in the 1990s. The country experienced waves of violent attacks by armed Islamist
groups, including in densely populated areas in Cairo. In this context of high volatility,
“Legal opposition elites from both the right and the left . . . turned a blind eye to the re-
pression of Islamists. In some cases, they even appeared to welcome it” (Lust-​Okar 2005,
148). Violence against Islamists provoked little outrage and “[o]‌pposition papers seemed
to have a “blackout” on reports of human rights violations against Islamists” (148).
106   Jean Lachapelle

When such fears are high, an autocratic ruler can repress in order to signal to non-​
Islamists that he is serious about dealing with an Islamist threat. Repression may serve
a dual purpose for the regime: not only does it destroy a problematic opposition group,
it also corrals support among civilians who share a fear of that opposition group. As
I argue elsewhere, this is the situation that Egypt experienced in the summer of 2013,
when generalized hostility created opportunities for the new regime to legitimize itself
in the eyes of non-​Islamists. This combination provided the prime conditions for using
repression to build a support base after the coup, or what I call a “legitimation strategy”
of repression (Lachapelle 2020).
Understanding the role of non-​Islamist audiences, and how regimes seek to signal
their resolve to them, attunes scholars to the indirect effects of repression upon civilians
who are not its direct targets. Repression can help authoritarian regimes enhance their
popularity by signaling to non-​Islamist groups that the regime is willing and capable
of providing order and security during chaotic times. In Egypt, repression in 2013
demonstrated to non-​Islamist audiences that the new post-​coup regime was capable of
eliminating the Islamist threat.
This argument builds on important studies that have underscored the significance of
divides between Islamists and non-​Islamists in the Arab world (Albrecht 2005; Buehler
2018; Lust-​Okar 2005; Shehata 2010). Notably, Ellen Lust (2005) demonstrated how
regimes have been able to prevent the joint mobilization of Islamists and non-​Islamists
through selective repression. By creating “divided structures of contention” that punish
non-​Islamists for joining hands with Islamists, autocrats can thwart the emergence of a
cross-​ideological front. These divide-​and-​conquer strategies were used by the Mubarak
regime to keep opposition forces divided (Albrecht 2005). My argument advances such
scholarship by theorizing the costs and benefits of these different repressive strategies
for the autocrat.
In the next section, I illustrate the role of non-​Islamist audiences and how an author-
itarian regime can use repression to build support, using a case study of Egypt. It shows
the limits of existing arguments that focus on direct threats to the regime for explaining
the scale of violence after the coup of 2013. Rather, as I will demonstrate, the regime
repressed the Muslim Brotherhood to build its popularity and support base among ci-
vilian groups that were fearful of the Islamist organization.

Case Study: Repression after the Coup

of 2013 in Egypt

The period after the coup of July 3, 2013, provides a particularly clear illustration of the
role of non-​Islamist audiences for repression. The coup overthrew Mohamed Morsi, a
democratically elected president and member of the Muslim Brotherhood, ushering in a
period of unprecedented levels of repression in Egypt that included mass incarceration,
 Repression of Islamists and Authoritarian Survival    107

the use of military trials against civilians, and enforced disappearances (Amnesty
International 2016; Human Rights Watch 2014; Rutherford 2018). The use of lethal force
against protesters became frequent after July 3, 2013, in contrast with the Mubarak era,
when such use of lethal force was rare. The most infamous episode of violence against
protesters after the coup of July 3, 2013, is the so-​called the “Rabaa Massacre” of August
14, 2013, when security forces dispersed protesters sitting in front of the Rabaa al-​
Adawiya mosque in Cairo, killing about a thousand supporters of the ousted president
Morsi in less than twelve hours.3
Understanding the extraordinarily high levels of repression that occurred in Rabaa
Square requires paying attention to the non-​Islamist audiences that the post-​2013 regime
faced after the coup. It was not the case that the Muslim Brotherhood posed a greater
threat to the new regime than to the Mubarak regime. This Islamist organization’s pop-
ularity was at a historical low. Nationally representative data from the Arab Barometer
surveys reveal that trust toward the Muslim Brotherhood had dramatically fallen be-
tween 2011 and 2013. Whereas 32 percent of respondents answered that they “absolutely
did not trust” the Muslim Brotherhood in the spring of 2011 (Arab Barometer Survey
Wave II), this number was 68 percent in the spring of 2013, shortly before Morsi was
ousted (Arab Barometer Survey Wave III). Compared with earlier periods of Islamist
political mobilization, the Muslim Brotherhood in 2013 was not in a position to mobilize
broad-​based support, as it had lost much of its popular appeal (Masoud 2014, 209–​215).
Indeed, if the Islamists’ popularity had rested upon a reputation for competence (Brooke
2017; Cammett and Jones 2014), then the deterioration of the economy and services that
Egyptians experienced under Morsi severely harmed the Islamists’ reputation. In other
words, at the time, the Muslim Brotherhood’s message did not resonate broadly across
Egyptian society; and it was not in a strong position to challenge the post-​coup govern-
ment and provoke a major wave of protests.
On the contrary, the regime was able to repress the Muslim Brotherhood largely be-
cause of widespread animosity against it. The regime could gain popularity in the eyes of
non-​Islamists by being tough on the Islamist organization. This strategy involved both
promoting a narrative portraying the Muslim Brotherhood as an existential threat to
Egypt and demonstrating the regime’s own ability to destroy this threat before worried
audiences. By emphasizing the threat that the Brotherhood posed, the new regime did
not invent narratives out of thin air, but rather drew upon a repertoire of already existing
anxieties about the Muslim Brotherhood as aiming to capture the Egyptian state and
establish a theocracy (Hellyer 2017; Nugent 2020). Indeed these anxieties had been con-
siderably heightened during Morsi’s presidency (2012–​2013).
Widespread fears of the Muslim Brotherhood among non-​Islamists and shared
concerns that its continued post-​coup mobilization was leading the country to the brink
of civil war helped people believe in the value of a strongman being in charge. In turn,
these popular sentiments affected the costs of repression for the regime. When the re-
gime committed the August 14, 2013, massacre, there was little public outrage, contrary
to what literature on backlash mobilization would lead us to expect (Francisco 2004). As
Kira Jumet accurately notes about the public reaction to the massacre, “Large segments
108   Jean Lachapelle

of the Egyptian public reacted to the Rabaa al-​Adawiya massacre with enthusiastic ap-
proval of the government’s action” (Jumet 2018, 205).
One way to examine the sentiments of non-​Islamists audiences during that period
is through popular opinion pieces written in the Egyptian press. I scraped all opinion
pieces published in the privately owned newspaper Al-​Masry al-​Youm between July 3
and August 14, 2013 (from the day of the coup to the day of the Rabaa massacre). Among
the 852 articles that I obtained, I identified those discussing the Rabaa sit-​in using an
automated textual analysis following the method developed by Roberts, Stewart, and
Airoldi (2016) and made available through the stm package in R. I ran a structural topic
model of seven topics on all unigrams and bigrams of stemmed words, and the stem-
ming was done using the arabicStemR package (Nielsen 2017). The anti-​coup mobi-
lization by the Muslim Brotherhood stood out as a topic, through the high frequency
of words such as “Rabaa” “sit-​in,” and “Ikhwan.” I kept articles with a topic proportion
greater than 0.5, for a total of 132 articles.
This analysis shows high levels of expressed support for repressing the Muslim
Brotherhood. Contributors to Al-​Masry al-​Youm made repeated calls for the govern-
ment to take a hard-​line approach against the Islamist protesters. They demanded that
security forces end the sit-​in and warned against seeking political reconciliation with
Islamists. Whereas the regime certainly promoted anti-​Brotherhood narratives in the
media, these declarations reflected genuine sentiment. Large swaths of the non-​Islamist
intelligentsia had supported the coup and celebrated the Muslim Brotherhood’s fall from
power. Such attitudes often extended to support for the government’s hard-​line position.
One article cautioned against seeking reconciliation with the Muslim Brotherhood and
warned the organization, “[I]‌f you seek a safe exit, I can assure you it will not happen.”4
Another columnist told people “not to cry” for “an armed group that has armed militias
and threatens the country’s safety.”5 Another wrote that “the police must deal with this
outpost [the sit-​in] as it deals with criminals and murderers. Such banditry is a violation
of the law, and it should not be tolerated,” adding that “the responsibility of the police
towards the people is to protect them from advocates of violence and bandits.”6
Similarly to Al-​Masry al-​Youm, most media outlets promoted a hard-​line approach
against the Muslim Brotherhood and protesters at the Rabaa sit-​in. Well-​known
intellectuals accused proponents of moderation of being unpatriotic. For instance, on
the day of the Rabaa massacre, famous novelist and Mubarak critic Alaa al-​Aswani
tweeted to his two million followers: “In Egypt now there is a people, a government,
a police force and an army in the confrontation with an armed terrorist group that is
committing the most heinous of crimes for power. There is no middle ground. Either
with Egypt or with terrorism” (quoted in Hefny 2018, 154–​155).
The regime also encouraged mass demonstrations in support of repression against
the Brotherhood. In late July, Sisi, who was still minister of defense at the time, had pub-
licly asked the population to give him a “mandate to confront terrorism.”7 Large num-
bers of people responded to this call, expressing their support by mobilizing in large
crowds across Egypt. These demonstrations gave visible expression to popular support
for forceful measures against the Islamist organization. Shortly after these protests, the
 Repression of Islamists and Authoritarian Survival    109

government approved a plan for the removal of protesters in Rabaa and Nahda Squares.
Public declarations made by Egyptian officials after the massacre suggest the authorities
expected the likely number of casualties to be in the thousands (Human Rights Watch
2014, 102).
There were few efforts to downplay the scale of the violence after the massacre. The
official number of casualties announced by the Health Ministry’s Forensic Medical
Authority was 627 deaths, a number comparable to the estimates of Human Rights
Watch.8 Prime Minister al-​Beblawy told a local newspaper that he estimated the
number of casualties from the clearing of key sites in Cairo where Muslim Brotherhood
supporters were repressed—​namely, the Rabaa and Nahda sit-​ins—​to be almost 1,000.9
Whereas many authoritarian regimes go to great lengths to try to hide their uses of vio-
lence, the Egyptian government in the summer of 2013 did not, which is consistent with
the signaling argument advanced in this chapter.
Since this incident, Egypt’s regime has detained tens of thousands of people, and the
courts have issued hundreds of death sentences in mass trials. In addition, local NGOs
as well as Amnesty International have documented hundreds of disappearances.10

Beyond Egypt

Understanding the role of non-​Islamist audiences in Egypt after the 2013 coup invites
scholars to rethink prominent episodes of repression across the Arab world, where au-
thoritarian regimes have targeted Islamist opposition groups. In Syria, the regime of
Hafez al-​Assad adopted especially drastic measures against the Muslim Brotherhood.
In the early 1980s, Assad’s regime made membership in the organization a criminal of-
fense that was punishable by death and killed large numbers in a crackdown to suppress
a Brotherhood-​led armed uprising in the city of Hama (Lefèvre 2013, 115). Indeed, the
Syrian regime besieged the city, and its repression resulted in over ten thousand civilian
casualties (Lefèvre 2013, 128). In Tunisia during the early 1990s, the regime of Ben Ali
rounded up several thousand Islamists and sentenced them to heavy prison sentences
(Shahin 1997, 101). Algeria has similarly witnessed the mass repression of Islamic op-
position groups, such as the Front Islamique du Salut (FIS) after a coup in 1992 that
annulled the legislative elections (McDougall 2017, 278). Likewise, in Iraq in 1982,
Saddam Hussein’s regime executed and imprisoned large numbers of the Daʿwa Party in
the city of Dujail after a failed assassination attempt against him (Shanahan 2004, 946).
At the same time, repression has not been the only way that regimes across the Arab
world have dealt with Islamists. In Jordan, the regime has allowed the Islamic Action
Front (IAF), the political arm of the Muslim Brotherhood, to participate in elections
since the legalization of political parties in 1992 (Boulby 1999, 117; Schwedler 2006,
52) and repression against the Muslim Brotherhood, the country’s most powerful
Islamist organization, has been rare. Events such as when the Jordanian regime arrested
four members of Parliament belonging to the IAF in 2006 have been the exception
110   Jean Lachapelle

Table 5.1: Repression against Islamist in the Arab World since 1980

Country Years Episode

Algeria 1982 Arrest of about 1,000 Islamist activists

1991 Arrest of leaders of the Front Islamique du Salut (FIS)
1992 Arrests of thousands of members of the FIS
1992–​2000 Counterinsurgency against armed Islamist groups
Egypt 1987 Hundreds of arrests and detentions against the Muslim
Brotherhood
1995 Over 900 arrests against the Muslim Brotherhood
1992–​2000 Over 50,000 arrests in counterinsurgency campaign against
the Islamic Group
2007–​2008 Hundreds of arrests against the Muslim Brotherhood
2013–​2014 Killing of hundreds, jailing of thousands against the Muslim
Brotherhood
Jordan 2006 Arrest of 4 parliamentarians from the Islamic Action Front
Morocco 1983 Arrest of leader of al-​ʿAdl wal-​Ihsan, Abdsalam Yassine
1984 Arrest of 71 members of al-​Shabiba al-​Islamiya
1992 Arrests of Guidance Bureau of al-​ʿAdl wal-​Ihsan and
followers
2006–​2008 Arrests of about 5,000 members of al-​ʿAdl wal-​Ihsan
Syria 1982 10,000 to 40,000 killed in Hama massacre
Iraq 1982 Over one hundred killed in Dujail massacre
Tunisia 1981 Arrest of over 60 leaders of the Mouvement de la Tendance
Islamique (MTI)
1987 About 3,000 arrests against the MTI
1990–​1992 About 8,000 arrests against Ennahda

Sources: Algeria: McDougall (2017, 278) and Hafez (2003, 80); Egypt: El-​Ghobashy (2005, 379), Hafez
(2003, 90), and International Crisis Group (2008, 10); Jordan: Hamid (2013, 548); Morocco: Shahin
(1997, 187–​88, 194–​95), Salgon (2012, 20); Syria: Lefèvre (2013, 128); Tunisia: Shahin (1997, 87,
101) and Jones (1988).

rather than the rule (Hamid 2013, 548). In Morocco, the regime has more often en-
gaged in targeted repression of Islamists, such as the arrests of rank-​and-​file and lead-
ership of the al-​ʿAdl wal-​Ihsan movement, a charity movement that has refrained from
participating in formal politics. However, the Moroccan regime has also allowed the
country’s main Islamist party, the Parti de la Justice et du Développement, to participate
in elections and to hold cabinet positions (Buehler 2018). Table 5.1 summarizes some of
 Repression of Islamists and Authoritarian Survival    111

these prominent episodes of state repression in Morocco, Algeria, Tunisia, Egypt, Iraq,
Jordan, and Syria since 1980.11
In short, repression against Islamists in the Arab world has not been a constant, but
has been subject to variations in timing and place across the Arab world. To understand
these patterns of repression, one must look beyond typical threat-​and state-​centered
approaches and consider a broader range of actors. When Arab autocrats repressed
Islamists, they dealt not only with their direct targets, but also indirectly with non-​
Islamists audiences. Pointing to the dangers that Islamists posed to civilian audiences
would often prove a powerful way of justifying repression and diminishing repression
costs. How much such messages resonated among non-​Islamists audiences has varied,
shaping the cost and benefits of repression for autocrats.

Conclusion

This chapter has demonstrated the role that non-​Islamist audiences play in shaping the
repression of Islamists through a case study of Egypt. Whereas existing studies focus
on threat perception, coercive state capacity, and institutional competition to explain
repression, I argue that non-​Islamist audiences shape repression against Islamists. The
experience of Egypt since 2013 illustrates how a new authoritarian regime that came to
power through a military coup has used repression to garner the support of non-​Islamist
audiences that feared the Muslim Brotherhood. Beyond Egypt, this argument helps ex-
plain patterns of violence against Islamists across the Arab world. When regimes could
not justify the repression of Islamists to domestic audiences, they tempered their re-
pression. However, when non-​Islamist audiences acquiesced to or encouraged the re-
pression of Islamists—​because they believed that the Islamists threatened the country’s
national character or security—​authoritarian regimes could repress Islamists at little
political cost.

Notes
1. For an insightful discussion of the Islamists’ electoral advantage see Cammett and Jones
(2014).
2. A party’s Islamist designation may also offer an “informational shortcut that communicates
something positive about a party’s policy intentions” and translates into electoral gains
(Pepinsky, Liddle, and Mujani 2012, 586).
3. Human Rights Watch, which conducted an in-​depth investigation of these events, counted
904 casualties in the two locations, but stresses that the true number is most likely “well over
1,000 protesters.” Human Rights Watch, 2014, “All According to Plan: The Rab’a Massacre
and Mass Killings of Protesters in Egypt,” August 12, 2014, https://​www.hrw.org/​report/​
2014/​08/​12/​all-​according-​plan/​raba-​massacre-​and-​mass-​killings-​protesters-​egypt.
112   Jean Lachapelle

4. https://​www.almasryalyoum.com/​news/​details/​198391.
5. https://​www.almasryalyoum.com/​news/​details/​198842.
6. https://​www.almasryalyoum.com/​news/​details/​198615.
7. https://​www.ft.com/​content/​41af5f60-​f435-​11e2-​a62e-​00144feabdc0.
8. Human Rights Watch (2014).
9. https://​today.almasryalyoum.com/​article2.aspx?ArticleID=396463&IssueID=2985, cited
in Human Rights Watch (2014).
10. https://​w ww.amnesty.org/​en/​latest/​news/​2016/​07/​egypt-​hundreds-​disappeared-​and-​
tortured-​amid-​wave-​of-​brutal-​repression/​.
11. I include campaigns against armed militant groups only if these groups established a
stronghold in urban areas, such as the Islamic Group’s insurgency in Egypt in the 1990s,
which managed to establish a stronghold in the Cairo suburbs of Imbaba, or the Muslim
Brotherhood’s uprising in the city of Hama in Syria in 1982. The list does not include mil-
itary operations against armed Islamist movements that occurred in border regions, such
as Egypt’s Sinai, or the clashes between the Algerian state and numerous armed groups op-
erating in the Sahara regions.

References
Ajami, Fouad. 1981. The Arab Predicament: Arab Political Thought and Practice since 1967.
Oxford: Oxford University Press.
Albrecht, Holger. 2005. “How Can Opposition Support Authoritarianism? Lessons from
Egypt.” Democratization 12, no. 3: 378–​397.
al-​Anani, Khalil. 2016. Inside the Muslim Brotherhood. Oxford: Oxford University Press.
Amnesty International. (2016). Egypt: “Officially , You Do Not Exist”: Disappeared and Tortured
in the Name of Counter-​Terrorism. London: Amnesty International.
Blaydes, Lisa. 2011. Elections and Distributive Politics in Mubarak’s Egypt. New York: Cambridge
University Press.
Blaydes, Lisa. 2018. State of Repression: Iraq under Saddam Hussein. Princeton, NJ: Princeton
University Press.
Boulby, Marion. 1999. The Muslim Brotherhood and the Kings of Jordan, 1945–​ 1993.
Atlanta: Scholars Press.
Brooke, Steven. 2017. “From Medicine to Mobilization: Social Service Provision and the
Islamist Reputational Advantage.” Perspectives on Politics 15, no. 1: 42–​61. doi:10.1017/​
S1537592716004126.
Buehler, Matt. 2018. Why Alliances Fail: Islamist and Leftist Coalitions in North Africa. Syracuse,
NY: Syracuse University Press.
Burgat, François. 1995. L’islamisme en face. Paris: La Découverte.
Cammett, Melani, and Sukriti Issar. 2010. “Bricks and Mortar Clientelism: Sectarianism and
the Logics of Welfare Allocation in Lebanon.” World Politics 62, no. 3: 381–​421. doi:10.1017/​
S0043887110000080.
Cammett, Melani, and Pauline L. Jones. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17: 187–​206. doi:10.1146/​annurev-​polisci-​071112-​221207.
Clark, Janine. 2004. “Social Movement Theory and Patron-​ Clientelism: Islamic Social
Institutions and the Middle Class in Egypt, Jordan, and Yemen.” Comparative Political
Studies 37, no. 8: 941–​968.
 Repression of Islamists and Authoritarian Survival    113

Clarke, Killian, and Korhan Kocak. 2020. “Launching Revolution: Social Media and the
Egyptian Uprising’s First Movers.” British Journal of Political Science 50, no. 3: 1025–​1045.
Davenport, Christian. 2007. “State Repression and Political Order.” Annual Review of Political
Science 10, no. 1: 1–​23.
El-​
Ghobashy, Mona. 2005. “The Metamorphosis of the Egyptian Muslim Brothers.”
International Middle East Journal 37, no. 3: 373–​395.
Francisco, Ronald. 2004. “After the Massacre: Mobilization in the Wake of Harsh Repression.”
Mobilization 9, no. 2: 107–​126.
Gerges, Fawaz. 2018. Making the Arab World: Nasser, Qutb, and the Clash That Shaped the
Middle East. Princeton, NJ: Princeton University Press.
Greitens, Sheena Chestnut. 2016. Dictators and their Secret Police: Coercive Institutions and
State Violence. New York: Cambridge University Press.
Hafez, Mohammed M. 2003. Why Muslims Rebel: Repression and Resistance in the Islamic
World. Boulder, CO: Lynne Rienner.
Hamid, Shadi. 2013. “The Islamic Action Front in Jordan.” In The Oxford Handbook of
Islam and Politics, edited by John L. Esposito and Emad El-​Din Shahin. Oxford: Oxford
University Press.
Hefny, Mostafa. 2018. “The Material Politics of Revolution and Counter-​Revolution: Labor
Organization, Autonomy and Democratization in Egypt (2011–​2016).” PhD diss., Columbia
University.
Hellyer, H. A. 2017. “A Revolution Undone: Egypt’s Road beyond Revolt.” Oxford: Oxford
University Press.
Human Rights Watch. 2014. “All According to Plan: The Rab’a Massacre and Mass Killings of
Protesters in Egypt.” Human Rights Watch, August 12, 2014. https://​www.hrw.org/​report/​
2014/​08/​12/​all-​according-​plan/​raba-​massacre-​and-​mass-​killings-​protesters-​egypt.
International Crisis Group. 2008. “Egypt’s Muslim Brothers: Confrontation or Integration?”
Brussels: International Crisis Group, June 2008. https://​www.crisisgroup.org/​middle-​east-​
north-​africa/​north-​africa/​egypt/​egypt-​s-​muslim-​brothers-​confrontation-​or-​integration.
Jones, Linda G. 1988. “Portrait of Rachid al-​Ghannouchi.” Middle East Report 153 (July/​August).
https://​merip.org/​1988/​07/​portrait-​of-​rachid-​al-​ghannouchi/​.
Jumet, Kira D. 2018. Contesting the Repressive State: Why Ordinary Egyptians Protested during
the Arab Spring. Oxford: Oxford University Press.
Kandil, Hazem. 2014. Soldiers, Spies and Statesmen: Egypt’s Road to Revolt. 2nd ed.
New York: Verso.
Kandil, Hazem. 2015. Inside the Brotherhood. Cambridge: Polity Press.
Kepel, Gilles. 2000. Jihad: Expansion et déclin de l’islamisme. Paris: Gallimard.
Ketchley, Neil. 2014. “‘The Army and the People Are One Hand!’ Fraternization and the 25th
January Egyptian Revolution.” Comparative Studies in Society and History 56, no. 1: 155–​186.
Kuran, Timur. 1991. “Now Out of Never: The Element of Surprise of 1989 European Revolution.”
World Politics 44, no. 1: 7–​48.
Lachapelle, Jean. 2020. “Repression Reconsidered: State Violence and Legitimation in
Authoritarian Regimes.” Working paper.
Lefèvre, Raphaël. 2013. Ashes of Hama: The Muslim Brotherhood in Syria. Oxford: Oxford
University Press.
Levitsky, Steven, and Lucan A. Way. 2010. Competitive Authoritarianism: Hybrid Regimes after
the Cold War. New York: Cambridge University Press.
Lewis, Bernard. 1988. The Political Language of Islam. Chicago: University of Chicago Press.
114   Jean Lachapelle

Lust-​Okar, Ellen. 2005. Structuring Conflict in the Arab World: Incumbents, Opponents and
Institutions. New York: Cambridge University Press.
Masoud, Tarek. 2014. Counting Islam: Religion, Class, and Elections in Egypt.
New York: Cambridge University Press.
McDougall, James. 2017. A History of Algeria. New York: Cambridge University Press.
Nielsen, Richard. 2017. Deadly Clerics: Blocked Ambition and the Paths to Jihad.
New York: Cambridge University Press.
Nugent, Elizabeth R. 2020. “The Psychology of Repression and Polarization.” World Politics 72,
no. 2: 291–​334. doi:10.1017/​S0043887120000015.
Pepinsky, Thomas B., R. William Liddle, and Saiful Mujani. 2012. “Testing Islam’s Political
Advantage: Evidence from Indonesia.” American Journal of Political Science 56, no. 3: 584–​
600. doi:10.1111/​j.1540-​5907.2011.00570.x.
Policzer, Pablo. 2009. The Rise and Fall of Repression in Chile. South Bend, IN: University of
Notre Dame Press.
Roberts, Margaret E., Brandon M. Stewart, and Edoardo M. Airoldi. 2016. “A Model of Text
for Experimentation in the Social Sciences.” Journal of the American Statistical Association
111: 988–​1003. doi:10.1080/​01621459.2016.1141684.
Roy, Olivier. 1994. The Failure of Political Islam. Cambridge, MA: Harvard University Press.
Rutherford, B. K. 2018. “Egypt’s New Authoritarianism under Sisi.” Middle East Journal 72, no.
2: 185–​208.
Salgon, Jean-​Michel. 2012. Dictionnaire de l’islamisme au Maghreb. Paris: L’Harmattan.
Sassoon, Joseph. 2016. Anatomy of Authoritarianism in the Arab Republics.
New York: Cambridge University Press.
Schwedler, Jillian. 2006. Faith in Moderation: Islamist Parties in Jordan and Yemen.
New York: Cambridge University Press.
Shahin, Emad Eldin. 1997. Political Ascent: Contemporary Islamic Movements in North Africa.
Boulder, CO: Westview Press.
Shanahan, Rodger. 2004. “Shiʿa Political Development in Iraq: The Case of the Islamic Daʿwa
Party.” Third World Quarterly 25, no. 5: 943–​954.
Shehata, D. 2010. Islamists and Secularists in Egypt: Opposition, Conflict and Cooperation.
London: Routledge.
Sullivan, Christopher M., and Christian Davenport. 2017. “The Rebel Alliance Strikes
Back: Understanding the Politics of Backlash Mobilization.” Mobilization 22, no. 1: 39–​56.
Tessler, Mark. 2011. “The Origins of Popular Support for Islamist Movements: A Political
Economy Analysis.” In Public Opinion in the Middle East: Survey Research and the Political
Orientations of Ordinary Citizens, edited by Mark Tessler, 25–​55. Bloomington: Indiana
University Press.
Wickham, Carrie R. 2002. Mobilizing Islam: Religion, Activism, and Political Change in Egypt.
New York: Columbia University Press.
Wickham, Carrie R. 2013. The Muslim Brotherhood: Evolution of an Islamist Movement.
Princeton, NJ.: Princeton University Press.
 Chapter 6

Regime T y pe s , Re g i me
Transitions , a nd
Re ligion in Pa ki sta n

Matthew J. Nelson

How does religion shape regime types, and regime transitions, in Muslim-​majority
states? In this chapter I examine the effect of religion on transitions between military
and civilian-​led regimes in Pakistan, noting that both types of regimes have generally
embraced the same religious constitutional provisions. In Pakistan, these provisions
outline different rights for Muslims and non-​Muslims, and since 1973 they have also re-
stricted speech said to infringe on “the glory of Islam.” Although Pakistan has experi-
enced six major regime transitions since 1947, few of these transitions have been driven
by religious groups or religious ideas. Religion rarely drives the process of regime tran-
sition in Pakistan. At the same time, however, I note that ostensibly religious constitu-
tional provisions have underpinned a series of “exclusionary” and “illiberal” regimes.
Since the formation of Pakistan in 1947, civilian-​led regimes have been removed in
three military coups—​in 1958, 1977, and 1999. Only one of these (1977) highlighted reli-
gious concerns.1 Military-​led regimes were later removed following widespread protests
on three occasions—​in 1969–​1970, 1988, and 2007–​2008. These protests, however,
featured nonreligious more than religious demands. While each protest-​based tran-
sition away from military rule suggested a push in the direction of “democracy,” the
regimes that subsequently emerged were characterized by (a) reserved domains of mil-
itary decision-​making (e.g., security policy) and (b) religious constitutional provisions
restricting free speech and diminishing the rights of non-​Muslims. As such, I do not
describe the civilian-​led regimes that emerged in the wake of each transition away from
military rule as broadly inclusive liberal democracies.
An inclusive liberal democracy is characterized by civilian power obtained via reg-
ular, competitive, free-​and-​fair elections based on universal adult suffrage as well as
a firm constitutional commitment to (a) the separation of powers (i.e., horizontal ac-
countability) and (b) fundamental individual rights without exceptions for particular
116   Matthew J. Nelson

groups. Within the civilian-​led regimes of Pakistan, however, policymaking domains

reserved for noncivilians (e.g., enduring military control over security policy) have
pushed against these parameters. Furthermore, limitations on freedom of religion and
freedom of speech (so-​called “illiberal” restrictions), as well as limitations targeting
particular groups (i.e., “exclusionary” restrictions), have remained in place with wide-
spread popular support. Even beyond persistent forms of military tutelage, in other
words, civilian-​led regimes in Pakistan have been detached from strong demands for
liberal democracy.
The presence or absence of popular pressure for an “inclusive” or “liberal” democracy
is, of course, an empirical question. In Azerbaijan and Kyrgyzstan, Kathleen Collins and
Erica Owen (2012) found that, while religious affiliation (e.g., Muslim/​non-​Muslim) did
not shape individual levels of attachment to democracy, higher levels of Muslim “religi-
osity” (i.e., higher levels of individual religious commitment or more extensive forms of
individual religious practice) did suggest a preference for “Islamic” more than “liberal”
democracy.2 Frédéric Volpi (2004, 1061, 1069) has articulated a similar argument. He
noted that Muslims who express an appreciation for democracy might simultaneously op-
pose “liberal” democracy, turning instead toward what Volpi called an Islamic “pseudo-​
democracy” defined by a focus on community or state-​based mobilization for religious
values (i.e., a “positive” focus on religious virtue) rather than widespread popular support
for the “negative” constitutional protections commonly associated with liberal democracy
(i.e., limitations on state power in favour of individual constitutional rights).
In this chapter, an inclusive “liberal” democracy differs from an “Islamic” democracy
insofar as the latter is associated with freely elected civilian-​led regimes in which var-
ious expressions of Islam (i.e., specific interpretations of Islam) are politically, legally,
and constitutionally salient even if those interpretations are neither liberal nor inclu-
sive with respect to individual rights and/​or nondiscrimination at the level of particular
groups. I do not argue that Islam is inevitably incompatible with liberal democracy. My
argument is simply that, in Pakistan, elected civilian-​led regimes have long embraced il-
liberal and exclusionary constitutional provisions broadly associated with that country’s
prevailing interpretations of Islam.
It is possible that grass-​roots electoral support for nonsecular politics could en-
courage leaders who articulate a liberal reading of their own religious tradition (i.e.,
religious leaders who use religious ideas to defend nondiscrimination as well as an ap-
preciation for fundamental individual freedoms vis-​à-​vis religion or speech). Fazlur
Rahman (1982) in Pakistan and Abdolkarim Soroush (Sadri and Sadri 2002) in Iran
famously stressed the possibility of pro-​liberal politics within the Islamic tradition.
But, historically, grass-​roots support for electoral but nonsecular politics in Pakistan
has rarely overlapped with widespread support for such leaders; in fact, Pakistan’s first
wave of pro-​democracy activism (1969–​1970) coincided with Fazlur Rahman’s forced
departure into exile. Support for civilian elected regimes in Pakistan has long coexisted
with popular support for illiberal and exclusionary forms of ostensibly “Islamic”
constitutionalism.
Assessing multiple features of Pakistan’s political and constitutional history, this
chapter locates Pakistan within a much broader literature on regime types and regime
 Regime Types, Regime Transitions, and Religion in Pakistan    117

transitions, arguing that while nonreligious grass-​roots protests figure prominently in

Pakistani transitions away from authoritarianism, broadly democratizing transitions
tend to preserve a “Pakistani” approach to Islamic constitutional provisions, pulling away
from an inclusive liberal democracy. As such, this chapter builds on Jan Teorell’s (2010,
156) call for a greater emphasis on “the subjective experience of . . . [democratizing] ac-
tors,” highlighting an appreciation for what elected civilian-​led government in Pakistan
has actually meant for those seeking to introduce and constitutionalize it.

Pakistan: Regime Types and

Regime Transitions

Tracking regime transitions requires a sense of the regime types from which, and to
which, those transitions occur. Increasingly, scholars with an interest in regime transi-
tion have moved beyond a simple account of “democracies” and “dictatorships” (Alvarez
et al. 1996) to examine a wider range of regime types. At the relatively open or com-
petitive end of the scale, social and even secular democracies are said to include sub-
stantive elements that press beyond the procedural core associated with Robert Dahl’s
definition of “polyarchy” (Dahl 1971, 1982; Diamond 1999). At the more constrained or
authoritarian end of the scale, Samuel Huntington (1991) and Barbara Geddes (1999)
have noted that transitions often involve a shift between several different types of auto-
cratic leaders, including (a) personalistic leaders (e.g., monarchs), (b) military leaders,
and (c) single-​party leaders. Most regimes, however, are no longer characterized as
“open” democracies or “closed” dictatorships. Instead, they are categorized as hybrid
regimes, including so-​called “defective” democracies (Collier and Levitsky 1997; Merkel
2004; O’Donnell 1994; Zakaria 1997) and “electoral” autocracies (Brumberg 2002;
Levitsky and Way 2002; Schedler 2006) that combine electoral procedures with var-
ious autocratic constraints (Bogaards 2009; Gilbert and Mohseni 2011; Morlino 2009;
Wigell 2008).
Making no a priori normative claims about secular/​nonsecular, liberal/​illiberal, or in-
clusive/​exclusionary regimes, I examine four different modes of regime transition in
Pakistan (1947–​2019), highlighting the analytical importance of various “hybrid” regimes:

1. outright military dictatorship → “hybrid” autocracy (aka electoral autocracy)

2. “hybrid” autocracy → outright military dictatorship (generally involving constitu-
tional abrogation/​suspension; no elections)
3. outright military dictatorship → “hybrid” democracy (e.g., exclusionary/​illiberal
democracy; generally including reserved domains of military power)
4. “hybrid” democracy → outright military dictatorship (e.g., military coup)

Outright military dictatorships originating in a military coup (No. 4) are easy to iden-
tify in Pakistan; they occurred in 1958, 1977, and 1999. Each coup, however, was followed
118   Matthew J. Nelson

by a further transition from outright dictatorship to some type of “hybrid” autocracy

(No. 1). For scholars like Levitsky and Way (2002, 16), or Andreas Schedler (2013, 29, 82),
“hybrid” electoral-​authoritarian regimes require competitive party-​based elections for
both the executive (e.g., the presidency) and the national legislature. But, in Pakistan,
presidential polls occasionally involve noncompetitive referenda rather than party-​
based elections. And, under both military and civilian-​led regimes, Pakistan has also
conducted several “indirect” polls in which nonparty local elections produce an elec-
toral college for subsequent National Assembly (NA) and/​or Provincial Assembly (PA)
elections. As such, none of Pakistan’s military regimes has met the definitional standard
set by Levitsky, Way, and Schedler for an “electoral-​authoritarian” regime.
Within Pakistan’s military-​led regimes, party-​based NA/​PA elections have also
appeared unevenly. Under General Ayub Khan (1958–​1969), nonparty elections were
held in 1959–​1960 and 1962. And, under General Zia-​ul-​Haq (1977–​1988), party-​based
activity was largely proscribed, notwithstanding electoral activity in 1979, 1983, 1985, and
1987 (Shah 2014, 158–​159). General Ayub Khan, however, did hold party-​based NA/​PA
polls in 1965. And, historically speaking, party-​based affiliations were also an open se-
cret under General Zia (1985). General Musharraf (1999–​2007) went on to hold party-​
based elections in 2002 and 2007–​2008. As such, I adhere to the convention among
Pakistan experts, highlighting transitions within each military-​led regime between
“outright military dictatorship” and hybrid forms of “electoral autocracy.”
My main interest, however, is not military coups or intra-​authoritarian transitions
from outright military dictatorships to electoral autocracies. Instead, I focus on a
two-​step process of regime transition from military rule back toward an “Islamic” de-
mocracy: specifically, I focus on (a) transitions from electoral autocracy back to out-
right military dictatorship (1969, 1988-​A, 2007),3 and, shortly thereafter, (b) broadly
“democratizing” transitions from outright military dictatorship to some type of civilian-​
led regime (1970–​1972, 1988-​B, 2008). As noted above, this two-​step process typically
begins with widespread nonreligious protests. But, when a re-​imposition of mar-
tial law fails to quell these protests, the second step involves military leaders handing
power back to civilians via elections.4 Again, the protests that play such a crucial role in
Pakistani transitions from military to civilian rule rarely stress religious demands. But,
in due course, the regimes they engender invariably retain the “exclusionary” and “illib-
eral” constitutional provisions of their predecessors.

“Hybrid Democracy” and

the Constitutional Status of Religion

Following the rise of each civilian-​led regime in Pakistan, extra-​constitutional but in-
creasingly formalized patterns of military power—​in effect, “reserved” domains of mili-
tary decision-​making—​have persisted alongside constitutional provisions that diminish
 Regime Types, Regime Transitions, and Religion in Pakistan    119

the rights of non-​Muslims (and, since 1973, restrict fundamental rights). These factors
ensure that Pakistan’s civilian regimes do not amount to an inclusive or a liberal democ-
racy. I have examined the origins of Pakistan’s exclusionary constitutional provisions
elsewhere (Nelson 2016); Maya Tudor also addresses them in this volume.
Few twentieth-​century states were created as territorial homelands explicitly associ-
ated with particular religious communities—​only Pakistan (Muslim, 1947) and Israel
(Jewish, 1948). In fact, even among Muslim-​majority states, few were deliberately
created to transform a national Muslim minority (e.g., Muslims in India) into a ma-
jority (i.e., Muslims in Pakistan). This unusual history directly underpins Pakistan’s
national identity and, therein, its constitutional approach to religion (specifically,
Islam). Since Pakistan’s first constitution was introduced in 1956, only three twentieth-​
century constitutions have actually placed the word “Islamic” in the name of the state
itself: Mauritania since 1958, Iran since 1979, and Afghanistan since 2004. Indeed,
Pakistan has never treated its majoritarian Muslim identity as an open constitutional
question; if it did, the state’s historically rooted raison d’être would unravel. The illiberal
and exclusionary features of Pakistan’s constitution, however, do not follow seamlessly
from Pakistan’s Muslim-​majority or “Islamic” constitutional identity; instead, these
features flow from historically and politically embedded processes privileging a specific
interpretation of that identity.
Pakistan’s most important constitutional provision concerning religion—​in many
ways its first constitutional provision—​is known as the Objectives Resolution. Initially
approved by the Muslim members of Pakistan’s first Constituent Assembly over the
objections of every non-​Muslim member within that Assembly (1949), then introduced
as a preambular provision within Pakistan’s first constitution (1956), this resolution was
subtly diluted in Pakistan’s second constitution (1962) before being restored to its orig-
inal form in a constitutional amendment one year later (1963–​1964). It has remained an
important feature of every constitution since then—​for example, in Pakistan’s third con-
stitution (1973)—​becoming a substantive article within that constitution (Article 2A) in
1985. Reflecting a nuanced compromise between Islamist and nationalist elites (Nelson
2016), this resolution states that, while “sovereignty over the entire universe belongs
to . . . Allah” and the state’s authority will be exercised “within the limits prescribed by
Him,” state authority must be exercised by “the people” of Pakistan via their “chosen rep-
resentatives” working alongside an “independent” judiciary.
Since 1956, this resolution has co-​existed with further constitutional elements ex-
plicitly focused on “Islam.” These elements specify that no Pakistani law may be repug-
nant to the injunctions of the Qurʾan and sunna (that is, the practice of the Prophet
and his closest companions), with Pakistan’s head of state (for example, the president)
appointing an “advisory” Council of Islamic Ideology to guide each NA/​PA in de-
fining the parameters of repugnancy (1956, Article 198; 1962, Article 204; 1973, Articles
227–​230). Together, these “Islamic” provisions aim to promote conformity with the
injunctions of Islam while, at the same time, framing Pakistan as an Islamic “republic”
in which parliament retains the power to frame the state’s interpretation of Islam
(Nelson 2016).
120   Matthew J. Nelson

As noted above, identifying a state religion—​as in the United Kingdom, Denmark,

or Costa Rica—​need not automatically restrict liberal rights (e.g., freedom of religion
or freedom of speech). But, in the UK, Denmark, and Pakistan after 1973 (Article 2),
the constitutional identification of a state religion has also been associated with “ex-
clusionary” provisions declaring that only those adhering to the state religion will be
permitted to serve as head of state (Pakistan: 1956, Article 32; 1962, Article 19; 1973,
Article 41). In fact, since 1973, related constitutional provisions in Pakistan have gone
on to clarify that, not unlike the British Queen vis-​à-​vis the Church of England (spe-
cifically) or Denmark’s strictly Lutheran head of state, Pakistan’s president must be,
not merely a “Muslim,” but a Muslim in a specific doctrinal sense—​i.e., not part of a
heterodox minority known as the Ahmadiyya (1973, Third Schedule). The Ahmadiyya
describe themselves as Muslims but, since 1974, they have been constitutionally
reclassified as “non-​Muslims” (Article 260) owing to certain views articulated by their
late-​nineteenth-​century founder Ghulam Ahmed—​namely, that he was not just a reli-
gious reformer but a “prophet” after Mohammad. (Typically, Muslims see the Prophet
Mohammad as the final “seal” of prophecy itself [Qurʾan 33:40].5)
In 1973, Pakistan combined these exclusionary constitutional provisions regarding
the country’s head of state with illiberal provisions removing free-​speech protections
for any statement that might be said infringe, not on the personal security of fellow cit-
izens, but rather on an abstract idea, namely, “the glory of Islam” (Article 19). Whereas
in countries like the UK, broadly related criminal laws concerning anti-​“Christian” blas-
phemy were removed in 2008, constitutional protection for laws specifically concerning
anti-​“Muslim” blasphemy remain in place in Pakistan.
In 1956, and again in 1973, the introduction of Pakistan’s exclusionary and illiberal
constitutional provisions was initiated by religious groups claiming special religious
authority to define the doctrinal boundaries of Pakistan’s Muslim majority—​above all,
groups that initially opposed the nationalism of Mohammad Ali Jinnah and his osten-
sibly “secular” anti-​colonial Pakistan movement, including (a) the urban middle-​class
Jamaʿat-​i Islami (JI, Party of Islam) and (b) the madrasa-​based Jamiat Ulema-​i-​Islam
(JUI, Party of Islamic Scholars).6 These religious groups boycotted the late-​colonial
provincial elections that underpinned the formation of Pakistan’s first Constituent
Assembly; as a result, they were not physically present when Pakistan’s first constitution
was drafted. But, even so, their relevance was recognized by the members of Pakistan’s
first Constituent Assembly through an ancillary body charged with considering reli-
gious views (namely, the Talimat-​e-​Islamia Board, i.e., the Board of Islamic Instruction).
And in due course some of the exclusionary and illiberal constitutional provisions
demanded by this board were formally adopted by Pakistan’s first Constituent Assembly.
Since the early inclusion of these exclusionary provisions (e.g., 1956, Article 32;
1962, Article 19; 1973, Article 41), subsequent military regimes as well as parliamen-
tary majorities containing very few JI or JUI members have also opted to retain them.
Indeed, alongside constitutional provisions explicitly excluding non-​Muslims from
 Regime Types, Regime Transitions, and Religion in Pakistan    121

the presidency (1973, Article 41), an exclusionary constitutional amendment formally

redefining Pakistan’s Ahmadiyya as non-​Muslims (Article 260) was passed unanimously
by a freely elected NA in 1974.7 And, in 2010, Article 91 was amended by another freely
elected civilian NA to state that NA members were only permitted to choose a “Muslim”
member—​i.e., not an Ahmadi—​to serve as prime minister (85 percent in favor). Initiated
by religious leaders but formally introduced and persistently retained by military as well
as civilian leaders, Pakistan’s exclusionary and illiberal constitutional provisions reflect
widespread views regarding the parameters of a “Muslim” (non-​Ahmadi) as opposed to
a more inclusive liberal regime.
To avoid conceptual slippage vis-​à-​vis the terminology commonly used for “hy-
brid” regimes, my discussion relies on two terms developed by Guillermo O’Donnell
and Philippe Schmitter (1986) to describe (a) “dictablanda” hybrid regimes at the rel-
atively closed end of the regime-​ type scale, including electoral autocracies, and
(b) “democradura” hybrid regimes at the relatively open end of the scale. In democradura
regimes, free-​and-​fair elections are the norm. But, at the same time, several different
limitations are possible, including (a) constitutional “brown areas” (i.e. territories where
constitutional provisions simply do not apply),8 (b) reserved domains lying beyond the
control of elected leaders (i.e. “tutelary” spheres of policymaking), (c) limitations on
constitutional checks-​and-​balances such that the power of elected leaders is partly des-
potic (i.e. “delegative” or “cæserist” democracies), and (d) barriers to equal participation
(“exclusionary” limits) as well as (e) restrictions on fundamental individual rights like
due process or freedom of religion and speech (“illiberal” constraints) (Adeney 2015,
120; Dahl 1971). I focus on the latter two: exclusionary and illiberal constraints.

Exclusionary Constraints
In Pakistan, “exclusionary” constraints targeting the notion of equal citizenship first
emerged in 1955 when the territories of West Pakistan (Punjab, Sindh, Balochistan, and
the Northwest Frontier Province or NWFP) were combined as “One Unit” with the same
number of elected representatives as East Pakistan (Bengal), despite the latter having a
larger population. Until this One Unit scheme was removed shortly before Pakistan’s
first national election in 1970, this provincial set-​up challenged inclusive notions of
equal participation by undermining electoral equality. In short, those in East Pakistan
(Bengalis) were “excluded” from equal citizenship.
The most persistent exclusions, however, have been associated with religion, in-
cluding (a) the presence of “separate electorates” for those defined as non-​Muslims
(1947–​1956, 1985–​2002, and with continuing relevance for the Ahmadiyya)9 as well as
(b) constitutional provisions stressing that, as noted above, Pakistan’s president and
prime minister must be “Muslims,” not members of any other religious groups (e.g.,
Hindus, Christians, or Ahmadis).
122   Matthew J. Nelson

Illiberal Constraints
With reference to the Ahmadis, illiberal measures also extend to Pakistan’s Second
Constitutional Amendment (1974, Article 260). This amendment unilaterally defined
the Ahmadis as “non-​Muslims,” thus undercutting their fundamental individual right
to religious self-​identification (Article 20A).10 Furthermore, as noted above, Pakistan
has illiberal constitutional provisions removing individual free-​speech protections for
otherwise peaceful statements deemed to infringe, not on the safety of other citizens,
but rather on “the glory of Islam” (Article 19).11 In addition, Pakistan offered no legal rec-
ognition for the marriage, divorce, custody, or inheritance rights of individual Hindus
until 2017. With respect to fundamental individual rights of legal equality (Article
25) and, with reference to inheritance, personal property (Articles 23–​24), this exclusion
was also clearly illiberal.
It is not merely that religious references have played a formal constitutional role in
every Pakistani regime. It is, rather, that Pakistan’s elected civilian regimes qualify as
“hybrid” democratic regimes precisely owing to Pakistan’s constitutional interpreta-
tion of that role. This is not a normative statement; it is an analytical statement growing
out of prevailing scholarly definitions of “exclusionary” and “illiberal” regimes, with
particular reference to the constitutional interpretation of Islamic ideas in Pakistan
(Bogaards 2009; Collier and Levitsky 1997; Wigell 2008; Zakaria 1997). One aspect of
hybrid democradura regimes—​namely, reserved domains of military power—​has been
widely discussed in the case of Pakistan (Adeney 2015; Mufti 2018; Shah 2014). I focus on
Pakistan’s constitutional treatment of religion (Islam) instead.

Religious Parties and the Shape

of “Electoral Autocracy”

Within Pakistan, transitions to outright dictatorship via military coups are well known.
But, as noted above, different types of elections also figure prominently within the
dictablanda regimes that have emerged in the wake of each coup. The first type of elec-
tion includes direct local elections held at a district and a subdistrict level on a nonparty
basis. The second consists of NA/​PA elections conducted with or without political
parties (including religious parties). There is no consistent pattern across these two
types of “military” elections: under General Ayub, local elections produced an elec-
toral college for nonparty NA/​PA elections in 1960, but as noted above, after Pakistan’s
NA passed a new “Political Parties Act” in 1962, further district-​and subdistrict-​level
elections under General Ayub produced an electoral college for party-​based NA/​PA
elections later that same year (and again in 1964–​1965);12 furthermore, under General
Zia-​ul-​Haq, party-​based elections were formally proscribed, but nonparty NA/​PA
elections were conducted in 1985.
 Regime Types, Regime Transitions, and Religion in Pakistan    123

A third type of military-​led voting includes indirect/​direct and nonparty-​/​party-​

based presidential polls. As noted above, General Ayub’s local elections produced an
electoral college for indirect (nonparty) presidential polls in 1960. This was followed
by direct (party-​based) presidential polls, also under General Ayub, in 1965. Thereafter,
however, voters returned for nonparty presidential polls under General Zia (1984 ref-
erendum, 98 percent in favor) and General Musharraf (2002 referendum, 98 percent in
favor; 2007 “indirect” polls, 57 percent in favor).
Crucially, elections under both military and civilian-​led regimes have produced
governments including key religious parties. In fact, despite their weak electoral per-
formance in every poll since 1970, politicians from the JI and/​or the JUI have managed
to join a ruling NA or PA coalition in almost every subsequent regime. In 1970, for in-
stance, despite these two religious parties winning just twenty seats—​JI 4 (NA); JUI 7
(NA) + JUI 9 (PA: NWFP)—​JUI leader Mufti Mahmud emerged as the NWFP’s pro-
vincial chief minister. This pattern, linking religious parties to some form of participa-
tion in government, has faltered only three times since 1970: in 1985, 1997, and 2018 (see
Table 6.1).
After General Zia’s military coup in 1977, JI and JUI members were appointed to Zia’s
(advisory) cabinet even as they continued to push for the NA/​PA elections that Zia re-
peatedly promised.13 The JUI largely boycotted Zia’s nonparty NA/​PA elections in 1985,
but 10 JI-​affiliated candidates still prevailed as “independents.”14 These independents,
however, remained outside the formal governing coalition created by Zia’s hand-​picked
prime minister, Muhammad Khan Junejo; in fact, as shown in Table 6.1, 1985 marked the
first occasion when no religious party was included in any NA or PA government (Nasr
1994: 196–​197). This is ironic, given Zia’s reputation as a key “Islamizing” dictator.
After Zia was killed in a plane crash (1988), the chairman of Pakistan’s Senate,
Ghulam Ishaq Khan, was constitutionally elevated to the presidency. At the same time,

Table 6.1: Pakistan: Religious Parties in Government since 1970

1970-​71 NWFP PA: JUI 1997 –​

1977 Advisory Cabinet: JI/​JUI 2002 NA + NWFP PA + Balochistan

PA: “MMA” **
1985 —​ 2008 NA + Balochistan PA: JUI-​F
1988 Balochistan PA: JUI-​F; Punjab PA: “IJI”* 2013 NA: JUI-​F; NWFP (renamed KP) PA: JI
1990 Balochistan PA: JUI-​F/​”IJI”; NA + Punjab 2018 –​
PA: “IJI” *
1993 NA: JUI-​F

* Within the Islami Jamhoori Ittehad (IJI; including JI and JUI-​S), religious parties did not hold
cabinet posts.
** Within the Muttahida Majlis-​e-​Amal (MMA; including JI, JUI-​S, and JUI-​F), the JUI-​F held cabinet
posts in Balochistan and, together with the JI, in NWFP.
124   Matthew J. Nelson

the Islamist JI and a faction of the madrasa-​based JUI known as the “JUI-​S” (Sami-​ul-​
Haq faction) worked with Pakistan’s Inter-​Services Intelligence Directorate (ISI) to
create a right-​wing alliance known as the Islami Jamhoori Ittehad (Islamic Democratic
Alliance, IJI) under the leadership of Punjab chief minister Mohammad Nawaz Sharif.
In the ensuing elections (November 1988), however, this alliance was defeated by a left-​
leaning NA coalition brought together by Benazir Bhutto, the leader a party known as
the Pakistan People’s Party (PPP). But, in due course, Bhutto’s coalition also incorpo-
rated a separate faction of the madrasa-​based JUI known as the “JUI-​F” (Fazlur Rahman
faction).15 In fact at a provincial level the JUI-​F also secured cabinet posts in Balochistan
even as the right-​wing IJI captured power under Punjab’s chief minister, Nawaz Sharif.
Again, this pattern of religious-​party involvement, both at a national level and at a pro-
vincial level, was the norm.
When Benazir Bhutto tried to control military appointments, however, the army’s
corps commanders instructed President Ghulam Ishaq Khan to dismiss her NA coali-
tion. The IJI went on to win the next round of elections (1990) both in Punjab (PA) and
in Islamabad (NA). (The IJI also created a provincial coalition with the madrasa-​based
JUI-​F in Balochistan.) But, when IJI prime minister Nawaz Sharif sought to review
the constitutional article that empowered Pakistan’s president to unilaterally dissolve
Pakistan’s NA (Article 58(2)(b)), his own NA was dissolved instead (1993). The Supreme
Court struck down Sharif ’s removal; but, almost immediately, the military persuaded
both Prime Minister Sharif and President Ghulam Ishaq Khan to step down, leading to
fresh elections.
This time, Nawaz Sharif ’s Pakistan Muslim League (PML-​N) lost to a PPP + JUI-​F
coalition led by Benazir Bhutto. But, after Pakistan’s new president, Farooq Leghari,
returned to the powers enshrined in Article 58(2)(b) to dismiss Prime Minister Bhutto
in 1997, the PML-​N was brought back into power with a massive single-​party majority—​
the second of three occasions when no religious party held power in any NA/​PA
government.
Sharif quickly built on his two-​thirds majority in parliament to pass Pakistan’s
Thirteenth Amendment (removing Article 58(2)(b)) before replacing Chief of the Army
Staff Jahangir Karamat with General Pervez Musharraf.16 And, after comprehensive
talks with Indian prime minister Atal Bihari Vajpayee, Sharif also signed a landmark
India-​Pakistan agreement known as the Lahore Declaration in 1999—​unaware that
General Musharraf had already launched an offensive military campaign across the Line
of Control in Indian-​administered Kashmir. US pressure forced Musharraf to retreat.
But when Prime Minister Sharif moved to sack General Musharraf in October 1999,
Musharraf simply deposed Sharif in Pakistan’s third military coup.
Despite marginalizing several seasoned PPP and PML-​N politicians by requiring
NA/​PA candidates to possess either a university degree or an advanced madrasa certif-
icate, Musharraf ’s own party, known as the Pakistan Muslim League “Quaid-​e-​Azam”
faction or PML-​Q, fell short of a parliamentary majority in the dictablanda NA elections
of 2002. In fact, with just 26 percent of the NA vote and only 37 percent of the seats,
Musharraf ’s PML-​Q was forced to construct a governing coalition with a collection of
 Regime Types, Regime Transitions, and Religion in Pakistan    125

religious parties known as the Muttahida Majlis-​e-​Amal or MMA. (The MMA or United
Council of Action included all of Pakistan’s major religious parties: the JI, the JUI-​S, and
the JUI-​F.) Yet, five years later, when Musharraf was ousted via mass demonstrations
and replaced with an elected government led by the PPP (2007–​2008), the madrasa-​
based JUI-​F returned to power as a junior partner in yet another ruling coalition led by
the PPP.17
Nawaz Sharif ’s PML-​N defeated this PPP-​led coalition during the NA/​PA elections
of 2013. But, again, the JUI-​F switched sides and secured a place in Sharif ’s new gov-
ernment. The Islamist JI also joined a ruling provincial coalition with Imran Khan’s
Pakistan Tehreek-​e-​Insaf (PTI) in Peshawar. The PTI or Pakistan Movement for Justice
returned to power—​both at a provincial level in Peshawar (PA) and at a national level in
Islamabad (NA)—​following elections in 2018. But, at this point, religious parties were
excluded from the NA and all provincial governments. Again, this is unusual. Since
1970, elections have routinely provided Pakistan’s religious parties with access to gov-
erning power. Religious parties have not governed on their own. But, since 1970—​across
both dictablanda and exclusionary/​illiberal democradura regimes—​religious parties
have rarely been excluded from government.
Scholars occasionally claim that adding religious parties to civilian-​led governing
coalitions will foster “more” democracy (or, at least, more inclusive forms of democ-
racy), whereas marginalizing such parties amounts to “less” democracy (Fair and Patel
2020). In Muslim-​majority contexts, however, these claims remain underspecified
without a better sense of the ways in which the governments that include these reli-
gious (Islamic) parties articulate their specific understanding of Islam, particularly with
respect to matters of individual and group-​based legal equality. There is, again, no in-
herent or inevitable link between the inclusion of these parties and the presence of an
inclusive liberal democracy.18

Regime Transitions and the Role

of Religion

Elsewhere in the Muslim world—​in the Middle East, for example—​secular militaries
and secular political parties tend to distance themselves from right-​wing Islamist
parties. But, in Pakistan, every leading party (PPP, PML-​N, PTI), including General
Musharraf ’s “king’s party” (PML-​Q), has invited these parties into a ruling coalition. As
such, it is difficult to say that Pakistan’s military is automatically inclined to place a check
on Islamist parties or, as noted above, that including Muslim religious parties within a
civilian-​led government might signal a shift toward a more inclusive liberal democracy.
Leonardo Morlino (2009, 283) uses a horizon of ten years to distinguish established
regimes from “transitional” polities.19 But, in Pakistan, every regime has collapsed
within this ten-​year horizon or just one year beyond it (e.g., 1947–​1958, 1988–​1999),
126   Matthew J. Nelson

placing Pakistan in something resembling a state of permanent transition (Hoffman

2011, 91–​92). In what follows, however, I do not locate Pakistan within a state of perma-
nent transition; instead I focus on electoral-​autocratic and exclusionary-​democratic hy-
brid regimes—​neither ever fully consolidated—​and, then, having done so, I focus on the
nonreligious factors that underpin periodic transitions between them.
In Pakistan, “hybrid” democradura regimes never collapse directly into “hybrid”
dictablanda regimes (Knutsen and Nygård 2015; Sanchez-​Urribari 2011). Instead, all of
Pakistan’s hybrid, civilian-​led, democradura regimes have ended with a military coup
and a stretch of outright military dictatorship. Similarly, hybrid dictablanda regimes
never transition, directly, into hybrid democradura regimes (Bunce and Wolchik 2010).
Instead, ostensibly “democratizing” transitions (1969–​1971; 1988A–​B; 2007–​2008) tend
to involve widespread nonreligious urban protests followed by an intermediate stage of
hardening military rule.20
In 1969–​1970, for instance, the electoral autocracy of General Ayub Khan (1962–​1969)
came to an end when nonreligious ethnic (Bengali) protests in East Pakistan combined
with smaller urban protests in West Pakistan—​in this case, ethnic (Sindhi, Pashtun,
and Baloch) protests demanding a restoration of the provinces that were merged under
West Pakistan’s One Unit scheme as well as left-​wing labor strikes and demonstrations
challenging the “modernizing” religious reforms of General Ayub’s Central Institute of
Islamic Research (Rahman 1976, 301; Sobhan 1969).21 These demonstrations eventually
prompted the exile of the institute’s leader, Fazlur Rahman, followed further protests
leading to an intra-​military transfer of power from General Ayub Khan to General
Yahya Khan. General Yahya Khan then proceeded to harden military rule by reimposing
martial law.
The same hardening of authoritarian rule unfolded in between the dictablanda re-
gime of General Zia-​ul-​Haq (1985–​1988) and the democradura regime of PPP prime
minister Benazir Bhutto (1988)—​partly owing to nonreligious protests led by a PPP-​
based movement known as the Movement for the Restoration of Democracy (MRD)
(1983) as well as ethnic protests (1985–​1986) in which Pashtun urban migrants battled
an ethnic formation known as the Mohajir Qaumi Movement or MQM supported by
General Zia in Karachi (Khan 2010, 39, 41).22 Finally, in 1988, Zia responded to these
protests with a return to outright military rule, dissolving his dictablanda NA.
In fact the same authoritarian hardening emerged between the dictablanda regime
of General Pervez Musharraf (2002–​2007) and the democradura regime of PPP prime
minister Yusuf Raza Gilani (2008)—​this time, in response to nonreligious protests
led by district lawyers demanding the reinstatement of Supreme Court Chief Justice
Mohammad Iftikhar Chaudhry, who was sacked after challenging General Musharraf ’s
effort to stand for re-​election as president without first holding the NA/​PA elections re-
quired to install a new Electoral College for that purpose (Kokab 2013). Once again, after
sacking Chief Justice Chaudhry, Musharraf cracked down with a return to martial law.
In Pakistan, the spasms of street-​level protest that end authoritarian regimes are rarely
framed by religious parties, religious groups, religious ideas, or religious demands. As
noted above, however, religious actors often intervene after each protest-​led transition
 Regime Types, Regime Transitions, and Religion in Pakistan    127

in favour of civilian rule to marshal support for the preservation of exclusionary and il-
liberal “religious” constitutional provisions—​provisions that obstruct the emergence of
a more inclusive liberal democratic regime.

Nonreligious Transitions: Economic and

Institutional Drivers
The case of Pakistan is an awkward fit for most of the literature addressing key drivers
of regime transition, including arguments focused on (a) economic shocks, (b) au-
thoritarian legislatures (and regime-​oriented political parties) that might serve as a
bridge during such transitions, as well as (c) anti-​regime decisions delivered by con-
stitutional courts. In what follows, I examine this awkward fit before moving beyond
these key drivers to stress an account of democratizing transitions rooted in (d) urban
protest—​not nonviolent or religious protests, but historically specific configurations of
nonreligious and occasionally violent protests.
As noted above, religious groups and religious ideas are not a key driver of regime
transition in Pakistan. Religious elements merely intervene post hoc to stress an ex-
clusionary interpretation of Pakistan’s Muslim identity (i.e., “Muslim,” not Ahmadi).
This pattern of post hoc religious intervention has ensured that broadly democratizing
transitions (otherwise rooted in nonreligious protest) pull away from the parameters of
a more inclusive liberal democracy.

Economic Drivers
Following Guillermo O’Donnell (1973), who argued that democracies faced with eco-
nomic shocks might shift in the direction of authoritarianism to advance unpalatable
reforms, some have noted that, even if economic growth tends to be associated with re-
gime survival, recessionary shocks can promote regime collapse (Markoff and Baretta
1990). Mark Gasiorowski (1995, 882), for instance, has traced the drivers of democratic
collapse to recessions, even as Milan Svolik (2008, 153) has found a role for recessions
in the collapse of what he calls “non-​consolidated” democracies (see also Haggard and
Kaufman 2012, 512). While Gasiorowski and Svolik linked recessions to transitions away
from democracy, however, Thomas Pepinsky (2009, 1–​2, 34) also found that destabilizing
debt crises could promote the collapse of authoritarian regimes (e.g., in Indonesia). Jan
Teorell (2010, 76) later extended this argument. Returning to the underlying features of
“democratization,” he used large-​n statistical analyses to argue that recessionary shocks
generally increase the possibility of so-​called “democratizing” transitions.
In Pakistan, however, these findings are difficult to replicate, mostly because, until
the global Covid-​19 pandemic in 2020-​21, Pakistan had never actually experienced a
recession.23 In fact, pulling away from Gasiorowski (1995), Svolik (2008), and Teorell
(2010), we see no clear macro-​level economic trends anticipating any transition toward,
or away from, democracy. Before the coup led by General Ayub Khan in October 1958,
128   Matthew J. Nelson

for instance, Amina Ibrahim (2009) noted that annual GDP growth rates in Pakistan
climbed from 2.0 percent (1955) to 3.5 percent (1956) before falling back to 3.0 per-
cent (1957) and, finally, 2.5 percent (1958). This might point to a link between declining
growth and authoritarian backsliding. But, before the coup led by General Zia in July
1977, World Bank figures show GDP growth climbing from 3.5 percent (1974) to 4.2 per-
cent (1975) and, finally, 5.2 percent (1976).24 In fact, before the coup led by General
Musharraf in October 1999, we see GDP growth rates falling from 4.8 percent (1996)
to 1.0 percent (1997) before climbing back to 2.6 percent (1998) and, finally, 3.7 percent
(1999).25
The same inconsistency can be seen in the years preceding each collapse of military
rule in Pakistan. Under Generals Ayub Khan and Yahya Khan (1958–​1970), for example,
GDP growth rates bounced from 5.4 percent (1967) to 7.2 percent (1968), then back to
5.5 percent (1969) before reaching their highest-​ever level in 1970 (11.4 percent) and their
lowest-​ever level (0.5 percent) during the civil war that preceded Pakistan’s democratic
transition in December 1971. Under General Zia, rates fell from 7.6 percent (1985) to
5.5 percent (1986) before climbing to 6.5 percent (1987) and, then, 7.6 percent (1988).
Yet under General Musharraf, GDP growth fell from 7.7 percent (2005) to 6.2 percent
(2006) to 4.8 percent (2007) and, finally, 1.7 percent (2008) when Musharraf actually
stepped down. Macroeconomic growth, or recession, does not reveal any consistent
patterns vis-​à-​vis regime transitions in Pakistan.

Institutional Drivers: Legislatures and Political Parties

In Pakistan, the meso-​level mechanisms through which economic shocks often ex-
press themselves—​ including legitimation crises rooted in fragmenting patronage
coalitions—​have also been uneven.
In Transitions from Authoritarian Rule: Prospects for Democracy (1986), Guillermo
O’Donnell and Philippe Schmitter focused on the role that elected dictablanda
legislatures might play in reinforcing authoritarian rule by accommodating competing
elites and facilitating redistributive pacts between them (see also Blaydes 2010; Gandhi
and Lust-​Okar 2009; Karl 1987; Lust-​Okar 2006). O’Donnell and Schmitter paid rela-
tively little attention to specific political parties within authoritarian legislatures (e.g.,
king’s parties), but this dimension was taken up by Jason Brownlee (2007) and Jennifer
Gandhi (2008). Whereas Brownlee argued that authoritarian parties help to sustain au-
thoritarian regimes, however, Gandhi stressed the value of authoritarian legislatures
without finding any link between party-​based legislatures and patterns of authoritarian
survival.
Yet again, the experience of Pakistan is an awkward fit for this literature regarding the
role of regime-​oriented political parties and elected legislatures in the context of author-
itarian survival, mostly because every military-​led dictablanda regime with an elected
national legislature in Pakistan has ended after just three to five years, with or without
the presence of (a) authoritarian king’s parties or (b) supportive religious parties: Ayub
Khan (authoritarian NA, authoritarian king’s party, no religious parties 1965–​1969); Zia-​
ul-​Haq (authoritarian NA, no parties at all 1985–​1988); Pervez Musharraf (authoritarian
 Regime Types, Regime Transitions, and Religion in Pakistan    129

NA, authoritarian king’s party, supporting religious parties 2002–​2007). In Pakistan,

neither authoritarian NAs nor authoritarian king’s parties (nor supportive religious
parties) seem to facilitate the extended survival of electoral autocratic regimes.
It may be that, in Pakistan and elsewhere, military regimes create regime-​friendly
parties to protect their interests both within the regime itself and after any subsequent
(election-​based) transition to civilian rule. But, again, this pattern is difficult to trace
in Pakistan, where prominent king’s parties—​the Convention Muslim League (CML)
under General Ayub Khan and the Pakistan Muslim League “Quaid-​e-​Azam” faction
(PML-​Q) under General Pervez Musharraf—​have not fared well in any subsequent
elections. During the “democratizing” elections of 1970, for instance, Ayub Khan’s
CML won just 7 out of 300 seats. And, during the transitional election of 2008, Pervez
Musharraf ’s PML-​Q won only 50 out of 341. Contra Joseph Wright and Abel Escribà-​
Folch (2012, 292–​294, 302), in other words, elite-​incorporating authoritarian parties
have not provided the sort of “exit guarantees” that might allow dictablanda regime-​
based elites to defect in favor of democradura-​oriented regime transitions. (In fact,
spanning both sides of most transitions, Pakistan’s autonomous religious parties have
generally fared much better.)

Institutional Drivers: Courts

Like those who track the possibility of a counterintuitive link between authoritarianism
and elected legislatures, scholars have also examined the ways in which patterns of ju-
dicial independence—​a familiar marker of democracy—​might coexist with patterns of
authoritarian survival (Moustafa 2014). In recent years, some scholars have also built
on this literature to consider the role that senior judges might play in driving regime
transitions.
With reference to regime survival, Christopher Larkins (1998, 424), focusing on
Argentina, noted that outright authoritarian governments and what he called anti-​rule-​
of-​law “delegative” democracies tend to use senior judges to preserve their regimes in
different ways: whereas authoritarian governments limited judicial independence
by restricting courts to particular types of cases, for instance, delegative democracies
tended to appoint biased judges.26 In short, different patterns of judicial (in)dependence
were associated with different types of regimes. But, again, Pakistan is an awkward fit,
particular insofar as both authoritarian and democradura leaders have tended to stress
similar judicial limitations.
Whereas General Zia-​ul-​Haq and General Pervez Musharraf introduced manda-
tory oaths of allegiance (stacking their courts with loyal judges) (Kokab 2013, 33), for
instance, they also expanded the role of military courts to limit the role of Pakistan’s ci-
vilian judiciary (Shah 2014, 153 fn13, 160 fn41; Tate 1993, 323, 325, 332). Pakistan’s civilian
leaders, however, opted for very similar strategies. Prime Minister Z. A. Bhutto, for ex-
ample, stacked Pakistan’s Supreme Court with loyal judges by amending the constitution
to retain a sitting Chief Justice beyond his original retirement age.27 And, in due course,
Prime Minister Nawaz Sharif permitted a constitutional amendment expanding the
role of Pakistan’s military courts (targeting those accused of “terrorism”).28 Divergent
130   Matthew J. Nelson

approaches to judicial independence—​or dependence (as the case may be)—​have not
distinguished Pakistan’s military and civilian-​led regimes.
Moving away from patterns of regime survival to matters of regime transition, how-
ever, Raul Sanchez-​Urribarri (2011, 878) and Jill Goldenziel (2013) highlight the ways in
which superior-​court judges might facilitate both the rise and fall of hybrid dictablanda
regimes. During Venezuela’s move away from democracy under Hugo Chavez, for in-
stance, Sanchez-​ Urribarri notes that judicial appointments were often reshuffled
to create a more “compliant” dictablanda judiciary. Goldenziel, in turn, examines
trajectories of Middle Eastern democratization, noting that, after the emergence of elec-
toral autocracy in Egypt, Turkey, and Kuwait, even ostensibly compliant judges began
to read (unfair) election results in ways that helped to identify moments of regime vul-
nerability or weakness—​moments the judges reinforced with decisions challenging each
country’s authoritarian government.
Regimes dislike anti-​regime judgments and, in many cases, they seek to constrain the
courts (or individual judges) to avoid them. But, following Goldenziel, it may be that
dictablanda regimes also produce illuminating electoral outcomes that open up space
for anti-​authoritarian judicial decisions (see also Helmke 2002). In 2004, for instance,
many Pakistani legislators refused to support General Musharraf ’s indirect re-​election
as president (44 percent abstained or failed to vote). And in 2005, local election turnout
scarcely exceeded 50 percent. Thereafter, Supreme Court Chief Justice Mohammad
Iftikhar Chaudhry used his suo motu powers to target General Musharraf (Kennedy
2012), focusing on the military’s post-​9/​11 “disappearance” of suspected militants
(2005) as well as instances of high-​level government corruption (2006). In other words,
anti-​regime Supreme Court judgments followed the poor electoral performance of
Musharraf ’s dictablanda regime. In 2007, Musharraf sacked Chief Justice Chaudhry.
But, shortly thereafter, widespread protests led by furious lawyers prompted Musharraf
to impose one last gasp of authoritarian martial law. When this move from electoral au-
tocracy to outright dictatorship failed to end the protests, Pakistan experienced yet an-
other authoritarian → democradura transition.

Nonreligious Transitions: Urban Protest

In each Pakistani transition away from authoritarian rule (1969–​1971, 1988, 2007–​2008),
the common denominator has involved provincially dispersed urban protests. With
the partial exception of the late 1960s, however, these protests have not drawn on re-
ligious rhetoric to target the incumbent regime. Indeed, if transitions away from de-
mocracy have been achieved via self-​serving military coups that, apart from 1977, rarely
refer to religion (but, pushing in the direction of electoral autocracy, incorporate NA/​
PA elections that almost invariably leave room for religious parties), transitions toward
civilian rule have generally unfolded via short-​lived reversions to outright authoritarian
rule that fail to overcome largely nonreligious grass-​roots urban protest.
 Regime Types, Regime Transitions, and Religion in Pakistan    131

During the 1990s, Terry Lynn Karl (1990) was pessimistic about the ways in which
grass-​roots protest might support a consolidation of democracy. Whereas Karl argued
that mass protests often generate an anti-​democratic backlash among frightened elites,
pulling away from democratic power-​sharing, however, Nancy Bermeo (1997, 315–​
316) saw this risk of elite defection rooted in mass-​based protests as relatively limited.
Specifically, Bermeo saw this risk as limited to cases in which violent protests directly
targeted prominent individual elites.
Dan Slater (2010) also focused on the relationship between grass-​roots protests and
particular types of regimes. Specifically, he argued that different state responses to protest
tend to produce different types of “authoritarian consolidation,” with (a) urban protests
demanding radical economic redistribution along ethnic or communal lines facilitating
elite domination underpinned by democradura “protection pacts” (e.g., Singapore,
Malaysia), and (b) rebellions threatening the territorial integrity of the state producing
dictablanda forms of bureaucratic authoritarianism (e.g., Burma, Indonesia).
If we consider urban protests demanding economic redistribution and political re-
structuring in West Pakistan alongside proto-​separatist calls for ethnic redistribution
in East Pakistan during the 1960s (or, for that matter, economic redistribution along
ethnic lines in both Sindh and Karachi throughout the 1980s), however, we see some-
thing different. In particular, departing from Burma and Indonesia, redistributive
ethnic protests demanding territorial autonomy in East Pakistan during the late 1960s
produced a call for NA/​PA elections followed by a civil war and, then, a return to civilian
rule. And, departing from Singapore and Malaysia, urban protests demanding economic
redistribution along ethnic lines during the mid-​to-​late-​1980s were followed by NA/​PA
elections in which the election of an ethnic minority (Sindhi) prime minister (Bhutto)
was accompanied, once again, by a return to civilian rule. Departing from Slater, in other
words, urban grass-​roots protest in Pakistan has not played a role in “authoritarian con-
solidation.” Instead it has tended to facilitate forms of authoritarian collapse.29

Urban Protest: Violent or Nonviolent?

A large-​N study conducted by Barbara Geddes, Joseph Wright, and Erica Franz (2014,
326) found that “coerced” transitions from authoritarian rule (i.e., violent protest-​based
transitions) were especially unlikely to push in the direction of democratization. In fact,
they argued, military dictatorships were more likely to collapse without coercion (e.g.,
via national elections) (see also Teorell 2010, 116).
But again, the experience of Pakistan is different. In Pakistan, large-​scale and often
violent protests in urban areas have often pushed in the direction of democratization,
with dictators repeatedly stopping short of the brutality required to crush the protests
that threaten them. In 1968–​1969, for instance, Ayesha Jalal (1990) noted that violent
urban uprisings were a crucial part of General Ayub Khan’s decision to step down.
And, in 1970, his successor General Yahya Khan promised national elections only if the
country’s ongoing violence came to an end (Jalal 1990, 308). (The violence continued,
but elections and, then, a transition to civilian rule went ahead.)
132   Matthew J. Nelson

In 1988, Paul Staniland (2008) noted that Pakistan’s corps commanders finally de-
cided to call for NA/​PA elections owing to “growing resentment . . . on the street,”
with Michael Hoffman (2011, 88) citing “difficulties presented by [the] suppression [of
protesters]” as “the major factor driving the military’s decision [to step aside].”30 In fact,
the only partial exception to this violent, urban, protest-​led pattern of democratizing
transitions involved the largely nonviolent protests led by lawyers in 2007–​2008—​
protests that eventually forced General Pervez Musharraf to relinquish power.
Geddes, Wright, and Franz (2014, 325) found that only one in five cases of “coerced”
authoritarian breakdown moved in the direction of democracy. But in Pakistan—​
tracking exclusionary and illiberal hybrid “democradura” regimes rather than non-​
hybrid “democratic” regimes—​three out of four cases belong to this 20 percent.31
If urban protests play a consistent role in Pakistan’s transitions away from authori-
tarianism, but neither economic recession, nor ethnic/​territorial mobilization, nor re-
ligious rhetoric, nor the use of violence is a consistent feature of these protests, it bears
asking why Pakistan’s authoritarian leaders—​so well known for their repressive capacity
in other contexts (e.g., Bengal, Balochistan, Karachi, and Khyber Pakhtunkhwa)—​
have failed to crush the demonstrators that eventually remove them from power. (On
decisions to repress or not, see Bellin 2004, 146; Davenport 2007; Escribà-​Folz 2013;
Ulfelder 2005.) Unfortunately, apart from Hoffman’s (2011) account of Pakistan’s
democratizing transition in 1988, I am not familiar with any study that successfully
exposes the decision-​making processes within Pakistan’s military or its associated
“king’s parties” to illuminate why, exactly, the army has generally avoided the brutal
measures required to end the protests that topple it.32
Bellin (2004, 144–​145) suggests that authoritarian regimes well versed in repression
are most likely to lose their “will and capacity to hold onto power” when, inter alia,
their “financial foundation is seriously compromised” or they lose “crucial interna-
tional support.” But, again, there is no indication that the financial position of Pakistan’s
dictablanda or authoritarian regimes faced any consistent pattern of economic decline
before their removal. Nor is there any indication that key foreign allies—​above all the
United States—​fully abandoned General Yahya Khan (1971), General Zia-​ul-​Haq (1988),
or General Pervez Musharraf (2008) before their regimes collapsed. Further research is
needed to understand why Pakistan’s army imposed far-​reaching restrictions on civil
liberties (1970, 2007) but, then, faced with urban protests that threatened the army’s cor-
porate interests (see Gartner and Regan 1996), chose to hand over power.

Religious Influence after “Democratizing” Transitions

Although religious protesters criticizing the ostensibly “modernizing” religious reforms
of General Ayub Khan and Fazlur Rahman emerged within a wider mix of urban unrest
during the late 1960s, religious claims did not figure prominently in the democratizing
protests of 1988 or 2007–​2008. In fact, transitions away from dictablanda regimes have
typically proceeded via nonreligious protests followed by what might be described
as a temporary reversion to outright authoritarian rule that fails to end the protests.
Religious actors have then intervened after each protest-​led transition to preserve
 Regime Types, Regime Transitions, and Religion in Pakistan    133

the “exclusionary” and “illiberal” constitutional provisions (e.g., Articles 19, 41, 260;
Third Schedule) that, by definition, obstruct the formation of a more inclusive liberal
democracy.
Both the JI and the madrasa-​based JUI played a role in delegitimizing General Ayub
Khan during the late-​1960s (Rahman 1976, 288, 296). But, as noted above, the religious
concerns of these two parties were dramatically overshadowed by ethnic and economic
protests led by (a) Bengalis in East Pakistan and (b) Z. A. Bhutto’s left-​of-​center Pakistan
People’s Party (PPP) in West Pakistan. In fact, even as religious parties helped to oust
General Ayub Khan in 1969, a JI-​affiliated militia known as Al-​Badr actively supported
the army’s repression of Bengali protesters in East Pakistan throughout 1970–​1971.
Indeed, even after the separation of East Pakistan as Bangladesh in 1971, the same reli-
gious parties and protesters resurfaced during (West) Pakistan’s subsequent constitu-
tional debates to reinforce and extend Pakistan’s illiberal and exclusionary constitutional
provisions (ref. Articles 19, 41, 260) (Qasmi 2014, 175–​179). While religious protesters
worked alongside ethnic and economic protesters to initiate a “democratizing” transi-
tion from General Ayub to General Yahya Khan and, shortly thereafter, from General
Yahya’s military dictatorship to the civilian-​led regime of Prime Minister Z. A. Bhutto,
in other words, the same religious actors simultaneously ensured that, ultimately, this
transition toward a civilian-​led regime did not push in the direction of an inclusive lib-
eral democracy.
A similar pattern unfolded during the democratizing transition away from General
Zia-​ul-​Haq in 1988. In this case, nonreligious ethnic protests led by Sindhis, then also
Mohajirs and Pashtuns, helped to spur a two-​part transition from the dictablanda re-
gime of General Zia’s hand-​picked prime minister, Muhammad Khan Junejo, toward
an outright dictatorship led by Zia himself and, then, from Zia’s outright dictatorship
toward a “democradura” regime led, first, by Prime Minister Benazir Bhutto (PPP) and,
then, by Prime Minister Nawaz Sharif (IJI/​PML-​N) (1988–​1999).
In this context, the role of religious parties in stifling a transition toward a more in-
clusive liberal democracy was more complex. Under General Zia, a pair of JI and JUI-​
S senators introduced a bill aiming to make the Qurʾan and sunna Pakistan’s supreme
law. After Zia dismissed his dictablanda NA in 1988, however, Zia went on to intro-
duce a very similar ordinance of his own (which, as an executive “ordinance,” required
NA endorsement within four months to reappear as a piece of ordinary “legislation”).
This four-​month window, however, lapsed after Zia’s death in August 1988. So, after the
election of Prime Minister Benazir Bhutto in November 1988, a group of IJI-​affiliated
senators prepared a third bill to accomplish the same objective. This bill, however, was
not confirmed by the legislature before Bhutto’s own NA was dismissed by President
Ghulam Ishaq Khan in 1990.
Finally, following the election of Prime Minister Nawaz Sharif in November 1990, a
fourth version of the same bill was brought to a vote and passed—​ironically, without
any religious-​party support (Kennedy 1992, 774–​780). In fact, after the nonreligious
protests that prompted the demise of Zia’s outright military dictatorship (Hoffman
2011), followed by Pakistan’s democratizing transition in 1988, the JI refused to support
134   Matthew J. Nelson

Nawaz Sharif ’s “shariat bill” because, according to the JI, that bill failed to satisfy the
JI’s comprehensive understanding of shariʿa. Specifically, and despite serving in Prime
Minister Nawaz Sharif ’s IJI-​led governing coalition, the JI insisted that its own under-
standing of shariʿa should be treated as an even more exclusionary and illiberal “supra-​
constitutional” source of power—​one that constrained not only Pakistan’s legislature,
but also its constitutional courts. In other words, religious parties like the JI sought to
ensure that Pakistan’s post-​Zia “democratizing” transition was even further removed
from a rigorous defense of legislative power and individual constitutional rights.
The same sequence—​nonreligious protests, followed by a democratizing transi-
tion, followed by ostensibly “religious” support for illiberal and exclusionary constitu-
tional provisions—​emerged in 2007–​2008, when nonreligious protests by thousands of
lawyers spurred a two-​part transition from the dictablanda regime of General Pervez
Musharraf to a stretch of martial law (November-​December 2007) followed by the elec-
tion of a civilian-​led regime under Prime Minister Yousuf Raza Gilani (PPP, 2008–​2012),
then Prime Minister Nawaz Sharif (PML-​N, 2013–​2017) and Prime Minister Imran Khan
(PTI, 2018–​present). The omnibus Eighteenth Amendment (2010) that emerged shortly
after Pakistan’s restoration of civilian rule did not remove any of the constitution’s ex-
clusionary and illiberal clauses. Instead, it altered Article 91 to state even more explicitly
that Pakistan’s prime minister must be a “Muslim” (not an Ahmadi). Again, religious
parties did not play an important role in the protests that ousted General Musharraf.
They simply persisted after this protest-​led transition, as part of Prime Minister Yusuf
Gilani’s ruling coalition, to enhance Pakistan’s exclusionary constitutional provisions.33

Conclusion

Regime transitions in Pakistan, both toward authoritarianism and away from it, are
rarely driven by religious groups or religious ideas. Transitions pushing away from de-
mocracy toward authoritarianism are typically associated with self-​serving military
coups (1958, 1977, 1999)—​coups that have, invariably, moved beyond outright military
dictatorship in favor of “electoral autocracies.”
Pulling away from authoritarianism toward democracy (1969–​1970, 1988, 2007–​
2008), however, Pakistan also departs from much of the literature linking democratizing
transitions to the shock of economic recession, the bridging role of authoritarian
legislatures (or “king’s parties”), and the anti-​authoritarian decisions rendered by con-
stitutional courts. This chapter stresses, instead, the “democratizing” role of widespread
urban protest—​not nonviolent or religious protests, but broadly nonreligious protests
that often turn towards violence.
Beyond the nonreligious drivers that underpin broadly democratizing transitions in
Pakistan, however, I also stress the importance of illiberal, exclusionary, and ostensibly
religious constitutional provisions to account for the “hybrid” regimes that Pakistan’s
democratizing transitions produce. Introduced and retained by civilian as well as
 Regime Types, Regime Transitions, and Religion in Pakistan    135

military-​led regimes, Pakistan’s “illiberal” constitutional provisions restrict otherwise

peaceful forms of speech seen as infringing abstract principles like “the glory of Islam”
(Article 19). “Exclusionary” constitutional principles simultaneously prevent particular
groups like the Ahmadiyya and non-​Muslims from enjoying equal forms of political
participation (Articles 41, 260, Third Schedule). Even beyond the constraints associated
with extra-​constitutional patterns of “tutelary” military authority, in other words, ex-
plicit constitutional provisions situate Pakistan’s elected civilian-​led regimes within the
realm of illiberal and exclusionary (rather than “liberal”) democracy.
The exclusionary and illiberal constitutional provisions underpinning this categoriza-
tion grow out of widespread attachments to an “Islamic” rather than a “liberal” democ-
racy. In fact, all of Pakistan’s exclusionary and illiberal constitutional provisions were
introduced or retained by huge parliamentary majorities. And, when proposals seeking
to amend the provisions growing out of these articles have been introduced, enor-
mous protests have forced Pakistan’s NA to retreat. In 2017, for instance, changes seen
as relaxing the electoral registration procedures for Ahmadi parliamentary candidates
were met with significant protests led by religious activists. These protests prompted
a unanimous parliamentary vote to leave the country’s more exclusionary provisions
unchanged.
In 2015, Katherine Adeney examined the prospects for “democratic consolidation”
in Pakistan. But, avoiding the question of democratizing transitions, she argued for a
multidimensional measure of hybrid democratic regimes. Focusing specifically on
2013—​the first year in which Pakistan saw a direct transfer of power from one elected
NA to another—​Adeney stressed the presence of an elected regime characterized by
persistent military intervention (i.e., reserved domains of military decision-​making) as
well as a weak defense of (a) civil liberties and (b) legal equality. Departing from core
themes in this chapter, however, she did not examine the ostensibly religious constitu-
tional provisions that underpin Pakistan’s illiberal and exclusionary constraints. Nor did
she consider Pakistan’s history of political support for such provisions. In this chapter,
I stress the ways in which, spurred by religious groups like the JI and the JUI, Pakistan’s
elected civilian leadership has actively promulgated (often unanimously) the exclu-
sionary and illiberal constitutional provisions that underpin what Adeney describes as
Pakistan’s weak defense of “civil liberties” and “equality.”
Citizens may protest for, and succeed in achieving, a greater role in public decision-​
making. But in Pakistan, following largely nonreligious protests, religious parties like
the JI and the JUI have often taken the lead in reinforcing the illiberal, exclusionary,
and ostensibly religious provisions of Pakistan’s postcolonial constitutions. In doing so,
however, they have rarely faced significant popular or parliamentary resistance.34 On
the contrary, the exclusionary and illiberal constitutional amendments that underpin
Pakistan’s understanding of an “Islamic” democracy have passed with large parlia-
mentary majorities. Indeed, despite adjustments revising other religious ordinances
and statutes—​for example, ostensibly religious criminal laws concerning “rape” (pre-
viously conflated with “adultery”)—​Pakistan’s illiberal and exclusionary constitutional
provisions have remained firmly intact for decades.35
136   Matthew J. Nelson

As Steve Wilkinson (2000) points out, civilian regimes in Pakistan have gener-
ally “fail[ed] to revoke” what he calls “anti-​minority laws” because, turning to pre-
vailing strains of Islamic interpretation, the leaders of Pakistan’s civilian regimes have
feared widespread religious criticism. Specifically, Wilkinson argues, civilian regimes
have feared “being labelled [an] enemy of Islam” (221–​222). In Pakistan, I argue, pop-
ular pressure for a liberal and inclusive democracy does not follow from ostensibly
“democratizing” forms of anti-​authoritarian mobilization.

Notes
1. The coup led by General Zia-​ul-​Haq in June 1977 was precipitated by protests regarding alleged
rigging in the re-​election of Prime Minister Zulfiqar Ali Bhutto three months earlier. Protesters
associated with a coalition known as the Pakistan National Alliance led by the Jamaʿat-​i Islami
(Party of Islam) and the Jamiat Ulema-​i-​Islam (Party of Islamic Scholars) called for a polit-
ical turn to the “system of the Prophet.” The coups in 1958 and 1999 were motivated solely by
military self-​preservation—​on the part of General Ayub Khan (facing retirement in 1959) and
Pervez Musharraf (facing dismissal after a military misadventure in Kashmir).
2. The World Values Survey (2010–​2014) found that, across sixty countries, 66 percent of its
respondents described themselves as “religious.” That figure rose to 72 percent in Muslim-​
majority countries and nearly 100 percent (99.7 percent) in Pakistan.
3. This process usually involves imposing martial law (except 1988) and dissolving the NA
(except 2007).
4. Religious rioting in 1953–​1954 was followed by a stretch of martial law and the dissolution
of Pakistan’s first Constituent (National) Assembly. After several weeks, however, these
protests were controlled.
5. The Third Schedule specifies presidential and prime ministerial oaths stating that these
officials must believe in Mohammad as “the last of the Prophets.”
6. Within a 22-​Point manifesto drafted in 1951 by thirty-​one religious leaders meeting
in Karachi was a demand for a “Muslim” head of state; Pakistan’s first Constituent
Assembly accepted this demand in August 1952 (Binder 1961, 226). In 1973–​1974, religious
parliamentarians guided Pakistan’s attorney general, who oversaw the unanimous amend-
ment of Article 260 (Qasmi 2014, 185–​220).
7. Following General Zia-​ul-​Haq’s military coup in 1977, Pakistan’s constitution was held in
abeyance for nearly eight years. During this period Zia also introduced Article 203C creating
a Federal Shariat Court (FSC) and a Shariat Appellate Bench of Pakistan’s Supreme Court.
When the constitution was restored in 1985, Zia required his newly elected (nonparty) NA
to incorporate Article 203C as well as Article 62 (requiring Muslim parliamentarians to
“practice [the] obligatory duties prescribed by Islam” and remain “righteous” and “honest”
in a specifically Qur’anic sense).
8. In Pakistan, brown-​area restrictions pertaining to the country’s Federally Administered
Tribal Areas (FATA) were removed in a Twenty-​Fifth Amendment (2018) merging FATA
with NWFP (renamed Khyber Pakhtunkhwa in 2010).
9. Pakistan’s 1956 constitution allowed each province to adopt joint or separate electorates
for Muslims and non-​Muslims (Article 145). This provision was removed (restoring joint
 Regime Types, Regime Transitions, and Religion in Pakistan    137

electorates) in 1962 and 1973, then revived by General Zia in 1985 (Eighth Amendment),
then removed by General Musharraf in 2002 (Conduct of General Elections Executive
Order No. 7) except for the Ahmadiyya (Executive Order No. 15).
10. Article 20A notes that “every citizen shall have the right to profess, practice, and propagate
his religion.”
11. As a constraint on freedom of speech and religion, Pakistan Penal Code restrictions on
blasphemy (§295, 298) might be described as a “religious” component of illiberal democ-
racy, but, as a general rule, Pakistan’s courts have not justified these restrictions in religious
terms—​instead they rely on liberal notions of “public order” in constitutional provisions
concerning religious freedom (Nelson 2020, unpublished). Constitutional provisions
limiting liberal rights (e.g., for the sake of “public order”) are common; these provisions
draw attention to the narrow line between liberal and “hybrid” (e.g. illiberal) democracies
(Dyzenhaus 2011).
12. In a Cold War constraint on left-​wing politics, the Political Parties Act (1962 §3-​1) said that
no party could act against Pakistan’s “Islamic ideology”; this exclusionary statutory provi-
sion was removed in 2002.
13. Most JI and JUI representatives resigned from Zia’s cabinet in 1979. After 1981, however,
JUI-​S leader Sami-​ul-​Haq continued to serve in Zia’s hand-​picked Majlis-​e-​Shura.
14. In 1985 Sami-​ul-​Haq (JUI-​S) was elected to Pakistan’s Senate.
15. In this chapter, there are two Fazlur Rahmans: one led General Ayub Khan’s Central
Institute of Islamic Research; one led the “Fazlur Rahman” faction of the JUI (JUI-​F).
16. The Thirteenth Amendment restored the prime minister’s power to appoint provincial
governors, military service chiefs, and, in collaboration with the Supreme Court Chief
Justice, superior-​court judges.
17. In 2008, the JUI-​F also joined the ruling PA coalition in Balochistan.
18. Fair and Patel (2020) highlight the link between democracy and the inclusion of such
parties in urban, educated, middle-​class contexts in Bangladesh; for an account of such
contexts in Pakistan see Maqsood (2017).
19. Wright (2008, 329) prefers a three-​year horizon.
20. Bogaards (2009, 404) sees such reversions from a hybrid dictablanda regime to outright
military dictatorship as extremely rare; he cites only Pakistan, Peru, and Belarus.
21. See note 29.
22. A Pakistani “mohajir” is related to partition-​era (1947-​1948) Urdu-​speaking migrants
from North India.
23. See World Bank data (https://​data.worldbank.org/​indicator/​NY.GDP.MKTP.KD.ZG?
locations=PK) and IMF data (https://​data.imf.org/​regular.aspx?key=61545852); Amina
Ibrahim (2009, 19) uses different data to show negative GDP growth—​unrelated to
transitions—​in 1952 (−1.9 percent) and 1997 (−0.4 percent).
24. During this period, inflation fell from 26.7 percent (1974) to 20.9 percent (1975) to 7.2 per-
cent (1976).
25. During this period, inflation fell from 11.4 percent (1997) to 6.2 percent (1998) to 4.1 per-
cent (1999).
26. Larkins builds on Neal Tate (1993) regarding the Philippines, India, and Pakistan,
1977–​1979.
27. See the Fourth, Fifth, and Sixth Amendments (1975–​1976; removed under General Zia’s
Eighth Amendment 1985).
138   Matthew J. Nelson

28. On judicial appointments, see the Eighteenth and Nineteenth Amendments. The Twenty-​
First Amendment regarding military courts included a two-​year sunset clause; military
courts were not renewed in 2019.
29. On protest and regime transition, see also Collier and Mahoney (1997) and Wood (2001).
30. Paula Newberg (1989, 566) discounts the role of protest.
31. On the factors that push protesters seeking self-​governance towards violence (or not), see
also Lichbach (1987) and Pearlman (2011). Pearlman notes that movements with strong
leaders, a coherent ideological focus, and other features of organizational cohesion
are better able to sustain nonviolent protests. This may help to explain why anti-​regime
protesters in Pakistan embraced (a) violence during the organizationally and ideologically
disparate late-​1960s/​1980s and (b) nonviolence during the more cohesive (but equally ef-
fective) Lawyers’ Movement of 2007–​2008.
32. On military decision-​ making regarding the repression of territorially based reli-
gious insurgents (rather than nonreligious urban protesters) in Pakistan, see Staniland,
Mir, and Lalwani 2018 (incl. fn32 on the difficulty of accessing data regarding military
decision-​making).
33. The omnibus Eighteenth Amendment received Presidential assent with 85 percent NA
and 90 percent Senate votes.
34. Philpott (2007, 522) sees Pakistan as a “consensual” integrationist state seeking a strong
constitutional role for Islam alongside democratic electoral politics.
35. See (a) the 2006 Protection of Women (Criminal Laws Amendment) Act retaining re-
ligious punishments for “adultery” in General Zia’s Offence of Zina (Enforcement of
Hudood) Ordinance (1979) but shifting cases of “rape” to §375-​376 of the Pakistan Penal
Code, as well as (b) the 2016 Criminal Laws (Amendment) (Offenses Relating to Rape) Act
revising §151-​154 of Zia’s Qanun-​e-​Shahadat (Law of Evidence) Order (1984).

References
Adeney, Katherine. 2015. “How to Understand Pakistan’s Hybrid Regime: The Importance of a
Multidimensional Continuum.” Democratization 24, no. 1: 119–​137.
Alvarez, Michael, Jose Antonio Cheibub, Ferando Limongi, and Adam Przeworski. 1996.
“Classifying Political Regimes.” Studies in Comparative International Development 31, no.
2: 30–​36.
Bellin, Eva. 2004. “The Robustness of Authoritarianism in the Middle East: Exceptionalism in
Comparative Perspective.” Comparative Politics 36, no. 2: 139–​157.
Bermeo, Nancy. 1997. “Myths of Moderation: Confrontation and Conflict during Democratic
Transition.” Comparative Politics 29, no. 3: 305–​322.
Binder, Leonard. 1961. Religion and Politics in Pakistan. Berkeley: University of California Press.
Blaydes, Lisa. 2010. Elections and Distributive Politics in Mubarak’s Egypt.
Cambridge: Cambridge University Press.
Bogaards, Matthijs. 2009. “How to Classify Hybrid Regimes? Defective Democracy and
Electoral Authoritarianism.” Democratization 16, no. 2: 399–​423.
Brownlee, Jason. 2007. Authoritarianism in an Age of Democratization. Cambridge: Cambridge
University Press.
Brumberg, Daniel. 2002. “Democratization in the Arab World? The Trap of Liberalized
Autocracy.” Journal of Democracy 13, no. 4: 56–​68.
 Regime Types, Regime Transitions, and Religion in Pakistan    139

Bunce, Valerie and Sharon Wolchik. 2010. “Defeating Dictators: Electoral Change and Stability
in Competitive Authoritarian Regimes.” World Politics 62, no. 1: 43–​86.
Collier, David, and Steven Levitsky. 1997. “Democracy with Adjectives: Conceptual Innovation
in Comparative Research.” World Politics 49, no. 3: 430–​451.
Collier, Ruth Berins, and James Mahoney. 1997. “Adding Collective Action to Collective
Outcomes: Labor and Recent Democratization in South America and Southern Europe.”
Comparative Politics 29, no. 3: 285–​303.
Collins, Kathleen, and Erica Owen. 2012. “Islamic Religiosity and Regime
Preferences: Explaining Support for Democracy and Political Islam in Central Asia and the
Caucasus,” Political Research Quarterly 65, no. 3: 499–​515.
Dahl, Robert. 1971. Polyarchy: Participation and Opposition. New Haven, CT: Yale
University Press.
Dahl, Robert, 1982. Dilemmas of Pluralist Democracy. New Haven, CT: Yale University Press.
Davenport, Christian. 2007. “State Repression and Political Order.” Annual Review of Political
Science 10: 1–​23.
Diamond, Larry. 1999. Developing Democracy. Baltimore: Johns Hopkins University Press.
Dyzenhaus, David. 2011. “Emergency, Liberalism, and the State.” Perspectives on Politics 9, no.
1: 69–​78.
Escribà-​ Folz, Abel. 2013. “Repression, Political Threats, and Survival under Autocracy.”
International Political Science Review 34, no. 5: 543–​560.
Fair, C. Christine, and Parina Patel. 2020. “Rational Islamists: Islamism and Regime Preferences
in Bangladesh.” Democratization 27, no. 4: 547–​569.
Gandhi, Jennifer. 2008. Political Institutions under Dictatorship. Cambridge: Cambridge
University Press.
Gandhi, Jennifer, and Ellen Lust-​Okar. 2009. “Elections under Authoritarianism.” Annual
Review of Political Science 12: 403–​422.
Gartner, Scott, and Patrick Regan. 1996. “Threat and Repression: The Non-​ Linear
Relationship between Government and Opposition Violence.” Journal of Peace Research
33, no. 3: 273–​287.
Gasiorowski, Mark. 1995. “Economic Crisis and Political Regime Change: An Event History
Analysis.” American Political Science Review 89, no. 4: 882–​897.
Geddes, Barbara. 1999. “Authoritarian Breakdown: Empirical Test of a Game-​Theoretic
Argument.” Paper presented at the 1998 APSA meeting.
Geddes, Barbara, Joseph Wright, and Erica Franz. 2014. “Autocratic Breakdown and Regime
Transitions: A New Dataset.” Perspectives on Politics 12, no. 2: 313–​331.
Gilbert, Leah, and Payam Mohseni. 2011. “Beyond Authoritarianism: The Conceptualisation of
Hybrid Regimes.” Studies in Comparative International Development 46: 270–​297.
Goldenziel, Jill. 2013. “Veiled Political Questions: Islamic Dress, Constitutionalism, and the
Ascendance of Courts.” American Journal of Comparative Law 61, no. 1: 1–​50.
Haggard, Stephen, and Robert Kaufman. 2012. “Inequality and Regime Change: Democratic
Transitions and the Stability of Democratic Rule.” American Political Science Review 106, no.
3: 495–​516 (512).
Helmke, Gretchen, 2002 “The Logic of Strategic Defection: Court-​Executive Relations in
Argentina under Dictatorship and Democracy,” American Political Science Review 96, no.
2: 291–​303.
Hoffman, Michael. 2011. “Military Extrication and Temporary Democracy: The Case of
Pakistan.” Democratization 18, no. 1: 75–​99.
140   Matthew J. Nelson

Huntington, Samuel. 1991. The Third Wave: Democratization in the Late Twentieth Century.
Norman: University of Oklahoma Press.
Ibrahim, Amina. 2009. “Guarding the State or Protecting the Economy? The Economic Factors
in Pakistan’s Military Coups.” London: LSE, Development Studies Institute.
Jalal, Ayesha. 1990. The State of Martial Rule: The Origins of Pakistan’s Political Economy of
Defence. Cambridge: Cambridge University Press.
Karl, Terry Lynn. 1987. “Petroleum and Political Pacts: The Transition to Democracy in
Venezuela.” Latin American Research Review 22, no. 1: 63–​94.
Karl, Terry Lynn. 1990. “Dilemmas of Democratization in Latin America.” Comparative Politics
23, no. 1: 1–​21.
Kennedy, Charles. 1992. “Repugnancy to Islam: Who Decides? Islam and Legal Reform in
Pakistan.” International and Comparative Law Quarterly 41, no. 4: 769–​787.
Kennedy, Charles. 2012. “The Judicialization of Politics in Pakistan.” In The Judicialization of
Politics in Asia, edited by B. Dressel. London: Routledge.
Khan, Nichola. 2010. Mohajir Militancy in Pakistan: Violence and Transformation in the Karachi
Conflict. London: Routledge.
Knutsen, Carl, and Håvard Nygård. 2015. “Institutional Characteristics and Regime
Survival: Why Are Semi-​Democracies Less Durable than Autocracies and Democracies?”
American Journal of Political Science 59, no. 3: 656–​670.
Kokab, Rizwan Ullah. 2013. Lawyers’ Movement in Pakistan. Lahore: Pakistan Studies Centre,
University of the Punjab.
Larkins, Christopher. 1998. “The Judiciary and Delegative Democracy in Argentina.”
Comparative Politics 30, no. 4: 423–​442.
Levitsky, Stephen, and Lucan Way. 2002. Competitive Authoritarianism: Hybrid Regimes After
the Cold War. Cambridge: Cambridge University Press.
Lichbach, Mark. 1987. “Deterrence or Escalation: The Puzzle of Aggregate Studies of Repression
and Dissent.” Journal of Conflict Resolution 31, no. 2: 266–​297.
Lust-​Okar, Ellen. 2006. “Elections under Authoritarianism: Preliminary Lessons from Jordan.”
Democratization 13: 456–​471.
Maqsood, Ammara. 2017. The New Pakistani Middle Class. Cambridge, MA: Harvard
University Press.
Markoff, John, and Silvio Baretta. 1990. “Economic Crisis and Regime Change in Brazil: The
1960s and 1980s.” Comparative Politics 22, no. 4: 421–​444.
Merkel, Wolfgang. 2004. “Embedded and Defective Democracies.” Democratization 11, no.
5: 33–​58.
Morlino, Leonardo. 2009. “Are There Hybrid Regimes? Or Are They Just an Optical Illusion?”
European Political Science Review 1, no. 2: 273–​296.
Moustafa, Tamir. 2014. “Law and Courts in Authoritarian Regimes.” Annual Review of Law and
Social Science 10: 281–​299.
Mufti, Mariam. 2018. “What Do We Know about Hybrid Regimes after Two Decades of
Scholarship?” Politics and Governance 6, no. 2: 112–​119.
Nasr, Seyyed Vali Reza. 1994. The Vanguard of the Islamic Revolution: The Jama’at-​e-​Islami of
Pakistan. London: I.B. Taurus.
Nelson, Matthew. 2016. “Islamic Law in an Islamic Republic: What Role for Parliament?” In
Constitution-​Writing, Religion, and Democracy, edited by A. Bâli and H. Lerner, 235–​263.
Cambridge: Cambridge University Press.
 Regime Types, Regime Transitions, and Religion in Pakistan    141

Nelson, Matthew. 2020. “Fundamental Rights Lawfare: Religion, Constitutional Courts, and
the Construction of Majoritarian National Identities” (unpublished manuscript).
Newberg, Paula. 1989. “Pakistan at the Edge of Democracy.” World Policy Journal 6, no.
3: 563–​587.
O’Donnell, Guillermo. 1973. Modernization and Bureaucratic Authoritarianism: Studies in
South American Politics. Berkeley, CA: Institute of International Studies.
O’Donnell, Guillermo. 1994. “Delegative Democracy.” Journal of Democracy 14, no. 1: 55–​69.
O’Donnell, Guillermo, and Philippe Schmitter. 1986. Transitions from Authoritarian
Rule: Tentative Conclusions about Uncertain Democracies. Baltimore: Johns Hopkins
University Press.
Pearlman, Wendy. 2011. Violence, Nonviolence, and the Palestinian National Movement.
Cambridge: Cambridge University Press.
Pepinsky, Thomas. 2009. Economic Crises and the Breakdown of Authoritarian
Regimes: Indonesia and Malaysia in Comparative Perspective. Cambridge: Cambridge
University Press.
Philpott, Daniel. 2007. “Explaining the Political Ambivalence of Religion.” American Political
Science Review 101, no. 3: 505–​525.
Qasmi, Ali Usman. 2014. The Ahmadis and the Politics of Religious Exclusion in Pakistan.
London: Anthem.
Rahman, Fazlur. 1976. “Some Islamic Issues in the Ayyub Khan Era.” In Essays on Islamic
Civilization: Presented to Niyazi Berkes, edited by D. Little, 284–​302. Leiden: Brill.
Rahman, Fazlur. 1982. Islam and Modernity: Transformation of an Intellectual Tradition.
Chicago: University of Chicago Press.
Sanchez-​ Urribarri, Raul. 2011. “Courts between Democracy and Hybrid
Authoritarianism: Evidence from the Venezuelan Supreme Court” Law and Social Inquiry
36, no. 4: 854–​884.
Schedler, Andreas. 2006. Electoral Authoritarianism: The Dynamics of Unfree Competition.
Boulder, CO: Lynn Rienner.
Schedler, Andreas. 2013. The Politics of Uncertainty: Sustaining and Subverting Electoral
Authoritarianism. Oxford: Oxford University Press.
Shah, Aqil. 2014. The Army and Democracy: Military Politics in Pakistan. Cambridge,
MA: Harvard University Press.
Slater, Dan. 2010. Ordering Power: Contentious Politics and Authoritarian Leviathans in
Southeast Asia. Cambridge: Cambridge University Press.
Sobhan, Rehman. 1969. “Pakistan’s Political Crisis.” The World Today 25, no. 5: 203–​211.
Sadri, Mahmoud and Ahmad Sadri, eds. 2002. Reason, Freedom, and Democracy in
Islam: Essential Writings of Abdolkarim Soroush. Oxford: Oxford University Press.
Staniland, Paul. 2008. “Explaining Civil-​ Military Relations in Complex Political
Environments: India and Pakistan in Comparative Perspective.” Security Studies 17, no. 2: 322–​362.
Staniland, Paul, Asfandyar Mir, and Sameer Lalwani. 2018. “Politics and Threat
Perception: Explaining Pakistani Military Strategy in the North West Frontier.” Security
Studies 27, no. 4: 535–​574.
Svolik, Milan. 2008. “Authoritarian Reversals and Democratic Consolidation.” American
Political Science Review 102, no. 2: 153–​168.
Tate, Neal. 1993. “Courts and Crisis Regimes: A Theory Sketch with Asian Case Studies.”
Political Research Quarterly 46: 311–​338.
142   Matthew J. Nelson

Teorell, Jan. 2010. Determinants of Democratization: Explaining Regime Change in the World
1972–​2006. Cambridge: Cambridge University Press.
Tepe, Sultan. 2012. “Moderation of Religious Parties: Electoral Constraints, Ideological
Commitments, and the Democratic Capacity of Religious Parties in Israel and Turkey.”
Political Research Quarterly 65, no. 3: 467–​485.
Volpi, Frédéric. 2004. “Pseudo-​Democracy in the Muslim World.” Third World Quarterly 25,
no. 6: 1061–​1078.
Wigell, Mikael. 2008. “Mapping Hybrid Regimes: Regime Types and Concepts in Comparative
Politics.” Democratization 52, no. 2: 230–​250.
Wilkinson, Steven. 2000. “Democratic Consolidation and Failure: Lessons from Bangladesh
and Pakistan.” Democratization 7, no. 3: 203–​226.
Wood, Elizabeth. 2001. “An Insurgent Path to Democracy: Popular Mobilization, Economic
Interests, and Regime Transition in South Africa and El Salvador.” Comparative Political
Studies 34, no. 8: 862–​888.
World Values Survey. 2010–​2014. http://​www.worldvaluessurvey.org/​WVSOnline.jsp.
Wright, Joseph, and Abel Escribà-​ Folch. 2012. “Authoritarian Institutions and Regime
Survival: Transitions to Democracy and Subsequent Autocracy.” British Journal of Political
Science 42, no. 2: 283–​309.
Zakaria, Fareed. 1997. “The Rise of Illiberal Democracy.” Foreign Affairs 76, no. 6: 22–​43.
 Chapter 7

Regime Chan g e u nde r

t he Part y of J u st i c e
an d Devel opme nt ( A K P )
in Tu rk ey

Feryaz Ocaklı

Turkey has completed its transition to an authoritarian regime. After more than six
decades of multiparty politics, it no longer satisfies the minimal expectations of dem-
ocratic rule. Turkey has slid gradually from a majoritarian, illiberal democracy
constrained by military tutelage to an authoritarian regime in which a dominant po-
litical party, the Justice and Development Party (Adalet ve Kalkınma Partisi—​AKP),
and an even more dominant president, Recep Tayyip Erdoğan, monopolize the com-
manding heights of political power. Erdoğan’s claim to Turkey’s supercharged “execu-
tive” presidency is not subject to free and fair elections contested on an even playing
field. The political, economic, and institutional context in which elections take place are
heavily skewed in favor of incumbents, and the ballot box has become a tool to legitimize
the government rather than holding it accountable. How did this transition take place?
What were the factors that enabled Erdoğan to centralize power so significantly? What
role, if any, did religion play? This chapter analyzes the different sources and dimensions
of regime change in Turkey under the AKP.
Turkey fits into the broad pattern of democratic backsliding that is observable
around the world. Bermeo (2016) defines democratic backsliding as “the state-​led de-
bilitation or elimination of the political institutions sustaining an existing democracy.”
The clear distinctions between democracies and military, one-​party, or personalistic
regimes that were typical during the Cold War have given way to regimes in which
elected authoritarians regularly subject the highest political offices in the land to un-
fair, but competitive, elections (Schedler 2015). Within this framework, Turkey’s recent
experience most closely fits the model of executive aggrandizement. An elected govern-
ment has eroded the democratic characteristics of an imperfect but serviceable illiberal
144   Feryaz Ocaklı

democracy. Erdoğan and his allies have built a formidable institutional structure that,
although still open to some contestation, largely operates as a means to manage societal
problems through organization and regulation rather than coercion under an authori-
tarian regime (Gobel 2011).
The breakdown of Turkey’s already flawed democracy happened gradually. If we
conceive of political regimes as spectrums rather than binaries, as collections of rules
and practices that vary in terms of how democratic or authoritarian they are, then we
can visualize this transformation as a slide on a scale. The transition was not smooth
or linear, however. It was characterized by fits of crises and occasional reversals under
the pressure of social movements. The origins of regime change date back to the first
single-​party government the AKP established in 2002. The political, economic, and in-
stitutional foundations of authoritarian transition were laid early in the party’s tenure
in government. Significant events that accelerated regime transition, such as controver-
sial court cases, constitutional amendments, and a failed coup attempt, built on a preex-
isting framework of authoritarianization.
It would be counterproductive to categorize the AKP’s trajectory in terms of a demo-
cratic early phase and an authoritarian later phase, as many observers of Turkish politics
have done. A series of EU-​oriented reforms in early 2000s led Western journalists and
policymakers to conclude that the AKP started out its time in government as a force for
liberalization and democratization, only to transform into an authoritarian force much
later (Berlinski 2017). Although the AKP did enact reforms in line with EU accession
criteria, they failed to make formal institutions more democratic or inclusive. These
reforms set the stage for the removal of bureaucratic resistance to the AKP’s domination
over state institutions.
The shift in perceptions of the AKP, from a seemingly democratic early phase to
an authoritarian later period, poses a challenge to the study of democratic erosion.
How can we identify the early signs of democratic backsliding? Existing studies
have focused on populism (Levitsy and Loxon 2013), elected authoritarians’ rational
fear of being overthrown through bureaucratic and social pressure (Bermeo 2016),
and the voters’ willingness to tolerate authoritarian practices in polarized societies
when the incumbent delivers economic gains (Svolik 2020). The analysis of Turkey’s
transition to authoritarianism presented here bears out each of these expectations
in different ways—​the AKP delivered economic development in its early years in
office, Turkish society became increasingly polarized, and Erdoğan was nearly
overthrown by the military and the Constitutional Court. However, early signs of
democratic backsliding in Turkey predated massive polarization, coup attempts,
and the economic boom. The AKP leaders initiated programs that would eventu-
ally enable them to control Turkey’s political, economic, and institutional spheres
as early as their first year in office. What changed over time was the ability of other
power blocks, such as Kemalist bureaucratic institutions, conglomerations of big
business, and oppositional civil society to resist the AKP’s power. The cross-​class
 Regime Change under the AKP in Turkey    145

coalition led by the AKP defeated a series of oppositional resistance centers that
often lacked ideological or organizational cohesion. The AKP used strategies of
political control and institutional domination that were part and parcel of secular
Turkish politics even before Erdoğan’s rise. What distinguished AKP rule from its
predecessors was its ability to wield a variety of such tools consistently over time to
defeat its political opponents. Religion or Islamist ideology had little to do with the
state’s authoritarian turn.
Turkey’s transition to authoritarianism upended theoretical as well as common-​
sense expectations from an economically successful country led by moderate Islamists.
Significant economic growth after 2002 not only solidified electoral support for the
AKP, it also positioned Turkey as a middle-​income democracy. Modernization theory,
including its updated versions, predict wealthier democracies to be resilient to authori-
tarian transition. Furthermore, Turkey under the AKP was upheld by the international
community, and by Western media and policy circles in particular, as the leading ex-
ample for the coexistence of Islam and democracy. The “Turkish model” was an in-
spiration for democratically minded Islamist movements and parties such as Tunisia’s
Ennahda. The fall of the Turkish model challenged the optimistic expectations many
held for Islamist groups to lead democratization movements in authoritarian Muslim-​
majority states.
Research on contemporary Turkish politics has identified a plethora of factors that ac-
count for its authoritarian transition. This chapter organizes and builds on these findings
by focusing on political, economic, and institutional dimensions of regime change. The
purpose of this style of exposition is not to identify the most important causal variable
or the one that has the greatest explanatory value. Regime change in Turkey is likely
overdetermined by the convergence of a variety of factors, and this chapter examines
some of the most salient ones.
Religion played a limited role in this transition. The party and the leader responsible
for leading the authoritarian turn are Islamists in their ideology and political socializa-
tion. The AKP was one of two parties in the electoral arena that clearly represented an
Islamist choice—​in a moderate, or “Muslim-​Democrat,” form. Turkish society has also
become more Islamic over time, with overt signs of religiosity on display throughout the
country. However, it is not clear whether religious identity, institutions, and movements
enabled Erdoğan to centralize power, or if it was Erdoğan’s growing authoritarianism
that strengthened the former. The causal arrow seems to run in both directions. The
rest of this chapter elaborates on the complex relationship between religion and regime
change by clarifying how the two interacted within specific political, economic, and in-
stitutional domains.
The next section examines the political dimensions of regime change. It focuses
on politics with a small “p,” meaning party strategy, electoral competition, personal
conflicts among politicians, and co-​optation of rivals. The following sections expand the
analysis to economic and institutional spheres.
146   Feryaz Ocaklı

Political Dimensions
of Authoritarian Transition

The origins of the AKP can be traced back to student organizations and Sufi
brotherhoods that opposed Turkey’s state-​led secularism. The relatively liberal polit-
ical context that emerged following the ratification of the 1961 constitution gave rise to
a range of mass movements and political organizations that challenged the status quo.
These movements spanned the ideological spectrum from communist factions and
labor unions to religious brotherhoods and ultranationalist crack squads (Zurcher
2004). The Selamet Hareketi (Salvation Movement) emerged from a religious and con-
servative milieu, had close ties to the Sufi Naqshbandi Order, and established the Milli
Nizam Partisi (National Order Party) in 1970. The party leader, Necmettin Erbakan,
was a spiritual student of Naqshbandi Sheikh Mehmet Zahit Kotku (Sezgin 2011). He
became a central figure of the political Islamist movement, and went on to establish a
number of Islamist parties that played a crucial role in Turkish politics.
The movement’s Refah Partisi (Welfare Party—​ RP) formed grass-​ roots activist
networks in the growing shantytowns surrounding Turkey’s large cities throughout
the 1980s and early 1990s. It encouraged Islamic civil society and media, formed or-
ganic links to pro-​Islamist labor unions and business associations, and won over
municipalities in Istanbul, Ankara, and across much of the country.
One of Erbakan’s protégé’s, Recep Tayyip Erdoğan, came of age as an activist in this
environment. As a member of the younger cohort of Islamist politicians, Erdoğan was
fully enmeshed in the party’s voter outreach efforts, local organizing, and civil society
functions. His first major political achievement was to be elected mayor of Istanbul in
1994, which launched his reputation as an up-​and-​coming Islamist politician.
Despite finding success in municipal elections, the RP was constrained by two major
factors that prevented it from holding office at the national level in a meaningful way.
The first was the adamant opposition of the Kemalist military and judicial bureaucracies
to Islamist groups (Yavuz 1999). Although White (2002) found that they were able to
build public support through formidable grass-​roots mobilization, the Islamists were
squeezed out of national office by the intransigence of the state bureaucracy. The second
factor was the enduring power and popularity of the more traditional center-​right
parties. The Anavatan Partisi (Motherland Party—​ANAP) and the Doğru Yol Partisi
(True Path Party—​DYP) represented the tradition of the Demokrat Parti (Democratic
Party—​DP) that had initially staked a claim in the popular center-​right of the political
spectrum in the 1950s. Together they controlled upward of 50 percent of the vote in the
mid-​1990s (TUIK). The Islamists remained in the periphery of right-​wing politics.
Faced with external threats from secularist regime forces and internal dissent from
younger activists, the political Islamist movement experienced a generational split
in 2001 (Mecham 2004). The movement’s older leaders established the Saadet Partisi
(Felicity Party—​SP), which would continue the same Islamist tradition. The younger
 Regime Change under the AKP in Turkey    147

and reformist cohort formed the AKP. Recep Tayyip Erdoğan, Abdullah Gül, Bülent
Arınç, Abdüllatif Şener, and their colleagues established their new party in response
to the anti-​systemic ideology of the movement’s older leaders, as well as the illiberal
secularism of the Kemalist bureaucracy. In an effort to placate the military and the
Constitutional Court, they declared their intention to make peace with the secular prin-
ciples of the constitution.
It is tempting to look back at this moment in history as a breaking point for Turkish
politics. The AKP would eventually rise to dominate the political arena and usher in
an authoritarian regime. However, the future of the party in 2001 was far from certain.
The highest vote share received by the AKP’s immediate predecessors was 21 percent
in 1995, and that number had declined to 15 percent in 1999. The Constitutional Court
had banned two Islamist parties between 1998 and 2001. There was no guarantee that
the AKP could avoid being thrown into the ash heap of history by the same intransi-
gent Kemalist bureaucracy, or by the voters. What enabled the AKP’s dramatic rise
was a combination of shrewd political strategy and good fortune. The year 2001 saw
the collapse of the political and economic system that had emerged following the 1980
military coup.
The AKP was in the right place at the right time. The 2001 economic crisis eroded trust
in the political parties that had failed to avert the catastrophe (Öniș 2010). The parties
in power at the time of the crisis, a coalition of the Demokratik Sol Parti (Democratic
Left Party—​DSP), the Milliyetçi Hareket Partisi (Nationalist Action Party—​MHP),
and the ANAP suffered a severe defeat in the 2002 general elections (Öniș 2012). The
two center-​right parties, the DYP and the ANAP, were also roiled in a deep corruption
scandal (Çarkoğlu 2007). This gave the AKP an opportunity to expand its vote share at
the expense of center-​right parties at a time when the electorate was largely alienated
from all systemic parties.
The AKP adopted a strategy of targeting center-​right voters by recruiting local elites
who were previously integrated into the DYP and the ANAP’s patronage structures
(Ocaklı 2015a). Businesspeople, merchants, clan leaders, civil society leaders, and local
politicians enabled the AKP to penetrate local social networks, and to present itself as
a moderate alternative to the discredited parties of the system. Erdoğan and his allies
formed close links with local clan networks in the Kurdish provinces (Ocaklı 2017b),
and sought to distance themselves from Erbakan and his legacy in the heavily secu-
larist districts of Western Anatolia (Ocaklı 2017a). A formidable grass-​roots organiza-
tion enabled the AKP to become an enduring presence in the everyday lives of potential
voters (Eligür 2010). The party was able to mobilize its supporters effectively both in
and out of the election cycle through close cooperation with business associations and
migrants’ civil society organizations (Ocaklı 2016a, 2016b).
Apparent ideological moderation, the ability to present itself as an outsider, avoiding
blame for the economic failures that led to the 2001 crisis, and an uncanny ability to mo-
bilize grass-​roots networks at the local level enabled the AKP to win a series of elections
from 2002 onward. The AKP’s formative years in government corresponded with a sig-
nificant uptick in global commodity prices (Baffes and Haniotis 2010; Barua and Kumar
148   Feryaz Ocaklı

2014). Buoyed by a favorable global economy and the political stability of a single-​party
government, the AKP’s neoliberal economic program delivered significant economic
growth. The economy grew by almost 6 percent per year from 2002 to 2006, while pov-
erty and income inequality declined and the middle class expanded (Acemoglu and
Ucer 2015). The AKP had firmly established itself as the party of economic growth and
prosperity by the time the 2007–​2008 global financial crisis reached Turkey. The combi-
nation of political stability, economic growth, divided opposition, and an active voter-​
outreach strategy enabled the AKP to completely displace the traditional center-​right
parties and win a series of elections.
The AKP established an enduring support base during this period. The elections were
largely free and fair. The party’s popularity only increased when the military attempted
to meddle in the presidential elections in 2007 (BBC 2017). The AKP’s persistent elec-
toral victories throughout the 2000s were a significant factor in its ability later to change
the political regime (Bakiner 2017). They gave Erdoğan and his allies the time and the
ability to change power dynamics within the state institutions. The Gülen network was
able not only to infiltrate, but also establish significant power over the judiciary and the
police force. The stage was set to purge secularist officers from the military through a se-
ries of court cases beginning in 2008. The AKP created a codependent relationship with
a number of corporations that profited from lucrative state contracts in return for po-
litical and economic support. It also expanded social welfare programs that benefitted
current or potential supporters. Hence, it is plausible to suggest that if the AKP had to
share the government with a coalition partner, or had it outright lost an election during
this period, it would have been very difficult, if not impossible, for Erdoğan to change
the regime.
Electoral integrity declined in the 2010s. Although it is difficult to point to a smoking
gun, there is evidence to suggest that the two general elections in 2015, the constitutional
referendum in 2017, and the general election in 2018 were rigged. A forensic study of the
2015 election found significant fraud in eastern Turkey, particularly in the areas where
the pro-​Kurdish Halkların Demokratik Partisi (Peoples’ Democratic Party—​HDP) was
dominant (Mebane, Hicken, and Kollman 2016). The fraud strongly favored the AKP.
The constitutional referendum of 2017 ratified the country’s transition from a parlia-
mentary to a presidential government, concentrated power in the office of the “executive
presidency,” and weakened checks and balances (Çalışkan 2018). A forensic study found
systematic and significant support for vote rigging and ballot stuffing in 11 percent of the
stations (Klimek et al. 2018). The same study found further support for fraud of a similar
size and direction in the 2018 presidential elections. These findings were confirmed by
election observers who witnessed ballot stuffing firsthand (Sendika.org 2017).
Islamic identity was a factor in the AKP’s relationship to voters. There is some support
for the argument that religious voters may have preferred the AKP over other parties
(Gidengil and Karakoc 2016). It was widely known that the AKP’s principal leaders, in-
cluding Erdoğan, Gül, and Arınç, had an Islamist background. They were close associates
of Necmettin Erbakan, and had been active in Islamist politics throughout the 1980s and
1990s. They had moderated but not fully disowned their Islamist backgrounds in the
 Regime Change under the AKP in Turkey    149

2000s (Yilmaz et al. 2017). The AKP was a natural home for the Islamist base. However,
it had to share this base with the party of the Islamist old guard, the SP. Although the SP’s
fortunes had declined shortly after its establishment in 2001, it would re-​emerge as a
challenger later in the decade.
The Islamist old guard struggled unsuccessfully against the AKP in the early 2000s.
They lacked the charisma and vitality of Erdoğan or Gül. Islamist voters abandoned
them en masse. The response came in 2008, when the party leaders finally decided
to try a new strategy. They unanimously promoted a young firebrand named Numan
Kurtulmuş to party leadership. Kurtulmuş was an economic and Islamic populist. He
articulated a devastating critique of the AKP from within the Islamist camp, accusing its
leaders of selling out the Islamic cause in favor of materialism and corruption. Although
the SP never acquired the vote share that could threaten the AKP, the presence of a char-
ismatic young leader in the Islamist camp challenged Erdoğan’s grip on his base.
Before the SP could launch a credible offensive, however, it fell victim to internal con-
flict. According to Israfil Bayrakci, an SP official at the party’s Ankara headquarters, the
political ideology articulated by Numan Kurtulmuş “strayed too far towards liberalism”
(I. Bayrakci, personal interview, September 16, 2015). The party’s veteran leadership grew
unhappy with his ideological flexibility. The discontent within the SP led Kurtulmuş to
split the Islamist movement once more. He took over half of the party’s local organiza-
tions with him to establish the Halkın Sesi Partisi (People’s Voice Party—​HAS Parti)
in 2010. The SP’s division worked in Erdoğan’s favor, and he co-​opted Kurtulmuş by
recruiting him to the AKP in 2012. Kurtulmuş has been a deputy party leader and a close
confidant of Erdoğan since then.
Co-​opting rivals has become an important instrument of Erdoğan’s efforts to cen-
tralize power. By bringing Kurtulmuş into the fold, he silenced a potentially effec-
tive adversary who could threaten his standing among the Islamist base. He used the
same strategy toward Süleyman Soylu, a young politician in the center-​right DYP. He
was elected party leader in 2008 when the party changed its name to Demokrat Parti
(Democratic Party—​DP). Similar to Kurtulmuş, Soylu vociferously criticized Erdoğan
and the AKP for corruption. However, he also chose to join the AKP following Erdoğan’s
invitation in 2012. Since then, he has been a deputy party leader and a power broker
within the AKP organization.
By recruiting Kurtulmuş and Soylu, Erdoğan not only pacified credible politicians in
the Islamist movement and the center-​right, he also expanded the pro-​Erdoğan con-
tingent of elites within the AKP. Both of these recruits have been strong supporters
of Erdoğan against his rivals within the party organization. Strictly pro-​Erdoğan
politicians in the AKP are often referred to as “reisçi,” or “supporters of the chief.” The
rise of the reisçi and the decline of other AKP leaders is well illustrated by the story of
Ersönmez Yarbay.
I met Ersönmez Yarbay in 2015 in his real estate office in Ankara that overlooks a
busy avenue in the city center. A middle-​aged, effusive man, Yarbay was more politician
than real estate broker. Others came in and took their seats silently as he described his
achievements in a former life—​a founding member of the AKP, president of the party’s
150   Feryaz Ocaklı

Ankara organization, and two-​term parliamentarian. Once a close associate of Turkey’s

most powerful men, Yarbay’s political career came to an abrupt end in 2007. “The boss
wrote me off,” he said, referring to President Erdoğan. “My career was in his hands. You
have to butter up the boss if you want to prosper. This is what happens if you don’t.” He
did not mince his words as others entered his office: “I have no expectations and nothing
to hide. My political life is long over.”
Yarbay still commanded respect among veteran Islamist politicians. His old friends
referred to him as an honest man who made a mistake by throwing in his lot with the
younger cohort of the Islamist movement. His rise and sudden estrangement from the
AKP is not an exception. It is a story shared by a growing number of former Islamist
heavyweights. Abdullah Gül, Turkey’s eleventh president and former prime minister,
Bülent Arınç, the former Speaker of the Parliament, Abdüllatif Şener, the former min-
ister of finance, Yaşar Yakış, the former minister of foreign affairs, and Dengir Mir
Mehmet Fırat, all founding members of the AKP, were gradually and systematically
pushed out. Former prime minister Ahmet Davutoglu and former deputy prime min-
ister for economic and financial affairs Ali Babacan have also lost their positions in the
AKP. Erdoğan’s grip on the party organization tightened over the years he spent in gov-
ernment. Arınç summarized the change well: “The ‘we’ that founded the AKP has be-
come an ‘I’” (Al Jazeera 2015).
The AKP’s founders were replaced by co-​opted politicians like Kurtulmuş and Soylu,
Erdoğan’s son-​in-​law Berat Albayrak, and the conspiracy-​minded media figure Yigit
Bulut. This new cadre is united in their unwavering support for Erdoğan over anyone
else. However, conflicts between them in other areas regularly make the national news
(Evrensel 2018).
Creeping authoritarianism and the gradual elimination of anyone who could pos-
sibly pose a threat to Erdoğan’s supremacy has led some scholars to conclude that the
regime in Turkey could plausibly be called “Erdoğanism” (Yilmaz and Bashirov 2018).
Erdoğanism relies on an Islamist ideology and a populist political strategy. Indeed,
Erdoğan’s populist discourse has achieved much in the way of a deep polarization in
Turkish society that contributed to democratic breakdown (Aytaç and Öniș 2014; Bilgic
2018; Somer 2019).
Erdoğan’s populist rhetoric and Islamist ideological leanings could have played a
somewhat unifying role for Turkey’s Kurdish minority. Islamic identity was an im-
portant factor in the AKP’s relationship with Kurdish voters. Following its establish-
ment in 2001, the AKP became one of only two parties that had any real presence in
eastern and southeastern Turkey. The left-​wing pro-​Kurdish movement and the AKP
dominated the electoral arena. In this context, the AKP sought to build national har-
mony between ethnic Turks and Kurds by emphasizing their shared Islamic identity
(Yavuz 2009). It provided a plausible alternative to the pro-​Kurdish parties for the
Kurds who either assimilated into Turkish culture or based their opposition to the
state on a Kurdish-​Islamic identity. The AKP’s struggle with secular state institutions
early in its existence evoked sympathy among segments of the Kurdish society
(Grigoriadis et al. 2018).
 Regime Change under the AKP in Turkey    151

The appeal to Islamic brotherhood seemed to work well for the AKP until the
“peace process” between the government and the illegal wing of the Kurdish National
Movement, the Partiya Kerkaren Kurdistan (Kurdistan Workers’ Party—​ PKK),
collapsed in 2015. What followed was a vicious return of military conflict in the Kurdish
region between the PKK’s guerilla units and the Turkish military (Ocaklı 2015b). The
securitization of the conflict had immediate effects for the legal wing of the Kurdish
National Movement, the HDP. The AKP launched an unprecedented attack on
parliamentarians and local elected officials from the HDP, removing them from office
and imprisoning them (Kamer 2018). The HDP’s former leader and a popular candidate
for Turkey’s presidency, Selahattin Demirtas, remains in prison as of late 2020.
The breakdown of the AKP’s Kurdish strategy demonstrated the limits of “Islamic
brotherhood” as a bond that could salve ethnic tensions. The government’s severe crack-
down on the pro-​Kurdish politicians after 2015 also revealed the extent of the regime’s
authoritarian turn.

Economic Dimensions
of Authoritarian Transition

There is a significant economic dimension to Turkey’s regime change, without which

this political process cannot be fully understood. The previous section touched upon
one aspect of this economic story—​the attainment of impressive growth rates and pov-
erty reduction in the early years of AKP rule. Economic stability and growth during this
period bolstered electoral support for the party, which gave Erdoğan and his allies the
ability and the time required to degrade checks and balances.
Beyond aggregate growth and investment rates, however, lies a more complex re-
lationship between the AKP and capital. The party leads a power block in collabora-
tion with a particular faction of the business elite. Together, they have shaped Turkey’s
political economy in ways that have advantaged pro-​government capital against their
competitors, and Erdoğan against his. This alliance has played a crucial role in regime
transition.
The connection between the AKP and a faction of Turkish capital dates back to the
1980s. The military regime (1980–​1983) and the Özal administrations (1983–​1989)
implemented Turkey’s first sustained neoliberal reforms. Prior to these reforms, state-​
led development strategies had favored Istanbul-​ based big businesses. Economic
liberalization, privatization, and deregulation opened up new spaces for small and
medium-​sized businesses from across the country to join export markets. Some of these
new businesses were formed in provincial areas in Anatolia, and were led by devout
businesspeople who were previously shut out of privilege networks. The rise of the pious
bourgeoisie empowered Islamist movements (Öniș 1997; Gülalp 2001)—​particularly the
Gülen movement and the Welfare Party—​and may have been a cause in their gradual
152   Feryaz Ocaklı

moderation (Gumuscu 2010). These companies were later united in a series of local
associations and one national business association: the Müstakil Sanayici ve İşadamları
Derneği (Association of Independent Industrialists and Businessmen—​ MUSIAD)
(Tuğal 2002).
The AKP forged a particularly close relationship with MUSIAD and local business
associations. As described earlier in this chapter, the party purposefully recruited
businesspeople shortly after its founding in 2001. Local businesspeople often constituted
the leadership cadres of the AKP’s provincial and sub-​provincial organizations (Ocaklı
2015a, 2016a), and benefitted from the party’s subsequent control over municipal re-
sources (Ocaklı 2016b).
Shortly after coming to power in 2002, the AKP began to institute a series of legal
changes that would enable it to use public resources for the enrichment of pro-​
government businesses. First, it limited the jurisdiction of autonomous regulatory
agencies that provided economic oversight (Buğra and Savaşkan 2014), and it sought
to establish partisan control over them. It changed the scope and applicability of the
public procurement law through amendments that gave significant flexibility to the gov-
ernment to reward preferred businesses (Buğra and Savaşkan 2014; Gurakar 2016). It
restructured the government’s role in housing construction and the real estate sector by
overhauling the Toplu Konut İdaresi (Housing Development Administration—​TOKİ)
and the land office law (Ocaklı 2018). Through these and other legal changes, Erdoğan
and his allies laid the groundwork for enriching a particular faction of capital in return
for material support and political loyalty.
The legal changes soon bore fruit. Pro-​government businesses expanded their areas of
operation significantly to encompass construction, energy, media, tourism, foodstuffs,
and a plethora of other sectors. The AKP’s preferential treatment enabled them to rapidly
accumulate capital. Pro-​government businesses received state contracts, cheap credits,
favoritism in purchasing state-​owned enterprises, preferential access to state procure-
ment deals, tax forgiveness, construction and mining permits, and access to public
lands for construction purposes. At the same time, they avoided targeted tax audits and
confiscation through the use of the Tasarruf Mevduatı Sigorta Fonu (Savings Deposit
Insurance Fund—​TMSF). The latter institution systematically dispossessed financially
troubled companies that were seen by the AKP as anti-​government, and transferred
their remaining capital to pro-​AKP corporations (Esen and Gumuscu 2018). These
policies not only emboldened pro-​government capital, but also intimidated the rest.
The AKP’s use of state power and resources to create a dependent business class is best
illustrated through the example of TOKİ. In January 2003, less than two months after it
formed its first government, the AKP passed an “Emergency Action Plan for Housing
and Urban Development.” It envisioned a five-​year plan to build 250,000 housing units,
which required vast resources and capabilities on the part of TOKİ. Over the following
several years, the government completely overhauled state involvement in the housing
sector (Demiralp, Demiralp, and Gümüş 2016). TOKİ was already endowed with real-​
estate company shares and real-​estate stock due to the earlier dissolution of the Emlak
Bankası (Real Estate Bank) in 2001 (TOKİ 2006). The institution was then given the
 Regime Change under the AKP in Turkey    153

legal right to make zoning plans, nationalize privately owned land and buildings, pro-
duce urban renewal projects, and build or finance construction projects in 2004. The
amendment of the Arsa Ofisi Kanunu (Land Office Law) in the same year provided
TOKİ with the ability to develop a virtually limitless amount of state-​owned land for the
purpose of housing construction (TOKİ 2006). TOKİ immediately acquired 64.5 mil-
lion square meters of land (TOKİ 2013). As a result of these changes, TOKİ was a com-
pletely different organization by 2005. It boasted an impressive financial infrastructure,
the ability to nationalize private property, and a mandate to develop state-​owned land
for housing construction.
The AKP then set out to transform Turkey’s housing sector. TOKİ surpassed its own
plans, and by 2013 it had started the construction of 565,000 housing units (TOKİ 2013).
The institution does not physically build apartments. It contracts out the engineering
and building process to private companies. TOKİ contracts have become a significant
means through which pro-​government construction companies accumulate capital
(Ocaklı 2018). Because TOKİ projects do not require the engineering expertise that
would be necessary to build more complex infrastructure projects, they can typically be
undertaken by small and medium-​sized companies. This has enabled the AKP to culti-
vate a government-​dependent bourgeoisie that it can easily control.
The specific case of the housing construction sector illustrates a wider phenomenon—​
the elaborate system through which the AKP uses state power to enrich pro-​government
businesses. TOKİ appropriates publicly owned land, transfers it to private companies
for construction purposes, enables them to capture the rent, and effectively privatizes
public resources. Thus, it has become a crucial mechanism through which the AKP has
built a pliant business elite.
In return for these favors, the AKP receives financial contributions, material support
for its election campaigns, and pro-​government media coverage. While direct financial
contributions are not insignificant (Cumhuriyet 2014; Sozcu 2014), the most important
factor in business support for the AKP is in the form of media coverage.
Before the 2001 economic crisis, Turkey’s media landscape was dominated by five
holding companies—​the Bilgin, Çukurova, Doğan, and Uzan Groups, and the indi-
vidual owner Korkmaz Yiğit (Yeşil 2018). All of these companies were invested in other
economic sectors, such as banking and finance, construction, energy, and tourism,
which made them susceptible both to government influence and to the business cycle.
The economic crisis led to a series of bankruptcies in these holding companies, which
resulted in government expropriation of their media properties by the TMSF. A case in
point is the Uzan Group. Cem Uzan established a political party named the Genc Parti
(Young Party—​GP) shortly before the 2002 elections and competed against the AKP.
When the AKP won the elections, the TMSF took control of the Uzan Group’s media
holdings. Bilgin faced financial troubles, while the Çukurova and Doğan Groups had
their media corporations gradually expropriated by the TMSF. The same institution
would later auction the Star newspaper and Kanal 24 TV station to the Sancak Group—​
a holding company closely allied to the AKP (Yeşil 2018). Other pro-​AKP corporations
would gradually buy these media institutions from the TMSF, as well as setting up new
154   Feryaz Ocaklı

ones (Esen and Gumuscu 2018). By 2013, the Albayrak, Ciner, Demirören, Doğuş,
Sancak, and Kalyon Groups controlled most of the media landscape (Yeşil 2018). These
openly pro-​AKP holding companies owned the most widely viewed TV channels
(NTV, Star TV, ATV, A Haber, Show TV, and others), as well as the widest circulation
newspapers (Yeni Safak, Milliyet, Aksam, Star, Sabah, and others). All of these holding
companies have investments in a wide variety of economic sectors that benefit from a
close relationship with President Erdoğan.
It is not surprising, then, that the AKP has acquired an unprecedented level of con-
trol over the flow of information in Turkey (Arbatli 2014). The distinction between real
news and “fake news” or outright government propaganda has blurred. Journalists have
not only been fired or pressured to tow the government’s line, they also report wide-
spread self-​censorship (Susma 2018). Meanwhile, journalists who resist the dominant
government-​friendly discourse in the media risk unemployment or, very often, impris-
onment (Beiser, 2018).
The emergence of a distinct pro-​AKP business elite has skewed the playing field be-
tween the incumbents and the opposition parties (Bekmen 2014; Somer 2016). In this
crony-​capitalist patronage network, the AKP supports a faction of capital, and it returns
the favor through donations, loyalty, and, most importantly, media control.
The AKP supplemented its system of party-​business collusion by adopting social wel-
fare programs that favored its potential supporters. Turkey’s social welfare programs
have traditionally been highly bureaucratic, limited, and inegalitarian. Family and in-
formal charity played a large role. Official social welfare programs were employment-​
based, and hence did not cover people who worked in the vast informal economy and
agriculture (Eder 2010). Following the 2001 economic crisis, the newly minted AKP gov-
ernment initiated an overhaul of the social welfare regime. The new system was organ-
ized at two levels: (1) programs that were managed at the national level, and (2) programs
that were decentralized and operated by a variety of actors at the local level.
At the national level, the AKP sought to address social welfare needs in two main
ways. The first was to rely on economic growth and macroeconomic stability to re-
duce poverty. As discussed earlier, rapid economic growth early in the AKP’s tenure
corresponded with an expanding middle class, which to some extent decreased the
need for social assistance. The second was the rationalization and expansion of the ex-
isting pensions and healthcare systems. Three state-​based but different healthcare and
pensions systems (the SSK, Bag-​Kur, and Emekli Sandigi) were unified under the um-
brella of the Social Security Institution (SSI). Healthcare assistance through the “green
card program” was expanded. Education continued to be state-​funded from the primary
school to the university level, although the quality of education between public and
private institutions differed widely. The affordable housing program led by TOKİ also
benefited lower income groups as well as producing revenue for pro-​AKP construction
companies.
The AKP’s main innovation in social welfare policy came in the form of decentraliza-
tion. This did not entail revenue growth for local governments to be used for social wel-
fare purposes—​public revenues in Turkey are remarkable centralized (Cammett, Luca,
 Regime Change under the AKP in Turkey    155

and Sergenti 2019). Instead, social welfare functions were devolved to local partnerships
between municipalities, religious charities, corporate philanthropy, and civil society or-
ganizations. Municipalities have acquired a significant role in welfare distribution (Eder
2010; Kaya 2015). They provide soup kitchens, iftar dinners during Ramadan, as well as
more general in-​kind aid. Municipal resources for welfare are based not on taxation but
on charity. Corporate philanthropy plays a significant role. Companies that receive pref-
erential access to contracts, tenders, and construction permits are induced to contribute
to “charity funds,” which are then spent on assistance to the poor (Buğra and Candas
2011). The Islamic banks and devout capital had long been parts of the social networks
connected to the Islamist movement (Gocmen 2014). The AKP tapped into their po-
tential and encouraged these companies to engage in charitable giving directed toward
the urban poor in conjunction with Islamic NGOs and local governments (Apaydin
2015). Taken as a whole, the AKP’s decentralized model substituted charitable giving
for rights-​based social welfare (Bugra and Candas 2011; Gocmen 2014). The decline of
union density and the normalization of flexible forms of work also decreased collective
agency on the part of the workers and reconstituted them as consumers and social assis-
tance recipients (Bozkurt 2013; Bozkurt-​Güngen 2018).
There is evidence to suggest that the AKP uses state spending as well as decentralized
social welfare to boost its support. Cammett et al. (2019) demonstrated that the party
used different budget lines for strategic objectives—​a practice known as portfolio di-
versification. Between 2003 and 2014, the AKP targeted reliably pro-​AKP districts with
firm subsidies and investments in education, defense, and cultural affairs. Battleground
districts received significantly more investment in infrastructure spending—​the kinds
of non-​excludable goods that symbolize good governance. Districts that reliably sup-
port the secularist opposition party received the lowest amount of public spending.
Aytac (2014) reached a similar conclusion in his analysis of the conditional cash transfer
program. He showed that the AKP government spent more in districts where it faced an
ideologically close challenger, and less in districts where its challengers were ideologi-
cally distant. Strategic use of state expenditures, particularly in the field of social welfare,
served to further strengthen the AKP’s electoral advantage over its competitors.

Institutional Dimensions
of Authoritarian Transition

Regime transition in Turkey was cemented by the AKP’s domination of the state’s se-
curity, administrative, judicial, and ideological apparatuses. Kemalist bureaucratic
institutions put up serious resistance against AKP hegemony. It took a variety of
strategies on the part of the latter to bring the former to heel. The initial years of AKP
rule (roughly 2002–​2008) were characterized by outward accommodation and EU-​
oriented reforms. Then (roughly 2008–​2011) came the war of court cases (Ergenekon
156   Feryaz Ocaklı

and Balyoz), through which the combined forces of the AKP and the Gülen move-
ment purged secularist officers from the military. The 2010 referendum and subsequent
legal changes sealed the government’s control over the judiciary. The following period
(roughly 2011–​2016) saw the conflict between the former allies, the AKP and the Gülen
movement, escalate. The failed coup attempt of 2016 resulted in the elimination of the
Gülen movement from Turkey’s economic, social, and political life, and the complete
domination of state institutions by the pro-​Erdoğan faction.
The AKP adopted post-​Islamist language and sought Turkey’s EU membership in its
first term in government (Yilmaz and Barry 2020). In order to further the EU accession
process, it initiated a series of reforms in 2004 and 2010 that would liberalize Turkey’s
laws in line with the Copenhagen Criteria. The most serious reforms concerned the role
of the military in politics (Satana 2008). The reforms increased parliamentary oversight
of the military budget, abolished military courts, reduced military presence in the Milli
Guvenlik Konseyi (National Military Council), and increased civilian control over the
Yuksek Askeri Sura (Supreme Military Council). The overall impact of these reforms
was the removal of institutional mechanisms that made military tutelage over the ci-
vilian government possible (Bayulgen, Arbatli, and Canbolat 2018). The AKP was able
to garner support for this program from liberal circles by emphasizing its necessity
for democratization. The military seems to have acceded to a reduction in its institu-
tional privileges because EU accession would be the culmination of an eighty-​year-​long
Kemalist modernization project.
During the same period, the party formed an alliance with the Gülen network. An
Islamic pietistic community-​ cum-​global business and education movement-​ cum-​
secret society, the Gülenists were a natural ally for the AKP (Yavuz 2018). The two had
separate historical trajectories, leaders, and mobilization strategies, but they shared a
common orientation in their opposition to Kemalist-​secularist institutions and policies
(Tas 2018). The Gülen network had already established a formidable set of market and
civil society institutions: a successful school system, financial institutions, a media
empire (most prominently the Zaman newspaper and Samanyolu TV), and a loosely
connected series of business associations. Less obvious to observers were its efforts to
raise a “golden generation” of acolytes who would join, rise in the ranks of, and eventu-
ally exert control over vital state bureaucracies (Yavuz 2018). Its alliance with the AKP
enabled the movement to place its members in leading positions in the police force, the
judiciary, regulatory organizations, and the military.
The AKP’s promotion of Gülenist cadres in the police and the judiciary provided the
party with well-​educated and highly skilled human resources with which to confront
Kemalist hegemony in state institutions. This investment paid off with the launch of two
court cases: Ergenekon and Balyoz, in 2008 and 2010, respectively. In both of these cases,
prosecutors affiliated with the Gülen movement launched investigations into alleged
plots to overthrow the elected government (Yavuz 2018). The hundreds of indictments
targeted mainly secularist military officers, but they later grew to include politicians,
academics, and journalists. Between 2008 and 2016, the court cases, prison sentences,
and lengthy jail terms pending trial had a devastating impact on the secularist officer
 Regime Change under the AKP in Turkey    157

corps. This process spelled the end of Kemalist dominance over the military, and estab-
lished the AKP-​Gülen alliance’s victory over the most prominent veto player in Turkish
politics.
It is important to note that all of the defendants in the Ergenekon and Balyoz trials
were later cleared of charges by the higher courts. This was partly due to insufficient
and false evidence. Doğan and Rodrik’s (2010) exposé on the Balyoz trial demonstrates
a concerted effort by factions of the police, judiciary, and the media to frame the
defendants through the systematic production of false evidence. However, this reversal
of fortunes only came after the collapse of the AKP-​Gülen alliance, and not due to the in-
ternal workings of an independent judicial branch. Zekariya Öz, the prosecutor who led
the Ergenekon case, had to flee the country in 2015 when the AKP government turned
against him after falling out with the Gülen movement (Demiralp 2016).
Judicial independence has historically been a mirage in Turkey. Similar to the military,
the high courts had a system of promotion that privileged Kemalist-​secularist judges
over others. The Constitutional Court in particular posed a serious threat to the AKP
during its first decade in office. The court had banned four parties that were the AKP’s
predecessors in the Islamist movement, and heard a case against the AKP itself in 2008.
The decision not to ban the AKP came down to a 6-​5 vote (Milliyet, 2008). The confron-
tation between the party and the high courts reached a decisive end with the 2010 con-
stitutional referendum. In conjunction with further legal changes in 2014, the AKP was
able to pack the highest institutions of the judiciary with its supporters. According to
Özbudun (2015), the constitutional amendments and additional legal changes enabled
the AKP to restructure the composition of the Constitutional Court, the Hakimler ve
Savcilar Yuksek Kurulu (Supreme Board of Judges and Prosecutors), and the other high
courts. These transformations effectively ended not only the judiciary’s resistance to the
AKP, but also any remnant of judicial autonomy from the executive branch (Akkoyunlu
2017; Somer 2016).
What followed this impressive process of institutional engineering was the deteri-
oration of the AKP-​Gülen alliance. The first signs of conflict began emerging in 2009
and 2010 following Erdoğan’s quarrels with Israeli leaders. The Gülen movement and
Erdoğan expressed conflicting views on Israel. Erdoğan grew more critical following the
2008 Operation Cast Lead in Gaza, and especially after the violent confrontation be-
tween the Israeli Navy and the Mavi Marmara flotilla that attempted to break the Gaza
blockade in 2010. Gülen sought accommodation between Turkey and Israel, and pub-
licly disapproved of Erdoğan’s outbursts (Tee 2018). The tensions deepened in 2012 when
the chief of the Milli Istihbarat Teskilati (National Intelligence Agency—​MIT) Hakan
Fidan was called in for questioning by Gülenist prosecutors. This escalation reflected
another disagreement between Erdoğan and Gülen. While Fidan was conducting secret
talks with the imprisoned PKK leader Abdullah Öcalan, Gülen opposed a potential rap-
prochement between the Turkish state and the Kurdish National Movement (Uras 2013).
The rift between the allies reached crisis levels in December 2013. A division of the
Istanbul Security Directory initiated a corruption probe targeting the highest levels
of the AKP government. More than ninety people were taken into custody, including
158   Feryaz Ocaklı

AKP-​affiliated businessmen, a mayor, and sons of cabinet ministers (Tee 2018). The
AKP responded by denying all allegations, passing legislation that gave the government
new powers over the judiciary, and initiating a purge of Gülenists in the police force
and the judiciary (Shively 2016). This crisis exposed not only the deep corruption of the
AKP leaders, but also the extent to which the Gülen movement had been able to infil-
trate the state and wiretap almost everyone of significance in Ankara. Tapes leaked on
YouTube in February 2014 allegedly contained private conversations between Erdoğan
and his son, as well as a classified discussion of Turkish foreign policy toward Syria be-
tween MIT chief Hakan Fidan, the minister of foreign affairs, and high-​ranking gen-
erals. Many other tapes were leaked later, allegedly containing damaging excerpts from
Erdoğan’s private phone conversations.
Throughout the growing conflict between the AKP and the Gülen movement, the
government’s almost reflexive response was to pass new legislation that increased its
grip on the police force and the judiciary. These were the two institutions most heavily
dominated by the Gülen movement (Yavuz 2018). Erdoğan’s attempt to wrest control
over them in order to weaken the Gülen movement increasingly subordinated the bu-
reaucracy to the will of the AKP.
The crisis between the former allies metastasized during the failed military coup at-
tempt on July 15, 2016. Although there is no conclusive evidence that the coup was or-
ganized with the explicit blessing of Fethullah Gülen, the leader of the Gülen movement,
all available evidence points toward the central role played by a network of Gülenist mil-
itary officers. The Kemalist-​secularist officer corps was already purged and thoroughly
demoralized by the Ergenekon and Balyoz trials. This had made possible the promotion
of Gülen-​affiliated officers to higher ranks during the height of the AKP-​Gülen alliance.
By July 2016 the crisis between the former allies was severe, and numerous newspapers
(both pro-​and anti-​government) were circulating rumors about a possible purge of
Gülenist officers during the August summit of the Supreme Military Council (Likoglu
2016; Ozturk 2016). The coup attempt was most likely the final gasp of a movement
under siege.
The coup itself was bloody. The plotters occupied a state-​owned television station,
one of the bridges over the Bosporus, and a number of police stations and military
bases. They bombed the Turkish Parliament and the MIT headquarters. They attacked a
Marmaris hotel where Erdoğan was vacationing, only to miss him by 15 minutes (Yavuz
and Koc 2018). Overall, the coup attempt led to 248 deaths and 2,196 injuries (Sabah
2017). The number of casualties was high because civilians across the country mobilized
against the coup attempt.
The failure of the coup marked the beginning of a new chapter in Turkish politics.
Erdoğan’s fierce reaction against not only the coup plotters, but also a much wider spec-
trum of opposition, initiated a process that may appropriately be termed a counter-​coup
(Tas 2018). The immediate declaration of emergency rule empowered the government
to start a wide-​ranging wave of investigations, arrests, dismissals, and closures. Overall,
more than 150,000 people were dismissed from their jobs. These were mostly teachers,
police officers, bureaucrats at the Ministry of the Interior, military officers, academics,
 Regime Change under the AKP in Turkey    159

health personnel, and members of the judiciary (BBC 2017). More than 500,000 people
were investigated, and around 97,000 were arrested (TurkeyPurge). By 2019 there were
around 31,000 people in prison connected to the investigations, and a further 22,000
had warrants for arrest (Cumhuriyet 2019). In addition, thousands of dormitories and
schools were shut down, and Gülen-​affiliated financial, commercial, and media organi-
zations were expropriated.
Erdoğan’s purge of state institutions and consolidation of power following the coup
attempt culminated in the 2017 constitutional referendum and the 2018 presidential
elections. The referendum officially transitioned Turkey from a parliamentary gov-
ernment to an “executive presidency.” It confirmed what was already clear: Erdoğan
had successfully centralized power, marginalized potential rivals within the AKP,
and defeated his secularist and Gülenist enemies. The re-​engineering of Turkish state
institutions was complete.

Conclusion

Turkey gradually transitioned from an illiberal democracy to an authoritarian regime

between 2002 and 2016. The main driver of the change was the elected government.
Threats to democracy came from the center—​entrenched powers in the bureaucracy,
the growing influence of the Gülen movement in vital state institutions, and the AKP
itself. Erdoğan’s centralization of power was underwritten by an unwavering electoral
base in a thoroughly polarized society.
Persistent electoral victories throughout the 2000s provided the AKP with the time
and ability to shift the balance of power in state institutions. The AKP was able to take
advantage of a favorable economic climate and displace the traditional center-​right
parties from the political arena. Erdoğan preempted the emergence of potential rivals in
the Islamist movement and the center-​right by co-​opting promising politicians into the
AKP. Furthermore, he gradually marginalized the other founders of the AKP in a suc-
cessful effort to establish his supremacy in the party organization.
Starting in its first year in office, the AKP laid the groundwork of a highly effective pa-
tronage scheme that would distribute state resources to pro-​government business elites.
State contracts, particularly in the areas of procurement and construction, became the
main vehicles through which pro-​AKP companies accumulated capital. Meanwhile,
rival business groups were targeted with punitive tax audits as well as outright expro-
priation in times of financial trouble. In return for state contracts and preferential treat-
ment, government-​friendly corporations made campaign contributions, gave charity to
municipal social welfare funds, and, more importantly, invested in the media sector. The
government gradually gained control over information flows by making sure that the
media landscape was controlled by pro-​AKP capital. This gave the party an invaluable
advantage over its competitors, as opposition parties and movements received very little
favorable coverage.
160   Feryaz Ocaklı

The competition over the control of state bureaucracies was less visible. Kemalist mil-
itary officers and judges resisted AKP hegemony. The party partnered with the Gülen
movement to advance their mutual goal—​the elimination of Kemalist-​secularists from
the high echelons of the state bureaucracy. To that end, the AKP promoted Gülenist
cadres in the police force and the judiciary. The AKP-​Gülen alliance was then able to
purge secularist officers from the military via the Ergenkon and Balyoz court cases. They
later packed the courts with AKP supporters after the 2010 constitutional amendments.
This partnership broke down in the early 2010s, leading to a number of crises that
pitted the Gülen movement against the AKP. The crisis culminated in the 2016 coup at-
tempt. The coup’s failure gave Erdoğan the upper hand, and the Gülen movement was
eliminated from Turkey’s social, economic, and political life. As a result, Erdoğan gained
uncontested dominance over all state institutions.

References
Acemoglu, Daron, and Murat Ucer. 2015. “The Ups and Downs of Turkish Growth, 2002–​
2015: Political Dynamics, the European Union and the Institutional Slide.” Working Paper
No. 21608 (October). Cambridge, MA: National Bureau of Economic Research. https://​
www.nber.org/​papers/​w21608.pdf.
Akkoyunlu, Karabekir. 2017. “Electoral Integrity in Turkey: From Tutelary Democracy to
Competitive Authoritarianism.” In Authoritarian Politics in Turkey: Elections, Resistance and
the AKP, edited by Bahar Baser and Ahmet Erdi Öztürk, 47–​63. New York: I.B. Tauris.
Al Jazeera. 2015. “Eskiden ‘biz’dik, şimdi ‘ben’e döndü” [We used to be “us,” now this turned
into “I”]. Al Jazeera Türk, September 11, 2015. http://​www.aljazeera.com.tr/​haber/​
eskiden-​bizdik-​simdi-​bene-​dondu.
Apaydin, Fulya. 2015. “Financialization and the Push for Non-​ state Social Service
Provision: Philanthropic Activities of Islamic and Conventional Banks in Turkey.” Forum for
Development Studies 42, no. 3: 441–​465.
Arbatli, Ekim. 2014. “Turkey’s New Path: The Rise of Electoral Authoritarianism.” Centre for
Research and Policy on Turkey 3, no. 12: 76–​92.
Aytaç, S. Erdem. 2014. “Distributive Politics in a Multiparty System: The Conditional Cash
Transfer Program in Turkey.” Comparative Political Studies 47, no. 9: 1211–​1237.
Aytaç, S. Erdem, and Ziya Öniș. 2014. “Varieties of Populism in a Changing Global Context: The
Divergent Paths of Erdoğan and Kirchnerismo.” Comparative Politics 47, no. 1: 41–​59.
Baffes, John, and Tassos Haniotis. 2010. “Placing the 2006/​08 Commodity Price Boom into
Perspective.” Policy Research Working Paper No. 5371. Washington, DC: World Bank. http://​
documents.worldbank.org/​curated/​en/​921521468326680723/​pdf/​WPS5371.pdf.
Bakıner, Onur. 2017. “How Did We Get Here? Turkey’s Slow Shift to Authoritarianism.” In
Authoritarian Politics in Turkey: Elections, Resistance and the AKP, edited by Bahar Bașer
and Ahmet Erdi Öztürk, 21–​46. London: I.B. Tauris.
Barua, Akrur, and Navya Kumar. 2014. “The Boom and Beyond: Managing Commodity Price
Cycles.” In Global Economic Outlook: 1st Quarter 2014. Deloitte University Press. https://​
www2.deloitte.com/​content/​d am/​D eloitte/​de/​D ocuments/​f inancial-​s ervices/​Global-​
Economic-​Outlook-​Q1-​2014.pdf.
Bayrakci, Israfil. 2015. Interview with the author, September 16, 2015.
 Regime Change under the AKP in Turkey    161

Bayulgen, Oksan, Ekim Arbatli, and Sercan Canbolat. 2018. “Elite Survival Strategies and
Authoritarian Reversal in Turkey.” Polity 50, no. 3: 333–​365.
BBC. 2017. “KHK ile işten atılanlar: Yaşadıklarımız sosyal idam” [Those dismissed via execu-
tive action: What we are living though is a social death sentence]. BBC News, April 13, 2017.
https://​www.bbc.com/​turkce/​39580439.
Beiser, Elana. 2018. Hundreds Of Journalists Jailed Globally Becomes the New Normal.
New York: Committee to Protect Journalists, December 13. https://​cpj.org/​reports/​2018/​12/​
journalists-​jailed-​imprisoned-​turkey-​china-​egypt-​saudi-​arabia.php.
Bekmen, Ahmet. 2014. “State and Capital in Turkey during the Neoliberal Era.” In Turkey
Reframed: Constituting Neoliberal Hegemony, edited by Ismet Akça, Ahmet Bekmen, and
Barıș Alp Özden, 47–​74. London: Pluto Press.
Berlinski, Claire. 2017. “Guilty Men.” The American Interest, April 24. https://​www.the-​
american-​interest.com/​2017/​04/​24/​guilty-​men.
Bermeo, Nancy. 2016. “On Democratic Backsliding.” Journal of Democracy 27, no. 1: 5–​19.
Bilgic, Ali. 2018. “Reclaiming the Turkish National Will: Resilience of Turkish Authoritarian
Neoliberalism after Gezi.” South European Society & Politics 23, no. 2: 259–​280.
Bozkurt, Umut. 2013. “Neoliberalism with a Human Face: Making Sense of the Justice
and Development Party's Neoliberal Populism in Turkey.” Science & Society 77, no.
3: 372–​396.
Bozkurt-​Güngen, Sümercan. 2018. “Labour and Authoritarian Neoliberalism: Changes and
Continuities under the AKP Governments in Turkey.” South European Society and Politics
23, no. 2: 219–​238.
Buğra, Ayșe, and Ayșen Candaș. 2011. “Change and Continuity under an Eclectic Social
Security Regime: The Case of Turkey.” Middle Eastern Studies 47, no. 3: 515–​528.
Buğra, Ayșe, and Osman Savaşkan. 2014. New Capitalism in Turkey: The Relationship between
Politics, Religion and Business. Cheltenham, UK: Edward Elgar.
Çalışkan, Koray. 2018. “Toward a New Political Regime in Turkey: From Competitive toward
Full Authoritarianism.” New Perspectives on Turkey 58: 5–​33.
Cammett, Melani, David Luca, and Ernest Sergenti. 2019. “Inclusive Distribution as
Electoral Strategy: The Politics of Turkey’s Central Government Spending under AKP
Rule.” Unpublished manuscript, March 2019. https://​www.researchgate.net/​publica-
tion/​323800418_​%27Inclusive%27_​distribution_​as_​electoral_​strategy_​The_​politics_​of_​
Turkish_​central_​government_​spending_​under_​AK_​Party_​rule.
Çarkoğlu, Ali. 2007. “The Rise of the New Generation Pro-​Islamists in Turkey: The Justice
and Development Party Phenomenon in the November 2002 Elections in Turkey.” South
European Society and Politics 7, no. 3: 123–​156.
Cumhuriyet. 2014. “İşte Erdoğan'a yapılan bağış” [Here is the donation made to Erdoğan].
Cumhuriyet, August 2, 2014. http://​www.cumhuriyet.com.tr/​haber/​turkiye/​100997/​iste_​
Erdoğan_​a_​yapilan_​bagis.html.
Cumhuriyet. 2019. “İşte yargıdaki FETÖ bilançosu” [Here is the FETO balance sheet in the
judiciary]. Cumhuriyet. March 4, 2019. http://​www.cumhuriyet.com.tr/​haber/​turkiye/​
1276109/​iste_​yargidaki_​FETO_​bilancosu.html.
Demiralp, Seda. 2016. “The Breaking Up of Turkey’s Islamic Alliance: The AKP-​Gülen Conflict
and Implications for Middle East Studies.” Middle East Review of International Affairs 20, no.
1: 1–​7.
Demiralp, Seda, Selva Demiralp, and Inci Gümüş. 2016. “The State of Property Development in
Turkey: Facts and Comparisons.” New Perspectives on Turkey 55: 85–​106.
162   Feryaz Ocaklı

Dogan, Pınar, and Dani Rodrik. 2010. Balyoz: Bir Darbe Kurgusunun Belgeleri ve Gerçekler
[Sledgehammer: The documents of a fictional coup and facts]. Istanbul: Destek Yayinevi.
Eder, Mine. 2010. “Retreating State? Political Economy of Welfare Regime Change in Turkey.”
Middle East Law and Governance 2, no. 2: 152–​184.
Eligür, Banu. 2010. The Mobilization of Political Islam in Turkey. Cambridge: Cambridge
University Press.
Esen, Berk, and Sebnem Gumuscu. 2018. “Building a Competitive Authoritarian Regime: State-​
Business Relations in the AKP’s Turkey.” Journal of Balkan and Near Eastern Studies 20, no.
4: 349–​372.
Evrensel. 2018. “Veliahtlık kavgası'nda son perde: Soylu, Albayrak'a omuz attı”
[Last act in the struggle to be the crown prince: Soylu hit Albayrak with his
shoulder]. Evrensel, August 3, 2018. https://​www.evrensel.net/​haber/​358429/​
veliahtlik-​kavgasinda-​son-​perde-​soylu-​albayraka-​omuz-​atti.
Gidengil, Elisabeth, and Ekrem Karkoç. 2016. “Which Matters More in the Electoral Success of
Islamist (Successor) Parties—​Religion or Performance? The Turkish Case.” Party Politics 22,
no. 3: 325–​338.
Gocmen, Ipek. 2014. “Religion, Politics and Social Assistance in Turkey: The Rise of Religiously
Motivated Associations.” Journal of European Social Policy 24, no. 1: 92–​103.
Goebel, Christian. 2011. “Authoritarian Consolidation.” European Political Science, 10, no.
2: 176–​190.
Grigoriadis, Ioannis N., and Esra Dilek. 2018. “Struggling for the Kurdish Vote: Religion,
Ethnicity and Victimhood in AKP and BDP/​HDP Rally Speeches.” Middle Eastern Studies
52, no. 2: 289–​303.
Gülalp, Haldun. 2001. “Globalization and Political Islam: The Social Basis of Turkey’s Welfare
Party.” International Journal of Middle East Studies 33, no. 3: 433–​448.
Gumuscu, Sebnem. 2010. “Class, Status, and Party: The Changing Face of Political Islam in
Turkey and Egypt.” Comparative Political Studies 43, no. 7: 835–​861.
Gurakar, Esra Çeviker. 2016. Politics of Favoritism in Public Procurement in Turkey: Reconfigurations
of Dependency Networks in the AKP Era. London: Palgrave Macmillan.
Kamer, Hatice. 2018. “HDP'nin 9 Milletvekilinin Tutuklanmasının Ikinci Yılı” [Second year
of the imprisonment of 9 HDP parliamentarians]. BBC Turkce, November 4, 2018. https://​
www.bbc.com/​turkce/​haberler-​turkiye-​46088912.
Kaya, Ayhan. 2015. “Islamisation of Turkey under the AKP Rule: Empowering Family, Faith
and Charity.” South European Society and Politics 20, no. 1: 47–​69.
Klimek, Peter, Raul Jiménez, Manuel Hidalgo, Abraham Hinteregger, and Stefan Thurner.
2018. “Forensic Analysis of Turkish Elections in 2017–​2018.” PLoS ONE 13, no. 10: 1–​14.
Levitsky, Steven, and James Loxton. 2013. “Populism and Competitive Authoritarianism in the
Andes.” Democratization 20, no. 1: 107–​136.
Likoğlu, Hüseyin. 2016. “Şakird subayların altın vuruşu” [The golden hit of the cadets].
Yeni Safak, June 27, 2016. https://​www.yenisafak.com/​yazarlar/​huseyinlikoglu/​
akird-​subaylarin-​altin-​vuruu-​2030019.
Mebane, Walter R., Jr., Allen Hicken, and Ken Kollman. 2016. “Were There Irregularities in
Turkey’s 2015 Elections? We Used this New Forensic Toolkit to Check.” Washington Post,
February 15, 2016. https://​www.washingtonpost.com/​news/​monkey-​cage/​wp/​2016/​02/​15/​
were-​there-​irregularities-​in-​turkeys-​2015-​elections-​we-​used-​our-​new-​forensic-​toolkit-​to-​
check/​?utm_​term=.0798f2bee5d0.
 Regime Change under the AKP in Turkey    163

Mecham, Quinn. 2004. “From the Ashes of Virtue, a Promise of Light: The Transformation of
Political Islam in Turkey.” Third World Quarterly 25, no. 2: 339–​358.
Milliyet. 2008. “AK Parti kapatılmasın kararı çıktı” [The AKP will
not be closed]. Milliyet, July 30, 2008. http://​www.milliyet.com.tr/​
karar—​font-​color-​red-​bugun-​mu—​font—​aciklanacak—​siyaset-​972729.
Ocaklı, Feryaz. 2015a. “Notable Networks: Elite Recruitment, Organizational Cohesiveness,
and Islamist Electoral Success in Turkey.” Politics & Society 43, no. 3: 385–​413.
Ocaklı, Feryaz. 2015b. “Why Is Turkey Attacking the Kurdish Militants Instead of
ISIS?” Small Wars Journal, August 28, 2015. https://​smallwarsjournal.com/​jrnl/​art/​
why-​is-​turkey-​attacking-​the-​kurdish-​militants-​instead-​of-​isis.
Ocaklı, Feryaz. 2016a. “Islamist Politics in Migrant Cities: Local Identity Cleavages and Intra-​
Party Politics in Turkey.” British Journal of Middle Eastern Studies 43, no. 4: 661–​682.
Ocaklı, Feryaz. 2016b. “Political Entrepreneurs, Clientelism, and Civil Society: Supply-​Side
Politics in Turkey.” Democratization 23, no. 4: 723–​746.
Ocaklı, Feryaz. 2017a. “Islamist Mobilisation in Secularist Strongholds: Institutional Change
and Electoral Performance in Turkey.” South European Society and Politics 22, no. 1: 61–​80.
Ocaklı, Feryaz. 2017b. “Politics in the Kurdish Periphery: Clan Networks and Local Party
Strategies in a Comparative Perspective.” Middle Eastern Studies 53, no. 4: 571–​584.
Ocaklı, Feryaz. 2018. “Reconfiguring State-​ Business Relations in Turkey: Housing and
Hydroelectric Energy Sectors in Comparative Perspective.” Journal of Balkan and Near
Eastern Studies 20, no. 4: 373–​387.
Öniș, Ziya. 1997. “The Political Economy of Islamic Resurgence in Turkey: The Rise of the
Welfare Party in Perspective.” Third World Quarterly 18, no. 4: 743–​766.
Öniș, Ziya. 2010. “Crises and Transformations in Turkish Political Economy.” Turkish Policy
Quarterly 9, no. 3: 45–​61.
Öniș, Ziya. 2012. “The Triumph of Conservative Globalism: The Political Economy of the AKP
Era.” Turkish Studies 13, no. 2: 135–​152.
Özbudun, Ergun. 2015. “Turkey’s Judiciary and the Drift toward Competitive
Authoritarianism.” International Spectator: Italian Journal of International Affairs 50, no.
2: 42–​55.
Özturk, Saygı. 2016. “Genelkurmay’dan ‘ayıklanacaklar’ listesi hazırlandı”
[A list of the “to be winnowed” has been prepared by the General Staff].
Sozcu, June 26, 2016. https://​www.sozcu.com.tr/​2016/​yazarlar/​saygi-​ozturk/​
genelkurmaydan-​ayiklanacaklar-​listesi-​hazirlandi-​1291347.
Sabah. 2017. “15 Temmuz darbe girişiminin acı bilançosu! Kaç kişi şehit oldu?” [The
painful balance sheet of the July 15 coup attempt! How many people were mar-
tyred?] Sabah, July 10, 2017. https://​www.sabah.com.tr/​gundem/​2017/​07/​10/​
15-​temmuz-​darbe-​girisiminin-​aci-​bilancosu-​kac-​kisi-​sehit-​oldu.
Satana, Nil S. 2008. “Transformation of the Turkish Military and the Path to Democracy.”
Armed Forces and Society 34, no. 3: 357–​388.
Schedler, Andreas. 2015. “Electoral Authoritarianism.” In Emerging Trends in the Social and
Behavioral Sciences, edited by Robert Scott and Stephen Kosslyn, 1–​16. Hoboken, NJ: John
Wiley & Sons.
Sendika.org. 2017. “Referandumda Yaşanan 8 Usulsüzlük” [Eight Irregularities
in the Referendum]. Sendika.org, April 17, 2017. http://​sendika63.org/​2017/​04/​
16-​nisan-​bitti-​referandumda-​yasanan-​8-​usulsuzluk-​415160.
164   Feryaz Ocaklı

Sezgin, Ipek Gencel. 2011. “Political Engagement Patterns of Islamist Movements: The Case
of the Nizam/​Selamet Movement.” PhD diss., İhsan Doğramacı Bilkent University, Ankara.
Shively, Kim. 2016. “Pragmatic Politics: The Gülen Movement and the AKP.” In The Turkish
AK Party and Its Leader: Criticism, Opposition and Dissent, edited by Umit Cizre, 183–​204.
New York: Routledge.
Svolik, Milan W. 2020. “When Polarization Trumps Civic Virtue: Partisan Conflict and the
Subversion of Democracy by Incumbents.” Quarterly Journal of Political Science 15: 3–​31.
Somer, Murat. 2016. “Understanding Turkey’s Democratic Breakdown: Old vs. New and
Indigenous vs. Global Authoritarianism.” Southeast European and Black Sea Studies 16, no.
4: 481–​503.
Somer, Murat. 2019. “Turkey: The Slippery Slope from Reformist to Revolutionary Polarization
and Democratic Breakdown.” ANNALS of the American Academy of Political and Social
Science 681, no. 1: 42–​61.
Sozcu. 2014. “Patronlara ‘bağış yapın’ baskısı” [Donation pressure on the bosses]. Sozcu, July
28, 2014. https://​www.sozcu.com.tr/​2014/​gundem/​patronlara-​bagis-​yapin-​baskisi-​565980.
Susma (Speak Up) Platform against Censorship and Self-​Censorship. 2018. Censorship and
Self-​Censorship in Turkey: September 2016-​December 2017. http://​susma24.com/​wp-​content/​
uploads/​2018/​02/​Tu_​rkiye_​de_​Sansu_​r_​-​ve_​Otosansür_​ENG.pdf.
Tas, Hakkı. 2018. “A History of Turkey’s AKP-​Gülen Conflict.” Mediterranean Politics 23, no.
3: 395–​402.
Tee, Caroline. 2018. “The Gülen Movement and the AK Party: The Rise and Fall of a Turkish
Islamist Alliance.” In Turkey’s July 15 Coup: What Happened and Why, edited by Hakan Yavuz
and Bayram Balci, 150–​172. Salt Lake City: University of Utah Press.
TOKİ (Toplu Konut İdaresi). 2006. Turkiye’de Konut Sektoru ve T.C. Basbakanlik Konut
Idaresi’nin (TOKİ) Konut Uretimindeki Yeri [Housing sector in Turkey and the role of TOKİ
in housing production]. TOKİ Arastirma Dizisi 2 [TOKI Research Series 2]. Ankara: TOKİ.
TOKİ (Toplu Konut İdaresi). 2013. Due Diligence Document. Ankara: TOKİ.
Tuğal, Cihan. 2002. “Islamism in Turkey: Beyond Instrument and Meaning.” Economy and
Society 31, no. 1: 85–​111.
TurkeyPurge. “Turkey’s Post-​Coup Crackdown.” Accessed April 13, 2019. https://​turkeypurge.
com.
TUIK (Türkiye Istatistik Kurumu [Turkish Statistical Institute]). “Adalet ve Seçim İstatistikleri”
[Justice and Election Statistics]. Accessed April 13, 2019. https://​data.tuik.gov.tr/​Kategori/​
GetKategori?p=Adalet-​ve-​Secim-​110.
Uras, Umut. 2013. “Turkish Probe Marks AKP-​Gülen Power Struggle.” Al Jazeera, December
24, 2013. https://​www.aljazeera.com/​indepth/​features/​2013/​12/​turkish-​probe-​marks-​akp-​
Gülen-​power-​struggle-​2013122473646994231.html.
White, Jenny B. 2002. Islamist Mobilization in Turkey: A Study in Vernacular Politics.
Seattle: University of Washington Press.
Yavuz, Hakan. 1999. “Towards an Islamic Liberalism?: The Nurcu Movement and Fethullah
Gülen.” Middle East Journal 53, no. 4: 584–​605.
Yavuz, Hakan. 2009. Secularism and Muslim Democracy in Turkey. New York: Cambridge
University Press.
Yavuz, Hakan. 2018. “The Three Stages of the Gülen Movement: From Pietistic Weeping
Movement to Power-​Obsessed Structure.” In Turkey’s July 15 Coup: What Happened and
Why, edited by Hakan Yavuz and Bayram Balci. Salt Lake City: University of Utah Press.
 Regime Change under the AKP in Turkey    165

Yavuz, Hakan, and Rasim Koc. 2018. “The Gülen Movement vs. Erdoğan: The Failed Coup.” In
Turkey’s July 15 Coup: What Happened and Why, edited by Hakan Yavuz and Bayram Balci.
Salt Lake City: The University of Utah Press.
Yeşil, Bilge. 2018. “Authoritarian Turn or Continuity? Governance of Media through Capture
and Discipline in the AKP Era.” South European Society and Politics 23, no. 2: 239–​257.
Yilmaz, Ihsan, and James Barry. 2020. “Instrumentalizing Islam in a ‘Secular’ State: Turkey’s
Diyanet and Interfaith Dialogue.” Journal of Balkan and Near Eastern Studies 22, no. 1: 1–​16.
Yilmaz, Ihsan, Greg Barton, and James Barry. 2017. “The Decline and Resurgence of Turkish
Islamism: The Story of Tayyip Erdoğan’s AKP.” Journal of Citizenship and Globalisation
Studies 1, no. 1: 48–​62.
Yilmaz, Ihsan, and Galib Bashirov. 2018. “The AKP after 15 Years: Emergence of Erdoğanism in
Turkey.” Third World Quarterly 39, no. 9: 1812–​1830.
Zurcher, Erik J. 2004. Turkey: A Modern History. London: I.B. Tauris.
 Chapter 8

Isl am , Nationa l i sm, a nd

Demo cracy i n Asia
Nations under Gods or Gods under Nations?

Maya Tudor

Islam is no more dangerous to democracy than any other religion. Yet research into the
Islamic world has generally assumed rather than interrogated a prominence of religious
beliefs in explaining social and political outcomes. As part of a volume that theoretically
explores whether religiosity structures politics across the Muslim world, this chapter
argues that religiosity is significantly more likely to undermine democracy when national
narratives historically centralized religious identities. This is because when national
narratives centralized religion, religious actors subsequently became powerful actors in
political struggles, creating another set of interests to accommodate, and because the
majoritarian ideas of citizenship that allowed subsequent political entrepreneurs to sub-
ordinate minority rights gained broad acceptance. Empirically, this chapter argues that
the varying centrality of religious identity in founding nationalisms has critically condi-
tioned the prospects for democracy in India, Indonesia, and Pakistan.
Nations have been perhaps most famously defined as “imagined communities.” And
while much attention has been paid to the word imagined, less noticed is that nations
are still important sources of community in a globalized world. Nations, and the nation-
alism that celebrates them, transforms solitary individuals into homelands, fulfilling an
abiding desire for belonging in the modern world (Tamir 2019). All communities can
spawn discrimination against outsiders. But the tendency to denigrate outsiders is no
more inherent to the community of nations than it is to communities such as families or
neighborhoods (Li and Brewer 2004).
What distinguishes nations from other kinds of communities is not that celebration
of nations is especially likely to turn against outsiders. Instead, it is that the nation is a
distinctly political community, intimately tied to the legitimation of state power. Nations
168   Maya Tudor

so centrally underpin our modern system of sovereign states that an eminent scholar
of nationalism has gone so far as to say that it is the nation that, above all, defines mo-
dernity (Greenfeld 1992). The celebration of the nation helps to affirm the power of the
state, which is after all that entity with the legitimate monopoly on the use of force. It is
for this reason that nations and nationalism are here to stay (Wimmer 2019).
If national communities legitimate the use of state power, then a natural next ques-
tion is whether different types of national communities have affect outcomes of in-
terest within political science. But the comparative scholarship of nationalism has
historically been surprisingly quiet on this topic, tending instead to privilege re-
search into the origins and spread of nationalism (Mylonas and Tudor Forthcoming).
Because nationalism’s potential to undermine democracy prominently figured in the
experiences of twentieth-​century Europe, comparative scholarship on nationalism has
tended to minimize the potential for nationalism to promote freedom and democracy,
as it has for many twentieth-​century ex-​colonies that adapted European nationalism to
gain colonial independence.
Across the colonial world, indigenous elites imagined and mobilized on behalf of new
nations in the middle decades of the twentieth century. When a close identification be-
tween religion and nation was forged in order to legitimate a new regime in the name
of a new people, this ideological marriage was subsequently difficult to disentangle
and dampened the possibilities for postcolonial democracies to endure. There are two
reasons for this.
First, when national narratives are narrowly married to religion, those institutions
representing religious identities are typically emboldened to adjudicate on who is a
good follower as well as who is a good citizen. After independence, when new patterns
of power-​sharing are first beginning to gel, the addition of a “veto player” makes it more
difficult to consolidate democratic rules of the game, resulting in political instability
during the critical regime-​founding moment (Tsebelis 1995). The centrality of religious
institutions to political power also, over time, embeds questions of religious identity
into everyday political debates, creating a politics that is at once relevant to governments
and state institutions for legitimating purposes, and yet not entirely controlled by them.
Second, when founding national identities are conceptualized in terms of a religion,
status inequalities are invariably created among citizens. This is because individuals
without the core religious identity are definitionally less central to the nation and are
thus more vulnerable to political targeting. These status inequalities are often used
by political entrepreneurs seeking to gain power. A national narrative tied to religion
enables, and can even encourage, politicians to target those citizens who do not possess
the core religion. Especially in times of crisis, politicians and would-​be politicians can
more easily propose that such second-​class citizens are less deserving of state protection.
I explore this argument through the historical experiences of three postcolonial
countries—​Pakistan, Indonesia, and India—​whose elites adopted different bases of na-
tionalist mobilization. In all three countries, elites waited until the middle decades of
the twentieth century to engage in anticolonial mass mobilization, garnering little in the
way of sustained mass support until the closing decades before independence. When
 Islam, Nationalism, and Democracy in Asia    169

the nationalist movements did engage in mass mobilization, they did so by using dif-
ferent ideational building blocks.
Pakistan’s independence movement popularized its call for independence by
instrumentalizing religion, whereas India’s and Indonesia’s independence movements
espoused a set of creedal principles to mobilize the anticolonial movement. In
Pakistan, the regular invocation of religion as a means of legitimating state policies and
governments has strengthened the relationships between religion and nation and has
regularly provided fodder for debilitating political debates. By contrast, the relative dis-
tance between religion and nation in Indonesia and India prevented religion from as-
suming a pivotal place in political debates at fragile political moments. Indonesia’s and
India’s national narrative eschewed religion and “settled” into an inclusive narrative of
nation (Bonikowski 2016). The force of religion was thus, until the recent decade, sur-
prisingly minimal within political debates. It is not religiosity per se, but rather the cen-
trality of religion in the national imagination, that has debilitating consequences for
democracy.

State-​Building and Nation-​

Building: The Twentieth-​C entury
Postcolonial Context

In the colonial world, the years before, during, and just after national independence
were a critical period in which regime patterns were formed. The national narratives
adopted to pursue colonial independence would prove formative for ensuing regime
dynamics because nationalist elites used these narratives to mobilize support and
embedded them in the national imagination in commemorative sites such as maps,
museums, national holidays, festivals, and school textbooks (Anderson 1983, vom Hau
2009, Zubrzycki 2016). It was through these avenues that the decisions of national elites
about whom to mobilize with what set of identities influenced whether newly sovereign
governments could create a stable regime. These decisions had enormous consequences
for such outcomes as regime stability, party systems, economic development, and geno-
cide (Yashar 1997, Slater 2010, Tudor 2013, Riedl 2014, Liu 2015, Straus 2015).
Despite the fact that many societies in Asia were deeply religious societies, there was
a great deal of variation in whether and how national elites embraced religion. To be
sure, religious motifs and symbols were almost everywhere employed as the “basic cul-
tural and ideological building blocks for nationalists” (Smith 2003, 254–​255). But un-
like the European experience with state formation, Asian countries typically lacked
centralized and highly organized religious forces that claimed sovereignty over states.
Consequently, when nationalism was imported and adapted to local contexts, elites had
greater choice about whether to centralize religion (as well as ethnicity and other salient
social cleavages) when defining a new nation.
170   Maya Tudor

Pakistan: Religious Nationalism Feeds

Autocratic Instability

For all of its seventy-​plus years since colonial independence, Pakistan has been an un-
stable regime with autocratic tendencies. This outcome has been critically enabled by
a founding nationalism that explicitly embraced Islam as the basis of the nation, a fu-
sion that proved difficult to subsequently repudiate, and that enabled religious leaders
to become powerful political voices in key political struggles and encouraged political
entrepreneurs to target religious minorities for legitimation.

The Call for Pakistan Fuses Nation and Islam

Pakistan’s nationalist movement—​the Pakistan Muslim League—​was a movement
created in 1906 primarily to protect the privileges of a group of culturally and socioeco-
nomically distinct Muslims who were well-​entrenched in British India’s colonial power
structures. While colonially entrenched landed aristocrats everywhere were threatened
by the mobilization of the urban, educated middle class and its calls for political reform,
they were generally too divided by hierarchy and language to create a unified political
movement. In the United Provinces however, language, religion, and government pa-
tronage cleavages all coincided and the prospect of reforms were thus threatening in
multiple ways. This threat incentivized sufficient mobilization to engender a political
movement, the Muslim League. Once quotas for the political representation of UP
Muslims was guaranteed in the 1909, the Muslim League remained a marginal political
movement until the 1937 colonially sponsored pre-​independence elections in colonial
provinces.
The Indian National Congress’s overwhelming success in those provincial elections
posed an existential threat to the League, leading the latter to eventuall announce sup-
port for the new nation of Pakistan. Politically, those elections clearly signaled the
imminence of national independence under Congress hegemony, and with it the disap-
pearance of the colonial patronage that had underwritten Muslim power in the United
Provinces. Those elections demonstrated that, even with the guarantee of quotas or ‘sep-
arate electorates’, the League could not retain substantial influence at a national level
unless it also governed one of the two large Muslim-​majority provinces of colonial
India—​Bengal and Punjab. The League thus forged alliances with those two regions.
For these reasons, and because UP Muslims perceived a threat to a long history of cul-
tural dominance (Brass 2005), the Muslim League began to espouse a national identity
based upon the claim that Indian Muslims formed a separate nation starting in 1940
(Jalal 1995, Tudor 2013). This served to legitimate the creation of Pakistan, because the
 Islam, Nationalism, and Democracy in Asia    171

movement successfully created alliances with existing political movements in two geo-
graphically separate wings of the country.
The rallying cry for a new state of Pakistan heavily utilized religion. Though he was
personally secular, Mohammed Ali Jinnah invoked Islam in order to provide “a sem-
blance of unity and solidity to his divided Muslim constituents” (Jalal 1990, 16). But
while playing upon Islam provided the basic ideological glue together diverse geo-
graphic, linguistic, and ethnic regions, Jinnah did not specify whether and how Islam
conditioned citizenship in this new political entity. In order to legitimate its participa-
tion in independence negotiations, Jinnah forged a nationalism based upon the rallying
cry of “Islam in danger” that was “specifically ambiguous and imprecise to command ge-
neral support, something specifically Muslim though unspecific in every other respect”
(Jalal 2007 [1995], 20).
Throughout the seven years before independence, Pakistani nationalism became in-
extricably bound up with religion. Traditional religious elites were mobilized in order
to ensure that the political interests of the landed Muslim elites were protected, either
through the creation of an entirely separate state or through the creation of weak cen-
tral government presiding over strong sub-​national units. In the United Provinces,
for example, regional Muslim League leader Nawab Ismail Khan convened a confer-
ence of ulama and intellectuals in order to draft an Islamic constitution for Pakistan
that formed the basis of recommendations for the constitution after independence
(Haqqani 2005, 67). In Punjab, pirs, or the religious leaders of shrines, gained votes for
the 1945–​1946 regional elections by explicitly linking Pakistan with Islam. The British
governor of Punjab wrote in summer 1945 than “Muslim Leaguers are doing what they
can in the way of propaganda conducted on fanatical lines; religious leaders and re-
ligious buildings are being used freely in several places for advocating Pakistan and
vilifying any who hold opposite views” (Glancy to Mountbatten 1945, 67). By spring
1946, the same governor wrote that “Maulvis and Pirs and students travel all around the
Province and preach that those who fail to vote for the League candidates will cease to
be Muslims, their marriages no longer valid and they will be entirely excommunicated”
(ibid., 68).
While the Muslim League sometimes attempted to position Pakistan as a secular
project to audiences with different interests, such prevarication was ultimately insuf-
ficient to temper the close marriage between Pakistan and Islam in the minds of both
those who voted for Pakistan and the key religious leaders whose mobilizing prowess
turned out to be crucial in delivering Pakistan. In an oft-​cited speech just days before
Pakistan’s independence, Jinnah hoped to draw a line in the sand between personal reli-
gion and citizenship by stating, “In the course of time Hindus would cease to be Hindus,
and Muslims would cease to be Muslims, not in the religious sense because that is the
personal faith of each individual, but in the political sense as citizens of the State.” Yet as
many historians have noted, this one speech could not undo nearly a decade of linking
Islam with the national identity of Pakistan.
172   Maya Tudor

The Settling of Islamic Nationalism

Upon Pakistan’s independence on August 14, 1947, the tight association between Islam
and Pakistani citizenship provided ready fodder for derailing Pakistan’s nascent de-
mocracy. The close relationship between Pakistan and Islam legitimated religious
elites as they asserted hierarchical distinctions between citizens. The marriage between
Pakistani citizenship and Islam, deliberately evoked in the lead-​up to independence,
provided a difficult but possibly surmountable stumbling block in the early debates over
how power was to be organized.
Yet the decisions taken by Pakistan’s early governments only strengthened the sense
among both leaders and the public that Pakistan was defined through Islam. Pakistan’s
first official statement of citizenship was the 1949 Objectives Resolution. The resolution
stipulated that Pakistan would be guided by such principles as democracy, freedom,
equality, and social justice “as enunciated by Islam,” again making the relationship be-
tween the political and the religious interdependent. The Objectives Resolution also
stated that “sovereignty over the entire universe belongs to Allah Almighty alone.” The
Constituent Assembly, which met in different forms between 1950 and 1956, also reg-
ularly discussed religious matters such as collecting zakat, establishing institutions of
religious learning, and arrangements facilitating five daily prayers. Beyond the constitu-
tional framing debates that invoked Islamic matters, the close association between the
state and religion was seen in how quickly centers of religious learning grew in the new
country, from 137 in all of Pakistan in 1947 to 244 in West Pakistan alone a decade later
(Haqqani 2018, 76). All these decisions deepened the link between Islam and Pakistan
into a newly settled conception of Pakistan as a state of, by, and for Muslims.
The first major democratic setback in Pakistan centralized questions of both religion
and power-​sharing and reflected the first instance of what soon became a characteristic
pattern in Pakistani politics—​that religion was used as a legitimating tool for political
gain and thereby strengthened the relationship between citizenship and Islam. The first
instance of martial law in Pakistan occurred in January 1953, the first step toward the
1958 military coup that definitively ended Pakistan’s democratic experiment. On this oc-
casion, the central government called in the military to quell riots that ensued when an
Islamic movement sought to remove the foreign minister from power on the grounds
that he was a non-​Muslim.
It is important to note that this conflict would not have escalated into an existential
threat for the government in the absence of political entrepreneurs who exploited reli-
gious divisions to gain power. Punjabi landlords engineered the crisis and utilized the
Islamic narrative of nation in order to destabilize a central government attempting to
pass a constitution unfavorable to provincial power. The central government, led by
the Bengali Khwaja Nazimuddin, had previously rejected a demand from the Punjabi
ulama to consider religious affiliation a prerequisite for serving in a key position of
power. The ulama argued that the core beliefs of the Ahmadis, which included the then-​
foreign minister Chaudhri Zafrullah Khan, did not align with mainstream Islam. The
central government refused to do so and had wide support among the Pakistani public
 Islam, Nationalism, and Democracy in Asia    173

for its refusal. The 1953 conflict, the first major public showdown over how closely reli-
gion would be tied to citizenship, saw the secular hand briefly winning out.
The victory was short-​lived however, because state-​level politicians chose to align
with religious institutions to oppose a proposed constitution that centralized power at
the national level. It was in the midst of a tussle over the constitutional distribution of
powers that the chief minister of Punjab, who was also a powerful provincial landlord,
publicly opposed the government on the Ahmadi issue. He announced that the central
government needed to accede to the demands of the religious institutions, with the ex-
plicit intention of undermining the central government (Noon 1993, 234). When riots
broke out, Prime Minister Nazimuddin responded by arresting the leaders of the anti-​
Ahmadi movement. This led to major riots in Lahore in February, which lasted 70 days
and required military intervention to resolve, Pakistan’s first invocation of martial law.
Shortly after martial law was withdrawn, a civilian bureaucrat serving as governor-​
general dismissed the prime minister, thus removing his proposed constitution from
consideration, ostensibly because of the central government’s economic ineptitude
(Khan 1967). Democracy in Pakistan thus faltered on the anvil of a state-​centered power
struggle, in which state politicians weaponized Islamic institutions and the ideas of cit-
izenship to remove a national government. In this process and the ensuing political in-
stability, the military increasingly became politicized.
This fusion between religion and citizenship deepened during Pakistan’s first decade.
The short-​lived 1956 constitution pronounced “The Islamic Republic of Pakistan” and
specified that the president must be a Muslim, and that no law must be passed that was
deemed contrary to the teachings of the Qurʾan. Barring the non-​Muslim fifth of the
population, and even certain kinds of Muslims, from equal citizenship rights, far from
laying to rest debates over the nature of citizenship, continued to fuel divisive political
issues in subsequent decades (Shaikh 2009).
The next major phase in Pakistan’s political history gelled the basic pattern that has
endured since: a five-​step dance in which the Pakistani military assume de jure and
sometimes de facto executive power. Between 1958 and 1969, General Ayub Khan ruled
the country as a virtual military dictator, during which time he tightened the association
between Islam and citizenship. As the power-​sharing struggle between the eastern and
western wings of Pakistan deepened, political struggles in Pakistan began to resemble a
politics of musical chairs, with each successive government’s grip on power less certain
(McGrath 1996, Tudor 2013). These developments caused the civilian bureaucracy and
military leaders, especially General Ayub Khan, to assume greater powers of governing,
culminating in the 1958 military coup.
Upon his assumption of governing power, General Ayub Khan felt Pakistan needed
an ideology to provide the “tremendous power of cohesion” and thus turned toward
Islam to provide this glue (Khan 1967, 196). Within a year of assuming power, Ayub Khan
wrote a paper entitled “Islamic Ideology in Pakistan,” which was circulated among army
officers and which mainstreamed Islam as defining the nation among Pakistan’s military
(Gauhar 1996, 93). The key mechanism through which Ayub transmitted an Islamic ide-
ology more generally was the educational curricula, which communicated that Pakistan
174   Maya Tudor

was conceived during Islam’s arrival on the subcontinent, in which Muslim Mughal
rule was heavily glorified, and which highlighted differences between Hindu India and
Muslim Pakistan at the expense of the syncretic religious tradition that had dominated
the areas constituting Pakistan for many centuries (Nayyar and Salim 2005). Ayub also
created the Ministry of Information and the Bureau of National Reconstruction, which
standardized the Islamic messages taught through schools.
Recognizing that invoking Islam to legitimate state power could potentially endanger
state authority, state leaders attempted to make Islamic teachings compatible with state
sovereignty over religion. Ayub claimed the authority to interpret Islamic citizenship
by creating a Central Institute for Islamic Research. Led by Oxford-​educated Fazlur
Rahman, the institute interpreted Islamic teachings in ways that were compatible with
the ultimate sovereignty of the state. Unsurprisingly, Rahman was opposed by ulama,
who were able to force Rahman to resign when dissatisfaction grew against Ayub’s
rule in the late 1960s. On issue after issue—​the definition of the Muslim calendar for
announcing Ramadan, population control, the practices of polygamy and divorce—​the
ulama successfully determined government policy, marking a growing power of Islamic
authority as the state settled. By the end of Ayub Khan’s rule, Pakistani citizenship was
clearly defined by Islam.

The Deepening of Islamic Nationalism

Military rule under Ayub Khan witnessed a further strengthening of links between
religion and the nation through two main channels. First, after East Pakistan seceded
in 1971, a newly shrunken Pakistan was overwhelmingly Muslim, which loosened any
demographic constraints on political leaders appealing to Islam for political legitima-
tion. Second, colonially trained military generals, bureaucrats, and judges who had been
taught to abstain from religious legitimation were gradually replaced by those who had
been immersed in the increasing religious fervor of the Ayub Khan era. During the next
decades then, Pakistani politics became less focused on questions of whether to cen-
tralize Islam, as this was now a largely settled question. Instead, the political agendas in-
creasingly became focused upon questions of how one should enforce Islamic precepts
within and through the state.
When Ayub Khan stepped down in 1969 and transferred power to another mili-
tary general, Yayha Khan, Yayha’s short two years in power also saw him appealing
to Islam (e.g., by banning any published materials that were offensive to Islam) for
political legitimation, though he was not particularly known for this religious de-
votion (Feldman 1975, 47). When the first national elections were held in December
1970, the strongly secular Awami League in East Pakistan had won an overwhelming
majority in half of the country, while the “Islamic socialist” Zulfikar Ali Bhutto won
a majority in West Pakistan. Official accounts of events in Pakistan maintain that a
military-​sponsored campaign of demobilization began in the early months of 1971 only
after power-​sharing talks between the two victorious parties broke down. As has been
 Islam, Nationalism, and Democracy in Asia    175

well documented, however, the bloody genocide that occurred was planned well in
advance and effectively ignored by the United States (Bass 2013). In June, addressing
a country engaged in civil war, Yayha asserted that “Every one of us is a mujahid”
(Hornsby 1971). With India’s intervention escalating the domestic war into a short-​
lived international war, Yayha presided over the breakup of the country and the seces-
sion of Bangladesh. Within days of the West Pakistani troops surrendering, Yayha was
removed from power.
The subsequent period of democratic politics under Zulfikar Ali Bhutto did not,
and increasingly could not, decouple religion from state matters. Because Bhutto had
campaigned upon a platform of Islamic socialism and increasingly came under attack
from ulama, who insisted that socialism was anti-​Islamic, Bhutto’s rhetoric became
ever more Islamic. Already at the start of his government, he unequivocally stated that
he was “first a Muslim and then a Pakistani,” a tendency that only became more pro-
nounced throughout his tenure. The 1973 constitution went beyond the 1949 Objectives
Resolution and now required the occupants of the two highest offices in the country—​
the president and prime minister—​to be Muslim.
In 1974, a near-​repeat of the 1953 riots over the Ahmadi community’s definition of
themselves as Muslims again led to wide-​scale religious riots. In this case, however, in
a telling sign of how far Islamization had progressed, there was no longer a prime min-
ister in power who was willing or able to resist the demands of the ulama. In September
1974, Bhutto ushered through the Second Amendment to the Pakistan Constitution,
stipulating that the Ahmadi community were not Muslims. The union between the state
and religion was now written into the founding template of the country. From this time
forward, Pakistan has unambiguously promoted Islam.
With another military coup in 1977, the Islamization policies of General Zia ul-​Haq
took center stage for the longest single period of government in Pakistan’s history to
date. Zia mainstreamed the Islamization of Pakistan in a much more comprehensive way
than had been done before. Zia criminalized all manner of the Ahmadi community’s
activities and created shariʿa courts with the power to strike down laws deemed un-​
Islamic. He also enacted Islamic laws, rendering mandatory Islamic education in
schools and in the military training academies, and requiring daily prayers for govern-
ment. The Hudood Ordinances of 1979 legalized flogging, stoning, and amputation for
crimes such as adultery and alcohol consumption. Blasphemy laws were introduced that
criminalized any speech deemed offensive to Islam. The scope for Pakistan to protect
the freedom of speech and association was now systematically impaired, preventing all
subsequent governments from delinking Islam and citizenship. Debates over national
policy no longer questioned whether Islam should be the central ideology of the state,
but instead focused upon questions of which interpretation of Islam was appropriate for
the state to pursue.
Yet, as is often the case when religions are adopted as central to the national imagining,
these debates have not settled questions of who is a citizen. As with all universalist faiths,
a range of pasts, ideologies, and sects serves only to unify in the case of an agreed-​upon
enemy (which, in Pakistan’s case, has been India). More commonly, however, since
176   Maya Tudor

foreign policy cannot long assume center stage in domestic politics, the centrality of
faith provides ample fodder for politically divisive debates that consume politics.
The marginalization of Ahmadis that took place in the first few decades was expanded
to mainstream the marginalization of the 15–​20 percent of Muslims who are Shiʿa, as
well as, increasingly, Sunni Muslims who speak out for the civil liberties of minorities
(Malik 2002, Nasr 2007). Democratic and military governments alike have conse-
quently borrowed from the Islamist playbook, with real or imagined indignities against
Islam derailing governments and distracting from substantive economic and political
questions that more prominently occupy politics in neighboring India. In 1984 General
Zia required all citizens applying for a Pakistani passport to swear an oath that Ahmadis
are not Muslims, epitomizing the relationship between mainstream Islam and Pakistan.
Groups such as the Anjuman-​i-​Sipah-​i-​Sahaba of Pakistan (ASSP) demanded that
Pakistan be declared a Sunni state, spurring on the creation of Shiʿa organizations such
as the Tehrik-​Nifaz-​i-​Fiqh-​i-​Jaafria (TNFJ), which instead demanded the implementa-
tion of a Shiʿa version of Islamic laws. Questions surrounding the role of religion have
been central concerns in every administration since Zia’s and have regularly been used
to legitimate fewer civil and political liberties for minorities. “If you want to destroy
someone in public life [in Pakistan] it is enough to drop a hint that they are Ahmadi”
(Hanif 2010).
After Zia’s death in 1988, civilian and military governments alike have, with little
choice, acceded to Islamic authorities on important issues and further circumscribed
what limited space there is for minorities to coexist as equal citizens. In 1989, Retribution
and Compensation Ordinances enabled victims of murder or assault to pay “blood
money” for “honor killings,” or cases where a male family member kills a woman or
her lover for bringing dishonor to the family. During the 1990s, Prime Minister Nawaz
Sharif ordered TV actresses to wear veils and made the death penalty mandatory for
those convicted of blasphemy. In 2011, when the mainstream politician Salman Taseer,
the governor of Punjab, spoke out against these laws by defending a Christian mother
of five sentenced to death for alleged blasphemy, he was assassinated by one of his own
bodyguards, a member of the traditionally syncretic Barelvi sect. President Zardari was
so concerned for his safety he was unable to attend the funeral of Taseer, a close friend
(Gall 2011).
Much like Indonesia, Pakistan had entered independent statehood with a political
party that had possessed little in the way of clear policies, power-​sharing agreements,
or organizational capacity to enforce governing agreements. Yet unlike Indonesia,
Pakistan’s early decades witnessed a continual linking of religion with national iden-
tity. The effect of this difference, over time, would be to provide regular fodder for the
state to target those citizens deemed second-​class, and for military intervention into
democratic politics (Shaikh 2009, Devji 2013). Military regimes have supported key re-
ligious leaders, who in turn demand strict adherence to Pakistan’s blasphemy laws and
persecution against sects. During ostensibly democratic governments, the military
has regularly used religious disturbances to destabilize civilian governments, which in
turn renders civilian governments ever more reluctant to protect the rights of religious
 Islam, Nationalism, and Democracy in Asia    177

minorities. This is Pakistan’s state of affairs to the modern day. The grounding of gen-
uinely democratic politics would require not only the military to stay in the barracks
and overcome its propensity to enact coups, but, just as importantly, a reimagining of
Pakistan that enables the country to pursue policies that are not subject to the whims of
Islamic ulama.

Indonesia: Inclusive Nationalism

and Party Weakness Promotes
Unstable Democracy

Like Pakistan, Indonesia possessed myriad social cleavages of region and language
within a Muslim-​majority.1 But unlike Pakistan, Indonesia forged a remarkably inclu-
sive founding national narrative in the decades preceding its colonial independence
in 1949. Inequalities of religion were strongly and deliberately eschewed in the found-
ling Indonesian nation-​state. But while Indonesia’s leading nationalist organization,
the Indonesian Nationalist Party (Partai Nasional Indonesia, or PNI), advanced an
inclusive founding national narrative, Indonesia also entered independent statehood
lacking a developed party organization, which was to prove critical in cementing dem-
ocratic rule in India. As in Pakistan, constitutional compromises over religion and re-
gion, as well as power-​sharing compromises, proved chronically difficult to forge, and
increasing political instability saw a military coup ending Indonesia’s tentative dem-
ocratic experiment. Yet in 1989, after the Indonesian military substantively retreated
to the barracks, Indonesia emerged as one of Asia’s steadiest democracies. Today,
that trend is primarily threatened by the increasing inroads Islamists have made in
redefining Indonesia.

Inclusive Anticolonial Indonesian Nationalism Bridges

Key Social Divides
Colonial Indonesia was ruled by dividing communities through maps, censuses, and
museum (Anderson 1998). The express intent and actual effect of colonial divide-​and-​
rule policies was to produce status inequalities between racial, ethnic, religious, and re-
gional communities. Indonesia’s Dutch colonial state forged particular communities
of support among non-​Muslim and non-​Javanese populations. As in many Southeast
Asian countries, colonial policies also enticed a sizable ethnic Chinese population to
emigrate. Indonesia’s new Chinese community prospered considerably and became the
first non-​European business community of any size in Dutch Indonesia by the dawn of
the twentieth century.
178   Maya Tudor

Just as in British India, the growing dominance of a new social group stimulated a
reaction in the form of a political organization protecting the status quo. In Indonesia’s
case, the Sarekat Islam (Islamic Union) was founded in 1912 and harnessed resentment
over Dutch favoritism toward the Chinese (Reid 2010). Just as in British India, the early
decades of the twentieth century also witnessed the burgeoning of religious revival or-
ganizations, which generally pursued various reform agendas that sought to influence
Islamic practices across Indonesia.
As in Pakistan, the two decades before independence were the formative period for
Indonesian nationalism. In 1927, Indonesia’s nationalist movement, the Indonesian
National Party (PNI) was founded. Just as in Pakistan, the party grew under the efforts
of a single charismatic leader—​the well-​educated, twenty-​six-​year-​old Sukarno. Unlike
in Pakistan however, Indonesia’s nationalist movement espoused an inclusive nation-
alism that transcended the major religious, regional and ethnic divides and managed
to bridge colonially-​emphasized social cleavages in an attempt to present a united front
against the Dutch colonial regime. The basis of Sukarno’s movement and the PNI Youth
Pledge focused upon “unity” across all of the most salient social divides (McVey 1970).
The PNI Youth Pledge in 1928 promoted unity across every one of Indonesia’s salient so-
cial cleavages.
One important dimension through which the PNI practiced inclusivity was through
its choice of language for mobilization. When the First Congress of Indonesian Youth
was held in 1926, some individuals suggested Malay be used as the language of an in-
dependent Indonesia. Just two years later, Malay was adopted as the language of the
Congress, and the name of the new Indonesian (Bahasa Indonesia) was introduced.
The PNI platform called for the language of Indonesia to be based on Malay. Malay was
chosen as the national language as a means to unify the 13,000 islands, despite of the fact
that the Javanese were by far the single largest ethnic grouping (nearly 50 percent of the
population) and made up a significant proportion of its educated elite. “Malay conveyed
a message of democratic inclusiveness, unlike the more hierarchical Javanese, in which
levels of language are used to reinforce status differences between the aristocracy and
the lowest class” (Bertrand 2003, 273).
In addition to choosing an inclusive language, The PNI-​led nationalist movement
used discursively egalitarian language. Sukarno fostered the use of an egalitarian
form of address, Bung—​loosely translated as “comrade/​brother”—​among both PNI
leaders and followers. “[The word] Bung was instrumental in bringing about a socio-​
political unity by reducing all to a commonly associated bond” (Sundstrom 1957, 135).
In contrast to Jinnah in Pakistan, Sukarno articulated an ideology that bridged the
ideological tension between Marxist and non-​Marxist nationalists. This broadly in-
clusive ideology of Marhaenism was, much like India’s trusteeship policy, an “attempt
to draw as many groups as possible into the revolutionary struggle along with the
proletariat” (Dahm 1969, 344). The ideology conceived of political relationships be-
tween PNI leaders and citizens in urban and rural areas in egalitarian terms, without
the mediation of traditional elites or the dominance of a particular regional or
linguistic group.
 Islam, Nationalism, and Democracy in Asia    179

Unlike in British India, however, the Dutch Indonesian colonial government offered
little space for the leading nationalist party to develop organizationally by competing
through regional elections. Instead, the Dutch colonial regime banned the PNI in 1931,
and it remained organizationally defunct for almost fifteen years, until Japan’s wartime
invasion led to the release of Sukarno and other nationalists from prison. By this time
Sukarno and the PNI were no longer the only nationalist forces on the scene. Though the
competition between an array of nationalist forces and the relative lack of programmatic
and organizational development within the PNI limited the PNI’s ability to comprehen-
sively imprint their vision onto the institutions of state, the inclusive vision it had forged
settled and institutionalized over time.

The Settling of Inclusive Nationalism

Indonesia fought to gain its independence for four years after the Second World War
concluded, during which time an inclusive narrative of Indonesia deepened. On August
17, 1945, after the Japanese were defeated, Sukarno and Mohammad Hatta declared
Indonesia’s independence from Dutch colonial rule and pronounced a 1945 constitu-
tion that enshrined the national philosophy of Pancasila. An armed struggle raged for
four bloody years before independence was recognized by the Dutch. The reason for the
Dutch defeat in 1949 had much to do with a highly localized popular resistance, concen-
trated in Sumatra and Java, that had unified around a common vision of a free nation
under Japanese occupation (Kahin 1952, Anderson 1983).
The concept of pancasila enabled Indonesia to be established as a religious but not
Islamic state. An effective compromise between the Islamist pressure to establish
Indonesia as a religious state, Pancasila, or “Five Principles,” stipulated that Indonesians
must express faith in one God, though not necessarily the Islamic Allah. Pancasila
also enabled Sukarno to recognize Hinduism, Buddhism, and Confucianism as fully
Indonesian. By making Pancasila the ideological core of the 1945 constitution, Sukarno
had creatively devised a formula that embraced the anti-​secularism of Indonesia’s
Muslims without excluding followers from other world faiths from full membership in
the Indonesian nation.
While the prolonged nature of the anti-​colonial struggle cohered together the na-
tionalist movement, the PNI was hardly a dominant political force. Instead, it competed
against myriad political and social organizations for political supremacy. Moreover, as
with Pakistan’s Muslim League, Indonesia’s PNI lacked the programmatic coherence,
organizational experience, and cross-​cutting alliances that India’s Congress party pos-
sessed, and that helped broker stability in early post-​independence India. “[T]‌he heter-
ogeneity of its constituent elements made the PNI an unwieldy political organization”
(Kahin 1952, 156). Even if the PNI had been less factionalized, it still lacked the organiza-
tional depth to lead the nation-​building process. Moreover, the PNI was just one of four
parties vying for political supremacy, a fact confirmed by the parliamentary elections
of 1955. Just as in Pakistan, the absence of a single, well-​organized political party fed a
180   Maya Tudor

politics of instability. Thus, Indonesia’s infant democracy remained vulnerable to au-

thoritarian challenges from the military.

Political Instability and Military Dictatorship

As governments came and went throughout the mid-​1950s, and as the PNI proved un-
able to broker political stability, Sukarno sought to single-​handedly create stability by
proclaiming emergency powers, declaring martial law in 1957, and proposing a new ide-
ology that combined three ideologies (nationalism, religion, and communism) into a
new Nasakom ideology as a means of defanging the ideologies of his main rivals (Feith
1962, McVey 1970). Eventually, in the face of what we now know to have been active US
support in fomenting regional rebellions, Sukarno claimed broad presidential powers,
disbanded the constituent assembly charged with formulating a new constitution, and
installed “Guided Democracy” in 1959 (Bevins 2017).
A mere six years later, General Suharto directly assumed political powers through a
bloody “New Order” revolution that is estimated to have brutally murdered up to a mil-
lion Indonesians for affiliation with the Indonesian Communist Party (Partai Komunis
Indonesia, or PKI). Suharto’s military government would stay in power until 1998. Just
as in Pakistan, where the country’s military was initially politicized and deployed to pro-
tect a weak government with little grass-​roots support but ended up assuming direct ex-
ecutive power with American support, Indonesia’s military moved beyond the barracks
and assumed the formal reins of power.
Suharto’s regime operated primarily through repression and exclusion, but on the
basis of an anti-​communist Pancasila-​inclusive nationalism that targetted communists
rather than upon the basis of an immutable social identity such as religion. The New
Order regime also “ began in a revived mood of great hostility to all things Chinese, pre-
sumably based on its sharp reaction against Sukarno’s closeness to Beijing and a virulent
anti-​communism which provided the legitimation for Suharto’s rise to power” (Reid
2010, 70). Yet the Indonesian military’s claim to legitimacy also rested on its leading
role in the anticolonial revolution, which had centralized inclusive nationalism. Thus,
Sukarno’s “unitary ideal was kept sacred and central by the military-​based regime of
General Suharto” (ibid., 45).
Suharto eventually permitted the creation of two weak opposition parties to contest
national elections in the early 1970s. These parties were direct incarnations of the PNI
(the Indonesian Democratic Party, or PDI) and Masyumi (the United Development
Party, or PPP [Partai Persatuan Pembangunan]). They were not democratic organi-
zations, engineered as they were for purposes of authoritarian control. The PDI even-
tually grew into the strongest challenger to the New Order during the mid-​1990s.
Sukarno’s daughter, Megawati Sukarnoputri, assumed leadership of the PDI in 1993. As
she criticized the government more vehemently, however, she was removed from the
 Islam, Nationalism, and Democracy in Asia    181

leadership of the party. The consequent public outcry prompted a violent crackdown
against pro-​Megawati protesters in Jakarta in November 1996. These dynamics would
provide the basis for democratic contestation once the Suharto regime fell in 1998.
While Indonesia’s military-​led authoritarianism was terribly repressive, it is crucial to
underline that Suharto’s authoritarianism was not created for the purposes of excluding
a particular ethnic, regional, religious, or class group from gaining power. Consequently,
when the military relinquished power in a context that no longer essentialized the ide-
ological cleavages that defined the Cold War, Indonesia’s inclusive founding national
narrative helped restore a regime that conceived of a broad range of ethnic, regional, and
religious groups as equal citizens, i.e. a democracy.

Military Retreat, Inclusive Nationalism, and

Democratic Stability
When Suharto’s regime was destabilized by the 1997–​1998 financial crisis that rocked
much of Southeast Asia, a broadly held view was that he would invariably be replaced by
another military leader (Crouch 2000). Yet the dynamics of Suharto’s fall revealed the
lingering influence of inclusive nationalism as Indonesia ushered in one of Asia’s most
unlikely democracies.
When the Asian financial crisis hit Indonesia, nationalist forces engaged in mass
protest. This reformasi mobilization transcended every imaginable categorical divide,
as the leading Islamic organizations Nahdlatul Ulama (NU) and Muhammadiyah
saw their leaders assuming anti-​regime stances. Though Suharto attempted to target
urban ethnic Chinese in an attempt to split the opposition, he was unsuccessful. Once
Suharto was removed from power, Indonesia’s inclusive nationalism promoted a de-
mocracy that readily accommodated ethno-​religious diversity and has thus been stably
remarried with democratic politics in Indonesia. At the same time, sanctioned as it is by
the founding Pancasila principle committed to godly nationalism, political targeting of
those without theistic beliefs readily recurs in Indonesia (Menchik 2016).
Since Suharto’s fall, Indonesian democracy has combined impressive consolida-
tion with low quality. The combination of dysfunctional parties that fail to promote
the interests of their followers and an inclusive founding national narrative together
provides a helpful lens through which to understand Indonesia’s otherwise puz-
zling outcome. Unlike in Pakistan, there are no powerful voices calling for a return to
military-​dominated regimes, providing clear evidence that Suharto’s military regime
was legitimated through its fight against communism. Once communism was no longer
a global threat, there was little ideational justification for authoritarian rule. However,
as elsewhere, authoritarian populism continues to be a democratic threat, as leading
elites can well be criticized for their failure to fulfill the promise of inclusive Indonesian
nationalism.
182   Maya Tudor

India: Inclusive Nationalism

and a Strong Party Brokers a
Stable Democracy

India’s eventual nationalist movement, the Indian National Congress, was set up in 1885
by a small group of English-​educated elite. Congress initially aimed to only promote
electoral reform within the colonial state and to hold civil service examinations in India,
two aims which would solely benefit the elite founders of the Congress—​a few hundred
individuals in a country with a population of over 300 million. Only when these goals
were frustrated over the subsequent three decades did this nationalist movement or-
ganize mass mobilization. Beginning in 1920, the Congress imported and adapted the
idiom of mass nationalism, claiming it represented a new nation of India. Crucially,
this nationalist movement did not embrace religion in defining India, but instead
promulgated a conception of the nation that was based on secularism and linguistic
inclusion.
During the nearly three decades of mass mobilization that followed, Congress
adopted and adapted the idea of a nation to the local social context. Unusually, India’s
nationalist movement evolved a national narrative that did not build on Hinduism,
the religion of approximately three-​quarters of the population. Because the social
fabric of precolonial and colonial India was defined by caste—​an endogamous and re-
ligiously sanctioned social category that Hinduism recognized and reinforced in most
interactions, including among Muslims—​the very conception of a nationalism that
treated religious minorities as equal citizens required a conceptual separation between
private and public identities (Rudolph and Rudolph 2006). Because Congress leaders
were striving to hold together a nationalist movement in a highly diverse society, they
rejected religion as the basis for the nation and also sought a limited degree of social re-
form that helped broker a public sphere of citizenship. Espousing a relatively inclusive
creedal nationalism was necessary not only in order to call a new nation into being, but
also, and just as importantly, to prevent the emerging national movement from frac-
turing in the face of the colonial regime’s persistent attempts to impede burgeoning na-
tionalism (Anderson 1983, Tudor 2013).
By experimenting with different approaches during the period between 1920 and 1947,
Congress leaders grew to understand that a united national movement necessitated a
clear separation between the dominant religion of colonial India, namely Hinduism,
and the Indian nation, because it would otherwise be prone to fracturing in the face
of colonial divide et impera policies. Basic secular principles were thus written into the
Congress’s founding charter. While many Congress leaders were also simultaneously in-
volved in Hindu reform movements, they drew a clear line between religious reform
and political reform.
 Islam, Nationalism, and Democracy in Asia    183

At its annual meeting in 1931, for example, the Congress adopted a formal policy that
prohibited it from adopting any policy to which a majority of either Hindu or Muslim
members objected—​thereby providing for a minority veto on anything that was deemed
sensitive. The same year, at a time when most European nations had yet to do so, India
adopted a universal adult franchise internally, and argued that the universal franchise
was an essential step in the fight for purna swaraj, or total independence. In 1935, once
electoral competition was normalized as the route to political power, local political
disputes regularly assumed religious overtones. Yet as a matter of formal policy and as a
matter of practice, India’s nationalist movement rejected Hinduism as the core building
block of the Indian nation.
Beyond demarcating a clear line between the nation and the religion of the demo-
graphic majority, the Congress’s nationalism was linguistically inclusive. As it engaged
in mass mobilization beginning in the early 1920s, it created twenty-​one regions in
which nationalist organizing would be done primarily in regional languages, rendering
the nationalist cause accessible to a broader class of society.
Finally, the Congress attempted to be socioeconomically inclusive, most notably by
using “cloth nationalism” to communicate the contours of the new nation to a largely
illiterate society. It was for this reasons that top Congress leaders universally traded in
their Western clothes by the 1930s, which set the English-​educated elite apart from their
countrymen, and began to wear khadi, the homespun clothing that became the uniform
of the nationalist movement (Trivedi 2003). Wearing homespun cloth was simultane-
ously a form of economic nationalism, hurting British cloth exports, as well as a means
of generating national solidarity in a deeply hierarchical society. Cloth nationalism
enabled a broad range of class actors to participate in the nationalist movement.
The Indian National Congress’s nationalism was also socioeconomically inclusive
insofar as its signature mobilizing tactic was nonviolent civil disobedience. The use
of nonviolence, which most clearly emerged during the 1930–​1931 Civil Disobedience
protest, enabled wealthy industrialists and landless laborers alike to participate in the
national movement without fear of violence. In particular, Gandhi’s theory of trustee-
ship neutralized the potential opposition of wealthy industrialists, many of whom had
benefitted from close relations with the colonial regime.
To be sure, conflict between cleavages of Indian society, especially along Hindu-​
Muslim lines during the 1930s, grew steadily as electoral politics grew in import. But
overall, and especially in relative terms across Asian colonies, India’s national move-
ment was notable for its institutionalization of an inclusive nationalism along the
dimensions of religion, language, and class within the purview of the well-​organized
Congress movement, which brokered the adoption of universal adult franchise within
the Indian constitution and restrained the military from political intervention. It was
because Congress leaders had popularized an inclusive Indian nation for decades as a
way of gaining political power that there was little question of reneging on the dem-
ocratic institutions of elections and broad civil rights for all once independence was
achieved (Tudor 2013).
184   Maya Tudor

Inclusive Nationalism Adjudicates an Early

Democratic Challenge
India’s relatively inclusive conception of nation not only helped to create, but also to
maintain, Indian democracy at crucial moments in its post-​independence history.
Notably, India’s inclusive nationalism was harnessed by regional politicians to ultimately
force the national government to reorganize subnational states upon a linguistic basis.
The anticolonial Congress movement conducted its party business along linguistic lines
starting in 1920.
During the pre-​independence mobilizations, Congress also adopted linguistic inclu-
sivity as official policy. Upon independence, as the horrors of Partition were unleashed
and several states tried to secede from the Union of India, Congress leaders reversed
their commitment to linguistic states because they felt doing so would boost the seces-
sionist fevers burning across India. It was for this reason that government commissions
in 1948 and 1949 recommended against the linguistic reorganization of states.
However, when regional language movements employed the ideals and tactics of
inclusive nonviolent civil disobedience, which nationalist leaders themselves em-
ployed, nationalist leaders ultimately found it necessary to accommodate the re-
gional movements for language autonomy. Throughout his tenure as India’s first prime
minister, Jawaharlal Nehru clearly opposed the linguistic reorganization of states.
Nevertheless, Nehru ultimately accepted the need to reorganize colonial states upon
a linguistic basis, both because of a long-​standing commitment to an inclusive nation
and because subnational movements employed the tactic of nonviolent mobilization
that had been long legitimated during the anticolonial agitations. At a time when Nehru
was at the height of his influence, he ultimately agreed to reorganize Indian states along
linguistic lines. The single most contentious policy issue in India’s post-​independence
decade was thus resolved, and India’s democracy consequently stabilized, in part due to
the inclusive nature of India’s national narrative.

Inclusive Nationalism Coalesces the DemocraticOpposition

India’s most severe democratic crisis occurred after Indira Gandhi assumed the posi-
tion of prime minister in 1966. Top Congress bosses chose Indira Gandhi to succeed
Lal Bahadur Shastri because of her pliable nature and dynastic legitimacy (Kidwai
1996). Indira’s introduction of independent and left-​leaning policies, despite the oppo-
sition of the party’s conservative regional leaders, ultimately led to a party split in 1969.
In the 1971 national elections, this split played out in national elections, with the older
Congress (O) leading a coalition that campaigned on a platform of “Indira hatao” (re-
move Indira), while Indira’s Congress (R) responded with a highly populist position that
instead promised “garibi hatao” (remove poverty). When Indira’s party returned with
352 of 518 parliamentary seats, Indira’s populist strategy was vindicated.
 Islam, Nationalism, and Democracy in Asia    185

Armed with an electoral landslide, Indira Gandhi sought to institutionally strengthen

the position of the prime minister. Amid a deteriorating economy following the 1973
oil crisis, popular movements against corruption led by former Gandhi associate
Jayprakash Narayan gained traction. When the JP movement eventually targeted Indira
Gandhi herself, she arrested her political opposition, clamped down upon civil liberties,
and suppressed media freedom, beginning on the evening of June 26, 1975.
During the subsequent twenty-​one months, a period that became known as the
Emergency, both the incumbent prime minister and her political opposition consist-
ently tried to appeal to the public in terms of an inclusive nationalism. In her speeches
announcing and justifying the Emergency, it is notable that Indira Gandhi attempted,
however thinly disguised this rhetoric may have been, to use inclusive nationalism to
justify her actions (Baloch 2017, 77). In her first television interview after the Emergency
was declared, Gandhi stated, “Indian democracy will be threatened when any party of
the extreme Right or extreme Left comes to power. It is being weakened by those who,
claiming to be non-​violent and democratic, give respectability to and ally themselves
with fanatic religious organizations and with parties wedded in terrorism. What holds
India together is the trust that all regions and all its religious groups will have a fair deal”
(Gandhi 1975, 63; emphasis added). Indira Gandhi also sought to refute any claim that
she was undermining the right to nonviolent protest (satyagraha) by arguing that she
was protecting a core tenet of the India’s nationalism-​secularism: “I have always said
that if somebody wants to do such a satyagraha, it does not matter because we are not
against satyagraha as such; we are not against criticism as such; we are not against oppo-
sition; what we are against is when a small minority [the RSS] tries to gag the vast ma-
jority of our country” (ibid.).
But Indira Gandhi’s opposition more successfully harnessed inclusive nationalism
during India’s most crucial election to date. On January 18, 1977, Indira Gandhi ended
the Emergency by releasing political prisoners and scheduling March elections. During
the eight weeks that followed, opposition across the political spectrum unified and co-
ordinated to put forth a single opposition candidate across most districts. The question
of civil liberties and the future of Indian democracy were the central issues of the 1977
national election. In its messaging, the Janata coalition emphasized the need to restore
civil liberties (Weiner 1977). The right to protest, so fundamental to the Indian nation,
as defined three decades earlier, helped the opposition to garner the largest percentage
of votes ever gained by a non-​Congress party. In perhaps democracy’s most pivotal mo-
ment, the founding right to protest was invoked and successfully claimed by opposition
forces.

Exclusive Nationalism Gains Traction

In recent decades, India’s inclusive national identity has been increasingly supplanted by
an exclusive national narrative that is making Hinduism central to the Indian nation. To
the extent that the Hinduism becomes a defining feature of India’s core national identity,
186   Maya Tudor

India’s democracy is likely to be significantly weakened. To be sure, some strands of na-

tionalism identified the Indian nation with Hinduism (Jaffrelot 1999). But until recently,
Hinduism could not be said to define the dominant conception of the Indian national.
This changed with the election of Narendra Modi in 2014, when Hindutva, a conserva-
tive ideology that identifies the Indian nation with a pure ideal of Hinduism, became
mainstreamed in Indian politics (Hansen 1999, Anand 2011).
When the Indian economy did not grow under Modi’s first term, the government’s
platform, especially in the 2019 election, unsurprisingly turned toward a celebration of
Hindu nationalism (Chandra 2017). The Modi government has systematically sought
to replace a largely syncretic religious history of the Indian subcontinent with a mon-
olithic narrative that positions Hindus as the only authentic citizens of India, and that,
importantly, selectively fabricates a Hindu-​centric history of India. Prior to the 2014
election, for example, the Modi campaign engaged in regular speeches that called the
devout Hindu Sardar Patel a more legitimate founding father than the avowedly secular
Jawaharlal Nehru. What at first glance may appear to be an obscure historical footnote
was instead only the most visible element of a systematic campaign to centrally focus
upon Hinduism as the marker of India’s national identity. The attempt to manufacture
a cultural primacy for Hindus impelled the elevation of Patel’s profile, which has in-
cluded building the world’s tallest statue at a cost of $430 million. But the singular focus
on Hinduism as defining India is also in evidence in the rewriting of classic historical
texts to elevate the historical role of Hinduism and omit the contributing roles of other
religions and minorities, including new holidays to honor Hindu figures, omitting “sec-
ular” from the preamble to the Indian Constitution on posters celebrating National Day,
the renaming of Muslim streets, the removal of Muslim names in textbooks, and the
omission of Muslim monuments (Tudor 2018, Ranganathan 2019).
To the extent that the Indian nation is seen as an exclusively Hindu nation, Indian de-
mocracy is likely to see the democratic backsliding on civil rights that has been typical
of religious minorities in Pakistan. In India, nine out of ten religiously motivated hate
crimes in the last decade occurred after Modi took power, and police have systematically
“stalled investigations, ignored procedures or even played a complicit role in the killings
and cover-​ups of crimes [against minorities]” (Human Rights Watch 2019). The “ob-
vious impunity for the string of crimes [against minorities] that have taken place, and
their shameful valorization by some leaders, is distinctly a strong factor in their contin-
uation” (Daruwala, in Human Rights Watch 2019). While it is too early to tell how en-
duringly the Modi government will have redefined the national narrative, to the extent
that this remains the case, Indian democracy is less likely to endure.

Conclusion

Across the postcolonial world in the middle of the twentieth century, indigenous
elites imagined, mobilized, and institutionalized new nations. In doing so, they made
 Islam, Nationalism, and Democracy in Asia    187

strategic decisions about what social cleavages, if any, these new nations build upon.
These decisions, often pivoting upon short-​term considerations such as how the eco-
nomic and political interests of this elite were contextually advanced, had far-​reaching
consequences for the likelihood and stability of the country’s post-​colonial democracy.
Across the Muslim world, as elsewhere in the postcolonial world, the choice of whether
to make religion the core identity of the nation was particularly impactful. This is because
the founding narratives of nation proved particularly sticky, alternatively facilitating
or limiting the political scope of religious institutions and emboldening political
entrepreneurs to use founding ideals to pursue political power (Gryzmala-​Busse 2015).
In all three Asian countries examined here, Islam was a major religion. Yet only in
Pakistan did the closeness of identification between Islam and nation serve to derail the
tentative democracy instantiated upon colonial independence. In Indonesia, a crucial
separation between Islam and nation was created through the principles of pancasila,
which helped to limit destabilizing debates about the role of Islam in politics. An in-
clusive national narrative was no guarantee of democracy. But when democracy broke
down in places with an inclusive national narrative such as Indonesia, it was crucially
not because religion became politicized. And in India, a clear line between religion and
nation has helped the unlikeliest democracy in the world to endure, a democratic re-
source that is increasingly threatened by a government seeking to closely identify the
nation with the religion of the demographic majority. To the extent that this is suc-
cessful, democracy is likely to weaken in India.
Together, these cases illustrate how narratives of nation that eschew religion, while in
no way guaranteeing democracy, can either create or remove a significant hurdle that
new democracies encounter in the struggle to stabilize. In and of itself, Islam poses no
particular threat to democracy. Rather, a close relationship between any religion and the
national identity can bedevil democracy.

Note
1. This section draws upon a previous paper coauthored with Dan Slater from the
University of Michigan. For related papers, see Slater and Tudor (2019) and Tudor and
Slater (2020).

References
Ali, Syed Husin. 2013. The Malay Rulers: Regression or Reform? Petaling Jaya, Malaysia: Strategic
Information and Research Development Center.
Anand, Dibyesh. 2011. Hindu Nationalism in India and the Politics of Fear. New York: Palgrave
Macmillan.
Anderson, Benedict. 1983. Imagined Communities: Reflections on the Origins and Spread of
Nationalism. London: Verso.
Anderson, Benedict. 1998. The Spectre of Comparisons: Nationalism, Southeast Asia, and the
World. New York: Verso.
188   Maya Tudor

Baloch, Bilal. 2017. Crisis, Credibility and Corruption. PhD diss., Oxford University.
Bass, Gary. 2013. The Blood Telegram. New York: Knopf Doubleday.
Bertrand, Jacques. 2003. “Language Policy and the Promotion of National Identity in
Indonesia.” In Fighting Words: Language Policy and Ethnic Relations in Asia, edited by
Michael E. Brown and Sumit Ganguly. Cambridge: MIT Press.
Bevins, Vincent. 2017. “What the United States Did in Indonesia.” The Atlantic. October
20, 2017.
Bonikowski, Bart. 2016. “Nationalism in Settled Times.” Annual Review of Sociology 42
(June): 427–​429.
Brass, Paul. 2005. Language, Religion, and Politics in North India. New York: Cambridge
University Press.
Chandra, Kanchan. 2017. “Who Is Narendra Modi? The Two Sides of India’s Prime Minister.”
Geopol Intelligence, April 3, 2017.
Crouch, Harold. 2000. “The Perils of Prediction: Understanding the Indonesian Transition
1989-​99.” Unpublished paper.
Dahm, Bernhard. 1969. Sukarno and the Struggle for Indonesian Independence. Ithaca: Cornell
University Press.
Devji, Faisal. 2013. Muslim Zion: Pakistan as a Political Idea. London: Hurst & Company.
Gall, Carlotta. 2011. “Assassination Deepens Divide in Pakistan.” New York Times, January
5, 2011.
Gandhi, Indira. 1975. Democracy and Discipline. New Delhi: Directorate of Advertising &
Visual Publicity.
Gauhar, Atlaf. 1996. Ayub Khan: Pakistan’s First Military Ruler. Karachi: Oxford
University Press.
Glancy to Mountbatten. 1945. India Office Records and Private Papers. 67. London: British
Library.
Greenfeld, Liah. 1992. Nationalism: Five Roads to Modernity. Cambridge, MA: Harvard
University Press.
Gryzmala-​Busse, Anna. 2015. Nations under God. Princeton, NJ: Princeton University Press.
Feith, Herbert. 1962. The Decline of Constitutional Democracy in Indonesia. Ithaca: Cornell
University Press.
Feldman, Herbert. 1975. “Pakistan in 1974.” Asian Survey 15, no. 2: 110–​116.
Hall, Peter, and Michele Lamont. 2013. “Why Social Relations Matter for Politics and Successful
Societies.” Annual Review of Political Science 16: 49–​7 1.
Hanif, Mohammed. 2010. “Why Pakistan’s Ahmadi Community Is Officially Detested.” BBC
News, June 16, 2010, http://​news.bbc.co.uk/​1/​hi/​8744092.stm.
Hansen, Thomas Blum. 1999. The Saffron Wave: Democracy and Hindu Nationalism in India.
Princeton: Princeton University Press.
Haqqani, Hussain. 2005. Pakistan: Between Mosque and Military. Lahore: Vanguard Books.
Haqqani, Hussain. 2018. Reimagining Pakistan: Transforming a Dysfunctional Nuclear State.
New Delhi: Harper Collins.
Hefner, Robert. 2000. Civil Islam: Muslims and Democratization in Indonesia. Princeton,
NJ: Princeton University Press.
Heng, Pek Koon. 1988. Chinese Politics in Malaysia: A History of the Malaysian Chinese
Association. Singapore: Oxford University Press.
Hirschman, Charles. 1980. “Demographic Trends in Peninsular Malaysia, 1947–​75.” Population
and Development Review 6, no. 1 (March): 103–​125.
 Islam, Nationalism, and Democracy in Asia    189

Hobsbawm, Eric, and Terence Ranger, eds. 1983. The Invention of Tradition.
New York: Cambridge University Press.
Hornsby, Michael. 1971. “President Yayha Dashes Hopes of Reconciliation.” The Times. London.
July 3, 1971.
Human Rights Watch. 2019. Violent Cow Protection in India. New York: Human
Rights Watch. www.hrw.org/​report/​2019/​02/​18/​violent-​cow-​protection-​india/​
vigilante-​groups-​attack-​minorities.
Inglehart, Ronald, and Christian Welzel. 2009. “How Development Leads to Democracy.”
Foreign Affairs, March/​April, 2019.
Ishak bin Tadin. 1960. “Dato Onn and Malay Nationalism, 1946–​1951.” Journal of Southeast
Asian History 1 (March): 56–​88.
Jaffrelot, Christophe. 1999. The Hindu Nationalist Movement and Indian Politics. New
Delhi: Penguin Books.
Jaffrelot, Christophe. 2003. India’s Silent Revolution. New York: Columbia University Press.
Jain, Rupam, and Tom Lasseter. 2018. “By Rewriting History, Hindu Nationalists Aim to Assert
Their Dominance Over India.” Reuters, August 9, https://​www.reuters.com/​investigates/​
special-​report/​india-​modi-​culture/​.
Jalal, Ayesha. 1990. The State of Martial Rule. Cambridge: Cambridge University Press.
Jalal, Ayesha. 2007 [1995]. Democracy and Authoritarianism in South Asia.
Cambridge: Cambridge University Press.
Kahin, George. 1952. Nationalism and Revolution in Indonesia. Ithaca: Cornell University Press.
Keyes, Charles. 2014. Finding Their Voice: Northeastern Villagers and the Thai State.
Seattle: University of Washington Press.
Khan, Ayub. 1967. Friends not Masters. London, Oxford University Press.
Kidwai, Anser. 1996. Indira Gandhi: Charisma and Crisis. Delhi: Siddhi Books.
Li, Qiong, and Brewer, Marilynn B. 2004. “What Does It Mean to Be an American? Patriotism,
Nationalism, and American Identity After 9/​11.” Political Psychology 25, no. 5: 727–​39
Liow, Joseph. 2016. Religion and Nationalism in Southeast Asia. Cambridge: Cambridge
University Press.
Liu, Amy. 2015. Standardizing Diversity. Philadelphia: University of Pennsylvania Press.
Low, David Anthony. 2004. Congress and the Raj: Facets of the Indian Struggle 1917–​1947. New
Delhi: Oxford University Press.
Mahoney, James. 2010. Colonialism and Postcolonial Development. New York: Cambridge
University Press.
Mainwaring, Scott, and Mariono Torcal. 2006. “Party System Institutionalization and Party
System Theory after the Third Wave.” In Handbook of Party Politics, edited by S. Katz and
William Crotty. New York: SAGE.
Malik, Iftikhar. 2002. Religious Minorities in Pakistan. London: Minority Rights Group
International, http://​citeseerx.ist.psu.edu/​viewdoc/​download?doi=10.1.1.440.4047&rep=re
p1&type=pdf
McGrath, Allen. 1996. The Destruction of Pakistan’s Democracy. Karachi: Oxford
University Press.
McVey, Ruth. 1970. Nationalism, Islam and Marxism. Ithaca: Cornell University Press.
Menchik, Jeremy. 2016. Islam and Democracy in Indonesia: Tolerance without Liberalism.
New York: Cambridge University Press.
Mill, John Stuart. (1861) 2004. Utilitarianism and Other Essays. London: Penguin Books.
Miller, David. 1995. On Nationality. Oxford: Oxford University Press.
190   Maya Tudor

Moore, Barrington. 1966. The Social Origins of Dictatorship and Democracy.

Boston: Beacon Press
Mylonas, Harris and Maya Tudor. Forthcoming. “Nationalism: What We Know and What We
Still Need to Know.” Perspectives in Politics.
Nasr, Vali. 2007. The Shia Revival: How Conflicts Within Islam Will Shape the Future.
New York: W.W. Norton.
Nayyar, A. H., and Ahmad Salim. 2005. The Subtle Subversion: The state of curricula and
textbooks in Pakistan Urdu, English, Social Studies and Civics. Islamabad: Sustainable
Development Policy Institute.
Noon, Firoze. 1993. From Memory. Islamabad: National Book Foundation.
Pauuw, Scott. 2009. “One Land, One Nation, One Language: An Analysis of Indonesia’s
National Language Policy.” University of Rochester Working Papers 5, no. 1, Summer.
Ranganathan, Maya. 2019. “Re-​scripting the Nation in ‘Post Truth’ Era: The Indian Story.” Asian
Ethnicity. December 2, 2019. Accessed online on January 9 2020 at: https://​www.tandfonline.
com/​ doi/​full/​10.1080/​ 14631369.2019.1698942?casa_​ token=COxjs_​IjVZUAAAAA%3A-​
zNo2vtJ lW1pMOfEs5jEMk90v9lmkmk ZwbV62dsCBZfcdTXGbEkUJrc1v
FJT7SFvC2uGGcgBMcg.
Reid, Anthony. 2010. Imperial Alchemy: Nationalism and Political Identity in Southeast Asia.
Cambridge: Cambridge University Press.
Riedl, Rachel. 2014. Authoritarian Origins of Democratic Party Systems in Africa.
Cambridge: Cambridge University Press.
Rudolph, Suzanne Hoeber, and Lloyd Rudolph. 2006. Postmodern Gandhi and Other
Essays: Gandhi in the World and at Home. Chicago: University of Chicago Press.
Shaikh, Farzana. 2009. Making Sense of Pakistan. Oxford: Oxford University Press.
Sisson, Richard, and Stanley Wolpert. 1988. Congress and Indian Nationalism: The
Preindependence Phase. Berkeley: University of California Press.
Slater, Dan. 2010. Ordering Power: Contentious Politics and Authoritarian Leviathans in
Southeast Asia. Cambridge, MA: Cambridge University Press.
Slater, Dan, and Maya Tudor. 2019. “Why Religious Tolerance Won in Indonesia But Lost in
India.” Foreign Affairs, July 3, 2019.
Smith, Rogers. 2003. Stories of Peoplehood: The Politics and Morals of Political Memberships.
Cambridge, UK: Cambridge University Press.
Straus, Scott. 2015. Making and Unmaking Nations: War, Leadership, and Genocide in Modern
Africa. New York: Cambridge University Press.
Sundstrom, Harold. 1957. Indonesia: Its People and Its Politics. Tokyo: Hokuseido Press.
Tamir, Yael. 2019. Why Nationalism. Princeton, NJ: Princeton University Press.
Tilman, Robert O. 1976. “The Centralization Theme in Malaysian Federal-​State Relations,
1957–​75.” Occasional Paper No. 39. Singapore: Institute for Southeast Asian Studies.
Trivedi, Lisa. 2003. “Visually Mapping the Nation: Swadeshi Politics in Nationalist India, 1920–​
1930.” Journal of Asian Studies 62, no. 1 (February): 11–​41.
Tsebelis, George. 1995. “Decision Making in Political Systems: Veto Players in Presidentialism,
Parliamentarism, Multicameralism and Multipartyism.” British Journal of Political Science
25, no. 3: 289–​325.
Tudor, Maya. 2013. The Promise of Power: The Origins of Democracy in India and Autocracy in
Pakistan. New York: Cambridge University Press.
Tudor, Maya. 2018. “India’s Nationalism in Historical Perspective: The Democratic Dangers of
Ascendant Nativism.” Indian Politics and Policy 1: 108–​132.
 Islam, Nationalism, and Democracy in Asia    191

Tudor, Maya, and Dan Slater. 2016. “The Content of Democracy: Nationalist Parties and
Inclusive Ideologies in India and Indonesia.” In Parties, Democracy and Movements in the
Developing World, edited by Nancy Bermeo and Deborah Yashar. New York: Cambridge
University Press.
Tudor, Maya, and Dan Slater. 2020. Nationalism, Authoritarianism, and Democracy: Historical
Lessons from South and Southeast Asia. Perspectives on Politics. In press.
vom Hau, Matthias. 2009. “Unpacking the School: Textbooks, Teachers, and the Construction
of Nationhood in Mexico, Argentina and Peru.” Latin American Research Review 44, no.
3: 127–​154.
Weiner, Myron. 1977. “The 1977 Parliamentary Elections in India.” Asian Survey 17, no.
7: 619–​626.
Wimmer, Andreas. 2019. “Why Nationalism Works and Why It Isn’t Going Away.” Foreign
Affairs, March/​April, 2019.
Yashar, Deborah. 1997. Demanding Democracy. New York: Cambridge University Press.
Zubrzycki, Genevieve. 2016. Beheading the Saint: Nationalism, Religion and Secularism in
Quebec. New York: Cambridge University Press.
 Chapter 9

Military P ol i t i c s i n
M u slim So c i et i e s

Nicholas J. Lotito

Authoritarianism is more prevalent in Muslim societies today than in other world

regions, but it has not always been so. Instead, Muslim societies have been left out of
global gains in civil and political rights over the past half-​century. A prominent argu-
ment attributes the “robustness of authoritarianism” in the Middle East to the strength
of the coercive apparatuses, including the police and military (Bellin 2004). Does this
argument apply to the broader Muslim world as well? And how does the military specif-
ically contribute to the institutionalization of authoritarianism and its survival?
Examining cross-​national data, I find that the democratic deficit observed in the
Middle East characterizes the broader Muslim world as well. By way of explanation,
I consider how the armed forces can contribute to the persistence of authoritarian rule.
In many Muslim-​majority countries, the military has sustained authoritarianism by
violently suppressing anti-​regime protests, thereby foreclosing the possibility of mass-​
based democratization. Examining global patterns of democratization since the 1970s,
I find that military repression of large-​scale protests has been more likely in Muslim-​
majority states than elsewhere. I illustrate the role of the military in responding to pop-
ular uprisings through the cases of the Arab Spring, then turn to explaining military
violence against protesters.
To explain the prevalence of military repression in the Muslim world, I focus on
three factors: security threats, foreign military aid, and organizational cultures within
the armed forces. I find the Muslim world has been characterized by chronic insecu-
rity, high levels of foreign assistance, and authoritarian tendencies embedded in the or-
ganizational cultures of the armed forces. These factors may all contribute to a higher
propensity for Muslim-​majority armies to intervene in politics, including by repressing
protests. By contrast, I do not find evidence that factors related to Islam as a culture or
religion have any systematic effect on military behavior. Several cases of successful de-
mocratization in the region demonstrate that authoritarianism is not an immutable fea-
ture of Muslim-​majority societies.
194   Nicholas J. Lotito

Persistent Authoritarianism

Many authoritarian regimes outside the Muslim world have seen substantial
improvements in civil liberties and political rights since the 1970s—​Huntington’s (1991)
“third wave” of democratization—​but Muslim-​majority states have been largely left be-
hind. Freedom House, the US-​based NGO, has been tracking civil liberties and political
rights around the world for the past half-​century. Figure 9.1 plots the average score in
each category (on a seven-​point scale) annually from 1972 to 2017, grouping more devel-
oped countries, Muslim-​majority developing countries, and other developing countries.
Over the period, full democracy (rated “Free”) has been limited mostly to wealthy
states with advanced industrial economies, represented here by OECD membership.
Among poorer countries, the level of freedom was quite low in 1972, and the gap be-
tween Muslim-​majority and other developing countries was small. Since the late 1970s,
however, there has been a marked divergence: the Muslim world has remained predom-
inantly authoritarian even as other states have realized democratic gains. While the av-
erage score for non-​Muslim developing countries is today near the top of the “Partly
Free” category, the average for Muslim-​majority countries remains “Not Free.” Among
Muslim-​majority states, the Middle East and North Africa (MENA) region has lower
average scores, but other Muslim-​majority developing countries still lag far behind the
non-​Muslim average.1

OECD*
Free

Non-OECD*
Partly Free

Muslim-Majority
Not Free

1970 1980 1990 2000 2010

Year
Civil Liberties Political Rights

*Excludes Muslim-majority

Figure 9.1: Freedom in the World Index, 1972–​2017.

Note: Excludes Muslim-​majority.
Source: Data from Freedom House.
 Military Politics in Muslim Societies    195

Why have Muslim societies been so resistant to democratization? As Eva Bellin (2004)
has argued, the answer lies not in the absence of democratizing forces or prerequisites,
but in the exceptional strength of authoritarianism in these countries. In other words,
Muslim-​majority states remain authoritarian because the regimes that have held power
since independence are exceptionally resilient. In particular, the robustness of author-
itarianism depends on the strength of the coercive apparatus: the civilian and military
forces that repress dissent, undermine civilian political control, and suppress mass
mobilization.
In societies where pent up demand for democratization has long been suppressed,
popular uprisings are a natural pathway to political change. Between 1946 and 2010,
17 percent of cases of authoritarian breakdown occurred due to popular uprisings,
which became more common after the end of the Cold War (Geddes, Wright, and Frantz
2018, 179). The most common causes of regime collapse were coups d’état (35 percent)
and elections (26 percent). While election-​related transitions often lead to democratic
regimes, such meaningful elections are unlikely to take place in the entrenched author-
itarian regimes of the Muslim world. Coups d’état, on the other hand, rarely led to dem-
ocratic rule during the Cold War, when they were common in the Muslim world. Coups
have more often led to democratization in the post–​Cold War period, but they have
become less frequent as authoritarian leaders have employed coup-​proofing strategies
(Marinov and Goemans 2014). If both meaningful elections and coups d’état have be-
come unlikely, popular uprisings may be the clearest pathway to democracy in many
Muslim-​majority societies.
Because political mobilization under authoritarianism is risky for individual
participants, we are less likely to see the emergence of large-​scale, pro-​democracy mo-
bilization where there is a high degree of routine political repression. Authoritarian
regimes employ various internal security forces, including police, gendarmes, and in-
telligence services, to monitor citizens’ political activities and limit mobilization. In
many cases, the military contributes to this “routine” repression as well, usurping or
duplicating the domestic roles of civilian forces. For example, Egypt has three main in-
telligence services.2 Rather than working together in a cohesive interagency process, the
components of Egypt’s fragmented intelligence community duplicate many of the same
roles in an effort to counterbalance each other and secure political influence (Kandil
2012; Sirrs 2010). This competition among civilian and military institutions is a common
feature of authoritarian governance and can exacerbate the degree of repression.
Beyond its role in routine repression, the military also plays a unique political role
as the regime’s last bulwark against revolution. In most cases of mass protest, the police
are up to the task of basic repression, dispersing crowds, and restoring calm. But if mass
mobilization grows sufficiently large, protesters will eventually overwhelm the capacity
of the civilian security forces. When protests escalate beyond the repressive capacity of
the police, leaders must choose whether to resort to the most severe form of physical
repression: military force against unarmed citizens. Because the military generally has
the physical capacity to violently disperse demonstrations, the military often becomes a
de facto arbiter between protesters and the regime. Military violence does not guarantee
196   Nicholas J. Lotito

that an anti-​regime uprising will fail, but the cessation of such violence is often a prereq-
uisite for the movement’s success. A revolutionary outcome is often marked by a loss of
the monopoly of force, which occurs when the regime or the military is unwilling to use
(sufficient) force against protesters (Tilly 1993, 241). As such, no process of political lib-
eralization can go forward without the support, or at least tolerance, of the armed forces
(Barany 2012).

Military Repression of Protests

Is military repression of nonviolent uprisings more likely in Muslim-​majority states?

I test this proposition using original data on military responses. First, I identify pop-
ular uprisings from 1946 to 2013 using the Nonviolent and Violent Campaigns and
Outcomes (NAVCO) 2.1 data set (Chenoweth and Shay 2019). I include protest
campaigns that were primarily nonviolent, called for regime change, and included at
least 100,000 participants. To focus on the role of the military, I exclude campaigns that
were suppressed by internal security forces alone, without requiring military backup.
For each of the resulting 86 campaigns, I code a categorical military repression variable
to indicate the level of violence employed against protesters by the armed forces.3
These data, summarized in Figure 9.2, illustrate three key findings. First, military vio-
lence is a common response to large protests, regardless of world region or time period.

Muslim-Majority Others
20

15

10

0
1940 1960 1980 2000 1940 1960 1980 2000

Military Violence No Military Violence

Figure 9.2: Military responses to popular uprisings, 1946–​2013.

Source: Data from NAVCO 2.1 (Chenoweth and Shay 2019); Lotito (2018b).
 Military Politics in Muslim Societies    197

In cases where protests overwhelmed the repressive capacity of internal security forces,
the military responded with violence 66.3 percent of the time. The high rate of military
involvement underscores the pivotal role the army plays in revolution. Many cases of
military violence ultimately resulted in success for the movement, but this does not sug-
gest the military was therefore unimportant. Instead, the decisive moment in many such
revolutionary situations is when the military stands down or switches sides.
Second, uprisings in the Muslim world have been more likely to face military repres-
sion. Nonviolent uprisings in Muslim-​majority countries are one-​third more likely to
face military violence than similar protests in other countries (82 percent vs. 61 per-
cent).4 Where the armed forces are more willing to use violence to defend the regime,
we are less likely to see mass challenges in the first place, because many protesters will be
deterred by the risk of violent repression. Because taking to the streets in this repressive
context is a strong signal of the movement’s robust popular support, we might expect
the military to be more reticent to use violence. Instead, the data show a low rate of large
protests, but a high rate of military repression, across the Muslim world.
Third, the pattern of military repression mirrors trends in democratization. The
“third wave” of democratization is clearly visible in the data from 1970 to 2010. During
this period, the rate of military repression declined substantially, contributing to the
success of so many democratization campaigns worldwide. However, the loosening of
military repression took place only outside the Muslim world. By contrast, the rate of
violent military repression remained high in Muslim-​majority countries throughout
the period. We may conclude that the continuity of authoritarianism in most Muslim-​
majority states is at least partially attributable to the robustness of military repression as
a counterbalance to democratizing pressures.
These data offer only a partial story of pro-​democratic mobilization in the Muslim
world. Notably, they leave out the many protests that never matured into large, sustained
campaigns, whether due to police repression, government concessions, or other causes.
As the data illustrate, there have been relatively few protest campaigns in the Muslim
world that have reached the threshold of potential military intervention. With fewer
uprisings, we can expect fewer successes, so the non-​emergence of large uprisings is an
important cause of authoritarian stability. It is likely that high levels of domestic repres-
sion, carried out primarily by police and intelligence services, played an important role
in preventing these uprisings from growing to revolutionary scale.
When the Arab Spring protests began in 2010, they quickly spread to virtually every
country in a region where previous mobilization was limited. The widespread mobi-
lization of the Arab Spring suggests that Muslim populations were not unmotivated
to demand change, but rather that structural obstacles to mobilization had prevented
the emergence of mass movements earlier. That most of these uprisings ended in mil-
itary repression starkly demonstrates the military’s power to sustain authoritarianism.
Where national armies did not engage in repressing the protests (Egypt and Tunisia),
old political leaders were swept from power.5 But where national armies brought to bear
their full military might against civilians (Bahrain and Syria), existing regimes remained
in power. Even where the military split apart and mutinous troops joined the ranks of
198   Nicholas J. Lotito

armed rebellion (Libya and Yemen), military repression kept leaders in power until for-
eign intervention tipped the balance.
Scholars have widely recognized the central role of the military in responding to the
Arab Spring, and many have attempted to explain variations in military responses. Most
of this research frames the question of military responses to protest in terms of defec-
tion and points to various factors that might influence the military’s loyalty to the re-
gime (Albrecht et al. 2016; Albrecht and Ohl 2016; Barany 2011; Bellin 2012; Hazen 2019;
Makara 2013; Nepstad 2013). In this view, the critical question for military officers is
whether to “defect or defend” (Lee 2015). Whether loyalty stems from officers’ profes-
sional values (Bellin 2012; Bou Nassif 2015b; Lutterbeck 2013), material interests (Bou
Nassif 2015a; Brooks 2013), or ethnicity (Bou Nassif 2015c), scholars assume that soldiers
will refrain from challenging the regime, and will even fight to defend it, as long as they
remain faithful to its leader. Although each of these variables—​professionalism, mate-
rial interests, and ethnicity—​offers some explanatory power in the Arab Spring cases,
none of them seem to vary systematically between Muslim-​majority societies and the
rest of the world. If some armies in the Muslim world suffer from poor professionalism,
patronage relationships, and sectarianism, many others are highly professional, well in-
stitutionalized, and untroubled by ethnic divisions.6

Explaining Military Responses

to Protest

While the Arab Spring cases help illuminate variations in military responses to protests,
as well as the importance of the military to political outcomes, existing theories do not
appear to explain the higher rate of military repression in Muslim-​majority countries.
Why have these armies been so likely to use repression against mass protests? This sec-
tion investigates three possible explanations for the difference: chronic insecurity, for-
eign military aid, and military culture.

Chronic Insecurity
As security organizations first and foremost, armies are deeply affected by the security
environment in which they operate. Whereas armies in the developed world are gener-
ally oriented toward external threats (i.e., foreign invasion), developing states often rely
on the military for internal security as well. Scholars of civil-​military relations have long
warned that the involvement of the military in internal security roles and missions tends
to politicize the armed forces, to the detriment of their professionalism and adherence
to norms of civilian supremacy (Desch 1999; Huntington 1957, 1995). The risk of mili-
tary intervention into politics is heightened when civilian institutions are weak and lack
 Military Politics in Muslim Societies    199

20%

15%
Civil War Rate

10%

5%

0%

1960 1980 2000 2020

Year

Muslim-Majority (42) Other (113)

Excludes countries with 2018 population under one million.

Figure 9.3: Incidence of civil war, 1946–​2018.

Note: Excludes countries with 2018 population under one million
Source: Data from UCDP/​PRIO Armed Conflict Dataset.

legitimacy (Staniland 2008). In this view, even relatively low-​intensity conflict or spo-
radic terrorist activity can have political consequences if the government turns to the
armed forces to manage the threat. Unfortunately, many countries experience chronic
violent conflict, so the involvement of the military in internal security becomes almost
inevitable.
Political violence is hardly unique to Muslim societies, but the prevalence of civil
conflict in the region has given the military an especially prominent role in domestic
security. In the post–​Cold War era, the Muslim world has experienced significantly
more internal violence than other regions of the world (Gleditsch and Rudolfsen 2016).
This divergence has been produced by a simultaneous increase in the rate of civil war
in Muslim-​majority countries and a decline in the rate for non-​Muslim countries, be-
ginning around 1990 (see Figure 9.3). By 2016 the incidence of civil war had become
extremely concentrated, with six of seven active civil wars located in Muslim-​majority
countries.7
Terrorism has also become increasingly frequent in Muslim-​majority countries.
Despite the common association of radical Islamism with contemporary terrorism,
until the 2000s, terrorism was more concentrated in democracies than in authori-
tarian regimes (Chenoweth 2010; Wilson and Piazza 2013). In the post–​Cold War era,
200   Nicholas J. Lotito

30.8
30

20

14.14

10

5.61

2.41

0
Incidents Lethality (per incident)
Islamist Non-Islamist

Figure 9.4: Islamist terrorism in civil war, 1970–​2012.

Source: Data from Terrorism in Armed Conflict (Fortna, Lotito, and Rubin forthcoming).

however, the correlation between regime type and terrorism has been reversed. One
possible cause of this shift is that much of the terrorism observed today occurs within
civil wars (Findley and Young 2012; Fortna, Lotito, and Rubin 2018; Stanton 2013). As
civil wars have become more frequent in Muslim-​majority countries, the terrorism rate
has increased accordingly. Exacerbating the problem, within civil wars, Islamist groups
tend to use terrorism more often than other nonstate combatants (see Figure 9.4).
Civil war and terrorism tend to have negative effects on economic and political devel-
opment, potentially undermining processes that might otherwise support democrati-
zation in Muslim-​majority countries (Gupta et al. 2004; Neumayer 2004; Nitsch and
Schumacher 2004).
The challenges of chronic insecurity are exacerbated when civilians enlist the mili-
tary as a bulwark against domestic political threats. Islamism—​a political ideology that
seeks to radically change government and society to conform with Islamic principles—​
has motivated both violent and nonviolent challenges to the authoritarian status quo.
In Muslim-​majority societies, Islamists are usually among the most prominent regime
opponents. Because Islamists are often quite popular, authoritarian leaders have a gen-
uine fear of Islamist-​led popular mobilization. Leaders often respond by empowering
the military in an attempt to weaken their Islamist opponents (Cook 2007). Their fear
also creates an interest in portraying even peaceful, everyday forms of Islamism as a na-
tional security threat. In this way, the question of religion in politics becomes securitized,
as governments use the language of threat to describe peaceful political participation
 Military Politics in Muslim Societies    201

(Buzan, Wæver, and Wilde 1998; McDonald 2008). Securitizing important political
debates leads to further deference to the military and provides justification for oppres-
sive national security measures. An example of a tangible legal consequence of this phe-
nomenon is the continuous state of emergency promulgated in Egypt during the entire
tenure of President Hosni Mubarak, from 1987 to 2011 (ICJ 2018).
In sum, the tendency of the military to intervene in domestic politics could be a con-
sequence of political violence, which has occurred at a higher rate in the Muslim world
than elsewhere in the post–​Cold War era. When governments rely on military forces for
internal security, it is likely to increase the politicization of the armed forces and their
encroachment into additional areas of political life. The problem is exacerbated when
authoritarian regimes extend the discourse of domestic security to include peaceful
forms of political participation, encouraging soldiers to use violence against peaceful
demonstrators.

Military Aid
A second factor that might help explain the high level of military repression in the
Muslim world is foreign aid. Scholars have long recognized that international support
can be critical to the survival of authoritarian regimes (Skocpol 1979). In fact, the risk
of sustaining authoritarianism has encouraged many donors to condition their assis-
tance on political or human rights standards (Carnegie and Marinov 2017; Dunning
2004; Resnick 2018). The likelihood that foreign aid will discourage democratization
may also be higher in countries where rulers rely on small distributional coalitions, as
in most Muslim-​majority countries (Wright 2009). Even when the donors are Western
democracies, aid recipients include some of the world’s most authoritarian regimes. For
example, the United States has been the primary international sponsor of authoritarian
regimes in Iran (until 1979), Egypt, Saudi Arabia, Pakistan, Indonesia (until 1998), and
other Muslim countries. As Jason Brownlee (2012) demonstrates in Egypt, American
support has often helped those regimes stave off domestic pressures for democratiza-
tion. “Elections and political reform,” Brownlee observes, “are welcome only insofar as
they impede extremists and enhance stability” (9).
However, the political effects of military aid specifically are less well understood.
During the Cold War and since, the US and other great powers have provided weapons,
training, and budgetary support to their allies’ armed forces. Military aid can provide
diplomatic benefits and improve security, but it also distorts civil-​military relations
in the target country by increasing the institutional strength of the military relative to
other state institutions (Lotito and Joyce 2020). Military aid is particularly significant in
the Muslim world: between 2010 and 2020, Muslim-​majority states represented seven of
the top ten recipients of US military aid.8 Owing to the geostrategic importance of many
Muslim-​majority states and the global threat of Islamist terrorism, the Muslim world
has received outsized amounts of foreign military aid. This pattern is clearly reflected in
US military aid data (see Figure 9.5).
202   Nicholas J. Lotito

Military aid comes in many forms. By far the largest category of expenditure relates to
the acquisition of weapons systems and other equipment. While this equipment serves
no direct political function, it may nevertheless produce political effects by bolstering
the organizational strength and capacity of the armed forces relative to other state
institutions. In more extreme cases, military aid can exacerbate violence (Boutton 2019;
Dube and Naidu 2015). Rather than provide minimal military resources to otherwise
defenseless nations, US military aid has supported exceptionally high military spending
across the Muslim world. In fact, Muslim-​majority states comprise ten of the world’s top
fifteen military spenders relative to national income.9 One consequence of extravagant
defense spending has been to strengthen the military relative to other state institutions.
This imbalance is reflected in public opinion surveys, which consistently find that public
trust in the armed forces far outstrips other state institutions in these countries (Lotito
2018a).
A relatively small portion of military assistance is dedicated to training and educating
foreign officers. This form of aid is hypothesized to create the possibility of norm change,
by exposing recipient military officers to democratic values. In a few cases, Western
training has apparently generated some adherence to democratic norms by supporting
an intergenerational shift in organizational culture (Soeters and Van Ouytsel 2014).
More generally, American training is not associated with democratization (Taylor 2014,
chap. 8). Foreign training may also have perverse effects. Savage and Caverley (2017)

$150
Total aid per capita

$100

$50

$0
Military Non-Military
Aid type

Muslim-Majority Other

Excludes states with population under 500,000. Amounts in constant 2018 US Dollars.

Figure 9.5: US foreign aid, 2001–​2019.

Note: Excludes states with population under 500,000. Amounts in constant 2018 US dollars.
Source: Data from USAID Greenbook.
 Military Politics in Muslim Societies    203

find that US military training increases coup risk because it strengthens the armed forces
relative to the civilian regime. Similarly, Casey (2020) finds that US patronage has done
nothing to reduce the risk of military coups.10 These studies do not offer definitive ev-
idence on the effects of military aid on responses to protest; however, they do suggest
that military aid does not typically foster pro-​democratic behavior and may instead en-
courage political intervention by the military.

Military Culture
A third factor that could help explain differences in military responses to protest in
Muslim-​majority states is military culture. Systematic differences in the organizational
cultures of the armed forces in Muslim-​majority states could explain general patterns
of civil-​military relations in these contexts (Lotito 2018b). The organizational culture
view of military behavior holds that, like any large organization, the military responds
to events based on its existing practices and knowledge. The constituent elements of or-
ganizational culture, like shared understandings and repertoires of action, are difficult
to observe ex ante, so their causal role in producing any particular action can be am-
biguous. Nevertheless, we can observe stable patterns of behavior over time and iden-
tify historical legacies and critical junctures in institutional development that may have
contributed to those patterns. Even if organizational culture does not fully determine
military behavior, it can powerfully condition soldiers’ responses to protests and, ulti-
mately, whether the army will engage in repression.
This view of organizational culture differs markedly from arguments based on cul-
tural essentialism. For example, a prominent argument holds that Arab culture explains
the poor battlefield performance of the Egyptian, Iraqi, Jordanian, Libyan, Saudi, and
Syrian armies in their wars against Israel (Pollack 2002). By contrast, the organizational
culture of the armed forces may be only tangentially related to the broader national cul-
ture. Nothing intrinsic to Muslim culture or the religion of Islam explains military inter-
vention into politics, the politicization of counterterrorism, or extraordinary foreign aid
flows. Instead, the armed forces of the Muslim world have often been indoctrinated with
modernist and nationalist, not religious, ideologies (Picard 2005, 121). Moreover, polit-
ical intervention by the armed forces has varied greatly across the region. For example,
coups d’état are commonplace in some Muslim-​majority countries, yet virtually un-
heard of in others (Powell and Thyne 2011, 255). Similarly, the armed forces have brutally
repressed anti-​regime mobilization in most cases but forced dictators from power in
others. Islamic culture, therefore, fails to explain any observed authoritarian tendency
in the military cultures of the Muslim world.
In many Muslim societies, the military plays a dual role, responsible for both na-
tional security and regime security. The ways these roles overlap and reinforce each
other have resulted in more politically influential armed forces than elsewhere. In the
postcolonial states of the Muslim world, the military played a critical role in defining
national identity by “erecting and defending its boundaries against external enemies
204   Nicholas J. Lotito

and internal separatist movements” (Picard 2005, 118). Intellectuals and governments
alike often viewed the military as “the ideal instrument to direct the industrializa-
tion, institutionalization, and reform necessary for a modern society” (Cook 2007,
2). In their role as nation-​building institutions, the armed forces have often been
indoctrinated with strict secularist ideologies, which encourage violent responses to
even nonviolent Islamist activism, encouraging soldiers to view religion through a se-
curity lens. In sum, processes of postcolonial nation-​building and chronic insecurity
in the independence era both tended to enhance the political role of the military in the
Muslim world.
A comparison of India and Pakistan illustrates how different military cultures can
create divergent patterns of civil-​military relations. Since Partition, India has not
witnessed a single military coup or other serious breach of military professionalism.
Meanwhile, Pakistan is a prime example of how a highly politicized military can serve
to sustain authoritarianism in a Muslim-​majority country. Superficially, the propor-
tion of Muslims in the two countries (13 percent in India vs. 96 percent in Pakistan)
corresponds to their divergent paths of civil-​military relations; however, the most
compelling explanations for the divergence have nothing to do with religion. Instead,
scholars point to historical factors like colonial inheritances (economic and military),
the strength of the ruling party in each country, and geopolitical insecurity to explain
why civil-​military relations have been far more democratic in India than in Pakistan
(Staniland 2008; Wilkinson 2015). From a cultural perspective, Shah (2014) argues that
the Pakistani army’s repeated interference in the political sphere is best explained by
norms within the officer corps. This “belief system” informs how soldiers perceive dem-
ocratic institutions, how they interpret the proper role and function of the military, and
how they respond to perceived civilian failures or threats to military interests (8). The
organizational culture of the Pakistani military, Shah argues, explains its use of “virtu-
ally genocidal” violence against nonviolent Bengali protesters in 1971, among other acts
of military repression (112). As Pakistan demonstrates, it is neither Islamic culture nor
a universal military mindset that dictates the military’s political behavior, but rather an
organizational culture specific to a particular national army.
The historical persistence of military organizational culture also underscores the
difficulty of democratic transition, particularly in countries where the military has
played a central political role. Among the most successful democratic transitions in the
Muslim world to date have been in Senegal and Tunisia, which transitioned to multi-
party democracy in 2000 and 2011, respectively. Senegal’s transition, which overcame
potential ethnic divisions within the country, was facilitated by the military’s long-​held
civic-​national loyalty (Harkness 2018, chap. 4). In Tunisia, the military’s decision not to
repress a massive popular uprising enabled a remarkably peaceful transition to democ-
racy (Bou Nassif 2015b; Brooks 2013; Jebnoun 2014; Lotito 2020). In both cases, mili-
tary culture portended a successful outcome: neither army had been politically central,
nor carried out a coup d’état. These factors made the Senegalese and Tunisian armies
exceptional by the standards of the Muslim world, and indeed the broader developing
world as well. Unfortunately, the military cultures of most of the region’s other armies
 Military Politics in Muslim Societies    205

still reproduce authoritarian tendencies, so we should not expect them to embrace de-
mocratization anytime soon.
Still, we might take comfort in the observation that even some long-​standing military
dictatorships have made successful transitions to democracy.11 In the Muslim world,
Indonesia is a striking example of a state that sidelined its formerly dominant military
(Lee 2009).12 If Senegal and Tunisia demonstrate that a long-​held culture of nonin-
tervention within the military is the optimal scenario for democratization, Indonesia
proves that cultural change is also possible. As in many military-​dominated regimes,
political power in Indonesia became highly personalistic under the thirty-​year rule
of Suharto. A military officer, Suharto seized power by coup d’état in 1967, then con-
centrated dictatorial powers in his own office. Having brutally repressed all dissent for
many years, Suharto lost power in 1998 when the armed forces ignored his clear and
public order to suppress mounting protests. The sudden and decisive shift in military
policy resulted from intense conflict and rivalries within army ranks, which created an
incentive for the commander of the armed forces, General Wiranto, to defect from the
regime (Lee 2015, chap. 4). Yet even as multiparty democracy flourished in Indonesia,
it took years of gradual reform to reduce the military’s role as a political “veto player”
(Mietzner 2013).

Conclusion

In many Muslim-​majority countries, the military continues to serve as a component of

the coercive apparatus that sustains authoritarianism. The military’s propensity to en-
gage in repression may be heightened by chronic insecurity, military aid, and organiza-
tional cultures favoring the repression of mass protests. As long as these factors persist,
the armed forces are likely to deter or suppress future protests. Nevertheless, cases like
Indonesia and Tunisia demonstrate that civil-​military relations can improve in the
longer term.
The Arab Spring transformed the political landscape of the greater Middle East,
leading to pervasive insecurity and the retrenchment of authoritarianism. The mili-
tary was a critical player in the uprisings and continues to gain political influence in
a region unsettled by terrorism, civil war, and great power competition. Much schol-
arship has focused on the unmet promise of the Arab Spring, leaving the prospects
for future mass mobilization in doubt. Having prevented revolution in 2011, will the
armed forces consolidate their political position, or will the high costs of civil war
deter future military repression? Will new economic challenges reinvigorate protest
movements?
Finally, when authoritarian regimes do fall, opening a window of opportunity for
democracy, the coercive institutions that underpinned the old regime typically re-
main in place. The continuity of authoritarian institutions represents a serious threat
to democratic consolidation, especially when political elites lack strong incentives
206   Nicholas J. Lotito

to carry out difficult institutional reforms (Lotito 2019). Not enough research has
been done to understand the circumstances under which these reforms might be ac-
complished in the Muslim world. Research on the role of the military in democratic
transitions in Europe and Latin America has identified critical factors like regional
organizations (e.g., NATO and the European Union) or elite pacts, which may not be
relevant to Muslim-​majority countries (Stepan 1988; Trinkunas 2006). The transitions
from military rule in East and Southeast Asia, where popular uprisings swept military-​
backed regimes from power, may offer more apt comparisons and useful lessons for
the Muslim world. If we are serious about understanding the prospects for democrati-
zation in the Muslim world, we cannot ignore the radical political transformations that
have occurred further east.

Notes
1. For 2017, political rights scores averaged 5.67 for MENA, 5.04 for Muslim developing coun-
tries outside MENA, 3.27 for non-​Muslim developing countries, and 1.22 for the OECD.
2. These are the General Intelligence Service, or Mukhabarat (under direct control of the
president), the military intelligence service (under the Ministry of Defense), and the
National Security Agency (under the Ministry of the Interior).
3. For coding details, see Lotito (2018b), 184–​186.
4. The difference is statistically significant at the 95 percent confidence level.
5. Egypt’s subsequent military coup of July 2013 underscores the pervasive political influence
of that country’s military.
6. Professionalism is usually defined subjectively, but Muslim-​majority countries like Egypt,
Indonesia, Turkey, and the UAE also rate highly on more objective measures, such as ex-
penditure per soldier and the development of national military academies or military
periodicals (Toronto 2017).
7. They are Afghanistan, Iraq, Libya, Sudan, Somalia, Syria, Turkey, and Yemen. The other war
was an Islamist insurgency in Nigeria, where the population is approximately half-​Muslim.
8. The others were Israel, Ukraine, and Russia (to which aid was eliminated following the an-
nexation of Crimea in 2014).
9. “Military Spending as a Share of GDP, The Top 15 Countries, 2019,” SIPRI, https://​www.
sipri.org/ ​ research/ ​ armament- ​ and- ​ d isarmament/ ​ arms- ​ and- ​ m ilitary- ​ e xpenditure/​
military-​expenditure. Of the remaining top spenders, two are great powers with major
current and historical military involvement in Muslim-​majority countries (US and
Russia), and each of the others borders a Muslim-​majority state with which it has a history
of armed conflict (Israel, Armenia, and South Sudan).
10. Soviet patronage, on the other hand, helped sustain client regimes through effective
coup-​proofing.
11. Notably, notorious military dictatorships in Spain, Brazil, the Philippines, and South
Korea gave way to democratic rule during the Third Wave.
12. Another Muslim-​majority country, Mali also underwent a remarkable transition from
military rule to democracy in just two years (1991–​1993); however, a coup d’état in 2012
underscored the fragility of civil-​military relations and democracy more broadly.
 Military Politics in Muslim Societies    207

References
Albrecht, Holger, Aurel Croissant, Fred Haley Lawson, and David Pion-​Berlin, eds. 2016.
Armies and Insurgencies in the Arab Spring. Philadelphia: University of Pennsylvania Press.
Albrecht, Holger, and Dorothy Ohl. 2016. “Exit, Resistance, Loyalty: Military Behavior during
Unrest in Authoritarian Regimes.” Perspectives on Politics 14, no. 1: 38–​52.
Barany, Zoltan. 2011. “The Role of the Military.” Journal of Democracy 22, no. 4: 24–​35.
Barany, Zoltan. 2012. The Soldier and the Changing State: Building Democratic Armies in Africa,
Asia, Europe, and the Americas. Princeton, NJ: Princeton University Press.
Bellin, Eva. 2004. “The Robustness of Authoritarianism in the Middle East: Exceptionalism in
Comparative Perspective.” Comparative Politics 36, no. 2: 139.
Bellin, Eva. 2012. “Reconsidering the Robustness of Authoritarianism in the Middle
East: Lessons from the Arab Spring.” Comparative Politics 44, no. 2: 127–​149.
Bou Nassif, Hicham. 2015a. “Generals and Autocrats: How Coup-​Proofing Predetermined the
Military Elite’s Behavior in the Arab Spring.” Political Science Quarterly 130, no. 2: 245–​275.
Bou Nassif, Hicham. 2015b. “A Military Besieged: The Armed Forces, the Police, and the Party
in Bin ‘Ali’s Tunisia, 1987–​2011.” International Journal of Middle East Studies 47, no. 1: 65–​87.
Bou Nassif, Hicham. 2015c. “‘Second-​Class’: The Grievances of Sunni Officers in the Syrian
Armed Forces.” Journal of Strategic Studies 38, no. 5: 626–​649.
Boutton, Andrew. 2019. “Military Aid, Regime Vulnerability and the Escalation of Political
Violence.” British Journal of Political Science, published ahead of print, September 21, 2019.
Brooks, Risa. 2013. “Abandoned at the Palace: Why the Tunisian Military Defected from the
Ben Ali Regime in January 2011.” Journal of Strategic Studies 36, no. 2: 205–​220.
Brownlee, Jason. 2012. Democracy Prevention: The Politics of the US-​ Egyptian Alliance.
New York: Cambridge University Press.
Buzan, Barry, Ole Wæver, and Jaap de Wilde. 1998. Security: A New Framework for Analysis.
Boulder, CO: Lynne Rienner.
Carnegie, Allison, and Nikolay Marinov. 2017. “Foreign Aid, Human Rights, and Democracy
Promotion: Evidence from a Natural Experiment.” American Journal of Political Science 61,
no. 3: 671–​683.
Casey, Adam E. 2020. “The Durability of Client Regimes: Foreign Sponsorship and Military
Loyalty, 1946–​2010.” World Politics 72, no. 3: 411–​447.
Chenoweth, Erica. 2010. “Democratic Competition and Terrorist Activity.” Journal of Politics
72, no. 1: 16–​30.
Chenoweth, Erica, and Christopher Wiley Shay. 2019. “NAVCO 2.1 Dataset.” https://​dataverse.
harvard.edu/​dataset.xhtml?persistentId=doi:10.7910/​DVN/​MHOXDV.
Cook, Steven A. 2007. Ruling but Not Governing: The Military and Political Development in
Egypt, Algeria, and Turkey. Baltimore: Johns Hopkins University Press.
Desch, Michael. 1999. Civilian Control of the Military: The Changing Security Environment.
Baltimore: Johns Hopkins University Press.
Dube, Oeindrila, and Suresh Naidu. 2015. “Bases, Bullets, and Ballots: The Effect of US Military
Aid on Political Conflict in Colombia.” Journal of Politics 77, no. 1: 249–​267.
Dunning, Thad. 2004. “Conditioning the Effects of Aid: Cold War Politics, Donor Credibility,
and Democracy in Africa.” International Organization 58, no. 2: 409–​423.
Findley, Michael G., and Joseph K. Young. 2012. “Terrorism and Civil War: A Spatial and
Temporal Approach to a Conceptual Problem.” Perspectives on Politics 10, no. 2: 285–​305.
208   Nicholas J. Lotito

Fortna, Virginia Page, Nicholas J. Lotito, and Michael A. Rubin. 2018. “Don’t Bite the Hand
That Feeds: Rebel Funding Sources and the Use of Terrorism in Civil Wars.” International
Studies Quarterly 62, no. 4: 782–​794.
Fortna, Virginia Page, Nicholas J. Lotito, and Michael A. Rubin. Forthcoming. “Terrorism
in Armed Conflict: New Data Attributing Terrorism to Rebel Organizations.” Conflict
Management & Peace Science.
Geddes, Barbara, Joseph Wright, and Erica Frantz. 2018. How Dictatorships Work: Power,
Personalization, and Collapse. New York: Cambridge University Press.
Gleditsch, Nils Petter, and Ida Rudolfsen. 2016. “Are Muslim Countries More Prone to
Violence?” Research & Politics 3, no. 2: 1–​9.
Gupta, Sanjeev, Benedict Clements, Rina Bhattacharya, and Shamit Chakravarti. 2004. “Fiscal
Consequences of Armed Conflict and Terrorism in Low and Middle-​Income Countries.”
European Journal of Political Economy 20, no. 2: 403–​421.
Harkness, Kristen A. 2018. When Soldiers Rebel: Ethnic Armies and Political Instability in Africa.
New York: Cambridge University Press.
Hazen, Timothy A. 2019. “Explaining Middle Eastern and North African (MENA) Military
Responses during the 2011–​2012 Arab Uprisings.” Democracy and Security 15, no. 4: 361–​385.
Huntington, Samuel P. 1957. The Soldier and the State: The Theory and Politics of Civil-​Military
Relations. Cambridge, MA: Harvard University Press.
Huntington, Samuel P. 1991. The Third Wave: Democratization in the Late Twentieth Century.
Norman: University of Oklahoma Press.
Huntington, Samuel P. 1995. “Reforming Civil-​Military Relations.” Journal of Democracy 6, no.
4: 9–​17.
ICJ (International Commission of Jurists). 2018. “Egypt: A Return to a Permanent State of
Emergency?” Briefing Paper. Geneva: ICJ. https://​www.icj.org/​wp-​content/​uploads/​2018/​
09/​Egypt-​Return-​to-​State-​of-​Emergency-​Advocacy-​Analysis-​brief-​2018-​ENG.pdf.
Jebnoun, Noureddine. 2014. “In the Shadow of Power: Civil–​Military Relations and the
Tunisian Popular Uprising.” Journal of North African Studies 19, no. 3: 296–​316.
Kandil, Hazem. 2012. Soldiers, Spies, and Statesmen: Egypt’s Road to Revolt. London: Verso.
Lee, Terence. 2009. “The Armed Forces and Transitions from Authoritarian Rule Explaining
the Role of the Military in 1986 Philippines and 1998 Indonesia.” Comparative Political
Studies 42, no. 5: 640–​669.
Lee, Terence. 2015. Defect or Defend: Military Responses to Popular Protests in Authoritarian
Asia. Baltimore: Johns Hopkins University Press.
Lotito, Nicholas J. 2018a. “Public Trust in Arab Armies.” Sada. Washington, DC: Carnegie
Middle East Center. https://​carnegieendowment.org/​sada/​77610.
Lotito, Nicholas J. 2018b. “Soldiers and Societies in Revolt: Military Doctrine in the Arab
Spring.” PhD diss., Columbia University.
Lotito, Nicholas J. 2019. “Trade-​ Offs and Public Support for Security Reform during
Democratic Transitions.” Middle East Law and Governance 11, no. 2: 180–​202.
Lotito, Nicholas J. 2020. “Military Organizational Culture and Protest Response.”
Working Paper.
Lotito, Nicholas J., and Renanah Miles Joyce. 2020. “Public Trust in Authoritarian and
Democratizing Armies.” Working Paper.
Lutterbeck, Derek. 2013. “Arab Uprisings, Armed Forces, and Civil-​Military Relations.” Armed
Forces & Society 39, no. 1: 28–​52.
 Military Politics in Muslim Societies    209

Makara, Michael. 2013. “Coup-​Proofing, Military Defection, and the Arab Spring.” Democracy
and Security 9, no. 4: 334–​359.
Marinov, Nikolay, and Hein Goemans. 2014. “Coups and Democracy.” British Journal of
Political Science 44, no. 4: 799–​825.
McDonald, Matt. 2008. “Securitization and the Construction of Security.” European Journal of
International Relations 14, no. 4: 563–​587.
Mietzner, Marcus. 2013. “Veto Player No More? The Declining Political Influence of the
Military in Postauthoritarian Indonesia.” In Democracy and Islam in Indonesia, edited by
Mirjam Künkler and Alfred C. Stepan, 89–​108. New York: Columbia University Press.
Nepstad, Sharon Erickson. 2013. “Mutiny and Nonviolence in the Arab Spring: Exploring
Military Defections and Loyalty in Egypt, Bahrain, and Syria.” Journal of Peace Research 50,
no. 3: 337–​349.
Neumayer, Eric. 2004. “The Impact of Political Violence on Tourism: Dynamic Cross-​National
Estimation.” Journal of Conflict Resolution 48, no. 2: 259–​281.
Nitsch, Volker, and Dieter Schumacher. 2004. “Terrorism and International Trade: An
Empirical Investigation.” European Journal of Political Economy 20, no. 2: 423–​433.
Picard, Elizabeth. 2005. “The Role of the Military.” In Modernization, Democracy, and Islam,
edited by Shireen Hunter and Huma Malik, 117–​132. Westport, CT: Praeger.
Pollack, Kenneth M. 2002. Arabs at War: Military Effectiveness, 1948–​1991. Studies in War,
Society, and the Military. Lincoln: University of Nebraska Press.
Powell, Jonathan M., and Clayton L. Thyne. 2011. “Global Instances of Coups from 1950 to
2010: A New Dataset.” Journal of Peace Research 48, no. 2: 249–​259.
Resnick, Danielle. 2018. “Foreign Aid and Democratization in Developing Countries.” In The
Oxford Handbook of the Politics of Development, edited by Carol Lancaster and Nicolas Van
de Walle. New York: Oxford University Press.
Savage, Jesse Dillon, and Jonathan D. Caverley. 2017. “When Human Capital Threatens the
Capitol: Foreign Aid in the Form of Military Training and Coups.” Journal of Peace Research
54, no. 4: 542–​557.
Shah, Aqil. 2014. The Army and Democracy: Military Politics in Pakistan. Cambridge,
MA: Harvard University Press.
Sirrs, Owen L. 2010. A History of the Egyptian Intelligence Service: A History of the Mukhabarat,
1910–​2009. New York: Routledge.
Skocpol, Theda. 1979. States and Social Revolutions: A Comparative Analysis of France, Russia
and China. New York: Cambridge University Press.
Soeters, Joseph, and Audrey Van Ouytsel. 2014. “The Challenge of Diffusing Military
Professionalism in Africa.” Armed Forces & Society 40, no. 2: 252–​268.
Staniland, Paul. 2008. “Explaining Civil-​Military Relations in Complex Political Environments:
India and Pakistan in Comparative Perspective.” Security Studies 17, no. 2: 322–​362.
Stanton, Jessica A. 2013. “Terrorism in the Context of Civil War.” Journal of Politics 75, no.
4: 1009–​1022.
Stepan, Alfred C. 1988. Rethinking Military Politics: Brazil and the Southern Cone. Princeton,
NJ: Princeton University Press.
Taylor, William C. 2014. Military Responses to the Arab Uprisings and the Future of Civil-​
Military Relations in the Middle East: Analysis from Egypt, Tunisia, Libya, and Syria.
New York: Palgrave Macmillan.
Tilly, Charles. 1993. European Revolutions, 1492–​1992. Oxford, UK: Blackwell.
210   Nicholas J. Lotito

Toronto, Nathan W. 2017. “Why Professionalize? Economic Modernization and Military

Professionalism.” Foreign Policy Analysis 13, no. 4: 854–​875.
Trinkunas, Harold A. 2006. Crafting Civilian Control of the Military in Venezuela: A
Comparative Perspective. Chapel Hill: University of North Carolina Press.
Wilkinson, Steven. 2015. Army and Nation: The Military and Indian Democracy since
Independence. Cambridge, MA: Harvard University Press.
Wilson, Matthew C., and James A. Piazza. 2013. “Autocracies and Terrorism: Conditioning
Effects of Authoritarian Regime Type on Terrorist Attacks.” American Journal of Political
Science 57, no. 4: 941–​955.
Wright, Joseph. 2009. “How Foreign Aid Can Foster Democratization in Authoritarian
Regimes.” American Journal of Political Science 53, no. 3: 552–​571.
 Pa rt I I

E L E C TOR A L
P OL I T IC S , PA RT I E S ,
A N D E L E C T ION S
 Chapter 10

Voting for I sl a mi sts

Mapping the Role of Religion

Ellen Lust, Kristen Kao, and Gibran Okar

Islamist parties performed well in the first elections after the Arab uprisings. In Egypt,
Islamists swept the parliamentary election, with the Muslim Brotherhood’s Freedom
and Justice Party (FJP) winning nearly 43 percent of the seats and the Salafist Nour Party
taking another 22 percent of them. In Tunisia, the Islamist Ennahda party won a plu-
rality (41 percent) of the seats in the Constituent Assembly, although Islamists did not
enjoy a majority. In Libya, Islamists took fewer than 20 percent of seats contested on
lists, although independent Islamist candidates performed well.
The transitional period thus raises questions and provides an important opportunity
to study the relationship between Islam and voting behavior in the Arab world. What
explains Islamists’ success? Is it related to factors unique to Islam, such as voters’ reli-
gious identity, practice, or preferences regarding the role of religion in the state? Or do
the material advantages, organizational strengths, or economic policies of these parties
give them the upper hand? Moreover, are the same factors responsible for support of po-
litical Islam across all countries, or do they vary given specific historical experiences and
social contexts?
In this chapter, we analyze support for Islamist parties in Egypt, Libya, and Tunisia
during political transitions.1 Overall, we find that religious factors are more consistently
important than economic or organizational factors. However, there are variations in
the particular religious drivers of support for Islamist parties across the three countries.
Preference for religious parties was strongly associated with Islamist support in Egypt
and Libya, but not in Tunisia; agreement with religious leaders’ influence in government
was associated with support for Islamists in Tunisia and Libya but remained insignifi-
cant in Egypt. Frequent religious practice was correlated with Islamist support only in
Tunisia, and identifying first and foremost as a Muslim did not seem to factor into voter
calculus in any of the three countries. Rarely do existing studies on voting for Islamists
214    Ellen Lust, Kristen Kao, and Gibran Okar

examine multiple measures of religious factors in this way. Moreover, despite the atten-
tion they are given in extant studies, economic and organizational factors appeared less
important overall during these transitional elections.
We argue that different historical legacies and social conditions that shape polit-
ical environments may explain variation in factors associated with Islamist support.
Authoritarian regimes’ ruling strategies left parties with different levels of organiza-
tional strength. The Egyptian regime allowed the Muslim Brotherhood to maintain a
political presence throughout the latter half of the twentieth century, while the Tunisian
and Libyan regimes clamped down on Islamists’ public activities. Consequently,
Egyptians developed a strong taste for religious parties over secular ones. In Libya, how-
ever, although Islamist movements were repressed, the Qaddafi regime promoted reli-
gious and social conservatism. In contrast, Habib Bourguiba implemented laic policies
to foster a secular society in Tunisia. Consequently, after the Arab uprisings, Tunisians
were much less likely to identify primarily as Muslims or to engage in religious practices
than were Egyptians and Libyans. Those who did practice, however, were more likely to
vote for Islamist parties. Although these historical legacies left significant differences in
the prevalence of religious identity and worship across these countries, all three saw the
politicization of Islam and the growth of an ideological cleavage within the state. This
was likely due to the heightened rhetoric around political Islam, both domestically and
on the international stage.
The chapter proceeds as follows. It first outlines the literature, considering how
debates over the success of Islamist parties have tended to focus on material and organ-
izational factors, downplaying the importance of religion. We then turn to the cases at
hand, describing the political and social landscape at the time of transitions. The third
section interrogates the relationship between organizational, material, and religious
factors and support for Islamist parties. We then present evidence that religious factors
were most highly associated with voters’ choices, although different dimensions of Islam
were salient in different contexts. We conclude by considering why religious factors
are more salient than material and economic ones. Throughout, we draw on surveys
conducted in Egypt, Libya, and Tunisia from 2011 to 2013,2 as well as the extant literature
on political Islam.

Explanations of Islamist Support

in the Arab World

Explanations for why voters support Islamist parties have largely overlooked the im-
portance of religion. Some scholars argue that organizational capacity and material
incentives are the key drivers of Islamist success—​a view that assumes Islamist parties
are not distinctive. A few exceptions exist, yet even these studies do not fully explore
 Voting for Islamists    215

whether party success is related to voters’ support for a role of religion in politics, or
individuals’ religious identity and practice.
One set of arguments explaining Islamists’ electoral advantage focuses on their or-
ganizational capacity. Established networks of worshipers and social programs proffer
Islamists an advantage when they proceed to the political stage. In the context of re-
pressive regimes, such as those in Iran and Egypt under Mubarak, the cellular nature of
religious institutions allowed Islamists to maintain activities (Wiktorowicz 2004). The
secretive and tight-​knit nature of these organizations endowed Islamists with a deter-
mined cadre of supporters following the Arab uprisings (Kandil 2015; Masoud 2014).
Religious and charity organizations also created a platform for community networking,
helping to mobilize voters, as in Turkey under the AKP (White 2012).
A second strand of the literature focuses more extensively on the service provision
that Islamist organizations provide. Many scholars document, for instance, extensive
networks of mosques and Islamic charities that create a space in which these parties can
spread their message and distribute material rewards, thereby fostering support (e.g.,
Blaydes 2011; Brooke 2019; Clark 2004; Wickham 2002; Wiktorowicz 2004; ). Similar
spaces often do not exist for non-​Islamist parties, giving Islamists a comparative ad-
vantage in elections relative to other groups (e.g., Cammett and Luong 2014; Tessler
2011). Islamists are thus viewed as more capable than other parties in meeting their
constituents’ demands, particularly among lower (Blaydes 2011) and middle classes
(Clark 2004).
This demonstrated ability to administer services, and the visibility that Islamist
movements gain by doing so, also translates into a reputational advantage: it lends cred-
ibility to their economic and social claims and garners them support in transitional
elections (Cammett and Luong 2014; Roháč 2012a). Steven Brooke (2017), for example,
argues that Egyptians’ experience with the Muslim Brotherhood’s service provision led
them to expect the Brothers would provide similar benefits if given the chance to govern.
Cammett and Luong (2014) argue that service provision promoted a reputational ad-
vantage for good governance and thus was a key driver of their widespread electoral suc-
cess. This explains why support for Islamists is not limited regionally (where services are
provided) or along sectarian lines (where ideological values are shared).
Other literature shifts attention from service delivery to policies. Some see voters’
preference for economic redistribution as fueling support for Islamists. Historically, re-
ligious movements have both shaped and co-​opted leftist, egalitarian economic and so-
cial platforms (Davis and Robinson 2012; Kahl 2005). Scholars therefore posit that voters
link Islamist parties with redistribution. Masoud (2014, 123–​145), for instance, argues
that the “secular-​Islamist divide” in Egypt is only a proxy for left-​right divisions over
economic futures. Others suggest similar dynamics as well: in Turkey, those negatively
affected by economic downturn were far more likely to vote for the AKP (Baslevent et al.
2005; Öniş 2004). In Tunisia, voting for Islamists was tied closely to middle-​income
households (Fourati et al. 2016). A second perspective puts foreign policy at the center
of Islamist voting. Supporters of the AKP, for instance, were far more likely to disagree
216    Ellen Lust, Kristen Kao, and Gibran Okar

with matriculation into the EU and had overall negative views of the West (Baslevent
et al. 2005).
In such views, support for Islamist parties has little to do with religion. Incumbent
regimes and secular parties can also offer material rewards to garner support, ben-
efit from networks and organizations, or enjoy reputational advantages. Religion
thus plays little role in a vast swath of the literature that explains support for Islamist
parties. Perhaps this is because scholars of the Arab world are sensitive to arguments
that exceptionalize the region or essentialize Islam (e.g., Huntington 1996; Pipes 2002).
Ironically, in doing so they only further the special treatment of the Muslim world;
studies of Christian Democrats in Europe and evangelical voters in the United States
(e.g., de Koster and van der Waal 2007; Raymond 2011), Pentecostals in sub-​Saharan
Africa (McClendon and Riedl 2019), and even Muslims in other parts of the world
(Hasan 2012) recognize the influence of religious values, identities, and practices on po-
litical behavior.
There are a few exceptions, of course. Wickham’s (2002) study of support for the
Muslim Brotherhood under Mubarak’s Egypt finds that the Brothers were able to attract
out-​group members through the “transvaluation of values.” The Brothers’ flight away
from the Mubarak regime was brought to represent a moral good through religious, in-
terpretive frameworks, encompassing the suffering of many disenfranchised and angry
Egyptian citizens.3 Before the uprisings, some scholars argued that Islamist parties
gained the lion’s share of their support due to their promises to incorporate more religion
into politics (Garcia-​Rivero and Kotze 2007), rather than their economic policies, which
they generally regarded as ill-​defined and weak (Robbins 2009, 27). Similarly, looking
at data collected during the transitional period (2011–​2014) by the Arab Barometer,
the World Values Survey, and the Afrobarometer, Wegner and Cavatorta (2019) argue
that a main factor of support was the ideological divide along the Islamist-​secular fault.
Empirical evidence to that effect was also seen in the Tunisian 2011 elections, wherein
the role of religion in public life was a statistically significant driver of voting behavior
(Dennison and Draege 2020). Grewal et al. (2019) argue that believers’ interests in “di-
vine rewards” shape voting patterns; however, their argument rests on the presumption
that only the poor would seek redemption by supporting Islamist candidates.
Even these studies, however, have failed to distinguish the various facets of religion
that can affect the ballot box. Religious support for Islamist parties can be understood
in terms of identity, practice, and politics. It can be associated with a voter’s identity as a
Muslim, much in the same way that ethnic identities or sectarianism can affect electoral
support (e.g. Cammett 2014; Cammett and Issar 2010; Chandra 2004; Posner 2005). It
can be associated with practice, the reflection of piety, and engagement in networks of
other believers. And it can be related to their political beliefs regarding the role of reli-
gion in the state. Understanding the relationship between identity, practice, and polit-
ical positions and voting for Islamist parties can help to shed light on the relationship
between Islam and elections.
We explore the relationship between support for Islamists, religion, and other factors
in the first elections following the Arab uprisings. These elections provided a unique
 Voting for Islamists    217

opportunity to study voter calculus. Voters’ choices were more clearly revealed in the
transitional elections than they were in previous polls; before the uprisings, voters
feared harassment for supporting Islamists (e.g. Hafez and Wiktorowicz 2004), elec-
toral institutions were structured to favor the regime’s ruling strategy (Kao 2015; Lust
2009; Lust and Jamal 2002), and Islamist party elites limited their engagement in order
to avoid provoking the ruling regime to crackdown on its members (Hamid 2011).
Moreover, during transitions from authoritarianism to free and fair elections, voter cal-
culus has been seen to shift away from analysis of incumbency, or record, and toward
the potential for reform (Fidrmuc 2000).

Voters and Parties after the Uprisings

There were notable differences in the social and political landscapes of Egypt, Libya, and
Tunisia after the Arab uprisings. In all three societies, majorities of the polity supported
a role of religious leaders in politics, although citizens of both Egypt and Libya were
more likely than Tunisians to engage in religious practices and to identify first and fore-
most as Muslim. Additionally, the societies varied regarding preferences for religious
political parties. The three countries also had different political infrastructures after
the uprisings. For instance, in Egypt, the Freedom and Justice Party benefitted from the
Egyptian Muslim Brotherhood’s organizational capacity, which dwarfed that of Islamist
organizations in both Tunisia and Libya.
Historical experiences shaped the societies in these three countries. Tunisia’s first pres-
ident, Bourguiba, was determined to “modernize” the country, and thus initiated a pro-
ject that banned traditional and religious dress from public workspaces, implemented a
Western workweek (thus working on Friday), and even encouraged Tunisians to refrain
from fasting during Ramadan (Boulby 1988, 568). By 2013, only 20 percent of Tunisians
identified first and foremost as Muslim, and only 35 percent of the population engaged
in formal religious practices. (See Figure 10.1.) In Egypt, the seat of the transnational
Muslim Brotherhood, religious movements inculcated a strong Muslim identity and
fostered religious piety. In 2012, 43 percent of Egyptians identified first and foremost
as Muslim and 41 percent often engaged in religious practices. Moreover, almost two-​
thirds of the sample in Egypt preferred a religious party. Religious identity and practice
were also high in Libya, where Qaddafi had promoted a conservative Islamic organiza-
tion, the Islamic Call Society (Ad-​Dawah). In 2013, 42 percent of Libyans identified first
and foremost as Muslim and 51 percent often engaged in religious practices.
Voters across the three countries also had different perspectives on policies regarding
religion. When respondents were asked to place their own preference regarding party
religiosity between completely religious (1) and completely secular (9), the median
voter in all three countries placed in the middle (4 and 5). (See left panel, Figure 10.2.)
Egyptians, however, were more likely to state a position on this question (just 6 percent
of respondents said they did not know, versus 11 percent in Libya and Tunisia), and the
(a) (b)
Often Engages in
35%
Islamic Practices 41%
57% 58% 51%

Non-Religious
80%
Identity

Rarely Engages in
65%
Islamic Practices 59%
43% 42% 48%

Muslim 20%

Tunisia Egypt Libya Tunisia Egypt Libya

(c) (d)

Strongly agree 44% 47% 45%

65% 36%
Prefers Religious
48%
Party

Agree 27% 23%

27%

10%
35% 64% 10%
Disagree 20%
Prefers Secular 52% 22%
16%
Party Strongly disagree 9%
Tunisia Egypt Libya Tunisia Egypt Libya

Figure 10.1: Religious identity, practice, and policies.

Note: The question regarding identity in the top left panel is: “What label describes you best,
and which is the second best? The choice set is nationality (e.g., Tunisian), Arab, Muslim, re-
gional, local, a world citizen, man/​woman, other” and is coded as “identifies as Muslim” if
“Muslim” is chosen as the first identity. Note that the exact options changed slightly across
different countries. The question for religious practice in the top right panel is: “On a scale
of 1–​9 which number better represents you, if 1 means a person who goes to a house of wor-
ship daily, takes religious lessons, listens to religious programs and reads religious books, and
9 means a person who rarely does these things (goes to a house of worship, takes religious
lessons, listens to religious programs and reads religious books)?” For religious party prefer-
ence in the bottom left panel, the question is: “On a scale of 1 to 9, please state your personal
preferences on the positions below: 1 prefer a religious political party and 9 prefer a strictly
secular party.” The question regarding the role of religious leaders in the bottom right panel
is: “Please tell me whether you strongly agree, agree, disagree, or strongly disagree with the
following statements: . . . Religious leaders should have no influence over the decisions of
the government.” For all 9-​point scales used in this chapter, we biased the results away from
existing hypotheses on correlates of Islamist voting, such that the middle category of 5 was
coded away from signaling religious practice, identity, or party preference. The data for Libya
presented here comes from the November 2013 data set.
Please note that we dropped non-​Muslims from all analyses presented in this chapter, not because
Christians are necessarily less likely to vote for Islamists, but because variables associated with re-
ligion would be expected to function differently for Christians than Muslims, and the theoretical
model explained in this chapter is not built to understand non-​Muslim voting for Islamists. There
were only six non-​Muslims in Tunisia and one in Libya, whereas in Egypt about 5% of the sample
is non-​Muslim.
 (a)

Libya
(80%)

Tunisia
(80%)

Egypt
(94%)

Completely 2 3 4 5 6 7 8 Completely
Religious Secular
Rating

(b)

Libya
(90%)

Tunisia
(80%)

Egypt
(94%)

Completely 2 3 4 5 6 7 8 Completely
Socialist Captalist
Rating

Figure 10.2: Individuals’ self-​placement on religious-​secular and socialist-​capitalist scales.

Note: Questions read as follows: (left panel) “On a scale of 1 to 9, please state your personal
preferences on the positions below: 1 prefer a religious political party and 9 prefer a strictly
secular party”; (right panel) “On a scale of 1 to 9, please state your personal preferences on the
positions below: 1 being a strict socialist system with the state running all economic affairs
and 9 being a full-​fledged free market capitalist system.” Figures in parentheses represent the
proportion of respondents who were able to state their positions. Islamist parties are denoted
with darkened distributions compared to non-​Islamist parties and mean response point
estimates are denoted with black diamonds. Data for Libya comes from the May 2013 data set.
220    Ellen Lust, Kristen Kao, and Gibran Okar

median voter, and the distribution of voters, skewed more toward the completely reli-
gious endpoint. This perhaps reflects the historical strength of the Muslim Brotherhood.
Libyans, too, were more likely to prefer a role of religion in politics than Tunisians.
This is unsurprising given Tunisia’s history of laicism. Despite these differences, more
than two-​thirds of citizens in every society agreed with the statement “religious leaders
should have no influence in governance.” (See bottom right panel, Figure 10.1.)
The populations also had distinct views on the role of the state in the economy. In
Egypt and Tunisia, where religion had been most politicized, respondents were less
likely to be able to place their own position on the economy than on religion: 80 per-
cent of Egyptians stated their position on the economy, versus 94 percent on religion;
80 percent of Tunisians stated an economic position, versus 89 percent on religion. Only
in Libya did positions on the economy and religion seem to be equally clear in voters’
minds. Egyptians, Libyans, and Tunisians also held quite distinct preferences. In Egypt,
which had liberalized its state-​led economy in the years leading up to the revolution,
there was considerable support for a strong role of the state in the economy. The same
was found in Tunisia. In contrast, Libya had greater support for a capitalist economy.
(See right panel, Figure 10.2.)
The countries’ historical experiences also shaped their political landscapes. Prior to
2011, Egypt’s regimes had stoked political tensions between Islamists and secularists
in an effort to weaken opposition (Lust 2011). The Egyptian Muslim Brotherhood had
been prevented from forming a political party and often suffered repression, although
Brotherhood members ran in parliamentary elections, either as independents or in
alliances with legal political parties. The Muslim Brotherhood won seventeen parlia-
mentary seats in the 2000 elections and eighty-​eight seats (or 20 percent of the total)
in 2005.4 The Tunisian Islamist movement fared far worse. Bourguiba’s extreme ver-
sion of laicité had softened in the 1970s, allowing a period of growth for Le Mouvement
de la Tendance Islamique (MTI). The movement gained considerable power, with
independents contesting the 1989 elections and winning an estimated 10–​17 percent of
the seats (Wolf 2017). Surprised, and perhaps terrified, by the Islamists’ success, the Ben
Ali regime clamped down on the movement, jailing approximately 25,000 members
and sending others into exile. The Muslim Brotherhood in Libya, likewise, suffered
from repression. Shortly after Muslim Brothers arrived in the country from Egypt, their
attempts to mobilize were met with violence from the regime. Brothers who refused
to refrain from political activity were publicly executed or sent to serve sentences in
Abu Salim prison. However, following overtures from Saif Qaddafi, the organization
reconciled with the regime. Thus, the Islamist movement was strong in Egypt, organiza-
tionally weaker in Tunisia, and at near standstill in Libya.
The countries also entered the transition with different endowments of secular
parties. Both Egypt and Tunisia had allowed opposition political parties to participate
in elections under authoritarianism, and thus had several non-​Islamist parties pre-
pared to engage in the transitional elections. Libya, in contrast, had eschewed polit-
ical parties altogether, going so far as to dictate that “Min tuhizb khan” (he who joins
 Voting for Islamists    221

a party is a traitor), sentencing to death those who dared to form political parties.
Party organizations were thus limited in Libya, and indeed a fear of party mobiliza-
tion remained strong as the first post-​Qaddafi elections approached. Consequently,
the constellation of parties varied substantially across the three countries in the first
transitional elections.5 Egypt and Tunisia had a fragmented party system, with a mix of
Islamist and non-​Islamist parties, both new and old. Libya had an even more splintered
system, composed primarily of small tribal and other social groupings running on
unified lists.
Authoritarian strategies thus left party leaders with different strengths and weak­
nesses at the outset of the transitions. The majority of Egyptian Muslim Brothers who
would play significant roles in the transition had been politically active inside Egypt
in the years preceding Mubarak’s downfall. The most prominent leaders of Tunisia’s
Ennahda party, in contrast, only began mobilizing for campaigns after they returned
from exile. In striking comparison, Mohamed Morsi, a Muslim Brother who would be-
come Egypt’s first democratically elected president, had taught at Zagazig University
and stood as a candidate for the 2000 elections, while Rached Ghannouchi, the leader
of Ennahda, had been twice imprisoned and in exile for twenty-​two years (Ashour 2016;
Benotman et al. 2013, 196–​197; Chemitsky 2012; Fitzgerald 2015, 178; Fredrich Ebert
Stiftung 2015, 5). Yet it is possible that Islamist leaders, particularly in Tunisia, could
have benefited from their repression. Kim Guiler (2020) maintains that “suffering for
one’s political cause” glorifies candidates juxtaposed to incumbent or nonvictimized
groups, finding evidence of this in Turkey and Tunisia. However, repression also signifi-
cantly weakened ties between leaders and constituents.
Likewise, repression may have undercut Tunisian and Libyan Islamists’ organi-
zational strength, particularly compared to the Egyptian Muslim Brotherhood. The
Egyptian Brothers played a central role in service provision after labor unions declined
following economic liberalization in the early 1980s. In Tunisia, the secular-​leaning
trade union, the UGTT, played a more important role in mobilization, balancing the
strength of the Islamist movement. Moreover, the unions took a leading role in the
revolution and maintained support throughout the transitional electoral campaigns
(Brownlee, Masoud, and Reynolds 2015; Yousfi 2018).6 Libya stood in stark contrast to
Tunisia. All parties lacked organizational strength, owing to years of harsh control over
associational life, so discerning any Islamist upper hand in mobilization would prove
difficult.
Finally, as the first elections following the Arab uprisings approached, voters and
parties in Egypt, Libya, and Tunisia navigated different electoral systems. These systems
partly reflected the nature of political parties and the distribution of power that existed
at the time of the uprisings; they also shaped the options available to parties and voters.
Egypt had a mixed electoral system for the 2011–​2012 parliamentary elections: Two-​
thirds (332 seats) of the country’s 498 parliamentary seats were elected on party lists, and
one-​third (166 seats) to individual seats in two-​seat districts (Eskandar 2011). Libya also
had a mixed electoral system, with 40 seats chosen in first-​past-​the-​post, single member
222    Ellen Lust, Kristen Kao, and Gibran Okar

districts, 80 seats for multi-​member districts with single nontransferable vote, and 80
through closed lists. The distribution of list seats was complicated by a gender quota that
conflicted with social norms regarding the role of women and ethnic divisions; conse-
quently, the majority of party list seats were allocated to the larger urban areas (Smith
2012). In Tunisia, all seats for the 2011 Constituent Assembly were elected on party lists
(Decree No. 35, 2011).

Islamist Parties: Organizational,

Material, and Religious Advantages?

The different contexts surrounding elections in Egypt, Libya, and Tunisia after the Arab
uprisings give us the opportunity to interrogate the association of organizational, ma-
terial, and religious factors with support for Islamists. We do so by drawing upon de-
scriptive evidence regarding how individuals view political parties, as well as regression
analyses that examine the association between indicators of material, organizational,
and religious positions on voter choice. We find evidence that religious factors were im-
portant across all three countries, although the specific facets of religion associated with
a higher propensity to vote for Islamist parties varied between cases.

Organizational Capacity
To examine the influence of party organization on Islamists’ success, we rely on voters’
perceptions. These measures are unable to determine the extent to which objective
organizational strength influences mobilization, but they do reflect citizens’ views
of organizational capacity. We find evidence that Egyptians widely viewed Islamist
parties—​Al Nour and the FJP—​as more capable than their secular counterparts, as
denoted by the mean scores (black diamonds) in Figure 10.3. Likewise, voters were
much more likely to have an opinion on the organizational capacities of Al Nour and
the FJP—​34 percent and 40 percent of respondents rated them, respectively—​compared
to the next closest party, Al Wafd, which only 26 percent of voters felt confident enough
to rate. In Tunisia, the Islamist Ennahda was comfortably in the middle of the spectrum
with regard to party organization. Tunisian voters appeared to have more information
on parties, on the whole, with more than 70 percent of them rating parties’ organiza-
tional capacity. However, the secular Congress for the Republic (CPR) party still elicited
a higher response than Ennahda (83 percent vs. 73 percent of respondents, respectively)
and was rated more organizationally capable. The Islamist claim to organizational excel-
lence fails to explain the success of Islamists alone.
We also use regression analyses to test the extent to which voters who value organi-
zational capacity are more likely to support Islamist parties. The dependent variable is
 Voting for Islamists    223

(a) Egyptian
Social
Democratic
(12%)
Al Wasat
(15%)

Free
Political Party

Egyptians Islamist Party

(16%)
Yes
Al Wafd No
(26%)

Al Nour
(34%)
Freedom
and
Justice
(40%)
Not at All A little Somewhat To a Large
Extent
Rating

(b)Progressive
Democratic
Party
(70%)
Al-Arida
(76%)
Political Party

Congress
for the
Republic
(83%)
Ennahda
(73%)

Ettakatol
(73%)

Not at All A little Somewhat To a Large

Rating Extent

Figure 10.3: Organizational capacity (Egypt and Tunisia).

Note: Egypt is presented in the left panel; Tunisia is in the right panel. Question wording: “For
each of the following parties, would you say that they are organizationally capable to a large
extent, somewhat, a little, or not at all?” Figures in parentheses represent the proportion of
respondents who were able to state their positions. Islamist parties are denoted with darkened
distributions compared to non-​Islamist parties, and mean response point estimates are denoted
with black diamonds. We did not ask this question in Libya.

having voted for an Islamist party list7 in past elections.8 We also include standard dem-
ographic variables expected to affect voting behavior (gender, age, education, and em-
ployment). We consider whether those for whom a party’s organizational capacity was
more important, in general, were more or less likely to support Islamist parties. We find
that voters for whom organizational capacity was important were 11 percentage points
(p < 0.01) more likely to vote for Ennahda. Organizational factors were not significant
in Libya, and the substantive estimate was near zero. (See Table 10.1.) Unfortunately,
 (a) Egyptian
Social
Democratic
(20%)

Al Wasat
(25%)

Free
Political Party

Egyptians
(29%)

Al Wafd
(46%)

Al Nour
(55%)

Freedom
and
Justice
(63%)
Completely 2 3 4 5 6 7 8 Completely
Socialist Capitalist
(b) Rating

Progressive
Democratic
Party
(51%)

Al-Arida
(52%)
Political Party

Congress
for the
Republic
(47%)

Ennahda
(60%)

Ettakatol
(54%)

Completely 2 3 4 5 6 7 8 Completely
Socialist Capitalist
Rating
(c)
Taghyer
(35%)

National
Centrist
(41%)
Homeland
(34%)
Political Party

Union for
Homeland
(37%)
National
Front
(53%)
Justice and
Construction
(67%)
National
Forces
Alliance
(80%)
Completely 2 3 4 5 6 7 8 Completely
Socialist Capitalist
Rating

Figure 10.4: Economic policies.

 Voting for Islamists    225

we lack the data to test this relationship between valuing organizational capacity and
Islamist support in Egypt.

Economic Platform and Material Benefits

There is little evidence across our cases that economic platforms or material benefits
were the primary drivers of support for Islamists. Indeed, many citizens had diffi-
culty placing parties’ economic positions. As shown in Figure 10.4, the proportions of
respondents who felt informed enough to rate parties on economic scales are notably
low in all three countries. With the exception of the National Forces Alliance (NFA) in
Libya, which was clearly perceived to be a neoliberal party, less than two-​thirds of the
sampled populations were able to place parties (including, importantly, Islamist parties)
on this economic scale. For many parties, this proportion was much smaller. Citizens
also only perceived small differences between party platforms on economic policies;
perceptions of economic policies are clustered around the center, from 4–​6, with only
the Libyan NFA, as a major exception, ranking at 7.
Moreover, there is no consistent evidence that citizens are better able to place the ec-
onomic positions of Islamist than non-​Islamist parties. In Egypt, a higher percentage
of respondents were confident enough to state economic positions of the Islamist FJP
and Nour parties (with 63 percent and 55 percent, respectively, placing the party) than
the next most-​recognizable party, the Wafd (46 percent). In Tunisia, the percentage of
respondents who felt they could place Ennahda, the main Islamist party, was higher
than the percentage who could place the non-​Islamist challenger Ettakatol, but not by
much (approximately 60 percent versus 54 percent). Similarly, in Libya, 13 percent more
of the sample could place the non-​Islamist NFA than the Islamist JCP (80 percent vs.
67 percent). That is, while the economic positions of Islamist parties may be more recog-
nizable than those of non-​Islamist parties in some countries, this is not universally true.
Using perceptions of the extent to which parties assist citizens in need as a proxy for
the potential material benefits parties might be expected to confer on supporters, we also
find no evidence of an Islamist advantage. Descriptive evidence from Egypt and Tunisia
demonstrates first that, overall, citizens do not believe that any of the major parties help
needy citizens to a large extent, and second, that respondents in Egypt were far less able
to position parties’ assistance to citizens (19 percent to 43 percent) than those in Tunisia

Note: Egypt is in the top left panel; Tunisia is in the top right panel; Libya is in the lower center
panel. Question wording: “On a scale of 1 to 9, 1 being a strict socialist system with the state run-
ning all economic affairs and 9 being a full-​fledged free market capitalist system, how would
you rate the following parties?” Figures in parentheses represent the proportion of respondents
who were able to state their positions. Islamist parties are denoted with darkened distributions
compared to non-​Islamist parties and mean response point estimates are denoted with black
diamonds. Note that the Libyan dataset for this figure comes from the May 2013 data set.
226    Ellen Lust, Kristen Kao, and Gibran Okar

(72 percent to 83 percent). On average, in Egypt, voters perceived the major Islamist
parties as helping citizens in need marginally more than their non-​Islamist counterparts.
Yet note that Islamists were seen as helping “a little” compared to the most prevalent
answer of “not at all” for non-​Islamists. (See Figure 10.5, left panel.) In Tunisia, despite
higher voter clarity on party actions, Ennahda is perceived to be at the bottom of this
spectrum, doing less on average for citizens in need than any other party. (See Figure
10.5, right panel.) Considering the Egyptian Muslim Brotherhood’s history of service

(a) Egyptian
Social
Democratic
(19%)
AlWasat
(21%)

Free
Political Party

Egyptians
(22%)

Al Wafd
(29%)

Al Nour
(38%)

Freedom
and
Justice
(43%)
Not at All A Little Somewhat To a Large
Extent
Rating

(b) Progressive
Democratic
Party
(72%)

Al-Arida
(83%)

Congress
Political Party

for the
Republic
(74%)

Ennahda
(74%)

Ettakatol
(76%)

Not at All A Little Somewhat To a Large

Extent
Rating

Figure 10.5: Helping citizens in need (Egypt and Tunisia).

Note: Egypt is in the left panel; Tunisia is in the right panel. Question wording: “For each of the
following parties, to what extent would you say that they help citizens in need? To a large extent,
somewhat, a little, not at all?” Figures in parentheses represent the percentage of respondents who
responded to the question. Islamist parties are denoted with darkened distributions compared to
non-​Islamist parties and mean response point estimates are denoted with black diamonds. Note
that this question was not asked in Libya.
 Voting for Islamists    227

provision, it is not surprising to find that Egypt’s Islamist parties held the high ground in
perceived service to the needy. The failure of Ennahda to capture the same accolades in
Tunisia indicates that economic and material arguments for Islamist support are more
likely reserved to specific social and historical contexts—​and certainly, in the transition
period, they had less universal weight on voter calculus than previously understood.
Our argument in this section is made clearer by regression analyses, which find little ev-
idence that economic preferences and material benefits drive Islamist support, or at least
not uniformly across the region. (See Table 10.1.) Across the three countries, there are no
significant associations between individuals’ economic preferences and reported vote for
Islamist parties. In Libya and Tunisia, we were also able to examine the role of service pro-
vision on voter choice. We asked respondents whether a party’s service provision to their
constituents was an important factor in vote choice. Our data show that Libyans who value
service provision from a party were 8 percentage points (p < 0.10) more likely to support
Islamists. (See Table 10.1.) There was no significant relationship in Tunisia. In Egypt, we
did not have the data to investigate service provision on electoral patterns.

Religious Factors: Identity, Practice, and Policy

Religious factors were significantly associated with support for Islamists. (See note on
Figure 1 for exact question wording.) We interrogate the role of identity, practice, and
policy. We find that identifying oneself first and foremost as a Muslim was not signif-
icantly associated with voting patterns in any of the countries. Religious practice was
significantly associated with support for Ennahda in Tunisia, but was not significantly
correlated with Islamist support in Egypt or Libya. However, religious policies were sig-
nificantly associated with voting behavior across the three countries.

• Identity: We examined the association between identifying as Muslim, first and

foremost, and Islamist support in our regression analyses. Identity was never sig-
nificant, and furthermore, the coefficient was near zero in Egypt and Tunisia. This
suggests that Islamist support is not driven by identity politics.
• Practice: We investigated the extent to which religious practice was associated with
voting patterns by first identifying individuals in our data who self-​identified as
frequent practitioners of Islam. In Tunisia, those who self-​identified as practicing
Muslims were 9 percentage points (p < 0.05) more likely to have voted for Islamist
party lists (Benstead et al. 2012). However, this association was both statistically
and substantively insignificant in Libya and Egypt.
• Policy: We interrogate the role of religious policy in three ways. Descriptively, we
analyze citizens’ ability to locate party positions regarding the role of religion in the
state. Additionally, we incorporate two measures of religious political preferences
in our model of voting. We regress individuals’ positions on their attitudes toward
the influence of religious leaders in government, as well as their preferences for a
religious party. We find strong evidence that citizens’ preferences for religion in
politics influence voting behavior.
Table 10.1: Individual Level Drivers of Support for Islamist Parties Compared
Tunisia Egypt Libya

Prefer Religious Party 0.0153 0.175* 0.0983+

(0.0295) (0.0175) (0.0520)
Religious Governance 0.0735* 0.0189 0.108+
(0.0352) (0.0162) (0.0574)
Religious Practice 0.0860* -​0.00844 -​0.0487
(0.0324) (0.0153) (0.0472)
Muslim ID 0.0317 0.0189 0.0518
(0.0386) (0.0147) (0.0518)
Organized Party 0.111* —​ -​0.00336
(0.0435) —​ (0.0439)
Service Party -​0.0461 —​ 0.0762+
(0.0356) —​ (0.0393)
Socialist Party 0.0163 -​0.00999 0.0449
(0.0283) (0.0155) (0.0575)
Age 0.0276+ -​0.0138* 0.00318
(0.0149) (0.00634) (0.0245)
Male 0.103* -​0.0508* 0.0675
(0.0290) (0.0198) (0.0431)
Employed 0.0562* 0.0281 -​0.0861+
(0.0286) (0.0200) (0.0504)
Higher Education -​0.0721* -​0.0811* 0.0426
(0.0312) (0.0156) (0.0435)
Constant -​0.0429 0.846* -​0.00617
(0.0582) (0.0275) (0.0694)
N 782 2021 353

Note: Standard errors in parentheses

+ p<0.10
* p<0.05
** p<0.01
*** p<0.001
Note: All variables are dichotomized with the exception of age, which is broken into age brackets of 18–​30,
and ten-​year increments thereafter. Higher education means the respondent has a secondary education or
above. We did not have variables in the Egypt data set for party organization or party focus on serving its
constituents. The N in Libya is low because candidates did not compete on lists in all districts.
Ordinary least squares regression models are displayed here, but all results are robust to logistical
models, with just a few caveats in Libya where all marginally significant outcomes are fully significant
according to traditional levels and being male becomes a significant variable in voting Islamist. The
data for Libya in this regression from the December 2013 data set.
 Voting for Islamists    229

Most citizens were able to rate party platforms with regard to the role of Islam in
the state. Moreover, party placement on this religious-​secular dimension was more
polarized than that of economic policy. Parties varied, on average, 2 to 3 increments
on a 1–​9 socialist-​capitalist scale; however, variation in perceived religious platforms
spanned 6 increments on a 1–​9 secular-​Islamist scale. (See Figures 10.4 and 10.5.) Voters
were also more confident in placing parties along this dimension; more than three-​
quarters of respondents in each country were able to position the top parties along the
religious-​secular dimension.
We also found evidence that citizens viewed Islamist parties’ platforms more clearly
than their secular counterparts, particularly with regard to the role of religion in the
state. As shown in Figure 10.6, 86 percent of Egyptians felt they were able to place the
Islamist FJP on the religious-​secular scale, and 80 percent were able to place the Nour
party. In contrast, only 60 percent were able to place the Wafd party, the most fre-
quently placed non-​Islamist party. Similarly, in Tunisia, 83 percent placed the Islamist
Ennahda party, compared to 71 percent placing Ettakatol. However, in Libya, 87 percent
of respondents placed the secular NFA, compared to 80 percent who placed the Islamist
JCP. Citizens’ greater capacity to place parties along the religious-​secular divide than the
socialist-​capitalist one, as well as the greater variance in parties on this religious dimen-
sion, counters claims that the secular-​Islamist divide served as a proxy for economic
divisions at the ballot box.
Regression analyses also find support for the argument that religious policy
matters. (See Table 10.1 below.) Different indicators, however, vary in importance
across countries. We find that voter preference for a religious party was significantly
and positively associated with voting for Islamist parties in Egypt and Libya, al-
though not in Tunisia. In Egypt, voters who stated that they preferred a religious
party were 17.5 percentage points (p < 0.05) more likely to report having voted for
an Islamist party list; in Libya, they were nearly 9 percentage points (p < 0.10) more
likely to do so. Our second measure, attitudes toward the role of religious leaders
in government, was significantly and positively associated with Islamist support in
Tunisia and Libya. Tunisians were 7 percentage points (p < 0.05) more likely to vote
for an Islamist party list if they preferred a strong role of religion in the state; Libyans
were nearly 11 percentage points (p < 0.10) more likely to do so. Recall that there was
no significant correlation between voters’ economic policy preferences and support
for Islamists.9

Discussion and Conclusion

In this paper, we introduced a nuanced understanding of the relationship be-

tween religion and electoral behavior, one that more closely resembles electoral
studies in the United States and Europe. Examining support for Islamists in the first
 Egyptian
Social
Democratic
(25%)
Al Wasat
(25%)

Free
Political Party

Egyptians
(39%)

Al Wafd
(60%)

Al Nour
(80%)

Freedom
and
Justice
(86%)
Completely 2 3 4 5 6 7 8 Completely
Religious Secular
Rating

Progressive
Democratic
Party
(64%)

Al-Arida
(65%)
Political Party

Congress
for the
Republic
(60%)

Ennahda
(83%)

Ettakatol
(71%)

Completely 2 3 4 5 6 7 8 Completely
Religious Secular
Rating

Taghyer
(37%)

National
Centrist
(44%)

Homeland
(41%)
Political Party

Union for
Homeland
(41%)
National
Front
(61%)
Justice
and
Construction
(80%)
National
Forces
Alliance
(87%)

Completely 2 3 4 5 6 7 8 Completely
Religious Secular
Rating

Figure 10.6: Placement of parties on religion and the state.

 Voting for Islamists    231

post-​revolutionary elections of Egypt, Tunisia, and Libya, we find strong evidence

that religious factors drove voters’ calculus during the transitional period. Moreover,
Islamist support appears to be primarily driven by attitudes toward religion and politics.
It is not identity voting; indeed, identifying first and foremost as Muslim is not signif-
icantly associated with Islamist support in any country. Religious practice was found
uniquely significant in Tunisia. Importantly, the role of these religious factors is inde-
pendent from and as strong as organizational and material factors, which have received
much attention in academic circles.
We expect this is due to the historical experiences that shaped the social and political
landscapes prior to and during the first post-​revolutionary elections in Tunisia, Libya,
and Egypt. Mapping our findings to their contexts, we find that complex social and po-
litical attitudes shaped the salience of issues—​particularly religious variables—​uniquely
in each country. In Tunisia, we find that private religious worship, as well as attitudes
towards religious leaders’ influence in government, were significantly correlated with
pro-​Islamist voting. This could indicate that societal cleavages around public-​private
engagement with religion were more likely to be politicized in contexts with a history
of secularization. Unlike Egypt and Libya, a strict system of laicité commanded the pre-​
revolutionary political sphere in Tunisia. Thus, religious practice in Tunisia, as well as
the attitudes toward religious influence within politics, could have driven more devoted
Muslims to support Islamist parties; we do not find this in the more conservative, reli-
gious societies of Egypt and Libya.
These results answer some questions, but also raise new avenues of inquiry. In Libya,
citizens were more likely to support an Islamist political party if they favored a party
that helped to provide services. These findings raise questions regarding why those who
wanted parties to provide services saw the Islamists as the better choice, particularly as
Libyan Islamists had less historical presence and record of service provision than their
Tunisian or Egyptian counterparts. Likewise, in Egypt, our findings regarding identity
are difficult to interpret. The Muslim Brotherhood sought to inculcate a transnational
“Muslim” identity. Therefore, we could have expected religious identity to be correlated
with Islamist voting patterns. However, our results do not support the common criti-
cism that Brothers and their supporters were lacking in Egyptian nationalism.
More general questions remain as well. The facets of religion that impacted voter
choice in the 2011–​2013 period are found to be unique across the countries, contingent
on the social and political landscapes of their time. So, too, we find that organizational

Note: Egypt is shown in the top left panel. Tunisia is shown in the top right panel. Libya is shown
in the bottom center panel. Question wording: “On a scale of 1 to 9, 1 being a strictly religious po-
litical party and 9 being a strictly secular party, how would you rate the following parties?” Figures
in parentheses represent the proportion of respondents who were able to state their positions.
Islamist parties are denoted with darkened distributions compared to non-​Islamist parties and
mean response point estimates are denoted with black diamonds. Note that the Libyan dataset for
this figure is based on data from May 2013.
232    Ellen Lust, Kristen Kao, and Gibran Okar

and economic factors are not generalizable across cases. This draws into question the
need not only to recognize the importance of religion in Islamist support, but also to
take into account how social and political conditions may shape the importance of var-
ious drivers of party success.

Notes
1. In Tunisia, Ennahda is coded as the Islamist party. In Egypt, the Islamist parties include Al
Nour, Al Wasat, and the Freedom and Justice Party (FJP). In Libya, the Islamist parties in-
clude the Justice and Construction Party (JCP) and the Homeland Party.
2. The household surveys included in this chapter were conducted face-​to-​face with
respondents, employing a random sample design and the Kish grid method at the door.
The Egyptian Post-​Election Survey (EPES) (Lust et al. 2012) was a household survey
conducted October 31–​November 10, 2012 (N=4471), and drawn from a PPS sample
of 21 governorates with a response rate of 67%. All governorates except the five border
governorates—​North and South Sinai, The Red Sea, Matrouh, and Al Wadi—​were
surveyed. These five governorates have less than 1% of the population in Egypt. The poll
was conducted by the Al Ahram Centre for Political and Strategic Studies (ACPSS) in col-
laboration with Ellen Lust (then at Yale University), Gamal Soltan (American University
in Cairo) and Jakob Wichmann (JMW Consulting), and with support from the Danish
Egyptian Dialogue Institute (DEDI). The Libyan Post-​Election Surveys (LPES) were
conducted in May 2013 (N = 1152) (Benstead et al. 2013a) and November 2013 (N = 1110)
(Benstead et al. 2013b) and were implemented by Diwan Marketing Research, in col-
laboration with the National Democratic Institute (NDI), JMW Consulting, Ellen Lust
and Lindsay Benstead, with funding from the Danish Ministry of Foreign Affairs. The
Tunisian Post-​Election Survey (TPES) (Benstead et al. 2012) was conducted in October/​
November 2012 (N = 1202) by Lindsay Benstead (Portland State University), Ellen Lust
(then at Yale University), and Dhafer Malouche (ESSAI), with support from the National
Science Foundation, Portland State University, Princeton University, and Yale University.
Probability proportional to size (PPS) sampling was used to select 73 urban and rural
communes located in 16 electoral districts. More information on these studies is available
at transitionalgovernanceproject.org.
3. Brown (2012) and Guiler (2020) make similar arguments regarding Islamists’ moral purity
or incorruptibility, but they do not focus specifically on religion.
4. Like most opposition parties, the Muslim Brotherhood was largely shut out of the 2010
elections, as tensions were rising surrounding, in part, attempts to pave the way for Hosni
Mubarak’s son, Gamal, to take power.
5. For a more detailed discussion of this, see Lust and Waldner (2016).
6. In the transition period, we also saw the emergence of competing unions, such as the UGET
(left-​wing, liberal) and UGTE (affiliate of Ennahda), which both held demonstrations and
counter-​protests.
7. We focused on party lists because individual candidates who ran on Islamist platforms
also ran as independents in these three countries. This makes it difficult to study voting for
Islamists using individual seats.
8. Comparison of the election outcomes and respondents’ voter choice reported in the survey
data lends confidence to our findings. In Tunisia, Islamist parties won 37% of the vote
 Voting for Islamists    233

(National Democratic Institute 2011), compared to 36% in our survey. In Libya, the Islamist
parties won 13% of the vote (Carter Center 2012) compared to 10% in our survey. In Egypt,
data for the percentage of votes cast for party lists was not available; however, Islamist party
lists won 71% of seats allocated to parties (Carnegie Endowment for International Peace
2015), compared to 88% reported support for Islamists reported in our survey.
9. VIF analysis showed that our regressed variables were not giving any redundant informa-
tion. This indicates the independence of our religious and economic measures.

References
Ashour, Omar. 2016. “Between ISIS and a Failed State: The Saga of Libyan Islamists.” Rethinking
Political Islam Series Working Paper. Washington, DC: Brookings Institution, July. https://​
www.brookings.edu/​wp-​content/​uploads/​2016/​07/​Libya_​Ashour-​FINALE.pdf.
Baslevent, Cem, Hasan Kirmanoglu, and Burhan Senatalar. 2005. “Empirical Investigation of
Party Preferences and Economic Voting in Turkey.” European Journal of Political Research
44, no. 4 (June): 547–​562. doi:10.1111/​j.1475-​6765.2005.00238.x.
Benotman, Noman, Jason Pack, and James Brandon. 2013. “Islamists.” In The 2011 Libyan
Uprisings and the Struggle of the Post-​Qadhafi Future, edited by Jason Pack, 191–​228.
New York: Palgrave Macmillan.
Benstead, Lindsay J., Ellen Lust, and Dhafter Malouche. 2012. Tunisian Post-​Election
Survey (TPES). Distributed by The Transitional Governance Project. http://​
transitionalgovernanceproject.org/​.
Benstead, Lindsay J., Ellen Lust, and Jakob Wichmann. 2013a. Libyan Post-​Election
Survey (LPES), May. Distributed by The Transitional Governance Project. http://​
transitionalgovernanceproject.org/​.
Benstead, Lindsay J., Ellen Lust, and Jakob Wichmann. 2013b. Libyan Post-​Election
Survey (LPES), November. Distributed by The Transitional Governance Project. http://​
transitionalgovernanceproject.org/​.
Blaydes, Lisa. 2011. Elections and Distributive Politics in Mubarak’s Egypt.
Cambridge: Cambridge University Press. doi:10.1017/​CBO9780511976469.
Boulby, Marion. 1988. “The Islamic Challenge: Tunisia Since Independence.” Third World
Quarterly 10, no. 2: 590–​614. doi:10.1080/​01436598808420073.
Brooke, Steven. 2017. “From Medicine to Mobilization: Social Service Provision and the
Islamist Reputational Advantage.” Perspectives on Politics 15, no. 1 (March): 42–​ 61.
doi:10.1017/​S1537592716004126.
Brooke, Steven. 2019. Winning Hearts and Votes: Social Services and the Islamist Political
Advantage. Ithaca, NY: Cornell University Press.
Brown, Nathan J. 2012. When Victory Is Not an Option: Islamist Movements in Arab Politics.
Ithaca, NY: Cornell University Press.
Brownlee, Jason, Tarek E. Masoud, and Andrew Reynolds. 2015. The Arab Spring: Pathways of
Repression and Reform. Oxford: Oxford University Press.
Cammett, Melani, and Sukriti Issar. 2010. “Bricks and Mortar Clientelism: Sectarianism
and the Logics of Welfare Allocation in Lebanon.” World Politics 62, no. 3 (July): 381–​421.
doi:10.1017/​S0043887110000080.
Cammett, Melani C. 2014. Compassionate Communalism: Welfare and Sectarianism in
Lebanon. Ithaca, NY: Cornell University Press.
234    Ellen Lust, Kristen Kao, and Gibran Okar

Cammett, Melani C., and Pauline Jones Luong. 2014. “Is There an Islamist Political
Advantage?” Annual Review of Political Science 17, no. 1 (May): 187–​206. doi:10.1146/​
annurev-​polisci-​071112-​221207.
Carnegie Endowment for International Peace. 2015. “2012 Egyptian Parliamentary
Elections.” Resource Page. Washington, DC: Carnegie Endowment for
International Peace, January 22, 2015. https://​carnegieendowment.org/​2015/​01/​22/​
2012-​egyptian-​parliamentary-​elections-​pub-​58800.
Carter Center. 2012. “General National Congress Elections in Libya.” Atlanta: Carter Center.
https://​www.cartercenter.org/​resources/​pdfs/​news/​peace_​publications/​election_​reports/​
libya-​070712-​final-​rpt.pdf.
Chandra, Kanchan. 2004. Why Ethnic Parties Succeed: Patronage and Ethnic Head Counts in
India. Cambridge: Cambridge University Press.
Chemitsky, B. 2012. “Libyan Muslim Brotherhood on the Rise.” MEMRI Inquiry and Analysis
Series, no. 828. Washington, DC: Middle East Media Research Institute (MEMRI).
Ciftci, Sabri. 2013. “Secular-​Islamist Cleavage, Values, and Support for Democracy and Shari’a
in the Arab World.” Political Research Quarterly 66, no. 4 (December): 781–​793. doi:10.1177/​
1065912912470759.
Clark, Janine A. 2004. Islam, Charity, and Activism: Middle-​Class Networks and Social Welfare
in Egypt, Jordan, and Yemen. Indiana Series in Middle East Studies. Bloomington: Indiana
University Press.
Clark, Janine A. 2012. “Patronage, Prestige, and Power: The Islamic Center Charity Society’s
Role within the Muslim Brotherhood.” In Islamist Politics in the Middle East: Movements and
Change, edited by Samer Shehata, 68–​87. Abingdon, UK: Routledge.
Corstange, Daniel. 2012. “Vote Trafficking in Lebanon.” International Journal of Middle East
Studies 44, no. 3 (August): 483–​505. doi:10.1017/​S0020743812000438.
Corstange, Daniel. 2018. “Clientelism in Competitive and Uncompetitive Elections.”
Comparative Political Studies 51, no. 1 (January): 76–​104. doi:10.1177/​0010414017695332.
Davis, Nancy Jean, and Robert V. Robinson. 2012. Claiming Society for God: Religious
Movements and Social Welfare Egypt, Israel, Italy, and the United States. Bloomington: Indiana
University Press.
De Koster, Willem, and Jeroen Van Der Waal. 2007. “Cultural Value Orientations and Christian
Religiosity: On Moral Traditionalism, Authoritarianism, and Their Implications for Voting
Behavior.” International Political Science Review 28, no. 4 (September): 451–​467. doi:10.1177/​
0192512107079636.
Decree No. 35, 2011 (Tunisia). http://​aceproject.org/​ero-​en/​regions/​africa/​TN/​
tunisia-​decree-​no.-​35-​dated-​10-​may-​on-​the-​election.
Dennison, James, and Jonas Draege. 2020. “The Dynamics of Electoral Politics after the Arab
Spring: Evidence from Tunisia.” Journal of North African Studies (Online, February 26): 1–​25.
doi:10.1080/​13629387.2020.1732216.
Eskandar, Wael. 2011. “How Are Seat Winners Determined in the Egyptian
Elections?” Jadaliyya, December 1, 2011. https://​www.jadaliyya.com/​Details/​24743/​
How-​Are-​Seat-​Winners-​Determined-​in-​the-​Egyptian-​Elections-​.
Fidrmuc, Jan. 2000. “Economics of Voting in Post-​Communist Countries.” Electoral Studies 19,
no. 2–​3 (June): 199–​217. doi:10.1016/​S0261-​3794(99)00048-​7.
Fitzgerald, Mary. 2015. “Finding Their Place: Libya’s Islamists during and after the Revolution.”
In The Libyan Revolution and Its Aftermath, edited by Peter Cole and Brian McQuinn, 177–​
204. London: C. Hurst.
 Voting for Islamists    235

Fourati, Maleke, Gabriele Gratton, and Pauline A. Grosjean. 2016. “Render unto Caesar: Taxes,
Charity, and Political Islam.” SSRN Electronic Journal, June 30. 2016. doi:10.2139/​
ssrn.2801892.
Fredrich Ebert Stiftung. 2015. Islamist Movements in Libya: Chances and Challenges of Political
Power. Bonn, Germany: Friedrich Ebert Foundation, Libya Office. https://​www.fes.org.ma/​
common/​pdf/​FES%20Libya%20English%20DEF.pdf.
García-​Rivero, Carlos, and Hennie Kotzé. 2007. “Electoral Support for Islamic Parties in the
Middle East and North Africa.” Party Politics 13, no. 5 (September): 611–​636. doi:10.1177/​
1354068807080088.
Grewal, Sharan, Amaney A. Jamal, Tarek Masoud, and Elizabeth R. Nugent. 2019. “Poverty and
Divine Rewards: The Electoral Advantage of Islamist Political Parties.” American Journal of
Political Science 63, no. 4 (July): 859–​874. doi:10.1111/​ajps.12447.
Guiler, Kimberly G. 2020. “From Prison to Parliament: Victimhood, Identity, and
Electoral Support.” Mediterranean Politics (Online, February, 20): 1–​ 30. doi:10.1080/​
13629395.2020.1721159.
Hafez, M., and Quintan Wiktorowicz. 2004. “Violence as Contention in the Egyptian Islamic
Movement.” In Islamic Activism: A Social Movement Theory Approach, edited by Quintan
Wiktorowicz, 61–​88. Bloomington: Indiana University Press.
Hamid, Shadi. 2011. “Arab Islamist Parties: Losing on Purpose?” Journal of Democracy 22, no. 1
(January): 68–​80.
Hasan, Noorhaidi. 2012. “Islamist Party, Electoral Politics and Da’wah Mobilization among
Youth: The Prosperous Justice Party (PKS) in Indonesia.” Journal of Indonesian Islam 6, no.
1: 17. doi:10.15642/​JIIS.2012.6.1.17-​47.
Huntington, Samuel P. 1996. The Clash of Civilizations and the Remaking of World Order.
New York: Simon & Schuster.
Kahl, Sigrun. 2005. “The Religious Roots of Modern Poverty Policy: Catholic, Lutheran,
and Reformed Protestant Traditions Compared.” European Journal of Sociology 46, no. 1
(April): 91–​126. doi:10.1017/​S0003975605000044.
Kandil, Hazem. 2015. Inside the Brotherhood. Malden, MA: Polity.
Kao, Kristen. 2015. “Ethnicity, Electoral Institutions, and Clientelism: Authoritarianism in
Jordan.” PhD diss., University of California, Los Angeles. https://​escholarship.org/​uc/​item/​
829340k4.
Lust, Ellen. 2009. “Competitive Clientelism in the Middle East.” Journal of Democracy 20, no. 3
(July): 122–​135. doi:10.1353/​jod.0.0099.
Lust, Ellen. 2011. “Missing the Third Wave: Islam, Institutions, and Democracy in the
Middle East.” Studies in Comparative International Development 46, no. 2 (April): 163–​190.
doi:10.1007/​s12116-​011-​9086-​z.
Lust, Ellen, Gamal Soltan, and Jakob Wichmann. 2012. Egyptian Post-​Election Survey (EPES).
Distributed by Transitional Governance Project. http://​transitionalgovernanceproject.
org/​.
Lust, Ellen, and David Waldner. 2016. “Parties in Transitional Democracies: Authoritarian
Legacies and Post-​Authoritarian Challenges in the Middle East and North Africa.” In
Parties, Movements and Democracy in the Developing World, edited by Nancy Bermeo and
Deborah Yashar, 157–​189. Princeton, NJ: Princeton University Press.
Lust-​Okar, Ellen, and Amaney Ahmad Jamal. 2002. “Rulers and Rules: Reassessing the
Influence of Regime Type on Electoral Law Formation.” Comparative Political Studies 35, no.
3 (April): 337–​366. doi:10.1177/​0010414002035003004.
236    Ellen Lust, Kristen Kao, and Gibran Okar

Masoud, Tarek E. 2014. Counting Islam: Religion, Class, and Elections in Egypt. Problems of
International Politics. New York: Cambridge University Press.
McClendon, Gwyneth H., and Rachel Beatty Riedl. 2019. From Pews to Politics: Religious
Sermons and Political Participation in Africa. Cambridge: Cambridge University Press.
National Democratic Institute. 2011. “Final Report on the Tunisian National Constituent
Assembly Elections.” Washington, DC: National Democratic Institute. https://​www.ndi.org/​
sites/​default/​files/​tunisia-​final-​election-​report-​021712_​v2.pdf.
Öniş, Ziya. 2004. “The Political Economy of Turkey’s Justice and Development Party.” SSRN
Electronic Journal, February 2, 2004. doi:10.2139/​ssrn.659463.
Pipes, Daniel. 2002. “Islam and Islamism.” Policy: A Journal of Public Policy and Ideas 18, no. 1.
Posner, Daniel N. 2005. Institutions and Ethnic Politics in Africa. Cambridge: Cambridge
University Press.
Raymond, Christopher. 2011. “The Continued Salience of Religious Voting in the United
States, Germany, and Great Britain.” Electoral Studies 30, no. 1 (March): 125–​135. doi:10.1016/​
j.electstud.2010.10.001.
Robbins, Michael. 2009. “What Accounts for Popular Support for Islamist Parties in the Arab
World? Evidence from the Arab Barometer.” In 104th Annual Meetings of the American
Political Science Associations. Cambridge, MA: Kennedy School of Government, Harvard
University. https://​w ww.belfercenter.org/​sites/​default/​f iles/​legacy/​f iles/​Robbins-​
Support%20for%20Islamist%20Parties.pdf.
Rohac, Dalibor. 2012a. “Religion as a Commitment Device: The Economics of Political Islam.”
SSRN Electronic Journal, September 12, 2012. doi:10.2139/​ssrn.2144864.
Rohac, Dalibor. 2012b. “The Secret of Islamist Success.” Foreign Policy, October 31, 2012. https://​
foreignpolicy.com/​2012/​10/​31/​the-​secret-​of-​islamist-​success/​.
Schwedler, Jillian. 2006. Faith in Moderation: Islamist Parties in Jordan and Yemen.
Cambridge: Cambridge University Press.
Shehata, Dina. 2013. Islamists and Secularists in Egypt: Opposition, Conflict, and Cooperation.
London: Routledge.
Singerman, Diane. 2004. “The Networked World of Islamist Social Movements.” In Islamic
Activism: A Social Movement Theory Approach, edited by Quintan Wiktorowicz, 143–​163.
Bloomington: Indiana University Press.
Smith, Ben. 2012. House of Commons Briefing Note—​Libya’s General Assembly Election. SNIA/​
6389.
Tessler, Mark A. 2011. “The Origins of Political Support for Islamist Movements: A Political
Economy Approach.” In Public Opinion in the Middle East: Survey Research and the Political
Orientations of Ordinary Citizens, edited by Mark Tessler, 25–​55. Indiana Series in Middle
East Studies. Bloomington: Indiana University Press.
Watanabe, Lisa. 2018. “Islamist Actors: Libya and Tunisia.” Zurich: Center for Security Studies
(CSS) ETH Zurich. https://​css.ethz.ch/​content/​dam/​ethz/​special-​interest/​gess/​cis/​center-​
for-​securities-​studies/​pdfs/​Watanabe-​06282018-​IslamistActors.pdf.
Wegner, Eva, and Francesco Cavatorta. 2019. “Revisiting the Islamist-​Secular Divide: Parties
and Voters in the Arab World.” International Political Science Review 40, no. 4 (August): 558–​
575. doi:10.1177/​0192512118784225.
White, Jenny. 2012. “Islamist Social Networks and Social Welfare Services in Turkey.” In
Islamist Politics in the Middle East: Movements and Change, edited by Samer Shehata, 59–​67.
Abingdon, UK: London: Routledge. doi:10.4324/​9780203126318.
 Voting for Islamists    237

Wickham, Carrie Rosefsky. 2002. Mobilizing Islam: Religion, Activism and Political Change in
Egypt. New York: Columbia University Press.
Wiktorowicz, Quintan, ed. 2004. Islamic Activism: A Social Movement Theory Approach.
Indiana Series in Middle East Studies. Bloomington: Indiana University Press.
Wolf, Anne. 2017. Political Islam in Tunisia: The History of Ennahda. London: C. Hurst.
Yousfi, Hèla. 2018. Trade Unions and Arab Revolutions: The Tunisian Case of UGTT. Routledge
Research in Employment Relations 39. New York: Routledge.
 Chapter 11

Part y Systems i n Mu sl i m
So ciet i e s

Elizabeth R. Nugent

This chapter explores whether and how existing theories of party system formation can
help us to understand the institutions that have developed in Muslim-​majority societies.
I ask the following questions: Are party systems in Muslim-​majority and non-​Muslim-​
majority societies systematically different from each other? If so, how—​and more im-
portantly, why? I rely on cross-​national time-​series data of party systems and draw on
the rich literature on politics in the Muslim world to answer these questions. The data
demonstrate that party systems in Muslim-​majority countries are consistently less
competitive, less open, and less institutionalized—​though I note important exceptions
located in Muslim democracies.
I argue that the traits of party systems in Muslim-​majority countries are better
explained by specific historical developments in the Muslim world and the path depend-
ence of party system institutions than by the political institutions prescribed by Islamic
tenets. More specifically, patterns of governance and institutions adopted by European
authorities during colonialism coincided with a consequential period of state-​building
in the Muslim world, and this in turn shaped the rules of the game in postcolonial coun-
tries. This finding is not unique to Muslim-​majority societies; scholars have made sim-
ilar arguments about postcolonial political institutions in other parts of the world, with
variation produced through the nature, timing, and duration of the colonial experi-
ence. I conclude the chapter by outlining remaining questions about party systems in
Muslim-​majority societies that might shape a research agenda on the topic.

Defining and Measuring Party Systems

The study of parties and the systems in which they are organized is central to the study
of politics. Drawing from the experience of early democratizing countries in Western
240   Elizabeth R. Nugent

Europe, scholars reached an early normative consensus that parties are of the utmost
importance in democratic politics (White 2006, 5–​15). As democracy spread beyond
Europe, political scientists employed measurements of the emergence of party com-
petition as a proxy for successful democratization. Scholars continue to deem parties
“absolutely essential for the proper functioning of representative democracy” (Montero
2003, 3).
Parties represent collective interests in political competition; in politics, “no idea has
ever made much headway without an organization behind it” (Sartori 1976, 85). In their
idealized form, parties simply reflect the shared, aggregated political preferences of
their constituent members; they “crystallize and make explicit the conflicting interests,
the latent strains and contrasts in the existing social structure” (Lipset and Rokkan
1967, 3). But, in practice, parties exert influence on the creation and nature of collec-
tive interests. Parties have agency in choosing which constituencies to mobilize and
under what banner, and in this way can exert influence on the dimensions of political
contestation. The political relevance of “class, ethnicity, religion, race or nation do[es]
not happen simultaneously . . . as a reflection of the objective conditions in the psyches
of individuals” (Przeworski 1985, 99–​101). Instead, parties may “establish themselves
as significant poles of attraction and produce their own alignments independently of
the geographical, the social, and the cultural underpinnings of the movements” (Lipset
and Rokkan 1967, 3). Neto and Cox (1997, 150) similarly argue that “politicians can take
socially defined groups and combine or recombine them in many ways for political
purposes.”
Parties are a function of the broader political context in which they exist. The amor-
phous concept of a party system refers to the interaction of parties competing with each
other within a given country.1 Scholars rely less on a clear definition of party systems and
instead used different criteria to identify and compare different types. Duverger (1954)
first discussed the importance of the interactions between parties as part of a larger
system, in order to understand the relative options offered to voters within a given con-
text. In addition to the characteristics of individual parties, he noted that it is important
to consider the “numbers, respective sizes, alliances, geographical localization, [and]
political distribution” of parties competing with each other within the same political
space. As the result of this interactive process, “a party system is defined by a particular
relationship amongst all these characteristics” (203).
Party systems are measured by several distinct attributes. These include the
number of parties contesting elections and the number of effective parties, or parties
that actually win seats in government (Blondel 1968; Laasko and Taagepera 1979;
Siaroff 2000). Party systems can also be categorized by the relative size, strength, and
balance of parties (Rokkan 1970; Wolinetz 2006). Party systems also differ on the
distribution and polarization of parties, measured by the distance separating parties
on salient axes of contestation and in affect.2 In addition, party systems differ in the
degree to which competition for government is open or closed—​meaning whether
all parties can enter elections without any constraints or whether contestation is
restricted only to certain parties or combinations of parties (Mair 1996, 2002). In
 Party Systems in Muslim Societies    241

terms of linkages with voters, party systems may revolve around programmatic, ide-
ological, clientelist, or charismatic competition (Kitschelt et al. 1999, 2010). The na-
ture of competition within a given party system determines the degree to which it is
institutionalized and structures the electorate (Mainwaring 1999; Mainwaring and
Scully 1995).
These measurements help us characterize and compare party systems. The criterial
approach to defining and measuring party systems bridges the divide across regime
types, linking democratic and authoritarian systems on a spectrum of competition and
institutionalization.

Comparing Party Systems

in Muslim-​M ajority and
Non-​Muslim-​M ajority Societies

Are party systems in Muslim-​majority and non-​Muslim-​majority societies different?

And if so, in what ways? To unpack this, I turn to data from the Varieties of Democracy
(V-​Dem) project (Coppedge et al. 2020). V-​Dem is an ideal data set to use for my
analyses because it includes an extensive set of indicators capturing different aspects
and dimensions of party systems, as well as important correlates. In addition to aggre-
gate scores measuring a party system’s level of competitiveness or democratic quality,
the data set includes fine-​grained constitutive measurements capturing indicators
central to comparing party systems, such as the number and relative strength of ef-
fective parties, the distribution of parties on salient axes of competition, and the na-
ture of voter-​party linkages. Moreover, the project’s expansive country and temporal
coverage is helpful for measuring trends and differences in party systems across time
and space. I subset version 10 of the V-​Dem data set from 1900 on, and separate coun-
tries with Muslim-​majority populations to those without a majority Muslim popu-
lation. The charts plot the raw data as well as fit lines, indicating trends in the data.
All data have been recoded so that the plots read as more “democratic” (with specific
meaning depending on the index or variable) as the y-​axis approaches 1. The x-​axis
measures year.
Two indices demonstrate how party systems in Muslim-​majority societies differ from
those in non-​Muslim societies. For each index, individual components trend the same
way as the index, with some variation in the degree of difference between party sys-
tems in Muslim-​majority and non-​Muslim-​majority societies, though not the level of
significance. The first index, electoral democracy, combines a number of subcomponents
measuring the openness and level of competition of party systems. The index ranges
from 0 to 1, with a higher score indicating more clean, fair, and competitive electoral
institutions. The index combines measurements of freedom of expression, universal
suffrage, and frequency of voting irregularities. In addition, the index includes a
242   Elizabeth R. Nugent

measurement of freedom of association, which captures whether parties are banned,

what if any barriers exist to party entry into electoral competition, the level of autonomy
of opposition parties, and whether multiparty elections occur. The index also includes a
component capturing pluralism in elected officials holding office, similar to the idea of
effective parties outlined in the previous section.
Figure 11.1 demonstrates that all party systems become more open, inclusive, and
competitive over time. However, party systems in non-​Muslim-​majority countries con-
sistently score higher than those in Muslim-​majority countries, suggesting they are rel-
atively more open to competition, more competitive, and fairer. The gap between party
systems in the two groups of countries is persistent but appears to increase starting in
the 1970s.
The second index, party institutionalization, combines a number of subcomponents
measuring the balance and function of parties within a given system. The index ranges
from 0 to 1, with higher scores indicating a more institutionalized party system that
structures political competition. The index includes measurements of the level and
depth of party organization and internal party cohesion in structure, platform, and
ideology. In addition, the index includes a measurement capturing the extent to which
party platforms are distinct from each other. The index also includes a measurement
of the nature of voter-​party linkages, whether these are clientelist, local collective, pro-
grammatic, or mixed between types.
Figure 11.2 demonstrates that, over time, party systems have become more institu-
tionalized. Party systems in non-​Muslim-​majority countries consistently score higher
than those in Muslim-​majority countries, meaning they are more institutionalized. The
gap between party systems in the two types of countries is persistent and consistent in
size from initial differences over time.

Electoral Democracy Index

1.00

0.75
Score

0.50

0.25

0.00
1900 1920 1940 1960 1980 2000 2020
Year
Muslim-Majority Other

Figure 11.1: Electoral democracy, 1900–​2019.

 Party Systems in Muslim Societies    243

Party Institutionalization Index

1.00

0.75
Score

0.50

0.25

0.00
1900 1920 1940 1960 1980 2000 2020
Year
Muslim-Majority Other

Figure 11.2: Party institutionalization, 1990–​2019.

Sociological Explanations for Party

System Divergence

What helps to explain the persistent differences between party systems in Muslim-​
majority and non-​Muslim-​majority societies? Explanations for how parties and party
systems form, and why they differ, fall into two broad categories: sociological or insti-
tutional (Boix 2009). Sociological explanations outline a bottom-​up process through
which societal divisions external to the political system shape the nature and func-
tion of political contestation. Underlying factors refer to the demographic distribution
along social, political, and economic cleavages. Broader historical processes of nation-​
building, state consolidation, development, modernization, and industrialization create
meaningful underlying divisions within society and variation across societies, which in
turn influence party formation, placement, and function (Deegan-​Krause 2007; Evans
and Whitefield 2000; Haggard and Kaufman 2016; Lipset and Rokkan 1967). Variation
in party systems across cases stems from different underlying cleavage structures, and
shifts within cases result from the shifting of underlying cleavages (Lupu and Riedl 2013;
McCarty, Poole, and Rosenthal 2006; Sigelman and Yough 1978; Ura and Ellis 2011).
How would a sociological explanation account for party system differences in Muslim-​
majority and non-​Muslim-​majority societies? It would mean that the nature of cleavages
in Muslim societies would both explain the initial divergence and the persistent difference
from party systems in non-​Muslim majority societies. V-​dem includes a proxy measure-
ment of polarization of society, asking coders to “characterize the differences of opinions
on major political issues,” presented in Figure 11.3. The data is limited from 2000 to 2019,
but does not indicate that Muslim-​and non-​Muslim-​majority societies are differently
244   Elizabeth R. Nugent

Societal Polarization

None

Limited
Score

Medium

Moderate

Serious

2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020
Year
Muslim-Majority Other

Figure 11.3: Societal polarization, 2000–​2019.

polarized; both, on average, score a medium level of polarization. While the data does not
include indicators for the early 1990s, the available contemporary data suggests there is no
relationship between societal polarization and party systems.

Institutional Explanations for Party

System Divergence

Institutional explanations for party system formation outline a top-​down process

through which formal institutions shape the nature and function of political contesta-
tion (Boix 2009). Most influential are those features that define the internal workings of
the political system, primarily electoral rules and institutions. Formal institutions can
complicate the translation of underlying interests into political competition, enhance
or diminish polarization, and shape what parties compete over by requiring certain
thresholds for participation, forcing certain partnerships and patterns of strategic place-
ment, and distributing public goods based on representation in elected bodies (Cox
1987; Shugart and Carey 1992; Taagepera and Shugart 1989).
How would an institutional explanation account for party system differences
in Muslim-​and non-​Muslim-​majority societies? It would mean that the nature of
institutions in Muslim societies would both explain the initial divergence and the persis-
tent difference from party systems in non-​Muslim-​majority societies. Historical institu-
tionalism suggests strong path dependence after an initial critical juncture establishing
institutions, so I spend more time theorizing why Muslim societies may have come to
possess different kinds of political institutions in the first place.
 Party Systems in Muslim Societies    245

There are at least two possible ways in which Muslim societies may have come to ini-
tially possess certain political institutions. The first is through Islamic culture. Scholars
have previously questioned whether the religion of Islam is the reason behind the pre-
dominance of authoritarianism in Muslim-​majority societies. The focus has largely
been on whether Islam prevents the development of democratic citizens, building on
the implications of canonical comparative politics texts on political culture, such as
those by Almond and Verba (1963) and Lipset (1969). Islam and its prescribed behaviors
are argued to create a culture that is inhospitable to democracy.3 While this claim re-
mains influential in academic, policy, and public debates about the Muslim world, it
largely lacks empirical support.4 Islam may also influence the nature of political
institutions through cultural prescriptions. Much debate has revolved around two spe-
cific institutions derived from the practices of the Prophet Mohammed and the early
Muslim community. The first is shura, translated as “consultation.” This concept refers to
a process of collecting and discussing different opinions on a particular subject in order
to reach a decision. The second institution is that of bayʿah, an oath of allegiance to a
leader.
There is no theoretical or empirical consensus that these concepts deterministically
translate into institutions lacking democratic qualities. Islam is a discursive tradi-
tion and is interpreted and practiced very differently by individuals across time and
space (Asad 2009). For example, shura has been argued to both be the basis for the
implementation of representative democracy as well as the basis for the tyranny of the
majority, a critique notably also levied against secular institutions. Further, both in
theory and in practice, Muslim polities have differed in whether consultation occurs
with experts, those affected by a decision, or representatives of these groups, resulting
in more and less direct representation. Similarly, interpretations of bayʿah differ in
whether it implies a social contract in which a leader is responsible to and limited
by the people he represents, or unquestioning loyalty to a leader in order to prevent
disorder and thus inhibits the establishment of institutional checks and balances of
executive power (Abou El Fadl 2004; Tezcür 2007). While many party systems in
Muslim-​majority societies demonstrate undemocratic characteristics, there are im-
portant exceptions in countries like Indonesia, Malaysia, and Tunisia, all of which
host highly competitive and open political systems that include Islamist parties and
draw on religious tenets in the constitution, undermining a deterministic relationship
between Islam and undemocratic political institutions. Quantitative studies demon-
strate that denser and more encompassing Islamic political, educational, and finan-
cial institutions correlates with political institutions of higher democratic quality
(Achilov 2010).
The second way in which Muslim societies may have come to possess different polit-
ical institutions is through colonial endowments. This argument rests on a close reading
of the historical political development of the Muslim world, with attention paid to when
and how party systems emerged, what their initial purpose was, and how this influenced
their shape and nature. European colonialism coincided with a consequential period
246   Elizabeth R. Nugent

of modern state-​building in Muslim-​majority societies for subsequent patterns of state

capacity, the nature of coercive institutions, and economic development. The develop-
ment projects that significantly shaped the institutions of states in Muslim-​majority
societies were guided by an overarching colonial strategy that favored the interests of the
colonizer. In line with their material and strategic interests, colonial powers created or
co-​opted a variety of governance institutions to control local populations and channel
dissent rather than to represent and respond to local interests (Acemoglu, Johnson,
and Robinson 2001; Anderson 1986; Méouchy and Sluglett 2004). The institutional in-
heritance with which independent states were endowed shaped post-​independence
institutions in significant ways through mechanisms of path dependence. Party systems
were no exception. Foreign colonial powers created legislative institutions that were un-
representative, promoting certain favored groups integral to colonial rule at the expense
of other groups. They also fostered, politicized, and institutionalized certain societal
divisions and certain modes of contestation and redistribution that were highly con-
tentious in order to divide and conquer local populations (Laitin 1985; Mahoney 2010;
Mamdani 1996).
This institutional inheritance predicts subsequent political institutions. The na-
ture of political institutions at the onset of liberalization accompanying independence
from colonialism significantly influenced the creation of electoral rules in the subse-
quent regime. For example, elites in postcolonial countries that had previously been
governed by a single-​party system systematically adopted electoral laws that concen-
trated legislative power in the hands of the dominant party. In contrast, postcolonial
countries that had formerly been governed by monarchies under colonialism adopted
electoral rules that supported a balance of power among multiple competing forces,
in a continued strategy of divide-​and-​conquer. In line with mechanisms of histor-
ical institutionalism, incumbents’ preferences over the distribution of domestic po-
litical power vary across different types of inherited regimes, and this shapes their
preferences for the distribution of power in the postcolonial period (Lust-​Okar and
Jamal 2002).
The legacy of colonialism on party system formation in the Muslim world means
that parties are less likely to serve a programmatic function. Parties may also serve
distributive, surveillance, and control functions unrelated to representation of collec-
tive interests (Blackman 2019; Hibou 2011). As a result, party systems are less open to
meaningful competition and less institutionalized in postcolonial periods. While the
colonial experience was globally consequential, it was not even or uniform throughout
the world, let alone in the Muslim world. Colonial systems of control differed in the
nature and form of the relationship between the colonizer and the colonized well be-
yond simplistic distinctions between direct and indirect rule (Crowder 1964; Méouchy
and Sluglett 2004). So while colonialism generally undermined the representative and
responsive functions of governance institutions, it also produced meaningful varia-
tion that is helpful in understanding and measuring its historical effect on subsequent
institutions (Ricart-​Huguet forthcoming).
 Party Systems in Muslim Societies    247

Conclusion

In this chapter, I have discussed whether, how, and why party systems in Muslim-​
majority countries differ systematically and significantly from those in non-​Muslim-​
majority countries. Cross-​national, time-​series data from V-​dem demonstrates that
while the party systems of the world are generally liberalizing, including more political
interests and becoming more institutionalized, the Muslim world lags behind its non-​
Muslim counterpart. In my reading of existing literature, this gap is better explained
by institutional endowments than by cultural ones. In thinking about party systems in
Muslim-​majority societies, it is necessary to interrogate the theoretical and practical
differences that render party systems in nondemocratic regimes fundamentally dif-
ferent from those in democratic contexts. What this means substantively is that scholars
of party systems in the Muslim world must recognize that political institutions do not
always reflect underlying cleavages or unfettered mobilization of collective interests.
Instead, these party systems reproduce inherited methods of control.
Future work on the subject might expand this line of inquiry in at least two important
ways. First, how do party systems persist through independence, an event that is often
considered a political rupture? Existing scholarship suggests the importance of initial
post-​independence leaders’ idiosyncratic preferences for governance and perceptions
of political threats. However, we might better understand the persistence of party sys-
tems by mapping those state employees and foreign consultants who continue to staff
these institutions through independence, as well as the relative balance of power be-
tween political interests as similarly shaped by colonial strategies of control.
Second, how might party systems be meaningfully reformed to become more repre-
sentative, more competitive, and more institutionalized? Neither decades of foreign aid
to promote democracy in Muslim-​majority countries nor countless domestic attempts
at mass mobilization and protest demanding democratization have significantly
changed the nature of these party systems. However, perhaps in further interrogating
the origins and persistence of these party systems, scholars might come up with inno-
vative solutions to overcome the challenges of institutional inheritance. Looking to suc-
cessful transitions from undemocratic to democratic party systems in Muslim majority
societies may shed light on best practices.

Notes
1. See Kitschelt (2007) for an extensive review.
2. For studies of preference polarization, see Sartori (1976), Ware (1996), and Mair (2002). For
studies on affective polarization, see Iyengar et al. (2019) and Nugent (2020).
3. Scholars alternatively point to the strength of [non-​nationalist] identities, the creation
of docile citizens, and a lack of respect for pluralism, gender equality, and protections for
minorities. See Gellner (1983), Huntington (1996), and Inglehart and Norris (2003).
248   Elizabeth R. Nugent

4. Public opinion data largely disconfirm the hypothesized relationship between high religi-
osity and intolerance. See work summarized in Tessler (2011).

References
Abou El Fadl, Khaled. 2004. Islam and the Challenge of Democracy. Princeton, NJ: Princeton
University Press.
Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2001. “The Colonial Origins of
Comparative Development: An Empirical Investigation.” American Economic Review 91, no.
5: 1369–​1401.
Achilov, Dilshod. 2010. “Can Islam and Democracy Coexist? A Cross-​National Analysis of
Islamic Institutions in the Muslim World.” PhD diss., University of Arizona.
Almond, Gabriel A. and Sidney Verba. 1963. Political Culture: Political Attitudes and Democracy
in Five Nations. Princeton, NJ: Princeton University Press.
Anderson, Lisa. 1986. The State and Social Transformation in Tunisia and Libya, 1830–​1980.
Princeton, NJ: Princeton University Press.
Asad, Talal. 2009. “The Idea of an Anthropology of Islam.” Qui parle 17, no. 2: 1–​30.
Blackman, Alexandra. 2019. “Ideological Responses to Settler Colonialism: Political Identities
in Post-​Independence Tunisia.” Working paper.
Blondel, Jean. 1968. Party Systems and Patterns of Government in Western Democracies.
Canadian Journal of Political Science/​Revue canadienne de science politique, 180–​203.
Boix, Carles. 2009. “The Emergence of Parties and Party Systems.” In The Oxford Handbook of
Comparative Politics, edited by Carles Boix and Susan C. Stokes, 314–​334. New York: Oxford
University Press.
Coppedge, Michael, John Gerring, Carl Henrik Knutsen, Staffan I. Lindberg, Jan Teorell,
David Altman, Michael Bernhard, et al. 2020. V-​Dem Dataset—​version 10. Gothenburg,
Sweden: Varieties of Democracy (V-​Dem) Institute. https://​www.v-​dem.net/​en/​data/​data/​
v-​dem-​dataset/​.
Cox, Gary W. 1987. The Efficient Secret. New York: Cambridge University Press.
Crowder, Michael. 1964. “Indirect Rule—​French and British Style.” Africa 34, no. 3: 197–​205.
Deegan-​Krause, Kevin. 2007. “New Dimensions of Political Cleavage.” In The Oxford Handbook
of Political Behavior, edited by R. Dalton and H.-​D. Klingemann, 538–​556. Oxford: Oxford
University Press.
Duverger, Maurice. 1954. Political Parties. New York: Wiley.
Evans, Geoffrey and Stephen Whitefield. 2000. “Explaining the Formation of Electoral
Cleavages in Post-​Communist Democracies.” In Elections in Central and Eastern Europe: The
First Wave, edited by Hans-​Dieter Klingemann, Ekkehard Mochmann, and Kenneth
Newton, 36–​70. Berlin: Ed. Sigma.
Gellner, Ernest. 1983. Muslim Society. Cambridge: Cambridge University Press.
Haggard, Stephan and Robert R. Kaufman. 2016. “Democratization during the Third Wave.”
Annual Review of Political Science 19: 125–​144.
Hibou, Béatrice. 2011. The Force of Obedience. Cambridge: Polity Press.
Huntington, Samuel P. 1996. The Clash of Civilizations and the Remaking of World Order.
New York: Simon & Schuster.
Inglehart, Ronald and Pippa Norris. 2003. Rising Tide: Gender Equality and Cultural Change
Around the World. Cambridge: Cambridge University Press.
 Party Systems in Muslim Societies    249

Iyengar, Shanto, Yphtach Lelkes, Matthew Levendusky, Neil Malhotra, and Sean J. Westwood.
2019. “The Origins and Consequences of Affective Polarization in the United States.” Annual
Review of Political Science 22: 129–​146.
Kitschelt, Herbert. 2007. “Party Systems.” In The Oxford Handbook of Comparative Politics, ed-
ited by Carles Boix and Susan C. Stokes, 522–​544. Oxford: Oxford University Press.
Kitschelt, Herbert, Kirk A. Hawkins, Juan Pablo Luna, Guillermo Rosas, and Elizabeth
J. Zechmeister. 2010. Latin American Party Systems. Cambridge: Cambridge University Press.
Kitschelt, Herbert, Zdenka Mansfeldova, Radoslaw Markowski, and Gabor Toka. 1999. Post-​
Communist Party Systems: Competition, Representation, and Inter-​ Party Cooperation.
Cambridge: Cambridge University Press.
Laakso, Markku and Rein Taagepera. 1979. “Effective Number of Parties: A Measure with
Application to West Europe.” Comparative Political Studies 12, no. 1: 3–​27.
Laitin, David. 1985. “Hegemony and Religious Conflict: British Imperial Control and Political
Cleavages in Yorubaland.” in Bringing the State Back In, edited by Peter Evans, 285–​316.
Cambridge: Cambridge University Press.
Lipset, Seymour M. 1969. “Some Social Requisites of Democracy: Economic Development and
Political Legitimacy.” American Political Science Review 53, no. 1: 69–​105.
Lipset, Seymour M. and Stein Rokkan. 1967. “Cleavage Structures, Party Systems, and Voter
Alignments: An Introduction.” In Party Systems and Voter Alignments: Cross-​National
Perspectives, edited by Seymour M. Lipset and Stein Rokkan, 1–​64. New York: Free Press.
Lupu, Noam and Rachel Beatty Riedl. 2013. “Political Parties and Uncertainty in Developing
Democracies.” Comparative Political Studies 46, no. 11: 1339–​1365.
Lust-​Okar, Ellen and Amaney Jamal. 2002. “Rulers and Rules: Reassessing the Influence of
Regime Type on Electoral Law Formation.” Comparative Political Studies 35, no. 3: 337–​366.
Mahoney, James. 2010. Colonialism and Postcolonial Development: Spanish America in
Comparative Perspective. Cambridge: Cambridge University Press.
Mainwaring, Scott. 1999. Rethinking Party Systems in the Third Wave of Democratization: The
Case of Brazil. Palo Alto, CA: Stanford University Press.
Mainwaring, Scott and Timothy R. Scully. 1995. Building Democratic Institutions: Party Systems
in Latin America. Stanford, CA: Stanford University Press.
Mair, Peter. 1996. What Is Different about Post-​ Communist Party Systems? Glasgow,
Scotland: Center for the Study of Public Policy, University of Strathclyde.
Mair, Peter. 2002. “Comparing Party Systems.” In Comparing Democracies 2: New Challenges in
the Study of Elections and Voting, edited by Lawrence LeDuc, Richard G. Niemi, and Pippa
Norris, 88–​107. Los Angeles: SAGE.
Mamdani, Mahmood. 1996. Citizen and Subject: Contemporary Africa and the Legacy of Late
Colonialism. Princeton, NJ: Princeton University Press
McCarty, Nolan, Keith T. Poole, and Howard Rosenthal. 2006. Polarized America: The Dance of
Ideology and Unequal Riches. Cambridge, MA: MIT Press.
Méouchy, Nadine and Peter Sluglett. 2004. The British and French Mandates in Comparative
Perspectives (Les Mandats Francais Et Anglais Dans Une Perspective Comparative).
Leiden: Brill.
Montero, Jose Ramon. 2003. “Introduction: Reviewing and Reassessing Parties.” In Political
Parties: Old Concepts and New Challenges, edited by Richard Gunther, Jose Ramon Montero,
and Juan J. Linz, 1–​38. Oxford: Oxford University Press.
Neto, Octavio Amorim and Gary W. Cox. 1997. “Electoral Institutions, Cleavage Structures,
and the Number of Parties.” American Journal of Political Science 4, no. 1: 149–​174.
250   Elizabeth R. Nugent

Nugent, Elizabeth R. 2020. After Repression: How Polarization Derails Democratic Transition.
Princeton, NJ: Princeton University Press.
Przeworski, Adam. 1985. Capitalism and Social Democracy. Cambridge: Cambridge
University Press.
Ricart-​Huguet, Joan. Forthcoming. “Colonial Education, Political Elites, and Regional Political
Inequality in Africa.” Comparative Political Studies.
Rokkan, Stein. 1970. “Nation-​Building, Cleavage Formation and the Structuring of Mass
Politics.” In Citizens, Elections, Parties: Approaches to the Comparative Study of the Processes
of Development, 72–​144. Colchester, UK: ECPR Press.
Sartori, Giovanni. 1976. Parties and Party System: A Framework for Analysis. Cambridge,
MA: Harvard University Press.
Shugart, Matthew Soberg and John M. Carey. 1992. Presidents and Assemblies: Constitutional
Design and Electoral Dynamics. Cambridge: Cambridge University Press.
Siaroff, Alan. 2000. Comparative European Party Systems: An Analysis of Parliamentary
Elections since 1945. New York: Garland.
Sigelman, Lee and Syng Nam Yough. 1978. “Left-​ Right Polarization in National Party
Systems: A Cross-​National Analysis.” Comparative Political Studies 11, no. 3: 355–​379.
Taagepera, Rein and Matthew Soberg Shugart. 1989. Seats and Votes: The Effects and
Determinants of Electoral Systems. New Haven, CT: Yale University Press.
Tessler, Mark A. 2011. Public Opinion in the Middle East: Survey Research and the Political
Orientations of Ordinary Citizens. Bloomington: Indiana University Press.
Tezcür, Güneş Murat. 2007. “Constitutionalism, Judiciary, and Democracy in Islamic
Societies.” Polity 39, no. 4: 479–​501.
Ura, Joseph Daniel and Christopher R. Ellis. 2011. “Partisan Moods: Polarization and the
Dynamics of Mass Party Preferences.” Comparative Political Studies 74, no. 1: 277–​291.
Ware, Alan. 1996. Political Parties and Party Systems. New York: Oxford University Press.
White, John Kenneth. 2006. “What Is a Political Party?” In Handbook of Party Politics, edited
by William S. Katz and William Crotty, 5–​15. Thousand Oaks, CA: SAGE.
Wolinetz, Steven B. 2006. “Party Systems and Party System Types.” Handbook of Party Politics,
edited by William S. Katz and William Crotty, 51–​62. Thousand Oaks, CA: SAGE.
 Chapter 12

Ide ol o gies, Bra nd s , a nd

De mo graphics i n Mu sl i m
Sou theast Asia
“Voting for Islam”

Thomas Pepinsky

In many Muslim-​majority countries, elections provide voters with an array of parties

and candidates, some of whom campaign explicitly on religious platforms. In studying
the electoral performance of Islamic parties, analysts frequently interpret votes for
Islamic parties as representing support for Islamist agendas or platforms. The under-
lying assumption in such analyses is that Islam is a programmatic identity shared by
voters and parties alike. But just as a wealth of scholarship has questioned the homoge-
neity of Islamist parties and their electoral platforms (Ayoob 2004; Kurzman and Naqvi
2010; Langohr 2001), so too must analysts of electoral behavior take a more sophisti-
cated approach to understanding Muslim voters at the ballot box.
In this chapter, I argue that voting for Islam is a more complicated and conditional
phenomenon than the assumption of Islam as a programmatic identity implies. To make
this argument, I distinguish among several different linkages between voters and polit-
ical parties that might generate votes for Islamic parties: religious ideology, party brand,
and demographic association. Voters face choices among religious and nonreligious
parties that bundle together various appeals, only some of which are directly tied to re-
ligion, and voters may vote for parties either out of policy concerns or as an expression
of their identity. The central implication of this argument is that voting for an Islamic
party is not always a vote for Islam, and voting for a non-​Islamic party sometimes is.
This warrants caution in interpreting popular support for religious parties as evidence
of popular support for religious agendas. It also warrants caution in interpreting the
success of non-​Islamic parties as a defense against religious agendas.
By taking seriously the linkages between voters and the parties that they choose, this
argument treats elections in Muslim-​majority countries as fundamentally comparable
252   Thomas Pepinsky

to elections in other countries with religiously motivated votes. In fact, the arguments
in this chapter apply—​mutatis mutandis—​to any electoral regime in which at least one
major political party either competes on a religious platform or tends to be identified
with a religious community. In much the same way that evangelical Christians may sup-
port the Republican Party in the United States, or Catholics may support the Law and
Justice Party in Poland, so too may religious Muslims vote for nonconfessional parties.
Inferring the role of individual religion or religiosity in explaining electoral behavior
requires attention to the complex array of nonreligious factors that shape party com-
petition. In the conclusion to this chapter, therefore, I outline a research agenda that
situates electoral behavior and voter characteristics (demographic attributes, religious
beliefs) within their broader sociological, historical, and institutional context.
I build this argument with special reference to Islam and Islamic parties in Indonesia
and Malaysia, two Muslim-​majority countries with long histories of both competitive
and uncompetitive elections among Islamic, nationalist, and other kinds of political
parties. Close engagement with these two cases allows for a careful consideration of how
religious and nonreligious factors shape party strategy and voter responses to it using
the wealth of qualitative and historical data available about religion and partisan compe-
tition in these two countries. That said, all of the conceptual arguments presented here
generalize beyond these two countries, to any election in a Muslim-​majority country
with elections that feature parties that identify as Muslim parties, either explicitly or
implicitly.
Before proceeding, a note on terminology is in order. In this chapter I will use the
term Islamic to refer to any party, organization, or movement that either implicitly or
explicit invokes Islam as part of its history, foundation, or platform. I will reserve the
term Islamist to refer to parties, organizations, or movements that explicitly seek to align
national politics with Islamic principles (compare to Masoud 2014, 1; Pepinsky, Liddle,
and Mujani 2018, 25). As an empirical matter, such labels are always contested, parties
may evolve over time, and partisan platforms may be opaque. But the conceptual dis-
tinction between religious identity and religious program is important for the argument
that follows.

Islam and Electoral Politics

The absence of competitive elections in many Muslim-​majority countries means that

the literature on religion and voting behavior in Muslim countries is comparatively
less developed than the literature on religion and voting behavior in Europe and North
America (for recent reviews, see Esmer and Pettersson 2007; Manza and Wright 2003).
There are many questions of interest to scholars of religion and voting behavior, but for
the purposes of this chapter, I confine my overview to perspectives that explain why
voters choose religious or nonreligious parties. Broadly speaking, there are three ways to
conceptualize the linkages between a party’s religious appeals and the choices that voters
 Ideologies, Brands, and Demographics in Muslim Southeast Asia    253

make: religious ideology, party brand, and demographic association. Each of these has
analogues in the literature on partisan competition beyond the Muslim world, and in
what follows I make brief mention of the wider literatures on electoral behavior from the
United States and Europe in order to link these perspectives to the general comparative
literature.
The most straightforward link between religiosity and partisan vote choice runs
through religious ideology. Here, the argument is that voters choose religious parties be-
cause they have religious preferences. Applied to voting behavior in the Muslim world,
this perspective holds that votes for Islamic parties depend on the beliefs that Muslims
have about the role of Islam in political life, and that voters choose parties that promise
to achieve such goals. In the same way that an individual who holds economically con-
servative views may vote for a conservative political party that campaigns on conser-
vative economic principles, so too may voters who hold religious views vote for a party
that espouses religious views. This has long been the standard view in the study of Islam
and democracy. Based on an impressionistic overview of the elections of Mahmoud
Ahmadinejad in Iran, Hamas in Palestine, and the Justice and Development Party
(AKP) in Turkey in the early 2000s, for example, Esmer and Pettersson conclude that
“Religiosity is a major factor influencing voting behavior throughout the Islamic world”
(2007, 497). In this, the effects of religiosity on voting behavior in the Muslim world do
not follow the predictions of the secularization thesis (see Norris and Inglehart 2004),
either because secularization is not happening in the first place, or because Muslims do
not respond to secularization in the way that Christians do.
It is surely true that some religious voters choose religious parties because they
support these parties’ religious platforms. But there is a wealth of evidence from be-
yond the Muslim world suggesting that voters do not vote for parties or for candidates
based on their policy platforms, or indeed based on any knowledge of what parties’
and candidates’ actual positions even are (for an early articulation of this view in the
US context, see Campbell et al. 1960). An alternative perspective is that voters choose
parties based a complex mix of policy beliefs and reputation that amount to a “party
brand” (see Lupu 2016). Applied to the case of religious parties and vote choice, this
linkage holds that parties may have religious identities, but voters choose them because
they believe that these parties represents religion as well as other things that they care
about beyond religion itself. This is most apparent with Christian Democratic parties in
Europe as parties of the center right, chosen by voters often with no reference to their re-
ligious heritage, even though that heritage exists and is explicit in their party platforms.1
Observing that a voter identifies as a Christian and votes for a Christian Democratic
party does not convey much information about what drives her or his vote choice, be-
cause Christian Democratic parties have broader economic and social platforms, with
positions on issues ranging from economic management to welfare policy to women’s
and minority rights and status. So, too, do parties in Muslim majority countries, and
this observation underlies a long line of research on the source of Islam’s political ad-
vantage (see Cammett and Jones Luong 2014 for a review). The implication is that voters
may support Islamic parties because of their party brands, which may be conservative,
254   Thomas Pepinsky

progressive, or reformist—​or business-​or social justice–​oriented. Equivalently, some

voters may support non-​Islamic parties because they believe that these parties have la-
tent Islamic brands.
Still another perspective draws attention not to parties’ policy positions, but rather to
the demographic characteristics of voters. In plural societies with political systems in
which confessional, linguistic, ethnic, or other kinds of parties represent different so-
cial groups, voters may choose parties because they represent their group. The simplest
articulation of this perspective emerges in the literature on consociationalism in the
Netherlands, where under the model of pillarization that predominated prior to the
1960s, religious communities voted for parties that represented them (Lijphart 1975).
As religious attendance and identification with a particular faith community waned, so
in turn did votes for confessional parties (van der Eijk and Niemöller 1987). Applied to
religious voting in Muslim-​majority (but still religiously heterogeneous) societies, this
argument implies that Muslims vote for Islamic parties they are the parties that repre-
sent their coreligionists.
These explanations for vote choice are obviously interrelated, and voters may have
complex and multifaceted explanations for the choices that they make. The distinc-
tion between demographic association and religious ideology, for example, may be
hard to discern among many voters. Are they voting for religious parties because
they believe these parties will implement religious policies that benefit their religious
group, or are they voting for religious parties because their existence represents the
group’s general interests regardless of the policies they enact? Likewise, the distinc-
tion between party brand and demographic association may be particularly subtle
when it comes to ethnic parties. If ethnicity and religion overlap, then Muslim voters
may choose ethnic parties because they know that their religious identities will also
be reflected in the policy priorities of such ethnically based parties. Despite these and
other empirical challenges in identifying the determinants of voting for Islamic and
non-​Islamic parties, these conceptual distinctions help to reveal the subtle issues in-
volved in assessing why voters choose particular political parties, both in the Muslim
world and beyond.
It is also important to distinguish religious ideology, party brand, and demographic
association from explanations focusing more narrowly on partisan identity. More than
simply a “running tally of party performance, ideological beliefs, and proximity to the
party in terms of one’s preferred policies,” partisan identity is an expression of identi-
fication with the party itself as the in-​group (Huddy, Mason, and Aarøe 2015). Fowler
(2020) provides a recent synthesis of the literature and some evidence on what he terms
the “partisan intoxication” hypothesis in the United States. This approach differs in that
it locates the motivation for vote choice in identification with the party rather than with
the community that that party claims to represent, as in the demographic association
perspective. Applied to other cases of confessional parties, it would imply that voters
support a party such as the Catholic People’s Party in the Netherlands out of identifica-
tion with the party rather than with the Dutch Catholic community; or in Israel, voting
 Ideologies, Brands, and Demographics in Muslim Southeast Asia    255

for Shas because of identification with the party rather than with the Haredi commu-
nity that it aims to represent. Although I do not rule out that there may be intoxicated
partisans in the Muslim world who vote for (or against) Islamic parties purely for
reasons of expressive partisanship, I do not entertain this approach any further in this
chapter.
Finally, note as well that the three competing perspectives outlined above ignore
completely the role of mobilization and party organization in explaining vote choice.
Plainly, voting for Islam does not occur in a vacuum: voters are mobilized by political
actors who make deliberate religious and other appeals, through organizational efforts
in mosques and elsewhere (Brooke 2019; Cammett and Issar 2010; Hamayotsu 2012;
Wickham 2002; Wiktorowicz 2003), and occasionally bought off directly (Aspinall and
Berenschot 2019). Moreover, as Kalyvas (1996) observes in his criticism of the early lit-
erature on Christian democracy in Europe, the very existence of religious parties also
requires explanation. Although a full account of voting for Islam would encompass all
of these steps, putting individual votes in their sociological and historical context, I keep
the analytical focus on individual decisions, given an existing set of partisan options. In
the conclusion, I revisit the roles of history, mobilization, and elite politicking in our ac-
counts of voting for Islamic parties.

Party Competition in the

Muslim World

As noted previously, the literature on religion and voting behavior in the Muslim world
is far less developed relative to the literature in Europe, the Americas, and other ad-
vanced industrial democracies. One simple explanation for this pattern is that elections
held in Muslim-​majority countries are far less likely to be free and fair than elections
in the rest of the world. To establish this point, Table 12.1 compares elections using data
on all elections around the world from the National Elections across Democracy and
Autocracy (NELDA) dataset (Hyde and Marinov 2012), as well as data on Muslim popu-
lation share across countries from the Pew Research Center (2017).
The NELDA dataset attempts to provide a standard metric of whether or not elections
are competitive or not, coding elections for national office across the world between
1945 and 2006. Data were coded using secondary sources by a team of trained research
assistants. More updated data from 2007 on would probably not change our inferences
given the continued concerns about democracy in the Muslim world (and beyond). For
the purposes of the analysis below, Muslim-​majority countries are all countries whose
Muslim population share exceeds 50 percent.2
I examine both the presence of multiparty elections (NELDA4) and whether or not
there are concerns about the freedom and fairness of elections (NELDA11). The dataset
256   Thomas Pepinsky

Table 12.1: Electoral Competitiveness in Muslim and Non-​Muslim Countries

Concerns about Electoral Freedom and
More than One Party Is Legal (NELDA4) Fairness (NELDA11)

Non-​Muslim Muslim Non-​Muslim Muslim

No 253 193 No 1,586 217

% 10.6 29.07 % 66.47 32.68
Unclear 5 2 Unclear 30 20
% 0.21 0.3 % 1.26 3.01
Yes 2,128 469 Yes 770 427
% 89.19 70.63 % 32.27 64.31
Total 2,386 664 Total 2,386 664

Source: Hyde and Marinov (2012).

contains other variables that capture electoral competitiveness, and a similar analysis
with those will produce comparable results, but I choose these two because they repre-
sent both an objective measure of whether partisan competition exists (NELDA4) and a
subjective evaluation of the state of electoral freedom (NELDA11). That these figures are
so comparable is reassuring.
From Table 12.1, it is evident that elections in Muslim-​majority countries are simply
less likely to be competitive referenda across multiple parties than are elections in non-​
Muslim-​majority countries.3 Table 12.2 tabulates the number of multiparty elections
(NELDA4 = “Yes”) in Muslim-​majority countries, differentiating among levels of con-
cern about electoral freedom and fairness (NELDA11).
The data in Table 12.2 have important implications for how we conceptualize voting
behavior in Muslim-​majority countries where voters have a choice among different
candidates. Even under these circumstances, voters frequently must anticipate that
there is some chance that their votes will not be fairly counted. Even in countries with
long histories of elections (such as Egypt, Iran, and Pakistan), restrictions on who can
compete, how parties may campaign, and the likelihood of electoral interference make it
difficult to ascertain linkages between voters and parties. In Egypt, for example, explic-
itly Islamist parties are popular, but under the regime of Hosni Mubarak, Islamists were
unable to form parties. In Iran, by contrast, an unelected council of religious figures
may prevent candidates from competing in elections. Among Muslim-​majority coun-
tries with extensive histories of elections, Burkina Faso, Indonesia, Jordan, Lebanon,
Malaysia, and Turkey are notable for the relative prevalence of competitive multiparty
elections. But even in these cases it is important to remember that these elections usu-
ally still fail to meet the basic standards of democratic competition in the minimalist
sense (Alvarez et al. 1996; Przeworski 1999).
 Ideologies, Brands, and Demographics in Muslim Southeast Asia    257

Table 12.2: Concerns about Electoral Freedom and Fairness in the Muslim World
Country No Unclear Yes Country No Unclear Yes

Afghanistan 1 0 1 Lebanon 11 0 4
Albania 0 2 10 Libya 0 0 2
Algeria 1 0 4 Malaysia 9 0 3
Azerbaijan 0 0 7 Mali 2 0 7
Bahrain 2 0 1 Mauritania 2 0 6
Bangladesh 3 0 6 Morocco 5 0 5
Burkina Faso 5 3 0 Niger 1 2 4
Chad 3 0 1 Nigeria 5 0 6
Comoros 4 0 3 Pakistan 2 0 7
Djibouti 0 0 4 Senegal 5 0 4
Egypt 2 0 17 Sierra Leone 4 0 4
Gambia 8 0 2 Somalia 2 0 0
Guinea 0 0 2 Sudan (2012–​) 2 0 2
Indonesia 7 0 3 Syria 2 2 3
Iran 0 1 18 Tajikistan 0 0 7
Iraq 1 0 7 Tunisia 1 0 8
Jordan 8 0 2 Turkey 15 0 7
Kazakhstan 0 0 9 Turkmenistan 1 0 0
Kuwait 1 0 1 Uzbekistan 0 0 6
Kyrgyzstan 2 0 6 Yemen 2 0 1
Total 119 10 190

Source: Hyde and Marinov (2012).

Party Competition in Muslim

Southeast Asia

With this background in mind, I turn to a closer analysis of elections and partisan com-
petition in Indonesia and Malaysia. I focus on these two countries for several reasons.
First, they are both representative cases of party politics in the Muslim world, each
having a long history of electoral competition under democratic, competitive authori-
tarian, and hard authoritarian rule. The two cases also differ usefully from one another,
both demographically and institutionally, which helpfully increases the variation in the
ways that Islamic, Islamist, and other parties compete for votes. And indeed, a common
258   Thomas Pepinsky

Table 12.3: Fractionalization Indices

Religion Ethnicity Language

Indonesia 0.23 0.74 0.77

Malaysia 0.67 0.59 0.60

Source: Alesina et al. (2003).

theme in the discussion of partisan competition and the determinants of vote choice is
that the very diversity of party competition makes them ideally suited for illuminating
the difficulty of inferring why people support Islamic and Islamist parties. Finally, a
focus on Southeast Asian Islam usefully expands the study of political Islam beyond the
Middle East and Arab world, and counters the implicit assumption among many who
work on Islam and politics that “Southeast Asian Islam is not real Islam” (Azra 2003, 39).
Both Indonesia and Malaysia have majority Muslim populations, but with different
ethnic and religious compositions. Indonesia is 88 percent Muslim, but the Indonesian
Constitution does not give special consideration to Islam over other religions. Malaysia
is 60 percent Muslim, and Islam is the country’s official religion. The two countries like-
wise differ in ethnic and religious heterogeneity. To capture this, I used fractionalization
indices, which provide a summary measure of how heterogeneous a population is across
some group dimension. They are calculated as

i
Fractionalization = 1 − ∑ si2
N

where si is the share of group i within the population. An entirely homogeneous pop-
ulation with only one ethnic group would have an ethnic fractionalization score of
1 – 12 = 0. A population with three equally-​sized religious groups would have a religious
fractionalization score of 1 – (.332 + .332 +.332) = .67. And a population where one lin-
guistic group comprises 80 percent of the population and three others comprise 10 per-
cent, 8 percent, and 2 percent of the population would have a linguistic fractionalization
score of 1 – (.82 + .12 +.082 + .022) = .34. Table 12.3 (using data from Alesina et al. 2003)
provides indices of fractionalization by religion, ethnicity, and language in each country.
These figures confirm that both countries are diverse, plural societies. But Indonesian
diversity is characterized by ethnolinguistic differences among Muslims (low religious
diversity, high ethnic and linguistic diversity), whereas Malaysia’s is characterized by the
overlap between religious and ethnolinguistic difference (common levels of religious,
ethnic, and linguistic diversity).4

Partisan Competition in Indonesia

Indonesia’s post-​independence political history can be divided into four periods: Liberal
Democracy (1949–​1957), Guided Democracy (1957–​1966), the New Order (1966–​1999),
 Ideologies, Brands, and Demographics in Muslim Southeast Asia    259

and the current democratic period (1999–​the present). A constant theme in Indonesian
party politics, both during periods of democratic competition and under authori-
tarian rule, is a cleavage between Islamic parties and movements and their nationalist
opponents. However, this cleavage is only one of many in Indonesian politics and so-
ciety, a fact that has implications for how to interpret votes for and against explicitly
Islamic parties. Table 12.3 summarizes the four periods and identifies the main political
parties, both Islamic and nationalist, in each.
Indonesia’s short experience with Liberal Democracy in the 1950s featured a complex
mix of partisan cleavages that reflected the social complexity of a newly independent
state that was still in the process of developing a national identity and establishing uni-
fied control over its territory. President Sukarno’s Indonesian Nationalist Party (PNI)
represented one centrist faction that united Muslims and non-​Muslims on a broadly
nationalist platform emphasizing development and territorial unity. On its left stood
that Indonesian Communist Party, one of the largest Communist parties anywhere
in the world, which represented a more secular vision for Indonesia’s political fu-
ture. Two separate parties represented Muslim interests explicitly. Masjumi (short for
Majelis Syuro Muslimin, or Council of Muslim Organizations) and Nahdlatul Ulama
(NU, or Awakening of the Ulama). The former traces its organizational lineage to the
Japanese occupation, as an organization representing Muslim elites and interests; the
latter emerged as a party in a split from Masjumi but does claim an organizational her-
itage that dates to the late Dutch colonial period. NU and Masjumi differed sociologi-
cally in that the former invokes traditionalist (and especially Javanese) Muslim beliefs,
whereas the latter is more modernist in orientation. Importantly, the process of writing
the Indonesian constitution had occasioned fierce debate about the role of Islam in
Indonesian politics, but the final outcome of this process saw no role for Islam itself, and
instead emphasized the generic religiosity of Indonesian politics under the banner of
belief in one god.

Table 12.4: Periodizing Indonesian Politics

Era Years Non-​Islamic Parties Islamic Parties

Liberal Democracy 1949–​1957 Communist Party of Nahdlatul Ulama (NU),

Indonesia (PKI), Indonesian Masjumi
Nationalist Party (PNI)
Guided Democracy 1957–​1966 PKI, Functional Groups NU marginalized, Masjumi
(Golkar) banned
New Order 1966–​1999 Golkar, Indonesian United Development Party
Democratic Party (PDI) (PPP)
Contemporary 1999–​present Golkar, PDI-​P, Democrat PPP, Prosperous Justice
Democracy Party, People’s Conscience Party (PKS), National
Party, Great Indonesia Awakening Party (PKB),
Movement Party, others National Mandate Party
(PAN), others
260   Thomas Pepinsky

Indonesia’s 1955 parliamentary elections saw PKI, PNI, NU, and Masjumi each taking
approximately 20 percent of the seats—​leaving no party with anything close to a ma-
jority or even a clear plurality (van der Kroef 1957). Amid general dissatisfaction with
the divided and fractious parliamentary order, Sukarno brought the Liberal Democratic
period to an end by proclaiming the establishment of Guided Democracy, a system that
would ostensibly resolve the contradictions inherent in multiparty democracy. Sukarno
banned Masjumi in 1960, and NU’s influence as a party declined during this period.
Guided Democracy, in turn, ended abruptly when Suharto seized power in 1966 in the
aftermath of the 1965 coup and the ensuing extermination of the PKI. Suharto’s New
Order regime used Golkar—​a corporatist mass organization created under Guided
Democracy to represent different groups in Indonesian society—​as its mobilizational
vehicle (see Reeve 1985). Following elections in 1971, Suharto ordered all opposition
parties to merge into one of two official opposition parties: the United Development
Party (PPP) to represent Islamic interests, and the Indonesian Democratic Party (PDI)
to represent nationalist (and non-​Muslim) interests.
Following Suharto’s resignation and Indonesia’s subsequent democratization, Golkar
formally became a political party, and was joined not only by PPP and the PDI-​P (suc-
cessor to PDI), but also by a wide range of parties across the political spectrum (aside
from the Communist Left). Today, the Prosperous Justice Party (PKS) campaigns on
Islamist and developmentalist principles, whereas the Crescent Star Party (PBB) claims
the heritage of Masjumi. Two other Islamic parties represent the interests of Indonesia’s
two largest Muslim organizations: the National Awakening Party (PKB) is associated
with (but organizationally independent from) NU, and the National Mandate Party
is similarly associated with but independent from Muhammadiyah, NU’s modernist
Muslim counterpart. Other nationalist parties have emerged as well, largely built as
the personal vehicles of aspiring presidential candidates: the Democrat Party for Susilo
Bambang Yudhoyono, the People’s Conscience Party for Wiranto, the Great Indonesia
Movement Party for Prabowo Subianto, and others.
This brief historical overview of party development in Indonesia immediately reveals
the different ways that Islamic parties capitalize on religious votes. Islamist parties
like PKS and PBB campaign on religious principles, although these are coupled with
developmentalist platforms. PKB and PAN have associational links with two large re-
ligious organizations representing distinct groupings of Indonesian Muslims. As I will
discuss later, all other parties of any national standing also have primarily Muslim
constituencies, and these nationalist parties may also appeal to religious Muslims.

Partisan Competition in Malaysia

Malaysia’s post-​independence political history also falls into distinct periods, although
the country’s form of electoral authoritarianism differs substantially from the New
Order and Guided Democracy models in Indonesia, as do the cleavage structures that
underlie Malaysian politics and society. Paralleling the discussion of Indonesia above,
 Ideologies, Brands, and Demographics in Muslim Southeast Asia    261

Table 12.5: Periodizing Malaysian Politics

Era Years Non-​Islamic Parties Islamic Parties

Alliance Rule 1957–​1969 United Malays National Pan-​Malaysian Islamic

Organisation (UMNO, Malaysian Party (PAS)
Chinese Association (MCA),
Malaysian Indian Congress (MIC),
Democratic Action Party (DAP),
Malaysian People’s Movement
(Gerakan)
National Operations 1969–​1971 Parliament suspended
Council
Barisan Nasional 1971–​1999 UMNO, MCA, MIC, DAP PAS
(Pre-​Crisis)
Barisan Nasional 1999–​2018 UMNO, MCA, MIC, DAP, People’s PAS
(Post-​Crisis) Justice Party (PKR)
Contemporary 2018–​present UMNO, MCA, MIC, DAP, PKR, PAS, National Trust
Democracy Malaysian United Indigenous Party (Amanah)
Party (Bersatu)

Table 12.5 summarizes the four periods and identifies the main political parties, both
Islamic and nationalist, in each.
The years immediately following independence saw the Federation of Malaya (then
excluding the states of Sabah and Sarawak in East Malaysia) ruled by the Alliance, a
coalition of parties representing each of Malaya’s three main ethnic groups: the United
Malays National Organisation (UMNO) for Malays, the Malayan Chinese Association
(MCA), and the Malayan Indian Congress (MIC).5 Each party’s membership is de-
fined ethnically, but religious affiliation is implicit, as the Malaysian Constitution
defines Malayness with reference to Islam (Hirschman 1987, 555). Since all Malays are
Muslims, and the party’s constituency is Malay,6 it is a party that represents Muslim
interests. In opposition sat three important parties: the Democratic Action Party (DAP)
representing social democratic principles, the Malaysian People’s Movement (Gerakan)
representing liberal principles, and the Pan-​Malayan Islamic Party (PAS). Importantly,
although none of these parties is defined by an ethnic constituency, each disproportion-
ately captured particular ethnic groups. DAP and Gerakan remained largely but not ex-
clusively Chinese in membership, and PAS largely but not exclusively Malay. A roster of
much smaller parties, some ethnic but most not, tended to capture votes of Malaysian
Indians and non-​Malay bumiputeras in East Malaysia.
After ethnic riots following the 1969 parliamentary elections (Goh 1971), UMNO
leaders suspended parliament and declared a state of emergency. Malaysia returned to
parliamentary democracy in 1971, with a newly expanded ruling coalition called the
Barisan Nasional (or National Front, BN)—​now including both Gerakan and PAS as
well as other smaller parties. In 1977 PAS left the BN, joining DAP in the opposition. The
262   Thomas Pepinsky

Malaysian partisan landscape would remain roughly fixed until 1999, when, in the after-
math of the Asian financial crisis, and a leadership struggle within UMNO, the National
Justice Party (later the People’s Justice Party, or PKR). This party is explicitly multiethnic
in character, and joined both PAS and DAP in the country’s first—​but short-​lived—​
opposition coalition (Weiss 1999).
Further factional struggles both within UMNO and within PAS led to the birth of
two new important parties in the 2010s: the National Trust Party (Amanah), an Islamist
party comprising progressives from PAS; and the Malaysian United Indigenous Party
(Bersatu), a Malay party comprising old guard UMNO politicians dissatisfied with the
leadership of Prime Minister Najib Razak. In the 2018 general elections, a coalition of
the DAP, PKR, Amanah, and Bersatu formed Pakatan Harapan (Alliance of Hope, PH),
which finally defeated the Barisan Nasional. PH currently forms the government (as of
2020), and the peaceful transition of power through elections from BN to PR was an
auspicious sign of democratization (Lemière 2018).
Malaysian partisan politics differs substantially from Indonesia, owing first and fore-
most to the explicit politicization of ethnicity, in turn reinforced through the party
system. But because ethnicity overlays religion in Malaysia, Malay parties, such as the
long-​ruling UMNO, are effectively parties that represent nearly exclusively Muslim
constituencies. Moreover, even though Malaysia has Islamist parties just as Indonesia
does, they also function effectively as ethnic parties in a way that Indonesia’s do not.

Voting for Islam in Muslim

Southeast Asia

As noted previously, the literature on voting behavior and partisan competition in the
Muslim world is underdeveloped. This is also the case for Indonesian and Malaysian
politics. Although both countries have been part of the East Asia Barometer since
Wave 2 (Welsh et al. 2007), I am aware of no publications using these data that have
sought to explain votes for Islamic versus other parties. Political conditions in each
country have also constrained research. In Malaysia, academic research on voting be-
havior has been politically difficult, although polling firms such as the Merdeka Centre
for Public Opinion Research have been able to produce reports that have implications
for research on Islam and politics (e.g., Martinez 2006). By contrast, since Indonesian
democratization, there has been an explosion of public opinion polling (Mietzner
2009), with increasing involvement by social scientists seeking to understand what
explains voter support for political parties (see, e.g., Mujani, Liddle, and Ambardi
2018). However, the observation that Islamic parties have not managed to grow in size
and influence relative to their position in 1955 (see Table 12.6) leads most Indonesia
specialists to conclude that the drivers of vote choice in Indonesia must lie outside of
religiosity.
 Ideologies, Brands, and Demographics in Muslim Southeast Asia    263

Table 12.6: Parliamentary Vote Share in Indonesia, 1955 and 2014

Year Islamic Parties Other Parties

1955 43.9 55.9

2014 31.4 68.6

Source: https://​bit.ly/​2FX2TxR and https://​bit.ly/​2K0hyMz.

The task of disentangling religious ideology from other drivers of vote choice is a dif-
ficult research design problem. A naïve approach would be to calculate the conditional
correlation between a measure of religious identity or religiosity and whether or not a
voter supports an Islamic party, but for the reasons described in the previous sections,
this is unlikely to be informative if parties differ for reasons other than just their re-
ligious ideology. Indeed, Liddle and Mujani (2007) adopt this approach and find evi-
dence that more religious Muslims in Indonesia are more likely to support some Islamic
parties (specifically, PPP or PKB but not PAN or PKS) relative to PDI-​P, the most na-
tionalist7 of democratic Indonesia’s political parties.
The first study to outline the inferential challenges associated with inferring
motivations for vote choice in the Muslim world was Pepinsky, Liddle, and Mujani
(2012), which developed an experimental approach to disentangle the effects of re-
ligious ideology and economic motivations for supporting political parties. Their
finding—​that under conditions of economic policy uncertainty, Indonesia’s Islamist
parties have a small but statistically significant advantage over non-​Islamist parties—​
shows that nonreligious factors drive support for Islamist parties. They explain this
finding by invoking a logic of cueing (similar to branding), in which voters associate
Islamic parties with normatively good features, but choose them only when they are
unable to evaluate parties on other dimensions. But isolating the effects of religious
ideology from other factors requires presenting survey respondents with hypothetical
parties rather than actual ones, and is incapable of assessing the relative weight of dif-
ferent linkages relative to one another in the aggregate, or even of inferring individual
voter motives at all.
In sum, the growing literature on religion and vote choice in Indonesia finds some
correlations between religiosity and religious identity and support for Islamic parties,
but more methodologically sophisticated approaches reveal that it is hard to disentangle
support for Islamic parties’ religious platforms versus their other platforms.
The interaction between ethnicity, religion, and partisan competition in Malaysia
makes this task still more difficult. Absent individual-​level data, most analyses of
electoral behavior have used ecological correlations that estimate the correlations
between constituency characteristics and party vote shares (Ng et al. 2015; Pepinsky
2009, 2015). But because of the definitional equivalence of Malay and Muslim under
the Malaysian Constitution, most research exploring the drivers of support for
264   Thomas Pepinsky

Malaysia’s incumbent and opposition parties focuses on whether ethnicity or urban-

ization (a proxy for broader processes of modernization) explains vote choice. There
is limited research on religiosity per se, and what exists is methodologically unso-
phisticated (Welsh 2014). The fact that UMNO and PAS tended—​prior to 2018—​to
run head-​to-​head in constituencies with large Malay majorities might suggest that
it would be possible to explore the correlation between religiosity and vote choice in
this subset of electoral contests, but even here, party lineages differ across states for
historical reasons (Funston 1980), confounding any comparison of religiosity among
Malays. And, moreover, PAS and UMNO differ across multiple dimensions aside
from their Islamist character. UMNO under BN rule was a hegemonic party with vast
resources of patronage at its disposal and an agenda that focused explicitly on Malay
rights and protection of the political and economic status quo, whereas PAS is an op-
position party with reformist ambitions, comparatively limited access to patronage,
and an agenda that emphasizes Malay rights only indirectly. A vote for PAS might be a
signal of northeast peninsular regional identity (evidence of the demographic associ-
ation linkage), a demand for reform with anti-​regime ambitions (evidence of the party
brand linkage), or a signal of support for its religious platform.
This discussion illustrates just how difficult it is to infer what linkages are at work
from observed votes for Islamic parties. Even when precisely articulated using a re-
search design that can identify the effects of party brands independently of religious ide-
ology, conclusions are necessarily limited in scope. For similar reasons, it is also difficult
to infer voter motivations from observed votes for non-​Islamic parties. As Baswedan
(2004) observed early in Indonesia’s contemporary democratic period, Golkar built a
constituency among religious Muslims in Indonesia beginning in the late New Order
period, with particular strength among religious Muslims outside of Java. Although the
party does not advocate for the implementation of Islamic law, as Buehler (2013, 2016)
has shown, at the subnational level, politicians from Golkar are frequently responsible
for implementing Islamic regulations. A voter seeking religious policy may therefore
choose Golkar instead of an Islamic party due to the party’s ability and willingness to
implement religious regulations. An implication of this finding is that we may observe
a negative correlation between individual religiosity and support for explicitly Islamist
parties in Indonesia, even though voters are explicitly voting for religious platforms by
voting for the officially non-​Islamist Golkar Party.
Likewise, in Malaysia, the fact that PAS sat opposition to the BN, and collaborated
with largely non-​Malay parties with multicultural and social democratic platforms,
raises the possibility that a strict Islamist voter concerned primarily with religious ide-
ology would vote against PAS and for UMNO. Liow (2003, 2009) and Weiss (2004)
describe how UMNO has moved to adopt the Islamic agenda held by PAS as a way to
preempt challenges to its religious credentials.
 Ideologies, Brands, and Demographics in Muslim Southeast Asia    265

An Agenda

This chapter has outlined a framework for conceptualizing the links between Islam and
voting behavior. Its main conclusions are negative. Because there are multiple linkages
between religion and voting behavior, when facing the complexity of partisan com-
petition in plural societies with multiparty elections in countries like Indonesia and
Malaysia, it is hard to conclude anything about voter intentions based on their decision
to vote for or against an Islamic party.
One might conclude from my argument that inferring voter intentions is impossible.
Although I do conclude that scholars interested in Islamic politics must take seriously
the research design challenges inherent in understanding voter intentions in Muslim
societies, this certainly does not imply that voting behavior is irrelevant for under-
standing Muslim politics. To conclude with a more constructive message, a research
agenda on the links between religion and voting behavior might build on these uncer-
tain foundations in three ways.
The first is to proceed systematically, breaking down the broader problem of
characterizing the different linkages between religion and voting behavior into coherent
and answerable questions. Although approaches such as that of Pepinsky, Liddle, and
Mujani (2012) cannot on their own resolve all of these issues, they can at least answer
one question properly. Assembling multiple studies that tackle different aspects of this
broader research area is the first step in a progressive research agenda, but each will need
to focus on precise questions and appropriate counterfactuals in order to distinguish
among the many different explanations reviewed above for why voters choose different
parties. Unfortunately, the methods developed by Pepinsky et al. are not automatically
portable to all such questions. To identify empirically the role of demographic associa-
tion, for example, will require a method to manipulate respondents’ demographic char-
acteristics, perhaps through priming or otherwise making more salient one particular
aspect of their social identity (see Sen and Wasow 2016).
The second direction to expand is to look comparatively, focusing on the countries
listed in Table 12.2. Turkey in particular has featured prominently in the literature
on Islam and democratic politics, with Islamism associated both with religiosity and
with a particular class formation (see, e.g., Çarkoğlu and Toprak 2006; Öniş 2006;
Yavuz 1997). And yet the conceptual problems outlined here have not met with sat-
isfactory solutions—​the ruling Justice and Development Party embodies, in var-
ious accounts, Islamic, reformist, developmentalist, and pro-​business principles. In
the case of Egypt, by contrast, Masoud (2014) shows that the Muslim Brotherhood’s
mobilizational capacity and other social forces explain its success during Egypt’s
266   Thomas Pepinsky

brief experience with democracy in the early 2010s. Further extending this analysis
to understand what votes for non-​Islamic parties mean in regions like West Africa,
where few explicit Islamic parties compete in multiparty elections, would also prove
enlightening.
A third way to extend this research agenda is to question the very assumption
that linkages between voters and Islamic parties matter. Scholars of Islamic politics
may more profitably focus on party strategies, elite choices, and social movement
entrepreneurs. As my coauthors and I concluded in our analysis of piety and public
opinion in Indonesia (Pepinsky, Liddle, and Mujani 2018), there is little evidence that
Muslim religiosity explains anything interesting about Indonesian politics. To con-
ceptualize the role of Islam in Indonesian politics is to appreciate the autonomy of
politicians and the broader social forces shaping party competition rather than the
mundane politics of vote-​getting and political behavior, taking seriously Islam as an
identity claim rather than a set of individual beliefs and policy preferences.
Such an approach need not reject the linkages between voters and Islamic parties.
After all, in electoral regimes, parties must win votes, so understanding the activi-
ties of elites and entrepreneurs is necessary but not sufficient for understanding elec-
toral politics. Instead, it would envision a more dynamic relationship between elites
and masses, or between movement entrepreneurs and voters, than a static analysis
of what individual factors predict vote choice. Rather than asking whether religi-
osity or other demographic factors predict vote choice, researchers might consider
the conditions under which religious messages appeal to voters, or whether religious
elites prove to be more trustworthy messengers for non-​Islamic parties. Rather than
looking at vote intention or party identification, one might instead investigate the sta-
bility of partisan attachment, the decision to turn out, or the willingness to sell votes.
Such approaches embrace voting behavior as essential to understanding the role of
religion in electoral politics, but redirect inquiry away from general statements about
voter intentions and toward what are more analytically tractable electoral questions
(and, arguably, more politically consequential electoral phenomena) in Muslim-​
majority countries.

Notes
1. See, e.g., the program of the Christian Democratic Union in Germany, https://​www.cdu.de/​
grundsatzprogramm-​2007: “Orientierungsmaßstab ist das christliche Menschenbild und
davon ausgehend die drei Grundwerte ‘Freiheit, Solidarität und Gerechtigkeit’” (emphasis
added) [accessed April 2, 2019].
2. This coding rule thus includes Nigeria (Muslim population share is 50.4 percent) and
excludes Guinea-​Bissau (42.2 percent) and Bosnia and Herzegovina (40 percent).
3. These figures account for all types of elections—​legislative, executive, and for constituent
assemblies—​and remain substantively identical when the analysis is restricted to only legis-
lative/​parliamentary elections.
 Ideologies, Brands, and Demographics in Muslim Southeast Asia    267

4. See Hamayotsu (2002) for a comparative perspective on Islam and political development in
these two cases.
5. Both MCA and MIC would change their names from Malayan to Malaysian after 1963,
when the country became the Federation of Malaysia.
6. Non-​Muslim indigenous Malaysians (bumiputeras) may join UMNO, but in practice their
numbers are small and their influence negligible. The currently leadership of UMNO is en-
tirely Muslim; see https://​umno-​online.my/​majlistertinggi/​ [accessed April 2, 2019].
7. They use the term secular, but it is more accurate to describe this party as non-​Islamic.

References
Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat, and Romain
Wacziarg. 2003. “Fractionalization.” Journal of Economic Growth 8, no. 2): 155–​194.
Alvarez, Michael E., José Antonio Cheibub, Fernando Limongi, and Adam Przeworski. 1996.
“Classifying Political Regimes.” Studies in Comparative International Development 31, no.
2: 3–​36.
Aspinall, Edward, and Ward Berenschot. 2019. Democracy for Sale: Elections, Clientelism, and
the State in Indonesia. Ithaca, NY: Cornell University Press.
Ayoob, Mohammed. 2004. “Political Islam: Image and Reality.” World Policy Journal 21, no.
3: 1–​14.
Azra, Azyumardi. 2003. “Bali and Southeast Asian Islam.” In After Bali: The Threat of Terrorism
in Southeast Asia, edited by K. Ramakrishna and S. S. Tan, 46–​47. Singapore: World
Scientific.
Baswedan, Anies Rasyid. 2004. “Political Islam in Indonesia: Present and Future Trajectory.”
Asian Survey 44, no. 5: 669.
Brooke, Steven. 2019. Winning Hearts and Votes: Social Services and the Islamist Political
Advantage. Ithaca, NY: Cornell University Press.
Buehler, Michael. 2013. “Subnational Islamization through Secular Parties: Comparing Shari’a
Politics in Two Indonesian Provinces.” Comparative Politics 46, no. 1: 63–​82.
Buehler, Michael. 2016. The Politics of Shari’a Law: Islamist Activists and the State in
Democratizing Indonesia. New York: Cambridge University Press.
Cammett, Melani, and Sukriti Issar. 2010. “Bricks and Mortar Clientelism: Sectarianism and
the Logics of Welfare Allocation in Lebanon.” World Politics 62, no. 3: 381–​421.
Cammett, Melani, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17: 187–​206.
Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960. The
American Voter. Chicago: University of Chicago Press.
Çarkoğlu, Ali, and Binnaz Toprak. 2006. Religion, Society and Politics in Changing Turkey.
Translated by J. Sanders. Istanbul: Turkish Economic and Social Studies Foundation.
Esmer, Yilmaz, and Thorleif Pettersson. 2007. “The Effects of Religion and Religiosity on
Voting Behavior.” In The Oxford Handbook of Political Behavior, edited by R. J. Dalton and H.
D. Klingemann, 481–​503. New York: Oxford University Press.
Fowler, Anthony. 2020. “Partisan Intoxication or Policy Voting?” Quarterly Journal of Political
Science 15: 141–​179. https://​drive.google.com/​file/​d/​1StjsBztpHTYDErcKbgjNk0ujWXXDx
b7O/​view.
268   Thomas Pepinsky

Funston, N. J. 1980. Malay Politics in Malaysia: A Study of the United Malays National
Organisation and Party Islam. Kuala Lumpur: Heinemann Educational Books (Asia) Ltd.
Goh, Cheng Teik. 1971. The May Thirteenth Incident and Democracy in Malaysia. Kuala
Lumpur: Oxford University Press.
Hamayotsu, Kikue. 2002. “Islam and Nation Building in Southeast Asia: Malaysia and
Indonesia in Comparative Perspective.” Pacific Affairs 75, no. 3: 353–​375.
Hamayotsu, Kikue. 2012. “Bringing Clientelism and Institutions Back In: The Rise and Fall
of Religious Parties in Indonesia’s Electoral Democracy.” In Party Politics in Southeast
Asia: Clientelism and Electoral Competition in Indonesia, Thailand and the Philippines, edited
by D. Tomsa and A. Ufen, 138–​159. London: Routledge.
Hirschman, Charles. 1987. “The Meaning and Measurement of Ethnicity in Malaysia: An
Analysis of Census Classifications.” Journal of Asian Studies 46, no. 3: 555–​582.
Huddy, Leonie, Lilliana Mason, and Lene Aarøe. 2015. “Expressive Partisanship: Campaign
Involvement, Political Emotion, and Partisan Identity.” American Political Science Review
109, no. 1: 1–​17.
Hyde, Susan D., and Nikolay Marinov. 2012. “Which Elections Can Be Lost?” Political Analysis
20, no. 2: 191–​201.
Kalyvas, Stathis N. 1996. The Rise of Christian Democracy in Europe. Ithaca, NY: Cornell
University Press.
Kurzman, Charles, and Ijlal Naqvi. 2010. “Do Muslims Vote Islamic?” Journal of Democracy 21,
no. 2: 50–​63.
Langohr, Vickie. 2001. “Of Islamists and Ballot Boxes: Rethinking the Relationship between
Islamisms and Electoral Politics.” International Journal of Middle East Studies 33, no.
4: 591–​610.
Lemière, Sophie. 2018. “The Downfall of Malaysia’s Ruling Party.” Journal of Democracy 29, no.
4: 114–​128.
Liddle, R. William, and Saiful Mujani. 2007. “Leadership, Party, and Religion: Explaining
Voting Behavior in Indonesia.” Comparative Political Studies 40, no. 7: 832–​857.
Lijphart, Arend. 1975. The Politics of Accommodation: Pluralism and Democracy in the
Netherlands. 2nd ed. Berkeley: University of California Press.
Liow, Joseph Chinyong. 2003. “Deconstructing Political Islam in Malaysia: UMNO’s Response
to PAS’ Religio-​Political Dialectic.” RSIS Working Paper no. 45. Singapore: Nanyang
Technological University. https://​dr.ntu.edu.sg/​bitstream/​handle/​10220/​4447/​rsis-​workpaper_​
53.pdf?sequence=1.
Liow, Joseph Chinyong. 2009. Piety and Politics: Islamism in Contemporary Malaysia.
New York: Oxford University Press.
Lupu, Noam. 2016. Party Brands in Crisis: Partisanship, Brand Dilution, and the Breakdown of
Political Parties. New York: Cambridge University Press.
Manza, Jeff, and Nathan Wright. 2003. “Religion and Political Behavior.” In Handbook
of the Sociology of Religion, edited by M. Dillon, 297–​314. New York: Cambridge
University Press.
Martinez, Patricia. 2006. “Democracy, the ‘Islamic State,’ and Imbedded Realities in Malaysia.”
In Religious Organization and Democratization: Case Studies from Contemporary Asia, ed-
ited by T.-​J. Cheng and D. A. Brown, 242–​266. Armonk, NY: M.E. Sharpe.
Masoud, Tarek. 2014. Counting Islam: Religion, Class, and Elections in Egypt. New York:
Cambridge University Press.
 Ideologies, Brands, and Demographics in Muslim Southeast Asia    269

Mietzner, Marcus. 2009. “Political Opinion Polling in Post-​authoritarian Indonesia: Catalyst

or Obstacle to Democratic Consolidation?” Bijdragen tot de Taal-​, Land-​en Volkenkunde
165, no. 1): 95–​126.
Mujani, Saiful, R. William Liddle, and Kuskridho Ambardi. 2018. Voting Behavior in Indonesia
since Democratization: Critical Democrats. New York: Cambridge University Press.
Ng, Jason Wei Jian, Gary John Rangel, Santha Vaithilingam, and Subramaniam S. Pillay. 2015.
“2013 Malaysian Elections: Ethnic Politics or Urban Wave?” Journal of East Asian Studies 15,
no. 2: 167–​198.
Norris, Pippa, and Ronald Inglehart. 2004. Sacred and Secular: Religion and Politics Worldwide.
New York: Cambridge University Press.
Öniş, Ziya. 2006. “The Political Economy of Islam and Democracy in Turkey: From the
Welfare Party to the AKP.” In Democracy and Development: New Political Strategies for the
Middle East, edited by D. Jung, 103–​128. New York: Palgrave.
Pepinsky, Thomas B. 2009. “The 2008 Malaysian Elections: An End to Ethnic Politics?” Journal
of East Asian Studies 9, no. 1: 87–​120.
Pepinsky, Thomas B. 2015. “Interpreting Ethnicity and Urbanization in Malaysia’s 2013 General
Election.” Journal of East Asian Studies 15, no. 2: 199–​226.
Pepinsky, Thomas B., R. William Liddle, and Saiful Mujani. 2012. “Testing Islam's Political
Advantage: Evidence from Indonesia.” American Journal of Political Science 56, no. 3: 584–​600.
Pepinsky, Thomas B., R. William Liddle, and Saiful Mujani. 2018. Piety and Public Opinion:
Understanding Indonesian Islam. New York: Oxford University Press.
Pew Research Center. 2017. “Interactive Data Table: World Muslim Population by Country.”
Washington, DC: Pew Research Center. https://​www.pewforum.org/​chart/​interactive-​data-​
table-​world-​muslim-​population-​by-​country/​.
Przeworski, Adam. 1999. “Minimalist Conception of Democracy: A Defense.” In Democracy’s
Value, edited by I. Shapiro and C. Hacker-​Cordon, 23–​55. New York: Cambridge University
Press.
Reeve, David. 1985. Golkar of Indonesia: An Alternative to the Party System. Singapore: Oxford
University Press.
Sen, Maya, and Omar Wasow. 2016. “Race as a Bundle of Sticks: Designs that Estimate Effects
of Seemingly Immutable Characteristics.” Annual Review of Political Science 19: 499–​522.
van der Eijk, Cees, and Kees Niemöller. 1987. “Electoral Alignments in the Netherlands.”
Electoral Studies 6, no. 1: 17–​30.
van der Kroef, Justus M. 1957. “Indonesia’s First National Election: A Sociological Analysis.”
American Journal of Economics and Sociology 16, no. 3: 237–​49.
Weiss, Meredith L. 1999. “What Will Become of Reformasi? Ethnicity and Changing Political
Norms in Malaysia.” Contemporary Southeast Asia 21, no. 3: 424–​450.
Weiss, Meredith L. 2004. “The Changing Shape of Islamic Politics in Malaysia.” Journal of East
Asian Studies 4, no. 1: 139–​73.
Welsh, Bridget. 2014. “Elections in Malaysia: Voting Behaviour and Electoral Integrity.” In
Routledge Handbook of Contemporary Malaysia, edited by M. L. Weiss, 11–​21. New York:
Routledge.
Welsh, Bridget, Ibrahim Suffian, and Andrew Aeria. 2007. “Malaysia Country Report: Second
Wave of Asian Barometer Survey.” Asian Barometer Working Paper no. 46. Taipei: Asia
Barometer Project Office. http://​asianbarometer.org/​publications/​612d4dc6f0bdd1f668bb9
a0b89df7d70.pdf.
270   Thomas Pepinsky

Wickham, Carrie Rosefsky. 2002. Mobilizing Islam: Religion, Activism, and Political Change in
Egypt. New York: Columbia University Press.
Wiktorowicz, Quintan, ed. 2003. Islamic Activism: A Social Movement Theory Approach.
Bloomington: Indiana University Press.
Yavuz, M. Hakan. 1997. “Political Islam and the Welfare (Refah) Party in Turkey.” Comparative
Politics 30, no. 1: 63–​82.
 Chapter 13

Religion an d Pa rt y
P olitics in I ndia a nd
Pakista n

Steven I. Wilkinson

Religion has influenced party politics in India and Pakistan since independence in
1947 in at least three different ways. First, religious parties have succeeded at winning
elections directly, forming governments and implementing their own political and
social visions. Only since 1990 has the BJP (Bharatiya Janata Party, or Indian People’s
Party) been able to do this at the state and national level in India, and only since 2000
have religious parties in Pakistan been able to win elections at the provincial level. The
BJP is now, in Prime Minister Narendra Modi’s second term, using its increased par-
liamentary majority and control of many major Indian states to remake India’s gov-
ernment and institutions in Hindu Nationalist form, including reducing minorities to
second-​class status; banning conversions and stigmatizing interreligious relationships;
abolishing the special constitutional status of the Muslim-​majority state of Kashmir;
punishing the media, civil society organizations, and institutions (such as universities)
that oppose Modi; and trying to institutionalize a particular form of Hindu nationalist
identity in government institutions, education, and public discourse.
Second, even if not elected themselves, religious parties, movements, and voters
can put pressure on and influence the behavior of those in government, using street
protests, media protests, and informal influence. They can also shape the bounds of
acceptable political debate, turning issues such as free speech, the status of religious
minorities, or an individual’s right to convert to another religion or preach their own
into majoritarian ‘valence’ issues, where any politician who challenges the orthodox
perspective or defends the unpopular minority does so at their peril. In Pakistan this
kind of influence has been much stronger, for much longer, than in India, although
the presence of Prime Minister Modi at the center has, since 2014, emboldened many
Hindu nationalist activists to take an increasingly strident role on the streets and on
social media.
272   Steven I. Wilkinson

Third, particularly in authoritarian or semi-​authoritarian states, leaders may some-

times choose to institutionalize religion in laws or state institutions, often as a way of
seeking legitimacy that their own regime may lack. In Pakistan, this pattern was par-
ticularly evident in the 1970s and 1980s, first under Zulfikar Ali Bhutto (President 1971–​
73, Prime Minister 1973–​77), whose 1973 Constitution affirmed Pakistan as an Islamic
Republic committed to institutionalizing Islamic principles in government and society,
and who then passed a constitutional amendment in 1974 marginalizing the Ahmadi
minority. The religiously conservative military dictator who displaced Bhutto, General
Zia-​ul-​Haq (1977–​1988), went even further and institutionalized his vision of Islam in
the education system, army, courts, and state administration.
In this paper I explore the reasons for the shift over time in the role of religion in
party politics in both countries. In India, religion’s role was initially much more limited
than in Pakistan. Congress’s top leadership, and in particular Prime Minister Nehru,
traumatized by partition, were deeply opposed to structuring politics around com-
munal identities, and those on the Hindu right or Muslim opposition who might have
challenged Nehru on this had either left for Pakistan, were discredited by their stance on
partition, or were tainted by the Hindu nationalists’ role in the assassination of Mahatma
Gandhi in January 1948. The Indian Constitution of 1950 firmly restricted the role of re-
ligion in the state and political system, and key clauses in the constitution helped in-
stead to maximize the political importance of competing cross-​cutting cleavages such
as language and caste among the 82% Hindu majority as well as India’s Muslim (11%),
Sikh (2%) and Christian (2%) minorities. The national Congress Party, however, became
increasingly associated over time with corruption and poor performance, especially
after it abandoned internal party elections in 1971, which made the party less respon-
sive to local groups and interests. Most of Congress’s regional opponents and newly
assertive parties from backward and lower castes were just as corrupt and ineffective,
however. These parties could win in their home states but their narrow social bases,
interparty rivalries, and lack of a broader unifying ideology made it very difficult to re-
place Congress in forming a stable national opposition. The BJP has succeeded where
these other opposition parties failed for two major reasons. First, as we might expect
from a party that grew out of a disciplined India-​wide Hindu-​nationalist organization
going back to the 1920s, the BJP is itself a relatively disciplined party, and it has managed
to avoid the open party splits that have plagued its competitors. Second, after suffering
early reversals such as the 1994 state elections, the party has also learned to strategically
downplay its hard Hindu nationalist agenda and opposition to caste politics, when nec-
essary, and win elections instead by focusing on a message of good governance, less cor-
ruption, and more development that appeals to a wider swath of the electorate.
In Pakistan, the Islamic religious parties started out in a difficult position, because
the major parties, the Jamaʿat-​i Islami (JI) and Jamaʿat Ulema-​i-​Islam (JUI), had offered
only lukewarm support for the Muslim League’s Pakistan project, and another party
(the Ahrars) had been allied with the Congress, but they came to exercise more indi-
rect influence than religious parties in India. They did this because of their ability to
generate large-​scale protests in mosques and in the street, and because the weakness of
 Religion and Party Politics in India and Pakistan    273

the mainstream parties (unlike in India) meant that secular-​minded politicians lacked
a firm base from which to defend either their own principles or the rights of unpop-
ular minorities. Only the military, in the 1950s and 1960s when it was dominated by
secular-​minded officers at the top, was willing to rein in the Islamic parties when they
threatened public order over anti-​Ahmadi agitations (Abbas 2013, 6). This weakness in
political party institutionalization, so crucial to democracy elsewhere (Mainwaring and
Scully, 1995), is due to two things: (1) the continuing power of large landlords and busi-
ness interests whose power base is independent of the parties (Chacko 2020); and (2) the
role of the military, which has actively worked to subvert any political party that looked
as if it might emerge as a strong counterbalance to its own power.
The decisive period in defining state-​religion relations in Pakistan was under the
regimes of Prime Minister Bhutto (1971–​1977) and then, after a military coup in 1977, his
successor General Zia-​ul-​Haq (1977-​1988). Bhutto made common cause with the Sunni
religious parties in 1973–​1974—​enshrining the Islamic character of the state in the 1973
constitution, which also in an Amendment in 1974 declared Ahmadis non-​Muslims—​
to solidify his own power, but he never built up a genuine party organization capable
of supporting his ambitions to transform the economy and society and increasingly
resorted to authoritarian methods. Zia, personally committed to Islamic ideals and
also in need of broader political legitimacy after his coup, quickly moved to institution-
alize his conservative vision of Islam throughout Pakistan’s courts, education system,
administration, and even the army. The loose military-​religious party alliances begun
under Zia continue to this day, in part because both have an interest in undercutting the
larger nonreligious political parties (Chacko 2020). The religious parties were unsuc-
cessful at the polls for decades (Nasr, 1994), but have grown in influence partly due to
the increasing religious conservatism of the country, and partly because they have be-
come effective focal points to protest issues such as corruption or Pakistan’s unpopular
foreign policies. Coalitions of religious parties have formed provincial governments in
Khyber Pakhtunkhwa (KP) and Balochistan, where they succeeded by capitalizing on
mass opposition in those provinces to Pakistan’s military alliance with the United States,
its policy in Afghanistan and US drone strikes. But the religious parties, splintered
into many factions and performing poorly in delivering effective governance and de-
velopment themselves when in provincial government, have not yet been able to break
through at the national level, their twelve-​party coalition winning only 8% of the vote in
the most populous province of Punjab, and only 17% and 19%, respectively, in their tra-
ditional strongholds of KP and Balochistan at the last 2018 national assembly elections
(Jahangir 2018).
In the following sections, I first review the ways in which religion was historically
institutionalized by the colonial state in pre-​independence India, and the reasons why
the Muslim League pushed for partition. Second, I explore the reasons why, after inde-
pendence, the Congress Party in India strongly opposed the role of religion in politics,
and the ways in which the constitution, the electoral system, and demographic factors
provided strong barriers to religious mobilization in India’s first decades. Then I explore
the reasons for the recent decline of the Congress, the rise of the BJP, and Prime Minister
274   Steven I. Wilkinson

Narendra Modi’s current efforts to remake India and its institutions along Hindu na-
tionalist lines. In the third section, I explore the role of Islam in Pakistan, especially the
military’s role in explaining why no strong secular party capable of challenging the role
of religion in the state could emerge.

The Colonial Institutionalization

of Religious Difference

The British colonial state in India deliberately structured representative politics and
administration in India along religious lines. This was part of a conscious divide-​
and-​rule policy in which Muslim, Sikh, and Christian minorities as well as landlords
and the rulers of India’s princely states (33% of the land area and 25% of the popula-
tion) were seen as vital allies against Indian opposition to British rule. The colonial bu-
reaucracy, police, and army recruited Indians from favored regions and groups along
religious and caste lines—​army units had separate companies and regiments from
different communities, and police forces and provincial governments generally had a
fixed Muslim percentage—​with Muslims, Christians, and Sikhs overrepresented in the
administration.1
When the colonial administration did grudgingly concede more democratic repre-
sentation at the provincial level to meet Indian demands, in 1909 and 1919, it made sure
that these institutions and elections were organized along religious lines to minimize
the chance of a united Indian political opposition.2 The devolved provincial legislatures
established under the 1935 Government of India Act were also structured along explic-
itly religious lines, with the 25% Muslim minority voting in separate Muslim reserved
constituencies in which only Muslims could stand for election. Most other Indian voters
voted in “General” constituencies in which Hindus were the overwhelming majority.3
Religious minorities in each of the provinces, whether Hindu, Sikh, or Muslim, were
granted a disproportionate number of seats (“weightage”), purportedly to respond to
their demands and help prevent majority domination, but also as part of this divide and
rule strategy.4
Many influential members of the Muslim community, concerned about safeguarding
their traditional political influence and because of long-​term concerns that Muslims
were being left behind in the race for higher education and government jobs, strongly
supported these “reservations” in politics, administration, and the army. But the effect of
this system of religious quotas in government administration and separate electorates,
just as the comparative ethnic politics literature would have predicted (Horowitz 1985),
was to intensify ethnic extremism and outbidding and diminish the likelihood of cross-​
community coalition building and inter-​ethnic moderation. Muslims may have been
overrepresented in provincial legislatures, but in almost all provinces the majority
Hindu community did not need their votes. One colonial governor complained to the
 Religion and Party Politics in India and Pakistan    275

viceroy about the fundamental weakness of the system: “The minority cannot get their
own way in the legislature, and as a permanent communal minority have no prospect of
ever getting it, and they are tempted inevitably to redress the weakness of their parlia-
mentary position by rousing religious feelings and emphasising the importance of the
community outside the legislature.”5
In the 1930s, the Muslim League, dominated by the barrister Mohammad Ali Jinnah,
and reflecting the interests of Muslim landlords, businessmen, and the professional
classes, became the main organization representing Muslim political opinion in most
provinces of British India. Importantly for the future, though the League did well among
Muslim voters in Hindu-​majority provinces in the 1936–​1937 elections, it did not do
well in Muslim-​majority provinces such as Punjab (dominated by the landlords of the
Unionist Party), Bengal (KPP), and North-​West Frontier (where Congress was victo-
rious). In Punjab, Bengal, and elsewhere, these rival parties won by mobilizing strong
support from landlords and traditional religious leaders (pīr) (Gilmartin 1990). The
League’s national dominance did not reflect broad membership or deep party organiza-
tional roots like that of the Congress, so much as the recognition by Muslims that Jinnah
and the League were better placed than rivals to represent their interests at the provin-
cial and especially at the all-​India level (Jalal 1985).
By the late 1930s, angered at what he saw as the sidelining of Muslim interests in
Hindu-​majority provinces won by Congress in the 1936 elections, Jinnah focused more
and more on the as yet ill-​defined concept of a separate “Pakistan” (“land of the pure”)
that could safeguard Muslim interests in the long term: to protect the status of Muslims
and the Urdu language in education and government employment, to block land re-
form that would threaten Muslim landlords (many of whom were leaders and financial
supporters of the League), to retain Muslim personal law, and to block efforts by Hindus
to regulate or ban the slaughter of cows for meat and sacrifice.
In India, although the Congress leadership at the top was secular and deliberately
included prominent members of minority communities, including Muslims, in the
provinces and especially in the north the party had many influential Hindu conserva-
tive members, who sought to promote such issues as the status of Sanskritized Hindi
language rather than Urdu, a ban on the slaughter of cows (sacred to Hindus), and no
“appeasement” on minority rights issues. At times, these politicians were either directly
associated with the right-​wing Hindu Mahasabha party or sympathetic to its positions.
In the United Provinces (now Uttar Pradesh), Bihar and Central Provinces, this right
wing was sufficiently powerful by the late 1930s and 1940s to advance anti-​minority
measures that were strongly opposed by national leaders such as Nehru and Gandhi
(Brennan 1993). Opposition from the Hindu right was very also important in 1945–​1946
in blocking the constitutional compromise with the Muslim League desired by Nehru
and Gandhi, which would have kept India together, but at what Hindu nationalists
regarded as an unacceptable cost. In addition to the Hindu Mahasabha, there was also
the Rashtriya Swayamsevak Sangh (RSS), a Hindu nationalist paramilitary organization
formed in the aftermath of Hindu-​Muslim riots in 1925, which carried out drills and
marches throughout the country to “protect” the Hindu majority. As relations between
276   Steven I. Wilkinson

India’s majority and minority communities became strained in the mid-​1940s, the
membership in the RSS rose to 270,000 in January 1947, with a further half million more
members in parallel Hindu, Muslim, and Sikh paramilitary organizations such as the
Muslim League National Guards and Akalis.6
The bloody partition of India in 1947–​1948 led to the forced movement of 14.5 million
people across India, especially in Punjab and North-​West Frontier Province, where the
ethnic cleansing was almost total and greatly embittered communal relations. But each
state was still left with substantial religious minorities in 1951: a 13% Hindu minority in
Pakistan, almost all in East Pakistan, while 82% Hindu India had a Sikh minority (2%)
mainly in Punjab, a Christian minority (2%) mainly in Kerala and the northeast, and a
Muslim minority (11% in 1951) concentrated mainly in the north and northeast, urban
areas, and in states that occupied territory that had formerly been part of Muslim-​ruled
princely states such as Hyderabad and Bhopal.7

Party Politics in India

The Congress Party that formed India’s first post-​independence government was
highly unusual in its degree of what Mainwaring and Scully term “party institutional-
ization,” compared to parties in most former colonies (Mainwaring and Scully 1995).
The party was long-​established (1885–​); internally democratic and competitive; eth-
nically and regionally diverse, with millions of members across the country; and
had a strong organizational structure. The party also had substantial experience in
running provincial governments in the devolved system set up in 1919 and reformed
in 1935. Congress was also strongly opposed, as a result of British divide-​and-​rule
policies and the trauma of India’s partition along religious lines, to institutionalizing
religion in state policy. Congress used its overwhelming majority to pass a secular
constitution (1950) that banned religious discrimination and overt religious ap-
peals in elections. Over the next few years, the party also institutionalized pow-
erful cross-​cutting cleavages of caste and linguistic identities through government
policies such as caste reservations in employment and politics, and the creation of
new states for India’s major linguistic minorities (Brass 1990; Wilkinson 2015). These
created substantial barriers to political mobilization along religious lines. India’s re-
ligious minorities faced two additional obstacles to mobilizing on the basis of reli-
gion: (1) doing so had been tainted as secessionist by Jinnah and the Muslim League’s
actions in the 1930s and 1940s; and (2) India’s 1950 electoral system did away with
separate electorates and institutionalized a mass franchise first-​past-​the-​post elec-
toral system, in which religious minorities, except in a few states and districts in
which they are highly concentrated (e.g., Kerala and Kashmir for Muslims, Punjab
for Sikhs), were now spread too thin to compete successfully on the basis of their
minority identities. Branches of the Muslim League in India were largely wound up
within a few years of independence.
 Religion and Party Politics in India and Pakistan    277

The national Congress Party under Nehru (1947–​1964), tried hard to integrate
Muslims, appointing Muslims to several prominent (though not particularly pow-
erful) cabinet positions in New Delhi, and making efforts to reduce discrimination
against them in the states. Congress leaders were largely committed to a broad secular
vision of Indian national identity, and to preserving Muslim personal law, even though
this was not broadly popular with most Hindu voters. Muslim economic advance-
ment lagged, however: Muslims were disproportionately concentrated in traditional
craft occupations such as weaving, brass, and glasswork, as mechanics and repairmen,
in transportation, and in insecure occupations such as vegetable sellers and laborers.
Economic surveys done from the 1950s until the 1980s found that Muslims were con-
sistently less well educated, poorer than Hindus, lived in worse housing, and had less
access to clean water and toilets. Muslims were worse off than most other communities
in India, excepting only the Scheduled Tribes and Scheduled Castes.8 The national
Congress government promoted various Muslim development schemes from time to
time, but these were ineffective in large part because most power to change Muslims’
lives lay not with the national but with the state governments—​which controlled agri-
culture and education—​and here Muslims continued to have less of a claim on resources
than Hindu upper and middle castes. Most Muslims by the early 1960s were voting for
the Congress, but this support was never universal.9
The Congress Party’s dominance at the all-​India level was challenged in the mid-​
1960s, especially in the 1967 national and state elections, by new regional parties,
often representing middle and lower Hindu castes. These parties grew out of farmers’
movements, unions, and associations of middle and lower castes, and continued to mo-
bilize voters through these organizations as well as the new parties. By the 1970s and
1980s, these opposition parties had fractured the dominance of the Congress Party at
both the state and the national levels, in what Christophe Jaffrelot (2003) has termed
the “Silent Revolution” of Indian politics. In 1989, Congress won only a plurality of seats
in the national parliamentary elections, single-​party dominance in New Delhi ended,
and India entered an “era of coalitions” (see Figure 13.1) in which no party was strong
enough to win power unaided in New Delhi. For the next twenty-​five years, power at
the center, as in many states, was decided through complex pre-​and post-​election deals,
involving horse-​trading and sometimes bribes to induce parties and individuals to stay
as supporters, or switch sides.
In many ways this proliferation of coalitions at the national and state levels and the
growing strength of parties defined by the intra-​Hindu cleavages of caste and regional
identities was very good for Muslim voters, because it enlarged the majority commu-
nity “market” for minority votes and allowed Muslims to demand more in return for
their political support. What had been happening in Kerala for several decades—​a sit-
uation where Muslims helped decide which of two broad coalitions would win political
power—​was now increasingly likely to happen in other states. In several states, lower-​,
middle-​, and upper-​caste Muslims split their support, allying with Hindu parties with
a similar caste profile and similar socioeconomic objectives. In Uttar Pradesh, for in-
stance, the middle-​caste Samajwadi Party has done well among Muslim voters from all
278   Steven I. Wilkinson

100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
1952 1957 1962 1967 1971 1977 1980 1984 1989 1991 1996 1998 1999 2004 2009 2014 2019

Congress Left Parties State Parties Others BJP/Jana Sangh (1977 Janata Coalition)

Figure 13.1: Party balance in the Indian Parliament (Lok Sabha 1952–​2019).
Source: Data from Election Commission of India.

sections of the community, while its lower-​caste BSP rival has effectively challenged it
for the support of the lowest-​caste Muslims (Verma and Gupta 2016, 113). This increased
political leverage helped particularly in increasing Muslim security at a time of growing
Hindu political mobilization. Wilkinson (2004) showed that, although anti-​Muslim
electoral polarization was often employed as an electoral tactic, communal violence was
much less likely in states with a larger number of competing Hindu parties, where mi-
nority parties and votes were more valuable, and where Muslims could demand protec-
tion in return for their support.
Despite the growing market for Muslim votes in the 1989–​2014 period, Muslims
continued to fall behind economically, in part because of other Hindu communities’
greater political clout, and in part because of persistent social and employment dis-
crimination. The 2006 Sachar Committee, charged by a Congress government with
looking into Muslims’ economic disadvantage and how to improve things, found that
for the first time since independence, Muslims in the lower age cohorts were even
behind Scheduled Castes in terms of education.10 The 2009–​2010 National Sample
Survey showed that Muslims’ per capita expenditure was only 87% of the national
average. Much of this was driven by large educational gaps between Muslims, 20%
of whom are illiterate and only 35% of whom have a secondary education, and other
communities.11 The massive recent growth of the private employment market has not
helped to close this gap, in part because of Muslims’ lower levels of education and
skills needed to compete for these new jobs, but also because of persistently high dis-
crimination in the labor and housing markets. Indian survey experiment studies of
the employment market have found, as in similar studies of anti-​minority discrimi-
nation in the United States, that Muslim applicants consistently receive far fewer calls
back from companies than Hindu candidates with the same credentials. If the call-
back rate for an upper caste applicant with a degree is 100%, the callback rate for a
Scheduled Caste applicant is 66%, and that for a Muslim applicant only 33% (Thorat
and Attewell 2010).12
 Religion and Party Politics in India and Pakistan    279

More recent work by labor economists using national survey data has confirmed that
Muslims now have declining prospects for social mobility compared to the past, and
relative to all other groups. Asher, Novosad, and Rafkin (2018, 6) find that “virtually all
of the upward mobility gains in India over recent decades have accrued to Scheduled
Castes and Tribes, groups with constitutional protections, reservations in politics and
education, and who have been targeted by many development policies.” Muslims, be-
cause they are deemed a religious minority, are excluded from most of the government
programs meant to help lower and middle castes.13 Even though discrimination is an
important cause of Muslim disadvantage, the political appetite for dealing directly with
this is lower than ever, as the BJP criticizes any effort to help Muslims as minority ap-
peasement and “vote bank politics.” The Congress Party, 60% of whose votes in 2014
came from minorities, also has no hunger for aggressively promoting religiously
targeted antipoverty policies, as it now tries to reposition itself as a party that is for the
majority as well as for the minorities (Palshikar 2014).14

What Influences Muslim Voting in India?

Indians, the old saying goes, don’t cast their votes so much as vote their castes. And it
is certainly true that, if we look at aggregate voting survey data, there does seem to be a
strong correlation between caste/​community and party choice. But the relationship be-
tween community and party ID has never been as clear at it seems. First, Indian voters
from all communities, like voters elsewhere in the world, have consistently said when
asked that they are motivated mainly by bread and butter economic issues rather than
religious and caste identities. In the 2014 and 2019 elections, economic issues were again
the top issues named by voters as influencing their vote choice. Table 13.1 shows the data

Table 13.1: Key Voting Issues in India’s 2014 National Election

ISSUE

Economic Growth 25%

Corruption 21%
Inflation 18%
Personal Income 14%
Law and Order 8%
Access to Government Benefits 7%
Leadership 5%
Identity 3%

Question: which of these 8 issues will most influence your voting choice?
Source: Lok Foundation/​CASI/​Carnegie survey (2014).
280   Steven I. Wilkinson

for 2014, with economic growth, corruption, and inflation as the top three issues, and
“identity” coming in eighth, named by only 3% of voters. Only at a few rare times in
Indian history, such as during the 1989–​1992 BJP-​led movement to build a Hindu temple
at a disputed mosque site at Ayodhya, has any identity issue broken into the top few
concerns of voters in surveys.15
Second, even if someone from a particular group votes for a particular party, that is
not necessarily a “group vote,” but may instead reflect the party’s position on some ec-
onomic or social issue, which happens to align with voter preferences.16 What seems to
be a religious or caste vote might be an economic or ideological or development vote.
Third, South Asia is well known to have high degrees of clientelistic politics, in which
parties and politicians secure votes from groups of voters in return for the promise to
distribute some goods, such as infrastructure, access to political or regulatory influ-
ence, or the direct delivery of money, food, liquor, etc.17 High levels of group voting may
therefore reflect the fact that some voters have a greater need for clientelistic transfers,
or are more reachable by politicians, rather than that they are voting for group reasons.
There is a lot of evidence, in fact, that minority Indian voters, like voters elsewhere in
the world, are highly strategic. Madhavi Devasher (2014), who has conducted a detailed
analysis of Muslim voting in India’s most populous state, Uttar Pradesh, finds that reli-
gion matters much less overall for Muslims’ voting patterns than people conventionally
think, and that Muslims are “pragmatic ethnic voter[s]‌who respond to local dynamics
far more than blindly voting along identity lines.” Her survey of 1,685 voters across forty-​
five constituencies finds that that Muslim voters—​operating in what Kanchan Chandra
(2004) has described as a “patronage democracy”—​are concerned with voting for likely
winners, who can deliver goods to their communities. But the ethnicity of a party’s can-
didate or the party’s ethnic profile is not always decisive to this choice. What matters is
ethnicity combined with a party’s chance of winning, which in turn depends on local
ethnic demography, the signals sent by parties in terms of candidate selection, and other
informational cues that allow voters to assess a party’s probability of victory in a local
constituency. As Devasher says, “1) parties match candidate identity to ethnic demog-
raphy in order to signal a winning ethnic combination; 2) Muslim voters in India do not
strictly prefer a Muslim candidate and 3) relative ethnic group size and candidate iden-
tity influence voters to pick one ethnic party over another” (Devasher 2014).
A mass of other recent evidence confirms this overall picture. Rahul Verma and
Pranav Gupta (2016) used data from Uttar Pradesh (see Tables 2 and 3) to examine
whether the stereotype that Muslims voted as a bloc when the BJP was powerful or the
Muslim share of the population in a constituency was high was true, and found there
was no relationship between either of these two main variables and the fractionaliza-
tion of the Muslim vote. Verma and Gupta also confirmed, as other surveys have done,
that there seems to be no difference in the way in which Muslim voters are contacted by
parties compared to other groups of voters. Nor was there a difference between Muslim
voters and non-​Muslim voters in the influence of political and community leaders on
vote choice, despite the stereotype that religious leaders have a great influence over mi-
nority vote choice (Verma and Gupta 2016, 112).
 Religion and Party Politics in India and Pakistan    281

The Rise of the BJP

The BJP was founded in 1980 on the ruins of its largely unsuccessful Hindu nation-
alist predecessor, the Jan Sangh. The new party, like the Jan Sangh, was headed up by
RSS cadres and party workers, such as Lal Krishna Advani and Atal Bihari Vajpayee.
But the new party was much more successful in broadening its appeal, especially in a
large-​scale movement (1989–​1991) to “liberate” what it claimed was the temple site of
the Hindu god Ram at Ayodhya, over which a mosque had later been constructed by
the Mughal emperor Babur. In large-​scale processions across India, the party’s leaders
mobilized support for their movement to liberate the site, with Advani in a motorized
chariot made to resemble those in Hindu tradition, and succeeded in drawing millions
of new supporters to the party, enabling it to form state governments in several impor-
tant North Indian states. In December 1992, thousands of Hindu nationalist supporters
converged on the Ayodhya site and, going back on commitments made to the govern-
ment, demolished it, despite the presence of large numbers of paramilitary troops and
police from the BJP-​controlled state government. The national Congress government
quickly dismissed four BJP state governments, but despite this setback the party estab-
lished itself during the 1990s as one of the main opposition parties. Crucial to this suc-
cess was the party’s pragmatic willingness to periodically trim its ideological sails, for
example by quickly dropping its perceived opposition to caste reservations once those
cost it support in the 1994 state elections. The party has never dropped its commitment
to Hindutva (“Hindu-​ness”) and majoritarianism, but it often relegates that part of its
platform to the background and face-​to-​face campaigning, and it has put much more
focus on economic development and governance in its election campaigns: the 2014
national platform mentioned the word “Hindu” once, and the word “development”
eighty-​four times.
The BJP, after establishing itself in some important state governments in the north in
the 1990s, and forming a couple of shaky national coalitions (1998–​1999, 1999–​2004),
during which it was constrained by its partners from pushing its Hindu nationalist
agenda, in the 2010s unquestionably emerged as India’s dominant national party. Its
leadership combines many “movement Hindu nationalists” with business elites and
middle classes attracted by its development agenda and many others attracted by its
electoral successes compared to its rivals. The BJP relies on a combination of social ser-
vice provision for poorer voters (provided by its allied Hindu nationalist organizations);
a pro-​economic development message; strong party organization; modern media com-
munication through cable, press, and social media; and an emphasis on Narendra Modi
as a charismatic strong leader.18
The dominant personality in the party since 2012–​13 has been Narendra Modi, a life-
long cadre of the RSS, who first came to national prominence when he took over as chief
minister of the important western industrial state of Gujarat in 2001. He was parachuted
into the state by the RSS to fix things, because the previous BJP chief minister of the state
had performed badly in rehabilitation work and economic recovery after a 2001 earth-
quake, and the BJP was behind Congress in the polls with a state election in the offing.
282   Steven I. Wilkinson

Table 13.2: Participation in Campaign-​Related Activities in Uttar Pradesh.

2012 2012 2014 2014

State Average Muslims State Average Muslims

Attended public meetings 36% 37% 30% 27%

Attended a roadshow 32% 34% 10% 11%
Engaged in door-​to-​door 17% 19% 13% 9%
campaigning
Collected or 9% 10% 6% 6%
contributed money
Distributed election leaflets 15% 18% 12% 12%
and posters

Source: UP 2012 Post-​poll and NES 2014, Lokniti-​CSDS in Verma and Gupta (2016).

Table 13.3: Voter Contact by Canvassers and Campaigners in Uttar Pradesh

State Average Muslims

Candidate/​party worker/​canvasser came to your house to ask 64% 66%

your vote
Party/​candidate contacted you or a family member though a 19% 22%
phone call or recorded voice or SMS

Source: NES 2014, Lokniti-​CSDS in Verma and Gupta (2016).

Shortly after becoming chief minister, after a still-​mysterious deadly attack on Hindu
nationalist cadres on a train in the town of Godhra, serious Hindu-​Muslim riots broke
out across the state from February to April 2002, in which perhaps 2,000 people (over-
whelmingly Muslims) died, and for which many independent observers blamed the
Modi government.19 In the aftermath of these riots, the delayed Gujarat state elections
were finally held in December 2002. Before the elections, BJP election workers and
candidates explicitly framed the contest between the incumbent BJP and the opposi-
tion Congress as one between India and Hindus, on the one hand, and Congress, foreign
interests, and Muslims on the other. In introducing one BJP candidate, Kalubhai Maliwad
(just released from months in prison for his alleged role in the riots), in Lunawada con-
stituency, one BJP worker put this issue clearly: “The election is being fought not on is-
sues like development, but on security” (Sharma 2004a). At another election meeting, in
Halol constituency, a party worker and VHP official told voters that Muslims in the con-
stituency had celebrated with firecrackers and a procession when the news of the murder
of Hindu nationalist workers came out, and he emphasized that “This is not a Congress
versus BJP election. It is Hindus versus Muslims” (Sharma 2004b).
 Religion and Party Politics in India and Pakistan    283

The BJP won a crushing victory in the December 2002 Gujarat state elections, with
observers noting that the party did especially well in constituencies where violence
had taken place—​which tended to be constituencies where the race was predicted to be
close—​and among Hindu voters (Dhattiwalla and Biggs 2012; Wilkinson 2004). Chief
Minister Modi, now secure in power, established himself over the next dozen years as
a highly effective and authoritarian chief executive who delivered high growth to the
state. Modi was the BJP’s national leader and prime ministerial candidate in the 2014
elections, when he easily defeated the Congress and its lackluster dynastic head Rahul
Gandhi, by promising a pro-​development agenda and the expansion of “the Gujarat
model” to the nation as a whole. Against a very weak Congress Party, led by the unin-
spiring Rahul Gandhi, Modi appealed to many as just the kind of strong leader needed
to carry out much needed economic reforms, root out corruption and government
inefficiencies, and deliver broader development.
Once in office, though, and most especially in Modi’s second term, which began in
2019, the BJP substantially dialed up its Hindu nationalist and anti-​Muslim rhetoric.
The party nominated Hindu nationalists to important positions in the education min-
istry (HRD), where they attacked “anti-​national” universities by limiting them in terms
of governance and student numbers, and selected vice chancellors friendly to a Hindu
nationalist agenda. The HRD also supervised the rewriting of textbooks to erase the por-
trayal of a syncretic Hindu-​Muslim past in favor of one that highlighted India’s Hindu
identity. A Hindu nationalist monk, Yogi Adityanath, was appointed chief minister of
India’s most populous state, Uttar Pradesh, where he proceeded to rename places with
Islamic names with names drawn from Hindu tradition (e.g., Allahabad became Prayag,
Faizabad district became Ayodhya). And nationally, Hindu nationalists campaigned
against traditional Muslim divorce (declared illegal in July 2019), cow slaughter, “love
Jihad” (cases where Muslim men allegedly tried to lure Hindu women away from their
faith), and various alleged anti-​national provocations by Muslims. Some members of
Adityanath’s party in UP have been explicit about their desire to “re-​assimilate” Muslims
into the Hindu fold. BJP state representative Surendra Singh Ballia has announced that
“[o]‌nce India becomes a Hindu rashtra (Hindu nation), Muslims who assimilate into
our culture will stay in India. Those who will not are free to take asylum in any other
country.”20 Feeling supported by BJP state administrations and politicians—​BJP Union
Minister of State Jayant Sinha garlanded convicted Hindu cow slaughter vigilantes in the
summer of 2018—​Hindu activists have acted with a growing sense of impunity across
northern and northwestern India.21 One eyewitness to a Hindu cow slaughter vigilante
attack in Mewat (Rajasthan) said that activists who lynched a Muslim man announced,
“The MLA is with us, nobody can do anything to us. Put him (Rakbar) on fire.”22
After the massive 2019 election victory, Modi and his right-​hand man Amit Shah,
whom he appointed as home minister, felt secure enough to push hard on several
highly divisive anti-​Muslim policies, such as abolishing the special constitutional status
of Muslim-​majority Kashmir; the building of the Ram temple at the disputed site in
Ayodhya at which the Babri Masjid had stood until it was destroyed by Hindu militants
in 1992; and a new law and census to identify and deport illegal immigrants, mainly
284   Steven I. Wilkinson

from Bangladesh, which was directed only against Muslims. The most nationally divi-
sive of these so far has been the Citizenship Amendment Act, which to many sought
to unravel India’s guarantees of religious neutrality, and was directed at intimidating
Muslims in general rather than just those from Bangladesh. It prompted widespread
civil society demonstrations across India in the fall and spring of 2019–​2020, in which
Muslim students, previously reluctant to engage in mass public protests, took a signif-
icant role. The BJP government and Home Minister Shah crushed these with the full
array of the powers at their disposal (as did the many BJP-​controlled state governments),
including preventive arrests, trumped up charges against activists, and sending police
and paramilitaries in to break up demonstrations. Although civil society organizations
and opponents have filed legal cases maintaining that many of the BJP’s actions have
violated the constitution, an increasingly compliant court system has generally refused
to review these cases.
In the National Capital Territory of Delhi, where an opposition government is in
power, but the central government controls the police, police were sent into the major
universities to arrest activists, and several local BJP politicians were accused of inciting
a serious Hindu-​Muslim riot in Delhi in January 2020, the first in the city for many
decades (Minorities Commission 2020). Before the riots, one BJP state representa-
tive was filmed saying, “This is our country. If you want to live here, you will have to,
like the Australian Prime Minister said, follow the country’s traditions . . . . So, I warn
you that CAA and NRC are made by Modi and Amit Shah. If you will go against these
acts, it won’t be good . . . . If you wish, you can go to Pakistan. We don’t have any issues.
Intentionally, we would not send you . . . . If you will act as enemies, we should also react
like enemies.”23
None of this seems to have hurt Modi’s or the BJP’s standing with their core Hindu
supporters. Most Hindu voters may not very often vote based on religion themselves,
but the flip side of that is that they are also unwilling to vote against the BJP because
of its anti-​Muslim stance, and many are generally sympathetic to anti-​Muslim appeals.
BJP national leaders generally stand aloof from directly targeting Muslims, but on the
ground, in the cable and print media, and on social media things are very different. In
the 2019 campaign, media sympathetic to the BJP have not been shy about portraying
images of veiled Muslim voters lining up to (implicitly) cast their votes against the
Hindu majority. Samajvadi candidate Ali Khan Mahmudabad, a member of a prom-
inent UP Muslim family, complained about the frequent portrayals of lines of veiled
Muslim voters in the Hindi press, next to data on turnout figures, as an implicit means of
warning Hindus that they must turn out to vote.24
Muslim voters are in a real predicament at present, because they have few places to
turn to express their opposition to Modi and his agenda. The secular opposition parties,
for the most part, are worried that if they identify themselves too much with minority
voters, they risk losing support from members of the Hindu majority. So a whole host of
parties have, like the Congress, played up some Hindu themes and do not talk too much
about their support for Muslims. Even the Trinamool Congress in West Bengal, which
had substantial Muslim support in a state with a 28% Muslim minority, felt it necessary
 Religion and Party Politics in India and Pakistan    285

to respond to Hindu nationalist attempts to appropriate the Hindu Ram Navami festival
by organizing their own party’s Hindu devotions during the festival, with the tacit ap-
proval of the chief minister (Roy 2017, 18–​20). Most Hindu-​led parties with significant
Muslim support have responded to this ascendant Hindu nationalism by emphasizing
their majority community credentials and reducing their public support for any issue
that could be labeled “pro-​minority” or “vote bank politics.” Because of widespread
Hindu prejudices and unwillingness to vote for Muslim candidates, and increasing voter
uncertainty about which parties will win, major parties are putting up fewer Muslim
candidates for election than ever before, and far fewer are being elected (Gudavarthy
2014). None of the twenty-​seven Muslim MPs elected to the Lok Sabha in 2019 (4.4% of
MPs, compared to 14% of the population) were from the governing BJP. Political leaders
of all stripes feel increasingly compelled to portray themselves as in step with the reli-
gious majority—​often through visits to shrines and the use of religious symbols—​and
are increasingly anxious to avoid association with pro-​minority stances. Congress
leader Rahul Gandhi rediscovered his previously invisible personal Hindu religiosity
during political campaigns in 2018 and 2019, when he visited dozens of Hindu temples,
in addition to quoting from the Gita in his speeches (Venkataraman 2018).
Only a few parties with very substantial Muslim support, such as the Samajwadi Party
and Bahujan Samaj Party in UP, or with a clear minority identity, such as the AIUDF
in Assam and the AIMIM in Hyderabad, are forceful in representing Muslim interests.
The AIMIM, with its base in several Muslim majority urban areas in the former princely
state of Hyderabad, is led by a charismatic politician, Asaduddin Owaisi, who is skilled
at both English media and playing traditional politics in the Muslim areas of the old
city of Hyderabad. He has tried over the past few years to extend his base from the state
of Telangana, campaigning in Maharashtra and Karnataka, as well as Bihar and Uttar
Pradesh in the north and West Bengal in the East. But apart from picking up a few seats
in Muslim-​majority areas of Maharashtra, and municipal seats in UP, basic electoral de-
mography (the fact the Muslims are not a majority of voters in most areas), as well as
the entrenched nature of local parties in these other states, has thus far prevented him
from carving out the all-​India leadership role of the Muslim minority that he seeks.25
He faces the same electoral constraints that faced previous Muslim leaders who tried
to reestablish the Muslim League as a national force: a country with a FPTP electoral
system, in which Muslims are a majority or a pivotal share of the vote in only a small
share of constituencies, and in which most Hindu voters will not vote for what they per-
ceive as a “Muslim” party. 26

Religion and Party Politics in Pakistan

Pakistan was founded explicitly as a country where Muslims and Muslim interests were
to be safeguarded, but its founder Jinnah did not see this as incompatible with rights for
religious minorities. In August 1947 he famously gave a speech in the Pakistan Assembly
286   Steven I. Wilkinson

to reassure minorities, stating, “You are free; you are free to go to your temples, you are
free to go to your mosques or to any other place or worship in this State of Pakistan. You
may belong to any religion or caste or creed—​that has nothing to do with the business
of the State.” Jinnah deliberately included a Hindu in his first cabinet, and he appointed
a prominent member of the Ahmadi minority, Zafarullah Khan, as Pakistan’s UN repre-
sentative and then his first foreign minister in December 1947.27 This tolerance from the
top, however, although in some ways reflective of Sufi traditions in much of the Pakistan
countryside, as in India always coexisted uneasily with the conservatism of much of the
population, which provided incentives for politicians and clerics to accuse the state of
supporting “anti-​Islamic” policies and groups. The forbearance toward the Ahmadis
lasted only a few years, with their meetings broken up in several places in 1952, religious
parties (seeking to rehabilitate themselves) organizing large-​scale demonstrations and
riots against them in 1953, and then the forced resignation of Zafarullah Khan in May
1954. The position of Christian and Ahmadi religious minorities in Pakistan has, over
the years, only become worse.
To understand why no equivalent party to Congress emerged that could have sharply
curtailed the influence of religion of politics, it is vital to understand three things. First,
the major party at independence, the Muslim League, had been based largely in the
Hindu-​majority areas that remained part of India, so it was organizationally very weak
in the Muslim-​majority areas that formed the new state. That lack of organizational
depth and strong local ties left the League ill-​prepared to deal with the many regional,
sectarian, and distributional conflicts that affected the new state, especially after the
death of the country’s unquestioned leader, Mohammad Ali Jinnah, in September 1948.
The lack of a strong party was also due to the continuing influence of local landlords
and pīrs in Pakistan. These landlords have a local power base that allows them to shift
parties when they deem it necessary, and this base offer poor voters much more in the
way of clientelistic benefits than the Islamist maslaki (doctrinal) parties that lack these
financial and political resources (Chacko 2020, 105). The lack of a strong party, and an
inattention to the need to “coup-​proof ” the military, created a political vacuum that was
filled by the powerful bureaucracy and, increasingly during the 1950s, by the army.28
General Ayub Khan mapped out a detailed plan to restrain what he saw as the “irrespon-
sible” aspects of mass democracy as early as 1954, and then implemented many of these
ideas after he took power in 1958.
The second thing to understand is that, since it became heavily involved in poli-
tics, the Pakistan military has had an incentive to ensure that no strong national party
emerges that can challenge its own power. The only time such a party (the Pakistan
People’s Party, or PPP) did emerge, after the army’s discrediting due to the 1971 de-
feat by India and the loss of East Pakistan, the leader of the party, Zulfikar Ali Bhutto,
took several legal and practical measures to rein in the power of the military, but ul-
timately wasted his opportunity by increasingly authoritarian methods within both
the party and the country (Wilkinson 2015). During the military’s three periods of di-
rect rule (1958–​1971, 1977–​1988, 1999–​2008), as well as during periods of democratic
 Religion and Party Politics in India and Pakistan    287

government, it has always promoted “loyal” politicians and parties to legitimize its
regime, draw support away from unreconciled opposition parties, and help chart a
pathway to a democracy that does not threaten its interests (Siddiqa 2020). The military,
which has its own powerful Inter-​Services Intelligence (ISI) agency, spies on politicians,
supports those politicians and parties it regards as allies, and tries to hamper or re-
move those it sees as threats to its interests, which it conflates with Pakistan’s interests.
It does the same with the civil society organizations that in most countries underpin
democratic politics, promoting some, while taking down others (Rizvi, 2011). The cur-
rent governing party, the Pakistan Tehreek-​e-​Insaf (PTI), led by the former cricketer
Imran Khan, is only the latest of these military-​backed parties (Siddiqa 2020). Other
examples include, at various times, the PPP in the 1960s, the MQM (Muttahida Qaumi
Movement) in the 1980s, and the Pakistan Muslim League (Nawaz) in the 1990s and
Pakistan Muslim League (Q) in the 2000s. These military-​party alliances are always
shifting, however, with the usual pattern being one in which, as the party gets stronger
or finds itself blocked from taking steps that might make it stronger, its leaders begin to
chafe at military influence, leading to the military’s efforts to then destabilize it and/​or
find other allies.
The third point is that nonreligious leaders, whether in the military or the major
­political parties, have with increasing frequency since the 1980s sought allies in the re-
ligious parties and initiated Islamist reforms as a way of seeking legitimacy and broad-
ening their regimes. This process was begun by Zulfikar Bhutto in the 1970s. Bhutto was
personally focused on an agenda of economic and social transformation, but sought
to protect his political flank by giving way to the religious parties’ demand to label
Ahmadis a non-​Muslim minority, and to include clauses in the new constitution that
committed the state (though not immediately) to enable Muslims to “order their lives in
the individual and collective spheres in accordance with the teaching and requirements
of Islam as set out in the Qurʾan and the sunna (Abbas 2013, 68). After Bhutto’s Chief
of Army Staff, General Zia, toppled Bhutto in 1977, Zia turned these clauses into insti-
tutional practice, motivated by both a mix of personal beliefs and a desire to legitimize
his regime. Zia’s regime pursued an Islamization agenda in education, the courts, legal
code, and administration that was largely successful, despite some pushback from the
Shiʿa minority.29 The constitutional clause allowing minorities to “freely practice” their
religions ominously removed the word “freely” (Abbas 2013, 69). Zia also co-​opted
members of the JI and JUI to his advisory council, and some of these party members also
gave his regime added legitimacy by standing for office in his rigged elections (Nelson,
this volume).
One reason why explicitly religious parties have not been successful at the national
level is that the major Pakistani parties, like the PTI, Pakistan Muslim League (N), and
PPP, either support the many ways in which the Islamist agenda has already been incor-
porated into the state and society, or else do not dare oppose it openly. The cumulative
impact of changes in the education system, courts, government, and the public sphere
has led over the years to a significant narrowing of debate and growing conservatism
288   Steven I. Wilkinson

among the population, with several generations now having received an explicitly
Islamist-​themed education either in the state sector or new madrasas (Rizvi 2011, 190),
with 67% of the population in 2017 saying shariʿa should be the only law of the land (up
from 51% in 2010), and many voters saying that religious figures should have an influence
on political life.30 That does not necessarily mean that voters subscribe to an extremist
agenda—​three-​quarters of voters in polls say that religious fanaticism is a problem, and
that Christians should be able to build churches.31 But any civilian or military leader
who can credibly be painted as “anti-​Muslim” by religious organizations continues to
run both a political and personal risk. This is especially true since Salman Taseer, the
secular-​minded governor of Punjab, was murdered by his bodyguard in 2011 over his
strong stance on the need to reform the anti-​blasphemy laws, which are often used to
settle personal scores or to dispossess the accused. Hundreds of Sunni clerics applauded
Taseer’s murder, and urged clemency for the killer, and three hundred lawyers said they
would defend him for free, as his supporters shouted “Allahu Akbar” and scattered rose
petals as the killer entered the court (Reuters 2011). Not surprisingly, other politicians
have not stepped up strongly to follow through on Taseer’s agenda. Those who are
tempted to do so, even in a small way—​as Prime Minister Imran Khan did in 2018 by
nominating a prominent member of the Ahmadi community, Princeton academic Atif
Mian, to an important economics advisory panel—​quickly give way once the protests
and demonstrations start (Chaudhry 2018). The military, determined to weaken civilian
political parties that might threaten its own powers, also help prevent any strong party
from emerging that might have enough support to take on the veto power of the reli-
gious parties.
In Pakistan, religious parties still perform poorly in national elections, though since
1970 they have often been included in national coalitions (Nelson, this volume). But
the Islamist maslaki parties have formed governments six times in the provinces of
Khyber Pakhtunkhwa and Balochistan over the past twenty years. In part the limited
electoral wins by religious parties in national elections is the product of their success in
influencing policy. Successive military and civilian regimes have, since the 1970s, already
institutionalized much of their agenda, including shariʿa courts, harsh blasphemy and
anti-​minority laws, zakat tithes, laws on interest, and financial support for madrasas and
recognition of their degrees. There are many sectarian Islamic parties, some of whom
have been patronized by civilian and military leaders seeking to draw support away
from other parties. These parties, allied with militant organizations or with the domi-
nant Deobandi and Barelvi sects, have in general been much more influential as parties
and movements of pressure on the administration of the day, by bringing thousands of
their supporters onto the streets, than they have been in winning elections and forming
governments (Chacko 2020). The strength of Islamic parties also reflects a broader in-
crease in religious conservatism among the population in Pakistan over the past three
decades, driven by foreign (especially Saudi) influence and money, widespread opposi-
tion to Western and especially US foreign policy in the region in which the military and
civilian leadership is complicit, and increasing membership in Islamic movements such
as Tablighi Jamaat.
 Religion and Party Politics in India and Pakistan    289

Conclusion

In India the BJP, having won an increased majority in 2019, and within reach of being
able to control the upper house (which would allow the party to amend the constitu-
tion), is now focused almost as much on implementing its Hindu nationalist vision as
on the economic development agenda that was the primary reason for its election vic-
tory. Modi and his right-​hand man, Home Minister Amit Shah, are using their majority
to try to remake India’s government and institutions in Hindu Nationalist form. The
BJP has rescinded the separate status of Kashmir, an act that may be unconstitutional
but which a newly compliant Supreme Court has so far declined to review. The party
has used its control of higher education to target freedom of speech in universities, and
to appoint professors whose Hindu nationalist credentials are stronger than their ac-
ademic qualifications. The BJP has introduced new citizenship and immigration laws
that in theory require all Indians to prove their citizenship, but in practice are directed
overwhelmingly against Muslims.32 In a growing number of ways—​from the renaming
of cities with Islamic names such as Allahabad to those with “authentic Indian” (i.e.,
Hindu) names, to efforts to police interreligious personal relationships, the state is
clearly portraying India’s Muslim minority as a threatening “anti-​national” group that
must be controlled and policed.
The Muslim minority in India, now 14% of the population, feels deeply threatened by
this rise of Hindu nationalism and recent government actions, but it currently has few
good electoral options, given the first-​past-​the-​post electoral system and the fact that
Muslims are a majority of the population in so few areas of the country. Many of the sec-
ular parties in India’s system, led by the much-​weakened Congress Party, have also been
reluctant to defend minority interests too strongly, for fear of being seen as too aligned
with minorities in an age where mass media represent them as threatening and “anti-​
national.” India’s greatest post-​independence achievement, in maintaining a broadly in-
clusive secular state, is under greater threat now than at any time since independence in
1947. One potential danger is that Muslims, if they feel they have no other option, may
turn to more radical and violent alternatives, which have so far had very little appeal
in India.
Most of the BJP’s voters, however, have not voted for this radical Hinduization of
state and society, even though they may currently not feel they bear many of these
policies’ costs. One question for the future, therefore, is whether the Modi regime
has sufficient time to make its majoritarian vision irreversible, as General Zia did in
Pakistan in the 1980s, before the BJP’s various policy missteps (e.g., the 2016 demon-
etization, increasing unemployment and economic slowdowns, farmers’ movements,
the 2020 border conflict reversals against China) combine to cause it to lose its cur-
rent electoral majority.
In Pakistan, much of the Islamist agenda is now well institutionalized within the state
(including the army), courts, and education system. It is difficult to see this changing,
290   Steven I. Wilkinson

because the three potential sources of reform—​change in popular opinion, in the stance
of the military, or in the mainstream parties—​are all unfavorable. Popular opinion has
become more conservative, the military has also become more conservative, with its
senior leadership now much more favorable to expressions of Islam than the British-​
trained military elites of the 1950s and 1960s, and the mainstream parties are unwilling
or unable to push back against the institutionalization of an orthodox Sunni view of
Islam within state and society. That does not necessarily mean, though, that the Islamist
parties will increase their share of the vote beyond the 10–​20% they have achieved in
recent elections, precisely because, with most Islamist issues “settled” and off the active
political agenda, most voters vote on bread and butter issues of governance and eco-
nomic development.

Notes
1. For instance, the police force in the most populous colonial province of UP, which was
around 14% Muslim, was 50% Muslim in 1947 (Wilkinson 2004). The exception to the ge-
neral picture of minority overrepresentation was the army where, as a result of Muslims’
support of the 1857 revolt, the Muslim proportion was held at roughly the same 25% propor-
tion as their share in the pre-​1947 population.
2. One British colonial official in charge of these early 20th-​century reforms remarked, in a
confidential note, that “[t]‌he object is to satisfy this demand [for Indian representation]
with the least possible disturbance of existing conditions and to do so in such a way as to
give a preponderating influence in the elections to the stable elements of society in prefer-
ence to the professional politicians and the middle class who are noisy in peaceful times and
would be useless in any emergency.” He then approvingly noted the various ways in which
the Austrian-​Hungarian and Prussian monarchies had rigged their own electoral systems
to divide and weaken their oppositions. Indian Council Reforms: reports and notes of a
committee established to consider the desirability of a council Part I (Calcutta: Office of
Superintendent of Government Printing, India 1907). National Library of Scotland, Minto
Papers, MS 12610.
3. The property-​based franchise at this time included about 14% of the adult population.
4. The British colonial governors of each province were also granted the right to veto any legis-
lation they felt threatened the “legitimate interests of minorities.”
5. Sir Harry Haig, Governor UP, to Viceroy Lord Linlithgow, May 13, 1938.” IOR L/​P&J/​5/​265
U.P. Governor’s Reports, January–​June 1938.
6. Volunteer Organizations in India Secret Memo MI2 22 Jan ’47, in IOR L/​ WS1/​ 745,
Intelligence Correspondence with War Office.
7. These proportions are for 1951. As of the 2011 Census, Hindus were 80% and Muslims 14% of
the population.
8. See, e.g., the findings of the 1983 Gopal Singh Report on Minorities, set up by the Congress
government that took office in 1980, and Shariff (1995).
9. Monthly Public Opinion Surveys by the Indian Institute of Public Opinion show Muslim
support for Congress at 58% in 1952, 53% in 1957 (Vol. 2, Nos. 16–​19, Jan–​April 1957), drop-
ping to 51% in 1969. Monthly Public Opinion Surveys of the Indian Institute of Public Opinion,
Vol. XIV, nos. 4–​6 (January-​March 1969).
 Religion and Party Politics in India and Pakistan    291

10. These data were subsequently analyzed by Rakesh Basant, a member of the Sachar
Committee, and other scholars in Basant and Shariff (2010). This collection made clear the
extent of disadvantage, and the great differences that exist in terms of Muslim outcomes
across states.
11. NSS Report No. 552: Employment and Unemployment Situation among Major Religious
Groups in India, Table 3.2.5.1; Community Access to Education (NSS 66th round survey
2009-​10).
12. Discrimination in the housing market increases these disadvantages; see, e.g., the Thorat,
Banerjee, Mishra, and Rizvi (2015) study of housing discrimination in the NCR.
13. There are a few exceptions, such as Kerala, where because of Muslims’ political influence, a
few Muslim castes have been designated Most Backward castes or Other Backward Castes.
Muslim India 81 (September 1989): 432.
14. In 2014, Suhas Palshikar pointed out that “almost six of every 10 Congress voters are SC/​
ST/​Muslim/​Sikh or Christian, while these social groups account for only about three in
every 10 BJP voters” (Palshikar 2014).
15. This movement reached its culmination on December 6, 1992, when, with the deliberate
connivance of the BJP state government in UP, the Babri mosque was destroyed by Hindu
militants, leaving a heap of rubble, and setting off religious riots across the country, and
tit-​for-​tat attacks on religious minorities in Pakistan and Bangladesh. Four BJP state
governments were dissolved, and President’s Rule imposed after the demolition.
16. It may also be the case that economic issues are seen through the prism of group identity.
17. See Kitschelt and Wilkinson (2007) and the Wilkinson article in that volume.
18. On the electoral effect of social service provision by religious organizations allied with the
Rashtriya Swayamsevak Sangh, see Thachil (2015).
19. One of the best researched portraits of Modi is by Jose (2012).
20. “BJP MLA says only those Muslims will stay in India who embrace Hindu culture,” Indian
Express, January 14, 2018.
21. “Jayant Sinha garlands Ramgarh lynching convicts, says ‘honouring law’ Hindustan Times,
July 7 2018.
22. Times of India, July 24, 2018.
23. Times of India, July 24, 2018, 26.
24. Ali Khan Mahmudabad Facebook post, April 18, 2019.
25. The party won only two parliamentary seats in 2019, one from a Muslim-​majority area of
the city of Hyderabad, and one from the city of Aurangabad in Maharashtra, which before
1947 was also part of the princely state of Hyderabad. The victory in Aurangabad, which
is around a third Muslim, was possible because of an electoral tie up with a Scheduled
Caste party.
26. For Bashir Ahmed’s 1970 (ultimately unsuccessful) attempts to restart the Muslim League
in the cities of Uttar Pradesh, see Times of India, June 19, 21, 1970.
27. Khan was the most prominent member of Pakistan’s Ahmadi community, which regards
itself as Muslim but is regarded by most Muslims as heretical, because its faith questions
the finality of the Prophet Muhammad.
28. On Pakistan’s institutional weaknesses at independence, see Jalal (1992), ­chapter 3.
29. The Shiʿa minority objected on religious grounds to the compulsory deduction of zakat
tithes from investment and bank accounts, and their 1980 protests resulted in the tithes
being made optional for Shias. See Rizvi (1988, 233–​235).
30. Gallup Pakistan Poll, May 29, 2017.
292   Steven I. Wilkinson

31. Gallup Pakistan Polls on “Minorities in Pakistan,” January 30, 2015, and “Religion-​
General” September 14, 2015.
32. The laws also allow minority non-​Muslim refugees from nearby countries to apply for
refuge in India, but do not apply to Muslims from those countries who may also face
persecution.

References
Abbas, Shemeem Burney. 2013. Pakistan’s Blasphemy Laws: From Islamic Empires to the Taliban.
Austin: University of Texas Press.
Asher, Sam, Paul Novosad, and Charlie Rafkin. 2018. “Intergenerational Mobility in
India: Estimates from New Methods and Administrative Data.” Working paper, September
2018. https://​www.dartmouth.edu/​~novosad/​anr-​india-​mobility.pdf.
Basant, Rakesh, and Abusaleh Shariff. 2010. Handbook of Muslims in India: Empirical and
Policy Perspectives. Delhi: Oxford University Press.
Brass, Paul R. 1990. The Politics of India since Independence. New York: Cambridge
University Press.
Brennan, Lance 1993. “The Illusion of Security: The Background to Muslim Separatism in the
UnitedProvinces,” in India's Partition: Process, Strategy and Mobilization, edited by Mushirul
Hasan, 318–​355. New Delhi: Oxford University Press.
Chacko, Johann. 2020. “Religious Parties: The Politics of Denominational Diversity in
an Islamic Republic.” In Pakistan’s Political Parties: Surviving between Dictatorship
and Democracy, edited by Mariam Mufti, Sahar Shafqat, and Niloufer Siddiqui, 105–​
121.Washington, DC: Georgetown University Press.
Chandra, Kanchan. 2004. Why Ethnic Parties Succeed? New York: Cambridge University Press.
Chaudhry, Fahad. 2018. Under Pressure, Govt Backtracks on Atif Mian’s Appointment;
Removes Economist from Advisory Council.” Dawn, September 8, 2018. https://​www.dawn.
com/​news/​1431495.
Devasher, Madhavi. 2014. “Masjid versus Mandal: Ethnic and Cross-​Ethnic Voting in India.”
PhD diss., Yale University.
Dhattiwalla, Raheel, and Michael Biggs. 2012. “The Political Logic of Ethnic Violence: The
Anti-​Muslim Pogrom in Gujarat, 2002.” Politics and Society 40, no. 4: 483–​516.
Gilmartin, David. 1990. Empire and Islam: Punjab and the Making of Pakistan.
Berkeley: University of California Press.
Gudavarthy, Ajay. 2014. “Muslims of Telangana: A Ground Report.” Economic and Political
Weekly 49, no. 17. https://​www.epw.in/​journal/​2014/​17/​election-​specials-​web-​exclusives/​
muslims-​telangana-​ground-​report.html.
Horowitz, Donald. 1985. Ethnic Groups in Conflict. Berkeley: University of California Press.
Jaffrelot, Christophe. 2003. India’s Silent Revolution: The Rise of the Lower Castes in North India.
New York: Columbia University Press.
Jahangir, Ramsha. 2018. “Religious Parties Clinch over 9 Percent of Votes
in NA.” Dawn, August 1 2018. https://​www.dawn.com/​news/​1424235/​
religious-​parties-​clinch-​over-​9pc-​share-​of-​votes-​in-​national-​assembly.
Jalal, Ayesha. 1985. The Sole Spokesman: Jinnah, the Muslim League and the Demand for
Pakistan. Cambridge: Cambridge University Press.
 Religion and Party Politics in India and Pakistan    293

Jalal, Ayesha. 1992. The State of Martial Rule: The Origins of Pakistan’s Political Economy of
Defence. New York: Cambridge University Press.
Jose, Vinod K. 2012. “The Emperor Uncrowned: The Rise of Narendra Modi.” Caravan, March
1, 2012.
Lok Foundation/​CASI/​Carnegie Foundation, 2014. Indian National Opinion Survey.
Kitschelt, Herbert, and Steven Wilkinson, eds. 2007. Patrons, Clients and Policies: Patterns of
Democratic Accountability and Competition. Cambridge: Cambridge University Press.
Mainwaring, Scott and Timothy R. Scully, 1995. “Introduction: Party Systems in Latin America,”
in Scott Mainwaring and Timothy R. Scully, eds. Building Democratic Institutions: Party
Systems in Latin America. 1–​34, Stanford, Stanford University Press.
Minorities Commission of Government of NCT. 2020. Report of the Fact-​Finding Committee
on the North-​East Delhi Riots of February 2020. New Delhi: Minorities Commission of
Government of NCT.
Nasr, Vali. 1994. The Vanguard of the Islamic Revolution: The Jama’at-​i-​Islami of Pakistan.
Berkeley: University of California Press.
Palshikar, Suhas. 2014. “The Defeat of the Congress.” Economic and Political Weekly 49, no. 39.
https://​www.epw.in/​journal/​2014/​39/​national-​election-​study-​2014-​special-​issues/​defeat-​
congress.html.
Reuters. 2011. “Salman Taseer Killing Stirs Mixed Emotions” Dawn, January 5, 2011. https://​
www.dawn.com/​news/​596299/​salman-​taseer-​killing-​stirs-​mixed-​emotions.
Rizvi, Hassan Askari. 1988. The Military and Politics in Pakistan. Delhi: Konark.
Rizvi, Hassan Askari. 2011. “At the Brink?” in The Future of Pakistan, edited by Stephen Cohen,
182–​198. Oxford: Oxford University Press.
Roy, Rajat. 2017. “Communal Politics gaining Ground in West Bengal.” Economic and Political
Weekly 52, no. 16: 18–​ 20. https://​www.epw.in/​journal/​2017/​16/​commentary/​communal-​
politics-​gaining-​ground-​west-​bengal.html.
Shariff, Abusaleh. 1995. “Socio-​Economic and Demographic Differentials between Hindus and
Muslims in India.” Economic and Political Weekly 30, no. 46: 2947–​2954.
Sharma, Rakesh. 2004a. “Hate Mandate.” Disc II of the film Final Solution (at1:05).
Mumbai: ActIndia Films.
Sharma, Rakesh. 2004b. “Hate Mandate.” Disc II of the film Final Solution (at 24:00–​27:00).
Mumbai: ActIndia Films.
Siddiqa, Ayesha. 2020. “The Kingmakers.” In Pakistan’s Political Parties: Surviving between
Dictatorship and Democracy, edited by Mariam Mufti, Sahar Shafqat, and Niloufer Siddiqui,
35–​51. Washington, DC: Georgetown University Press.
Thachil, Tariq. 2015. Elite Parties, Poor Voters: How Social Services Win Votes in India.
New York: Cambridge University Press.
Thorat, Sukhadeo, and Paul Attewell. 2010. “The Legacy of Social Exclusion: A Correspondence
Study of Job Discrimination in India’s Urban Private Sector.” In Blocked by Caste: Economic
Discrimination in Modern India, edited by Sukhadeo Thorat and Katherine Newman, 35–​51.
Oxford: Oxford University Press.
Thorat, Sukhadeo, Anuradha Banerjee, Vinod K Mishra, Firdaus Rizvi. 2015. “Urban rental
housing market: Caste and religion matters in access,” Economic and Political Weekly, 50,
nos. 56 & 57: 47–​53.
Venkataraman, Ramesh. 2018. “Opponents of Hindutva Should Encourage Rahul’s Temple
Visits. Here’s Why.” Business Standard, December 27, 2018. https://​www.business-​standard.
294   Steven I. Wilkinson

com/​article/​p olitics/​opponents-​of-​hindutva-​should-​encourage-​rahul-​s-​temple-​visits-​
here-​s-​why-​118122700361_​1.html.
Verma, Rahul, and Pranav Gupta. 2016. “Facts and Fiction about How Muslims Vote in
India: Evidence from Uttar Pradesh.” Economic and Political Weekly 2, no. 53: 110–​116.
Wilkinson, Steven I. 2004. Votes and Violence: The Electoral Incentives for Ethnic Riots in India.
Cambridge: Cambridge University Press.
Wilkinson, Steven I. 2015. Army and Nation: The Military and Indian Democracy since
Independence. Cambridge, MA: Harvard University Press.
 Chapter 14

Rel igion and E l e c tora l

C om petition i n Se ne g a l

Dominika Koter

Religion is a highly salient aspect of life in Senegal. Its citizens eagerly display their
religious affiliation and piety, but they pride themselves on tolerance toward non-​co-​
religionists. The country has been often lauded for its enduring peace and cooperation
across religious divides. Senegal is a secular republic, but religion permeates major so-
cial debates as well as politics. On the one hand, since Senegal’s independence in 1960,
religious parties have been few in existence and rather marginal. On the other hand,
religious leaders from different Muslim brotherhoods, especially the Mourides, have
played an important role in assisting politicians in voter mobilization, acting as vote
brokers, and occasionally issuing vote orders. Yet the analysis of electoral data over sev-
eral decades shows that this type of mobilization has not resulted in sectarian electoral
cleavages and religion is not a strong predictor of vote choice. Starting with the country’s
first president, Léopold Sédar Senghor, who was a Catholic, Muslim religious leaders
have supported politicians across religious and brotherhood lines, creating fluid polit-
ical competition. All four Senegalese presidents have drawn support across the country’s
religious divides, as have the main political parties. The case of Senegal thus stands in
contrast to many other Muslim-​majority states. Religious leaders are central to political
life in general, and political competition in particular, but the role they play is not polar-
izing or divisive. Why is this the case?
This chapter analyzes how religion impacts electoral competition but does not pro-
duce religious parties or divisive religious cleavages. It attributes this absence of sec-
tarian cleavages to the role of religious leaders on the eve of independence and the
existing incentive structures that made alliances with politicians across sectarian lines
more beneficial. Muslim religious leaders largely emerged out of the colonial period
in a strong social and material position and immediately became coveted partners for
politicians looking to strengthen electoral ties with the population. In the existing elec-
toral context, religious leaders stood to gain by supporting promising politicians even if
they were non-​co-​religionists. The cross-​sectarian partnerships forged in the early years
296   Dominika Koter

of mass electoral politics proved beneficial to politicians and religious leaders alike,
fostering a dynamic electoral marketplace, one that is structured by prospects of mate-
rial benefits, rather than religious identity.

Religious Composition in Senegal

Throughout the post-​ independence period, the share of Senegalese society that
is Muslim has been over 90 percent, whereas Christians (mainly Catholics) have
constituted around 5 percent (see Table 14.1). The Senegalese practice the Sufi form of
Islam, with the vast majority of believers belonging to one of several brotherhoods (ta-
riqa), including the Tijan, Mouride, Layene, and Quadriyya brotherhoods.
In Sufi Islam, as practiced in Senegal, individuals declare themselves followers and
pledge allegiance to a religious leader, a marabout. Among Senegalese Muslims virtually
everyone has a marabout. Whereas the Tijani brotherhood has the most members, the
Mouride brotherhood is understood to be the most influential and powerful because of
its long domestic roots and higher levels of control over its members (Behrman 1970;
Cruise O’Brien 1971; Villalón 1995). In contrast to the Tijani brotherhood, which is trans-
national, the Mouride brotherhood is homegrown, started by Cheikh Amadou Bamba
in Central Senegal in the 1880s (though there are now Mouride disciples throughout the
world).
Islam arrived in Senegal from across the Sahara and it spread throughout the land
largely during the colonial period in the 19th century. Catholicism was brought to
Senegal by missionaries and it took hold mostly in coastal regions, particularly among
the Serer and Diola groups. While some Serer are Catholic and others are Muslim, the
Diola are rarely Muslim, whereas the other main Senegalese ethnic groups, including
the Wolof and Peul, are almost exclusively Muslim. Thus, religious affiliation is not
evenly distributed across the country, but it does create some cross-​cutting ties. The
same can be said about brotherhood affiliation. While the Wolof, Senegal’s largest ethnic
group, are members of different brotherhoods, most Peul are Tijanis and, as a conse-
quence, most Mourides come from the Wolof ethnic group. Geographically, apart from
the religiously and ethnically diverse capital, Dakar, Mourides tend to be concentrated

Table 14.1: Religious Composition in Senegal

Religion Muslim 93.8%

Christian (mainly Catholic) 4.3%
Brotherhood (Sufi Order/​tariqa) Tijaniyya 41.6%
Mouridiyya 28%
Quadriyya 10%
 Religion and Electoral Competition in Senegal    297

in the so-​called Groundnut Basin of Central Senegal, whereas Northern and Eastern
Senegal have more Tijanis. Major concentrations of Catholics are in the Saloum and
Casamance areas.
The French colonizers had a rather inconsistent approach to religion in Senegal. At
different points in time, they saw Islam as either a threat or as having a stabilizing effect.
There were instances when they tried to limit its impact, as for example when they exiled
the founder of the Mouride brotherhood, Cheikh Amadou Bamba. At other times, the
French administrators seemed to resign themselves to the reality that Muslim religious
leaders held a lot of sways, and they tried to work with them and saw them as a way
to boost compliance with their policies. For example, Muslim religious leaders proved
to be very effective recruiters of volunteers during both world wars (Robinson 1999).
The colonizers also viewed Islam in sub-​Saharan Africa, which they often called Islam
noir, as a more tolerant and more amenable variant of Islam than that in North Africa
or the Middle East. At the same time, the French were often mistrustful of Christian
missionaries and did not wholeheartedly encourage their expansion in Senegal.
As is the case in many other religiously diverse areas, religious belonging has many
visible markers. Christians and Muslims can be distinguished easily by their first names,
and often by their dress (this is especially true of men and particularly on Fridays,
when many Muslim men wear a boubou, a Muslim robe). Affiliation with different
brotherhoods is not always obvious but can sometimes be inferred, because Senegalese
often wear a necklace with a picture of their religious leader, or they display such pictures
in their homes. They also go on different pilgrimages and on some occasions celebrate
religious holidays on different days.

The Role of Muslim Religious Leaders

What is characteristic of Senegalese Muslim religious leaders is that they have histor-
ically played a very important part in the lives of their followers, well beyond the re-
ligious realm. Their role in providing material support, distributing land, managing
agricultural production (especially in Central Senegal), and shaping social interactions
is well-​documented (Coulon 1981; Villalón 1995) and it has persisted over time. Large
number of Muslim religious leaders, though not all, emerged out of French colonial rule
in a strong position, benefiting to a large degree from the destruction by the colonizer of
traditional social orders, especially among the Wolof, leaving the marabouts as the dom-
inant local leaders in large parts of Senegal.1 The marabouts have further successfully
buttressed their position throughout the postcolonial period (Cruise O’Brien 2002).
Contemporary survey data from the Afrobarometer indicate that the Senegalese have
high levels of trust in their religious leaders, showing more confidence in them than
in many state officials (see Koter 2016, 77–​79). They are also very likely to turn to them
for help, at higher rates than elsewhere in Africa. As in Mali (see Bleck and Thurston,
this volume), religious leaders serve as important providers of material support and
298   Dominika Koter

play a role of social mediators.2 This outsize role in social and economic life turned the
marabouts into important political actors. The broad trust that they have enjoyed over
decades among the population gives them clout to influence political dynamics.

Understanding Political Competition

in Senegal

The most remarkable aspect of religion in Senegalese politics is that although it has been
central to political mobilization for several decades, it has not produced significant re-
ligious divisions. Why is this the case? As I explained in previous work (Koter 2016),
this is because politicians do not invoke identity directly; instead, they enlist religious
leaders across the religious spectrum to help mobilize voters because of the marabouts’
considerable clout. Religious leaders thus play only an indirect role: they parlay their in-
fluence to work as electoral intermediaries, but they avoid running for office. Religious
leaders are not the only figures who play such a role—​local notables or village chiefs are
often also involved—​but they are the main and most visible electoral intermediaries. 3
There are several factors that can help explain why this “division of labor” between
religious leaders and politicians is a common practice. First, religious leaders and
politicians have distinct but complementary qualifications and skill sets. This remains
the case to this day but was particularly true in the early years of political competition.
Many religious leaders are not fluent in the official language, French. While literacy in
French is not a formal requirement, the lack of it makes navigating formal institutions
difficult. The great religious families mistrusted French education and they eschewed
it for their family members, concentrating instead on their religious education. This
avoidance of formal state education on the part of religious leaders is thus part of a long
history. In contrast, most prominent politicians at the national level4 have formal educa-
tion and university degrees, making them better equipped to navigate state bureaucracy.
Although religious leaders were historically not well positioned to enter formal politics,
they could parlay their following into political influence.
I would argue that the symbiotic relationship between religious leaders and
Senegalese politicians has been so successful and lucrative for the former that there are
few reasons to change the winning recipe. This apparent separation of roles also seems
to have become the norm in Senegalese politics. Indeed, the exceptions to this rule, such
as Modou Kara Mbacké, who formed the Party for Truth and Development (known by
its French acronym, PVD), or Serigne Mamoune Niasse, a marabout leader of a small
party, People’s Rally (RP), are both rare and controversial. Many Senegalese see this as
a breach of the separation of religion and politics and they find this practice very prob-
lematic. Yet, importantly, Kara and the PVD are a marginal phenomenon with very lim-
ited electoral clout. For example, the PVD gained only 2.5 percent of the vote, and two
seats, in the 2012 legislative elections.
 Religion and Electoral Competition in Senegal    299

The relationship between religious leaders and politicians that emerged in this con-
text is mostly transactional, encouraging them to forge the most beneficial alliances,
even if it means mobilizing voters for non-​co-​religionists.5 The resulting electoral mar-
ketplace has been incredibly fluid in Senegal, to the exclusion of rigid religious divides
or permanent alliances between different religious groups and parties. This practice has
started at the onset of mass electoral politics in the 1950s, on the eve of independence,
and it has persisted, with some adjustments, for most of the post-​independence period.
In contrast to many other African countries, Senegal experienced a relatively short pe-
riod of one-​party rule and saw a high degree of continuity of electoral practices, with
long stretches of multiparty competition. Thus, the tradition of cross-​religious political
partnerships that emerged in the 1950s has been consistently renegotiated and sustained
throughout the post-​independence period.
I now turn to fleshing out important elements of the mechanisms tying political in-
volvement of the marabouts to nonsectarian voting patterns, first describing their
role in electoral politics and then showing how fluid their support has been, following
changing material incentives. It is this fluidity of marabout support that explains the
lack of religious electoral cleavages.

The Nature of Marabout Involvement in Politics

Religious notables play an important role in electoral politics by throwing their support
behind political candidates. One of the main ways they help mobilize voters for given
politicians is by issuing voting orders (consignes de vote), also widely known as ndigal, a
religious edict. 6 While voting orders used to be very public and obvious until the 1980s,
in recent years they have become much more discreet.7 Religious leaders do not articu-
late an explicit voting order, but rather show their unambiguous support for a particular
presidential contender, or offer prayers for his success.
A rich body of work documents the continuous use of religious leaders as electoral
intermediaries by Senegalese politicians and even the French administration since the
onset of mass politics in the 1950s. The marabout played a crucial role in rallying voters
against independence from France during the 1958 referendum (Cruise O’Brien 1971).
The two main parties in the 1950s, the Bloc Démocratique Sénégalais (BDS), led by
Senghor, and the French Socialist Party (SFIO), led by Lamine Guèye, relied on Mouride
and Tijan leaders to help garner votes (Boone 2003, 62). After independence, ties with
intermediaries, especially the religious leaders, were crucial for President Senghor and
his UPS, which succeeded the former BDS (Schaffer 1998, 107). The same practice was
continued in the 1980s and 1990s by Senghor’s successor, Abdou Diouf (Schaffer 1998,
107; Zuccarelli 1988, 165). As Young and Kanté (1991, 72) noted, at the basis of Diouf ’s
electoral success “lay networks of clientelistic linkages joining the president and party
to civil society through the intermediation of a host of marabouts and local patrons.”
Newspaper coverage of electoral campaigns is full of politicians’ visits to prominent re-
ligious leaders in search of political support. A headline in the popular daily Walfadjri
300   Dominika Koter

on the eve of the 2001 legislative election, which read “Everyone Has a Turn with the
Marabout,” aptly illustrates politicians’ electoral strategies. The article describes how
both incumbent president, Abdoulaye Wade, and the former ruling party, the PS,
met with all imams from the community and visited all the marabouts of all religious
localities (Diebakhate 2001).
During my previous research in the run-​up to the February 2007 presidential elec-
tion, many of my interlocutors, including members of the then ruling PDS, talked about
their reliance on local powerbrokers, especially religious leaders. Many politicians
resigned themselves to the reality that parties are obligated to look for support and culti-
vate links with religious leaders.
Arguably, the single most important intermediary in 2007 was the Khalifa Général
of the Mouride brotherhood. According to many, on the eve of the 2007 election, the
Khalifa Général sent a coded voting order (ndigal implicite) to his followers (taalibés)
in support of President Wade.8 In 2007 the Khalifa’s support for President Wade be-
came clear during an announcement of the date for the annual Mouride pilgrimage, the
Grand Magal. In 2007, the Magal was supposed to take place shortly after the presiden-
tial election. During the announcement, the Khalifa stated that President Wade pro-
vided all the necessary funds for the Magal and that right after the pilgrimage Wade
would personally oversee great state infrastructure projects in the holy city of Touba,
the center of the brotherhood. The fact that the great projects associated with the Magal
were supposed to take place after the election strongly indicated that the president
needed to be re-​elected in order to carry out his work. While to an outside observer this
might seem cryptic, or even convoluted, the Mouride voters whom I interviewed had an
unambiguous interpretation of this message.9 There is also evidence that after his 2007
re-​election, President Wade attributed his high score in the department of Mbacké, in-
cluding the capital of the Mouride brotherhood, the holy city of Touba, to the Khalifa’s
backing (Koter 2016: 103).
A very important dynamic structures the electoral engagement of religious leaders
in Senegal: almost all politicians are courting the same set of religious dignitaries. In
other words, politicians are approaching religious leaders from various brotherhoods,
most attracted to those with the biggest stature. It is not the case that President Wade
(in power 2000–​2012), a Mouride, sought support only in the Mouride brotherhood.
As the PDS’s campaign chief in the department of Mbour, Magatte Diop, summed up,
“Wade has tentacles in all religious families” in Senegal.10 Diop was personally visiting
local religious notables and other influential local leaders on behalf of Wade and “asking
for their prayers.” As Diop explained, “prayers” are a favorite campaign euphemism
describing electoral support. President Wade managed to acquire all types of marabouts
as intermediaries, ranging from the local to the national level.11 After reviewing weeks of
press coverage from several of the most popular newspapers for the 2007 and 2012 pres-
idential campaigns, it became clear that all major candidates visited religious leaders
from all brotherhoods. The religious capitals were bustling with visitors in the run-​up
to the elections, with different candidates often crossing paths as they crisscrossed the
Senegalese religious establishment.
 Religion and Electoral Competition in Senegal    301

Apart from the widespread and visible use of marabouts as intermediaries by

politicians, there is also evidence that religious leaders do indeed influence the voting
behavior of their followers. Schaffer’s excellent study of the 1988 election provides par-
ticularly clear examples. One of his interviewees explained his following the voting
order from the Khalifa Général of the Mouride brotherhood by saying, “If Serign Touba
orders us to do something, we do it. Even if it pains us, we have to do it” (Schaffer 1998,
111). Another voter interviewed by Schaffer said that he would have preferred in fact to
vote for the opposition candidate Abdoulaye Wade, but voted for his marabout’s choice,
the incumbent, President Diouf (110).
There are also survey data indicating that people are influenced in their electoral
choices by religious intermediaries. A study conducted in 1999 in the departments of
Thiès and Diourbel by Gercop, a group of researchers from the Senegalese University
of Saint-​Louis, found that 38 percent of their respondents admitted following voting
orders (consignes de vote) from religious leaders.12 The 2005 Afrobarometer survey in
Senegal also asked respondents “who (if anyone) can influence their political choices.”13
Twenty-​three percent of Wolof respondents reported that they are influenced “a lot”
by their religious leader. Since admitting to being told how to vote might be somewhat
shameful, the real number is likely to be higher.
There is broad agreement in Senegal that religious leaders from the Mouride broth-
erhood are more likely to issue voting orders and that Mouride voters are more likely to
follow them than members of other brotherhoods.14 This can be attributed to the nature
of the relationship between marabouts and their followers in different brotherhoods;
Mouride leaders have greater control of material resources and the brotherhood has
more stringent norms of obedience. When compared with other brotherhoods, Mouride
marabouts have greater control over their disciples (Behrman 1970). Mourides are more
dependent economically on their marabouts and the brotherhood extends to a greater
extent into the economic realm, in particular through their control of vast areas of land,
mainly in the Groundnut Basin, as well as their infiltration of the sectors of urban trade
and transportation (Villalón 1995). The Mouride marabouts also place a much higher
emphasis on personal obedience than their counterparts from other brotherhoods
(Coulon 1981). Cruise O’Brien (1971) argues that Mourides are more submissive and
more devoted to their marabouts than members of other brotherhoods. As Villalón
(1995, 119) explains, Mouride ideology “enshrined the act of submission to a marabout as
a central defining feature of the order.” The Mouride brotherhood also approaches edu-
cation differently than other brotherhoods; it has historically discouraged the spread of
literacy, fearing that state education could undermine the order’s control over their dis-
ciples (Cruise O’Brien 1971). The higher deference to Mouride leaders is also reflected in
the size of monetary contributions made by disciples to their marabouts; Mourides are
much more regular and lavish in their payments15 to their marabouts (Cruise O’Brien
1971, 91–​96).
This greater deference to religious leaders is also reflected in survey data. The 2005
Afrobarometer survey found that while only 11 percent of Tijanis say that religious
leaders affect them a lot in their electoral choices, the figure is more than double among
302   Dominika Koter

the Mourides (23 percent).16 The Gercop study also found that Mourides were more
likely that Tijanis to follow voting orders from their religious leaders.17
While there is variation in the degree of religious leaders’ influence in different
brotherhoods and regions of Senegal, and there is debate about the exact degree of this
influence, the fact that religious leaders are significant electoral players is beyond dis-
pute. I now turn to the most striking feature of marabouts’ involvement in electoral poli-
tics: the fluidity of their support. Documenting this fluidity is important because it helps
understand why marabouts’ engagement in electoral mobilization does not translate
into religious cleavages. The changing allegiance of marabouts, and their willingness to
switch support, allows Senegalese politicians to construct religiously (and ethnically)
diverse electoral bases.

Fluidity of Marabout Support

Religious leaders stand to gain monetary rewards for themselves and their followers
by pursuing their role of electoral intermediaries. They thus tend to follow those best
positioned to offer such rewards, often those already in power or most likely to get
elected, making marabout allegiance fairly fluid. Many scholars have documented how
the marabouts have been very successful in securing various infrastructure projects
in return for the support of their followers. The renovation of the Great Mosque in
Touba and electrification and road construction are some of the most salient examples
(Schaffer 1998, 112–​113). At a local level, roads to villages, electricity, wells, and bore-​
holes are some of the most frequently demanded goods (Cruise O’Brien 1971, 276).
At a higher level, during the 1960 general election, the UPS contributed over 100,000
pounds to the mosque construction fund in Touba, alongside promises of loans, cre-
ation of trading centers, attribution of subsidies, and higher prices for the main cash
crop, the groundnut (Cruise O’Brien 1971, 267). The government distributed roads,
deep wells, dispensaries, agricultural credit, as well as staple foods—​millet and rice—​to
communities with supportive marabouts (Cruise O’Brien 1975, 175–​176).
Schaffer shows that not only were many marabouts successful in securing public
goods for their communities, but also that the marabouts’ followers attributed the exist-
ence of these goods directly to their leaders. One of Schaffer’s interviewees, talking about
the Khalifa Général of the Mouride brotherhood, describes his prowess in extracting
communal benefits: “Water, electricity, paved roads—​he [Abdou Ahad] got everything
he wanted” (Schaffer 1998, 112). He then adds that based on the acquisition of these
goods, it was right for the Khalifa to support President Diouf and urge his followers
to do so.
Abdoulaye Wade has continued the tradition of promising public investment and in-
frastructure to the religious capitals. Even though the religious establishment in Touba
did not endorse candidate Wade in 2000, immediately following his election, President
Wade sought the support of the Khalifa Général of the Mouride brotherhood, offering
his city important infrastructure, including an airport (Beck 2002, 541).
 Religion and Electoral Competition in Senegal    303

Yet, importantly, intermediaries also stand to gain personal goods. Apart from
extracting funds for various infrastructure projects or other public goods, these prom-
inent local leaders often receive personal benefits, both money and gifts in kind, such
as sponsored pilgrimages to Mecca. Cruise O’Brien (1971, 263) argues that the flow of
money from political sources made it possible for certain marabouts, the Mourides
in particular, to “acquire wealth on a scale hitherto unimaginable.” In fact, he goes as
far as to say that politics became the greatest single source of revenue for the Mouride
elite. Cash is not the only form of benefit offered to intermediaries. President Diouf,
for example, gave marabouts diplomatic visas ahead of the 2000 election (Beck 2002,
541). During the 2007 election, marabouts who were heads of a lineage were receiving
around 50 million francs CFA (about 100,000 USD) and a 4-​by-​4 car.18 At more local
level, marabouts were sometimes rewarded with a pilgrimage to Mecca, whereas the
lowest-​level clerics, such as village imams, were given beautifully bound editions of the
Qurʾan.19
Because of the importance of material rewards, religious leaders are willing to switch
their support when they can find a more generous patron. Over the course of political
competition in Senegal, the marabouts had the tendency to support candidates on the
basis of prospective material rewards, regardless of their religious affiliation, rather
than candidates from their own brotherhood. Even before Senegal’s independence,
Muslim marabouts were courted by candidates across the political spectrum and faced
choices in their political allegiance. The marabouts were ready to support a candidate
that seemed to offer them the best deal. This dynamic was clear during the 1951 legisla-
tive election to the French National Assembly. In 1951, the election pitted Léopold Sédar
Senghor, a Catholic, against a Muslim, Lamine Guèye. To access the substantial rural
electorate, both candidates were seeking the support of powerful Muslim dignitaries.
Arguably, the most important among them was Serigne Falilou Mbacké, the Khalifa
Général of the Mouride brotherhood. An account of Serigne Falilou’s interactions with
both candidates sheds light on the decision-​making process in choosing to support a
political candidate. As the Senegalese historian El Hadj Ibrahima Ndao (2003) recounts,
on the eve of the election both candidates went to the Mouride capital, Touba, to seek
the Khalifa’s support. The Khalifa told his confidants that he would support the candi-
date who came to him and offered resources to continue the construction of the Great
Mosque of Touba. Ndao chronicles the Khalifa’s meetings with both candidates:

Lamine Guèye came to Touba during the course of his campaign. He promised to
help the countryside (le monde paysan) achieve development, in the event of his vic-
tory. But he didn’t make any reference to the construction of the mosque . . . . When
Senghor arrived for the same reasons, he repeated his objectives and ambitions for
Senegal. He reaffirmed his pledge (sermon) of never betraying Islam and above all
stressed his decision to help the Khalifa complete the construction of the Great
Mosque of Touba . . . . It is from that day that Serigne Falilou Mbacké took up
Senghor’s cause and launched his first voting order (ndigal) to the Mouride voters.
(Ndao 2003, 132)
304   Dominika Koter

A similar openness to the possibility of supporting either candidate in the 1951 elec-
tion is also evident in the case of the Khalifa Général of the Tijan brotherhood, Ababacar
Sy. The Khalifa was initially more predisposed to his fellow Muslim Lamine Guèye,
but after being courted much more by Senghor than Guèye, he decided to support the
former (Ndao 2003, 132).
With the support of the two most prominent marabouts in Senegal, Senghor beat
Lamine Guèye by more than a two-​to-​one margin (Ndao 2003, 138). In his interactions
with the Muslim marabouts, it would appear that the odds of gaining their support were
against him. If the marabouts were driven by affective ties they would surely support
Lamine Guèye.20 But in the contest for marabouts’ support, it was Senghor who seemed
more in tune with what the marabouts wanted and was able to make them more ap-
pealing promises.
Since then, during most, if not all, elections, the marabouts tended to support
those already in power, including the French administration, earning this practice of
legitimizing incumbents the term vote légitimiste. Cruise O’Brien (1971) asserts that in
the 1958 referendum, the three great local brotherhoods, which had patronage ties to
the French administration, sided with the French and helped postpone Senegalese in-
dependence, by convincing their followers to vote against it. After independence, the
marabouts overwhelmingly supported President Senghor and his party, UPS, which
subsequently became a single party. This support continued after a limited liberal-
ization under Senghor and then after restoration of multiparty politics under Abdou
Diouf. With a return to competitive politics, the marabouts could have supported dif-
ferent candidates, but they have thrown their weight mostly behind President Diouf.
Both in 1983 and 1988, the religious leaders issued a ndigal to their followers to vote
for Diouf (Schaffer 1998, 107). In 1993 and 2000, the main marabouts, namely heads
of brotherhoods, abstained from issuing a ndigal, but many others did. As electoral
contests became more competitive, there has been a certain diversification of voting
orders, with some marabouts supporting the main opposition candidate at the time,
Abdoulaye Wade. Yet the majority of marabouts involved still sided with the incumbent
(Coulibaly 2006, 235–​236).
The fact that some marabouts chose not to support the incumbent and abstain was
most likely caused by diminished patronage flowing from the Diouf administration.
Galvan (2001, 53) argues that the shrinking of the Senegalese state, due to the struc-
tural adjustment measures, reduced financial resources that could be channeled to
the marabouts, thus souring Diouf ’s relationship with the marabouts. At the time of
the 2000 election, some commentators thought that Senegalese society was no longer
susceptible to voting orders from religious leaders. Yet, in the light of the increased in-
volvement of religious leaders in the 2007 election, and to a lesser extent in the 2012
election, such analysis appears to be premature. The economic factors seem much more
convincing in explaining their less fervent involvement in the two previous elections
(Audrain 2004).
Intermediaries also showed a preference for the incumbent during the Wade pres-
idency. Whereas Wade struggled to acquire intermediaries as a challenger, he found
 Religion and Electoral Competition in Senegal    305

it much easier as an incumbent. When he was in opposition, Wade did not have the
backing of any high-​ranking religious leaders, even among his fellow Mourides. As an
incumbent, during the 2007 election, Wade enjoyed the support of some of the country’s
most prominent religious leaders, including the Khalifa Général of the Mouride
brotherhood.
However, by 2012 Wade’s ability to acquire intermediaries was much diminished
in comparison to the previous election, despite the fact that his campaign courted
intermediaries as vigorously as ever before. Wade had the backing of many secondary
marabouts,21 but the upper echelons of the religious hierarchies in Touba and Tivaouane
remained silent, despite multiple visits by the president and frequent “requests for
prayers.” In this respect, the differential behavior of marabouts between 2007 and 2012
mirrors their fluctuating support over time for the Diouf regime. When a president run-
ning for re-​election faces strong challenges, intermediaries are much less likely to em-
brace him because the value of promises of post-​electoral transfers is greatly diminished.
There are many instances in Senegalese politics when intermediaries switched their
allegiance apparently on material grounds. For example, Serigne Mamoune Niasse
used to support his fellow Tijan, Idrissa Seck, a former prime minister who later came
in second in the 2007 presidential election. Yet by November 2006 Niasse abandoned
Idrissa Seck for President Wade, and he was swiftly rewarded with a ministerial posi-
tion (Diop 2006). True to his interests, Niasse did not remain loyal to Wade for very
long. By the time of the 2012 election, he decided to support again Wade’s bitter rival
Seck, explaining publicly that he feared that Wade was going to lose (Naudé 2011). For
this reason, most observers of Senegalese politics see marabouts’ political support as
highly instrumental. As the famous investigative journalist Abdou Latif Coulibaly
(2006, 239) describes it, religious intermediaries are seen as “rentiers, keen on consumer
goods.”
In general, the behavior of the marabouts, especially the more politically involved and
influential Mourides, shows great inconsistency with the instrumental importance of
identity.22 As documented, the Mourides worked as intermediaries for the French, then
for a Catholic president, followed by a Tijan president, and only most recently a Mouride
incumbent. Even intermediaries who share identity with their patron, such as Modou
Kara Mbacké with President Wade, made their decisions based on material incentives.
In Kara’s case, the fact that he supported President Diouf, a Tijan, against Wade in the
2000 election before offering his support to President Wade in 2007, shows that shared
identity is not driving his behavior. 23
Another behavior that is inconsistent with religious affinity driving political sup-
port is the high frequency of split support for different candidates within maraboutic
families. The phenomenon of a father and son, or a set of brothers, supporting different
candidates is relatively common among Senegalese marabouts.
One explanation for this behavior is that intermediary families want to “hedge their
bets.” If members of the family support different candidates, they increase the chances
that one of them will be on the winning side. In addition, marabouts sometimes
align themselves with different politicians to play out their internal family rivalries
306   Dominika Koter

and potential succession battles (Behrman 1970, 73, 77, 88). Supporting opposition
candidates against other members of the family who back the incumbent is often a result
of jockeying for status within the brotherhoods (Beck 2008, 73).
In contrast to Muslim leaders, Christian dignitaries have largely stayed out of elec-
toral politics. Although they are shown respect by politicians, they have not issued
voting orders or made pronouncements that could be seen as influencing electoral poli-
tics. There are probably multiple reasons for this largely apolitical stance of the Catholic
establishment. First, Catholics are a small minority, and mobilizing Catholic voters
would not be numerically significant. Representing a minority religion, the Catholic
establishment was probably eager to appear politically neutral. Moreover, the Catholic
Church in Senegal is connected to a transnational institution, in contrast to the Muslim
brotherhoods, especially the Mouridiyya, that are centered on Senegal. Marabouts have
hailed from Senegalese families, whereas around the start of mass politics, many priests
were foreign.
It is this fluid allegiance of Muslim religious leaders, driven by monetary incentives,
that helps explains why religious cleavages did not materialize in Senegal.

Electoral Patterns

Despite the high salience of religion in everyday life, Senegalese politics have not been
marked by religious cleavages (Diouf 1994, 44–​45). This has consistently been the case
since the onset of mass politics in the 1950s. Even though we do not have fine-​grained
data for that period, the electoral outcomes are widely inconsistent with religious
cleavages. During the 1957 elections to the territorial assembly, Léopold Sédar Senghor,
a Catholic Serer, gained support for his BPS throughout the country, winning seats in
every constituency, even where there are no Catholics (and few co-​ethnics). After inde-
pendence, President Senghor actually enjoyed the highest levels of support in regions
dominated by non-​Serer Muslims. When political competition intensified, resulting in
the first alternation in power in 2000, even that most momentous election in Senegalese
political history was not affected by religious identity. Round 2 of the Afrobarometer
survey (2002) recorded respondents’ electoral choices from the 2000 election, showing
the lack of importance of identity factors, such as religion and brotherhood member-
ship (and ethnicity), as predictors of voting. Table 14.2 shows the religious composition
of the vote share for the three leading candidates, Abdoulaye Wade, Abdou Diouf, and
Moustapha Niasse. These candidates’ vote shares among Muslims and Catholics, as well
as among the main brotherhoods, appear evenly distributed and proportional to the
group size. When Wade, a Mouride, ran against Diouf, a Tijan, their levels of support
among the two groups were almost indistinguishable. Diouf, a Tijan, got 39 percent of
his votes from Tijanis, and 41 percent from Mourides. Wade, a Mouride, got 45 percent
of his votes from Mourides and 39 percent from Tijanis.
 Religion and Electoral Competition in Senegal    307

Table 14.2: Religious and Brotherhood Composition of the Electorate of Main

Presidential Candidates in Senegal
Candidates’ share of their total vote received
from a given group
Group’s share of
Religious group the population A. Wade A. Diouf M. Niasse

Muslims 95% 93% 91% 93%

Catholics 4% 5% 9% 5%
Muslim brotherhood
Mouride 40% 41% 45% 37%
Tijan 39% 39% 39% 49%

Source: Afrobarometer, Round 2 (2003). N = 1147.

Table 14.3: Religious and Brotherhood Composition of the Electorate of Main

Parties in Legislative Elections
Parties
Group’s share of
Religious group the population PDS PS AFP

Muslims 95% 95% 90% 91%

Catholics 4% 4% 4% 9%
Muslim brotherhood
Mouride 40% 43% 41% 32%
Tijan 39% 38% 42% 51%

Source: Afrobarometer, Round 2 (2003).

The results of the 2001 legislative elections similarly highlight the lack of discernible
religious voting patterns. The three main parties, PS, PDS, and AFP enjoyed compa-
rable level of support among the country’s Muslims and Catholics, in proportion to their
overall scores in that election (see Table 14.3). The same can be said about the breakdown
of their support among Mourides and Tijanis or among the main ethnic groups.
After Round 2, the Afrobarometer has not collected brotherhood affiliation data24
that would allow similarly detailed comparisons for the 2007, 2012, and 2019 presiden-
tial elections, but the coverage of those elections has never pointed to any visible re-
ligious voting patterns. Instead, one of the more discernible patterns throughout that
period has been an urban-​rural cleavage, with rural areas consistently siding with the
incumbent (Koter 2016, chap. 5). Indeed, one of the most striking features of Senegalese
308   Dominika Koter

politics is that all four Senegalese presidents have constructed remarkably similar win-
ning coalitions.

Conclusion

Religion has played a consistently influential role in Senegalese politics, but in a fairly
unique way. Religion has not become a political cleavage, as in countries such as Côte
d’Ivoire, nor has there ever been a need to manage inclusion of different religious
groups through consociationalism, as in Lebanon. Religious dynamics are essential
to understanding electoral politics in Senegal, but not as a simple determinant of vote
choice as might be expected through the prism of identity politics. Just as their Malian
counterparts (see Bleck and Thurston, this volume), Senegalese marabouts eschew run-
ning for office directly, preferring to exercise indirect influence on political leaders. In
contrast to Mali, the patterns of accommodation between religious leaders and the state
have not been disrupted by the rise of a jihadist presence in West Africa.
Muslim religious leaders in Senegal have a long-​standing history, originating before
the country’s independence from France, of mobilizing voters, but their involvement
has been instrumental, fluid, and pragmatic, rather than sectarian. Marabouts have a
pattern of supporting politicians across confessional lines, creating a dynamic electoral
arena. Their most striking tendency has been their support of incumbents, including
their willingness to switch to backing new incumbents following alternation in power.
In so doing they have contributed to the political stability in Senegal and helped avoid
divisive cleavages. At the same time, the marabouts’ influence on Senegalese democracy
is controversial. Their practice of converting social clout into political influence is open
to criticism, such as the potential exploitation of vulnerable voters. The marabouts’ will-
ingness to switch support between candidates prevents hardened cleavages but it also
appears to some as unscrupulous or even mercenary, especially since the marabouts
enjoy considerable material benefits from their endorsements (Coulibaly 2006). Their
preference for candidates with the most resources, typically incumbents, is skewing the
electoral playing field and constraining democratic competition.
The marabouts’ role has been consistent but not immutable. Undoubtedly,
marabouts played a more crucial and influential role in the early post-​independence
period than they do today. Their voting orders were more explicit and they carried
more weight. Yet their impact is still relevant. Scores of contemporary politicians
still believe that courting marabouts is an important part of their electoral strategies.
There are growing debates about the appropriateness of marabout involvement in
electoral politics, especially in urban areas, with an increasing social sentiment that
marabouts should stay away from politics. In addition to changing norms, rapid ur-
banization also plays a role, as there is evidence that marabouts exercise on average
greater control in the countryside than in urban areas (see Koter 2013 for discussion).
Changing control of material resources as larger segments of society move away from
 Religion and Electoral Competition in Senegal    309

agriculture also affect disciples’ economic dependence and the marabouts’ clout. Even
though many marabouts managed to make successful inroads into trade and urban
transportation, their influence is more uneven in urban settings and it competes with
other social networks. While debates about the appropriate role of marabouts in pol-
itics are vigorous and impact the style of marabout involvement, it is hard to imagine
that religious leaders will cease to play a role in Senegalese electoral politics in the
near future. As long as they remain influential in the lives of their followers, they will
continue to be seen as appealing electoral intermediaries from the point of view of
Senegalese politicians.

Notes
1. See Koter (2016), c­ hapter 3, which discusses the effect of French invasion and rule on the
religious establishment.
2. Yet, in contrast to the Malian case, they are not doing so in the context of state failure or
widespread insecurity.
3. There are important geographic differences in the repertoire and strength of intermediaries
across the country, but most regions of Senegal, with a notable exception of Casamance,
have influential local leaders (Beck 2008; Boone 2003; Diop and Diouf 2002) and, overall,
the marabouts are by far the most influential intermediaries in Senegal.
  The most prominent works on the topic include Beck (2001, 2008); Boone (2003);
Coulon (1988); Cruise O’Brien (1971, 1975); Diop, Diouf, and Diaw (2000); Galvan (2001);
Villalón (1995); and Young and Kanté (1991).
4. This is not always true at the local level.
5. By non-​co-​religionists I also mean members of different Sufi brotherhoods.
6. The religious order ndigal also appears in the literature spelled ndigel.
7. One of the reasons behind the changing nature of ndigal over the years is growing so-
cial objection to voting orders. Many activists have portrayed ndigal as anti-​democratic,
making overt voting orders less socially acceptable.
8. Author interview, Dakar, February 26, 2007.
9. Author interviews, Touba, March 2007, and Palmarin, February 2007.
10. Author interview, Mbour, February 20, 2007. A similar point also finds confirmation in
Riedl (2014). One of her interviewees, a constituency representative in Podor, stated that
during the single party era the PS was “open to all traditional chiefly families and all (reli-
gious) brotherhoods” (123).
11. Author interview, Dakar, April 14, 2007.
12. “Etude sur le comportement électoral dans les régions de Thiès et Diourbel,” May 1999.
The sample size was 4877 (2900 in the department of Thiès and 1977 in Diourbel).
13. Question 88a_​Sen, Round 3, 2005.
14. See Koter (2016, 107) for discussion.
15. An example of such payment is hadiyya, a personal gift to the leader.
16. Question 88a_​Sen, Round 3, 2005.
17. Forty percent of Mourides, but only 34.9 percent of Tijanis, said that they accept voting
orders from religious leaders. A previous Gercop study, conducted in Saint-​Louis, also
found that Tijanis are less influenced by religious leaders.
310   Dominika Koter

18. Author interviews with Madior Diouf, Dakar, March 1, 2007, and Madieyna Diouf, Dakar,
April 2007.
19. Author interview with Semou Pathé Guèye, Dakar, March 1, 2007.
20. The 1920 election to the French Assembly provides another example of a contest pitting a
Muslim against a Catholic candidate, albeit at a time of limited suffrage. The marabouts
supported a Catholic, Blaise Diagne, against a Muslim, Galandou Diouf (Beck 2008, 72–​3).
21. Among others, Wade gained open support from the Khalifa of Darou Mousty, Serigne
Cheikh Khady Mbacké (Le Soleil, February 15, 2012), Cheikh Béthio Thioune, Cheikh
Ndigel Fall (Sud Quotidien, February 21, 2012), Seigne Fallou Mbacké (Le Soleil, March
1, 2012), the marabout Pathé Kébé (Le Soleil, March 6, 2012), Serigne Bakhé Mbacké (Le
Soleil, March 9, 2012), and several religious dignitaries and imams in the department of
Kolda (Le Soleil, March 13, 2012).
22. For additional examples of prominent intermediaries supporting candidates from other
groups see Diouf (1994) and Beck (2008, 72–​73).
23. Kara’s support for Diouf is discussed in Audrain (2004).
24. All major candidates in those elections were Muslim, so Muslim versus Christian cleavage
is less relevant.

References
Audrain, Xavier. 2004. “Du ‘ndigel avorté’ au parti de la vérité.” Politique Africaine 96: 99–​118.
Beck, Linda. 2001. “Reining in the Marabouts? Democratization and Local Governance in
Senegal.” African Affairs 100: 601–​621.
Beck, Linda. 2002. “Le clientelism au Sénégal: Un adieu sans regrets.”In Sénégal Contemporain,
edited by Momar-​Coumba Diop, 529–​549. Paris: Karthala.
Beck, Linda. 2008. Brokering Democracy in Africa: The Rise of Clientelist Democracy in Senegal.
New York: Palgrave Macmillan.
Behrman, Lucy. 1970. Muslim Brotherhoods and Politics in Senegal. Cambridge, MA: Harvard
University Press.
Boone, Catherine. 2003. Political Topographies of the African State: Territorial Authority and
Institutional Choice. Cambridge: Cambridge University Press.
Coulibaly, Abdou Latif. 2006. Une démocratie prise en otage par ses elites: Essai Politique sur la
Pratique de la Démocratie au Sénégal. Dakar: Les Editions Sentinelles.
Coulon, Christian. 1981. Le Marabout et le Prince: Islam et pouvoir au Sénégal. Paris: Editions
A. Pedone.
Coulon, Christian. 1988. “Senegal: The Development and Fragility of Semidemocracy.” In
Democracy in Developing Countries: Africa, edited by Larry Diamond, Juan Linz, and
Seymor Martin Lipset, 141–​178. Boulder, CO: Lynne Rienner.
Cruise O’Brien, Donal. 1971. The Mourides of Senegal. Oxford: Clarendon Press.
Cruise O’Brien, Donal. 1975. Saints and Politicians: Essays in the Organization of a Senegalese
Peasant Society. Cambridge: Cambridge University Press.
Cruise O’Brien, Donal. 2002. “Les negotiations du contrat social sénégalais.” In La construc-
tion de L’Etat au Sénégal, edited by Donal Cruise O’Brien, Momar-​Coumba Diop, Mamadou
Diouf, 83–​93. Paris: Karthala.
Diebakhate, Ibrahima. 2001. “A chacun son tour chez le marabout.” Walfadjri, April 4. 2001.
 Religion and Electoral Competition in Senegal    311

Diop, George Nesta. 2006. “Un gouvernment de mission . . . électorale.” WalFadjri, November
11, 2006, 3.
Diop, Momar-​Coumba and Mamadou Diouf. 2002. “L’administration, les confréries religieuses
et les paysanneries.” In La construction de L’Etat au Sénégal, edited by Donal Cruise O’Brien,
Momar-​Coumba Diop, Mamadou Diouf, 29–​47. Paris: Karthala.
Diop, Momar-​Coumba, Mamadou Diouf and A. Diaw. 2000. “Le baobab a été déreciné.
L’alternance au Sénégal.” Politique Africaine 78: 157–​179.
Diouf, Makhtar. 1994. Sénégal: Les ethnies et la nation. Paris: L’Harmattan.
Galvan, Dennis. 2001. “Political Turnover and Social Change in Senegal.” Journal of Democracy
12, no. 3: 51–​62.
Koter, Dominika. 2013. “Urban and Rural Voting Patterns in Senegal: The Spatial Aspects of
Incumbency, c.1978–​2012,” Journal of Modern African Studies 51, no. 4: 653–​679.
Koter, Dominika. 2016. Beyond Ethnic Politics in Africa. Cambridge: Cambridge
University Press.
Naude, Pierre-​François, 2011. “Senegal: Quelle chances pour Wade a la présidentielle 2012?”
Jeune Afrique, September 23, 2011. https://​www.jeuneafrique.com/​179301/​politique/​s-​n-​gal-​
quelles-​chances-​pour-​wade-​la-​pr-​sidentielle-​de-​2012/​.
Ndao, El Hadj Ibrahima. 2003. Sénégal, histoire des conquêtes démocratiques. Dakar: Les
Nouvelles Editions Africaines du Sénégal.
Riedl, Rachel. 2014. Authoritarian Origins of Democratic Party Systems in Africa.
New York: Cambridge University Press.
Robinson, David. 1999. “The Murids: Surveillance and Collaboration,” Journal of African
History 40, no. 2: 193–​213.
Schaffer, Frederic. 1998. Democracy in Translation. Ithaca, NY: Cornell University Press.
Villalón, Leonardo. 1995. Islamic Society and State Power in Senegal. Cambridge: Cambridge
University Press.
Young, Crawford, and Babacar Kante. 1991. “Governance, Democracy, and the 1988 Senegalese
Elections. In Governance and Politics in Africa, edited by Goran Hyden and Michael Bratton,
57–​74. Boulder, CO: Lynne Rienner.
Zuccarelli, François. 1988. La vie politique sénégalaise (1940–​1988). Paris: CHEAM.
 Chapter 15

Clientel i sm,
C onstituency Se rv i c e s ,
an d Elections i n Mu sl i m
So ciet i e s

Daniel Corstange and Erin York

The stereotype of elections in the Muslim world is that they are contests between ruling
parties that survive on a mixture of patronage and repression, and an Islamist opposi-
tion that thrives on appeals to religion and privatized social services. This image belies
considerable variation across societies and elections, but, as with many stereotypes, it
does contain a kernel of truth. Many ruling parties do dip into state resources to reward
supporters and woo the uncommitted, while their hamstrung opponents make do with
whatever resources they can mobilize from nonstate sources while hammering home an
ideological message.
We have little to say here about the ideologies, but we have quite a lot to say about the
distributive politics. What drives this emphasis on patronage and other non-​ideological
appeals? What forms do these appeals take, who uses them, and how do they affect dis-
tributional outcomes? Do opposition parties that lack deep pockets or access to state
resources stand a chance of competing? How do Muslim societies compare to their non-​
Muslim counterparts elsewhere in the world?
There may be myriad ways that politicians try to cultivate voters around election time,
but politicians in new democracies and competitive autocracies tend to make greater
use of patronage and service-​based appeals than do their counterparts in richer, more
consolidated democracies (Keefer 2007). These strategies vary in a number of features,
and differ in the degree to which they disrupt the idealized form of accountability we
like to impute to democratic elections. Clientelism, in particular, is distinct from seem-
ingly related appeals by making the act of voting an explicit transaction with an enforce-
able quid pro quo—​voters and politicians both give and get something, and can punish
314    Daniel Corstange and Erin York

each other for failing to live up to the agreement (Cammett 2014; Corstange 2016; Stokes
et al. 2013).
Although we typically expect politicians to be accountable to voters in one form or an-
other, clientelistic exchanges create perverse accountability: voters are now accountable
to the politicians for whom they voted and from whom they extracted benefits (Stokes
2005). Accordingly, clientelism is a form of off-​the-​books politics that is almost always
denounced as a perversion of the democratic process. It contrasts with other forms of
on-​the-​books distributive politics that lack the quid pro quo, are perfectly legal, and are
possibly even desirable from the perspective of accountability.
In this chapter, we explore the prevalence and implications of clientelism in the
Muslim world, drawing on cross-​national and case evidence to illustrate our central
claims. We begin with context: how institutions and the quality of governance in Muslim
societies compare to other parts of the world. We show that the presence of clientelism
is linked more to political and economic factors than to cultural attributes. In particular,
we find evidence that Muslim and Arab polities are not, in fact, more clientelistic than
comparable non-​Muslim states.
We next turn to the distributive implications of this strategy. We first provide a
brief conceptual overview of clientelism and how it differs from seemingly similar dy-
namics such as pork-​barrel politics and constituency services. We then illustrate the
practice of service provision with evidence from Lebanon and Yemen. This reveals the
conditions that favor clientelistic strategies: ruling parties are the primary beneficiaries
of this approach, and poor and rural voters are the most likely to suffer the distributive
consequences.
Finally, we demonstrate that there are alternatives to clientelism, even in patronage-​
oriented societies. We use empirical evidence from Morocco to demonstrate that op-
position parliamentarians use on-​the-​books, institutionalized constituency service
provision to counteract the off-​the-​books exchanges favored by regime-​aligned depu-
ties. We end with a brief discussion of accountability relationships under different forms
of distributive politics: who is accountable to whom, and what this accountability means
for democratic representativeness.
Clientelism is common in Muslim (and comparable non-​Muslim) societies, and
this has important implications for distributive politics and the quality of goods that
voters receive. Yet clientelistic strategies are not the only means of attracting votes,
and other linkage strategies—​including those that do not require the same quid pro
quo agreement—​exist and are electorally successful even in settings where clientelism
dominates.
 Clientelism, Constituency Services, and Elections    315

Contextual Differences between the

Muslim and Non-​Muslim Worlds

Here, we provide a descriptive backdrop to the use of electoral clientelism and other
forms of distributive politics in the Muslim world. In particular, we provide comparisons
of basic institutions and governance between Muslim and non-​Muslim societies in
order to evaluate the degree to which there are, or are not, broad differences relevant
to the use of clientelism. In other words, are Muslim societies distinct on these issues
of governance? Our short answer is no. There are some contextual differences, but they
appear to follow largely from mundane disparities in levels of economic development.
Cross-​country comparisons of Muslim and non-​Muslim countries have long been
a challenge due to the lack of broadly comparable indicators about Muslim, and espe-
cially Arab, societies. This basic data problem has improved in recent years, however,
which means we can make general descriptive comparisons across countries and re-
gions that were impractical for an earlier generation of scholars. We make no claim that
the measures we use are free from conceptual roughness or measurement error, only
that we now have enough comparable data to make the attempt worthwhile despite the
imperfections.
At a basic level, of course, we have always been able to observe differences in regime
type across the world—​specifically, on the broad distinction between democracies
and autocracies. To the degree that clientelistic politics are more common under some
institutions than others, we can gain some context from the distribution of regimes in
the Muslim and non-​Muslim worlds. Although several measures are in widespread cir-
culation, the two most commonly used are from Freedom House and the Polity Project.
Figure 15.1 shows the distribution of 2018 Freedom in the World and Polity scores across
three categories: Arab Muslim, non-​Arab Muslim, and non-​Muslim majority states.1
As Stepan and Robertson (2003, 2004) once observed, and as these data confirm,
there are distinctions between Muslim and non-​Muslim societies—​but these differences
in favor of autocratic government are especially strong in the Arab world, which has
the lowest average scores on both metrics. We do not wish to dwell on the question of
a “democratic deficit” here. Instead, the prevalence of autocracies and unconsolidated
democracies in the Arab and Muslim worlds is relevant to the use of electoral clien-
telism, because these types of societies are the ones in which parties and voters find cli-
entelistic politics to be especially helpful linkage strategies (Keefer 2007).
It is also important to note that the autocratic nature of many of these polities may
also influence which parties are advantaged by this form of competition. As outlined
below, the distinguishing feature of clientelism is a quid pro quo exchange, but these
linkage strategies also rely on the availability of private goods to distribute. In autocratic
polities, because the regime typically has the deepest pockets, regime-​linked parties
may have the most access to the goods used to lure in supporters—​a possibility explored
more fully below.
316    Daniel Corstange and Erin York

(a) (b)
100 10
Freedom in the World score (2018)

75 5

Polity (2018)
50 0

25 −5

0 −10

Arab Muslim Non−Arab Muslim Other Arab Muslim Non−Arab Muslim Other

Figure 15.1: Political freedoms and democracy. Left panel shows distribution of aggregate
“Freedom in the World” 2018 scores across Arab, non-​Arab Muslim, and non-​Muslim-​majority
states and territories. The center lines in the boxes indicate the median country score by group,
and the boxes contain the middle 50 percent of the countries. Scores are out of 100, with higher
scores indicating greater political freedom. Right panel shows distribution of 2018 polity scores
across Arab, non-​Arab Muslim, and non-​Muslim-​majority states and territories. Scores range
from −10 to 10, with −10 indicating complete autocracy and 10 full democracy.

Differences in institutional type do not tell us much about institutional quality. Here,
we draw on a recent cross-​national effort by the World Bank to measure how well na-
tional institutions perform, regardless of type. As a descriptive matter, we care about
quality of governance to the degree that clientelism concentrates in poorly governed
societies, even as we make no causal claims about which one causes the other. The World
Bank collects data on six high-​level governance indicators; we generate an average
quality of governance measure by averaging across these six standardized indicators at
the country level.2
In terms of raw differences, there is a large gap in institutional quality between the
Muslim and non-​Muslim worlds—​with the former nearly a full standard deviation
worse than the latter on average. Yet Table 15.1 shows that at least some of this gap can be
explained by differences in country wealth. Including even a single standard control—​
log GDP per capita—​reduces the gap significantly, especially for the non-​Arab states.
In short, there are some regional differences, but at least a part of the observed differ-
ence is due to prosaic differences in country wealth—​wealthier countries enjoy better
institutions, and Muslim countries just happen to be poorer on average than their non-​
Muslim counterparts. One possible explanation for the residual gap—​particularly for the
Arab states—​is the composition of the economy or the sources of country wealth, with
rentier states that depend on oil or aid inflows enjoying lower-​quality institutions than
their counterparts.
We next turn more directly to measures of the prevalence of clientelism as a linkage
strategy. Here, we use data from a recent expert survey, the Global Party Survey, to
examine linkages across a wide range of countries, including far better representa-
tion of the Muslim and Arab worlds than had previously been available to scholars.3
 Clientelism, Constituency Services, and Elections    317

Table 15.1: Quality of Governance: Regression

Dependent variable:

Average WB Quality of Governance

(1) (2)

Arab Muslim -​0.938*** -​0.712***

(0.192) (0.105)
Non-​Arab Muslim -​0.807*** -​0.243*
(0.176) (0.099)
GDP Control ✓
Observations 202 189
R2 0.166 0.736
Adjusted R2 0.158 0.732

* p<0.05;
** p<0.01;
*** p<0.001.
Note: Table 15.1 presents coefficient estimates and standard errors (in parentheses) from OLS estimation
of average quality of governance metrics on state category alone (model 1) and with a control for log
per capita GDP (model 2). Data are from the World Bank country-​level data for 2018. Units are standard
deviations.

Table 15.2 examines the use of clientelism at the country level, with and without
controls.4 Descriptively, we find suggestive but imprecise evidence that parties in the
Arab world make more use of clientelism than their counterparts in the non-​Muslim
world—​that same difference is larger and more precisely estimated in the non-​Arab
Muslim world. Yet these data also suggest that, when taking into account other
factors that explain clientelistic behavior, parties in Muslim countries are not detect-
ably more likely to use clientelism than their counterparts elsewhere in the world.
What should we take away from this contextual comparison? At a basic level, Muslim
societies, and especially Arab ones, are less democratic than non-​Muslim societies.
Though this observation is not a new one, it is relevant to the extent that clientelism is
more widespread in autocracies and the new democracies that these autocracies may
someday become. Second, institutional quality is lower in Muslim societies—​again, rel-
evant to the extent that clientelism concentrates in weak institutional environments. Yet
a large chunk of the quality deficit is due to a very mundane difference: Muslim societies
tend to be poorer than non-​Muslim ones. Third, parties in Arab and Muslim societies
appear to use clientelism more than their counterparts in the non-​Muslim world. Yet
much—​perhaps even all—​of this difference is due to prosaic differences such as the
wealth and autocratic nature of the societies in question.
318    Daniel Corstange and Erin York

Table 15.2: 2018 Party Strategies: Regression

Dependent variable:

Use of Clientelism

(1) (2)

Arab Muslim 0.246 -​0.021

(0.253) (0.270)
Non-​Arab Muslim 0.677*** 0.310
(0.227) (0.243)
Controls: ✓
Observations 161 143
R2 0.055 0.208
Adjusted R2 0.043 0.179

* p<0.1;
** p<0.05;
*** p<0.01.
Note: Table 15.2 presents coefficient estimates and standard errors (in parentheses) from OLS estimation
of use of clientelistic strategies at the country level, with and without controls for economic and
political characteristics (polity score, log per capita GDP, and inclusion in OECD). Data are from the
Global Party Survey. The outcome variable is standardized, so units are standard deviations.

The analytic, and also normative, conclusion is therefore encouraging: there are some
contextual differences between Muslim and non-​Muslim countries for the purposes of
understanding clientelism, but we need not invoke exotic or essentializing factors to ex-
plain the differences. One can, of course, ask the historically prior question of why these
societies tend to be poorer and more autocratic in the first place (see the chapters by
Chaney and Rubin in this volume). Our point is more discrete: Muslim societies do not
appear to be atypically prone to clientelism much beyond what we would expect from
their levels of wealth and institutional configurations. In the next sections, we clarify the
concepts at hand before turning to an exploration of their implementation in practice.

Clientelism and Constituency Services

as Concepts

So far, we have spoken informally about clientelism. Let us now be more precise about
what we mean by that term. In the abstract, elections oblige candidates to compete for
voters’ support, whether by selecting one candidate over others, or simply by turning out
 Clientelism, Constituency Services, and Elections    319

to vote at all. Politicians have developed a variety of ways to woo voters—​some norma-
tively desirable, and some less so. An idealized version of these linkages envisions voters
choosing between parties based on their proposed policies to allocate public goods and
services across wide swaths of the population. On the other end of the normative spec-
trum, other linkage options target private inducements to supporters at the expense of
the rest of the population.
The conceptual challenge is that these strategies feature different bundles of charac-
teristics, and their individual components tend to be conflated—​programmatic versus
nonprogrammatic distribution, public versus private goods, targeted versus untargeted
beneficiaries, and so on. Here, we untangle several of the key components in order to
focus on what makes clientelism distinct from other linkage strategies. As explained
below, the quid pro quo, rather than the other features that get lumped in with it, is
what defines clientelism and makes it a potentially pernicious form of off-​the-​books
politics.
Some parties rely on programmatic politics to woo supporters. The responsible party
model characterizes the interaction of parties and voters over the course of the electoral
cycle (Kitschelt 2000): parties package policy proposals into a platform, voters choose
the party with the platform they prefer, the winning party fulfills its promises or not
while in office, and then voters decide to retain the party or not at the next election. This
is a stylized account of party strategies, yet it provides a normative ideal that serves as a
benchmark to compare actual electioneering.
The defining characteristic of programmatic politics is not the content of the
policies per se, but rather that they allocate benefits to voters via well-​ defined
and nondiscretionary rules. Put differently, the policies pre-​define a target set of
beneficiaries—​possibly the entire society, but often not—​and everyone eligible for the
benefits can access them, regardless of their political affiliations. These benefits in-
clude both public goods and more targeted private goods—​consider, for example,
Social Security and other state-​run pension plans that provide individual payments to
retirees—​that retain their programmatic appeal by virtue of the formulaic and objective
means to determine eligibility and access.
Programmatic redistribution is still redistribution, of course, and therefore creates
winners and losers. Yet the conceptual conclusion is the same: the appeal is not a clien-
telistic one if political support is not a criterion for access. Consequently, geographically
targeted redistribution—​sometimes called pork-​barrel politics, a term that is unlikely
to catch on in Muslim societies—​may be an appealing tool for politicians operating in
geographically defined districts. Yet it is not clientelistic insofar as benefits flow to all
members of the district regardless of their political leanings.
The same conceptual logic applies to constituency services: a hodgepodge of
intercessions and other services that representatives provide to residents of their
districts, usually on an ad hoc and individual basis. Examples include resolving pen-
sion problems, expediting passport applications, or pressuring bureaucrats to fix
potholes in a road. Although the representatives serve as an intermediary between
the citizen and various governmental offices, they typically offer these services to all
320    Daniel Corstange and Erin York

comers. Indeed, the lack of political criteria for help is what makes them constituency
services rather than clientelistic transactions. Although it may be true that supporters
feel more comfortable approaching a politician for these services, support is not nec-
essary; indeed, politicians may try to engage in more constituency services to culti-
vate nonsupporters by providing observable evidence of evenhandedness among all
members of the district.
Ultimately, the core distinction between clientelism and other forms of distributive
politics is not the nature of the goods offered, nor the degree to which they are targeted
or untargeted. Instead, the distinguishing feature is the existence of an explicit, enforce-
able quid pro quo: parties trade payoffs for votes, and they punish voters who fail to
fulfill their end of the bargain. At its core, clientelism is an exchange relationship with
two-​way accountability—​parties may be accountable to voters at the ballot box, but
voters are also accountable to parties from whom they accept inducements.
How, though, do parties enforce these quid pro quos? In order to police their bargains,
parties must monitor voters, and so act more as political machines than they do as
the preference aggregators in the responsible party model (Kitschelt 2000). In these
settings, party functionaries and independent contractors—​“brokers,” in the catch-​all
sense—​focus on distributing benefits to supporters and monitoring their behavior for
evidence of compliance (Stokes et al. 2013).

Clientelism and Constituency Services

in Lebanon and Yemen

There is, of course, much more to say about clientelism and distributive politics at the
conceptual level. To illustrate these concepts in action, however, we now walk through
a brief sketch of clientelistic politics in Lebanon and Yemen. These two Arab coun-
tries differ in their electoral institutions, but share an emphasis on distributive politics
and service delivery in their elections (for in-​depth treatments, see Cammett 2014 and
Corstange 2016). Like Morocco, which is discussed below, and unlike the Arab Gulf
states, neither country is a significant oil producer that enjoys high per capita wealth due
to oil rents. In that sense, the cases we examine should have patronage environments
more typical of the rest of the Muslim world—​where resource scarcity, not abundance,
is the norm.
Lebanon and Yemen are colorful in their political rhetoric, with a thin veneer of ideo-
logical language. Parties and candidates sometimes release electoral platforms, although
these documents are often long on recitations of desirable outcomes (“reduce illiteracy
by such-​and-​such a percent”) and short on implementation details. In general, a wide
array of politicians and civil society activists in both countries agree that party programs
are irrelevant in elections.5 We are not, of course, denying the existence of ideological
 Clientelism, Constituency Services, and Elections    321

parties; far from it, as we have interviewed leaders from many of them. Our point is that,
with few exceptions, such parties are electorally marginal. Easily derided as “parties on
paper,” they focus their efforts on a small constituency of true believers but win few votes
and fewer seats as a result.6
Currently, the most electorally viable parties that offer more than a token nod to
an ideology are, loosely speaking, Islamist parties: Hizbullah in Lebanon and Islah in
Yemen. No doubt some of their supporters are motivated by the party program or a ge-
neral appeal to religious probity, and for others their main attraction is that they offer a
viable alternative to the status quo. Nonetheless, such appeals are not sufficient to guar-
antee electoral viability. Studies of such movements are plagued by survivorship bias, as
the many failed or marginal Islamist parties—​such as Lebanon’s al-​Jamaʿa al-​Islamiyya
or Yemen’s al-​Haqq—​attest.
Simply repeating slogans that “Islam is the solution” does not, by itself, bring in the
votes, and many non-​Islamist parties campaign with religious language as well. Instead,
the Islamist parties that break through to electoral viability appear to owe their suc-
cess to their capacity to deliver social services to voters, at least in part (Brooke 2019;
Cammett and Luong 2014; Masoud 2014). Indeed, many of their detractors assert that
Islamist popularity is due primarily to handouts—​a well-​worn critique in Lebanon and
Yemen, and in many other societies with burgeoning Islamist movements.
Ruling parties, in contrast, tend to be big tents with little ideological glue holding
them together. While they might be integrators of widely differing visions for the
country, the more critical view derides them as “a collection of opportunists,” in the
words of a Yemeni opposition leader.7 One ambivalent reformer in the Yemeni ruling
party captured both of these views when he grumbled that “parties are just slogans and
banners and haven’t done anything for the people,” but also that his party “is not re-
ally a party, but rather a mosaic of the whole country.”8 In the absence of programmatic
appeals or ideological purity, ruling parties focus instead on the proverbial bread and
butter issues of service delivery.
Notwithstanding cloying rhetoric to the contrary, nearly all major parties attempt
to buy popular support through service delivery and other material payoffs. Despite
grumbling that they should be making laws rather than providing services to voters,
parliamentary deputies readily acknowledge that voters expect their deputies to focus
on the latter—​hence the term “service deputy” (Lust 2009). Thus, one Lebanese MP
complained that “the role of an MP is to make laws. It’s not to provide services. It’s the
role of the state to provide services.” He then acknowledged that

[t]‌here’s no accountability. No one asks what I’ve done for the last four years. It only
matters whose list I’m on, that I went to funerals, that I got someone out of jail, and so
on. No one cares about my voting record.9

Another MP explained that the state was too weak to provide the services that people
need, so they turn to their elected deputies instead:
322    Daniel Corstange and Erin York

Here, constituents count on MPs to render them services. If they need a job they
come to you, if they want to register their kids in school, if they’re pregnant and need
to get into a hospital. It’s everything, from A to Z.10

Put differently, some demands are for collective services that we expect every con-
temporary state to provide, such as electricity, piped water, or schools. Other demands
appear more idiosyncratic and personalized, such as jobs or preferential access to hos-
pital beds. The fact that politicians address these demands idiosyncratically rather than
formulaically offers an avenue for clientelistic exchange.
Despite this emphasis on service delivery, the services are not necessarily of high
quality. Figure 15.2, for example, lampoons the inadequate supply of electricity in Yemen
by comparing it to the reliability of the candles used in the daily blackouts, a visual motif
that appears regularly in opposition-​aligned newspapers (cf. Corstange 2007). Chronic
shortages and low standards pose a puzzle: Who is actually satisfied with the modest
services offered to them? Living in a poor country with limited resources is, of course,
one partial explanation for poor services, but surely not everyone is willing to accept
this excuse forever. The goal of the parties in power, in other words, is to find voters for
whom modest results are enough to win support.
Service-​oriented parties may patronize a wide variety of voters, but, on balance, they
have strong incentives to prefer poor voters over their better-​off counterparts. There
are various explanations for this preference, including the declining marginal utility of
income, shorter time horizons for poor people, and a greater distaste for selling votes

Figure 15.2: “Electricity in our country.”

Source: Al-​Sahwa (Yemen), August 25, 2005.
 Clientelism, Constituency Services, and Elections    323

among richer people. The unifying theme, however, is that it is generally cheaper to
buy a poor person’s vote than a rich person’s. Hence, one Lebanese election monitor
explained that, “to buy a poor person’s vote . . . [maybe] just offer them food and they’ll
be happy. Maybe a cellular card or fuel. For middle-​class people, they might require
jobs, refrigerators, and so on.”11 A member of a Lebanese electoral reform commission
explained in more detail:

Vote buying has a low impact in middle-​class districts, but a big impact in the poor
areas . . . poor people sell their votes. Slightly higher income groups . . . require dif-
ferent kinds of interventions, like getting a job. They won’t sell their votes so readily.
They’ll do it if you provide services for the village, which is more like constituency
services than direct vote buying. They require too high a price to make it econom-
ical . . . . Also, the middle class has a different discourse, and selling votes is shameful,
whereas the poor can’t afford that discourse.12

Poor voters offer appealing targets for service-​oriented parties due to their rela-
tively modest demands, but a particular type of poverty is especially attractive: that of
the rural areas. This claim, no doubt, seems implausible at first, given the neglect so
often documented in the countryside. The Yemeni political cartoon in Figure 15.3, for
example, illustrates the problem: a smarmy-​looking bureaucrat—​and they are always
smarmy in these cartoons—​laughs off a complaint about the lack of services in a village

Figure 15.3: Villager: “The village has no development projects, electricity, water, or roads . . . ”
Official: “All right, we’ll declare it a nature preserve!”
Source: Al-​Nida (Yemen), September 14, 2005.
324    Daniel Corstange and Erin York

by declaring it a nature preserve. The political paradox, of course, is that many villages
that support the ruling party nonetheless lack basic services and infrastructure, and the
latter that does exist often lacks “amenities” such as doors, windows, or roofs.13
Yet vote for the ruling party they do, and often overwhelmingly so. One reason
is simply because of their poverty, which makes them cheaper to satisfy with minor
payoffs; the introduction of even modest services where none had previously existed
amounts to a significant improvement in quality of life. A second, and more pernicious,
reason is that rural voters are more constrained and therefore more attractive as clients.
Compared to their urban counterparts, they tend to be less mobile, easier to monitor,
and more susceptible to peer pressure—​all of which combine to make it easier for parties
to enforce clientelistic quid pro quos and harder for voters to escape. Lebanese election
monitors explained that “in a small village, everyone knows everyone. Someone can
do a detailed report on what you did on Election Day. It’s easier to keep track of vote
buying in rural areas because it’s a smaller community.”14 Their Yemeni counterparts
highlighted the same issue: “[I]‌n the countryside, there’s no anonymity. Everyone knows
everyone, so you can expect . . . pressure to cast the ‘right’ vote.”15
In the abstract, any party could engage in these exchanges with rural residents. In
practice, however, ruling parties are much more able to do so, since they can exploit
state workers in the countryside in a way that their opponents cannot. Yemeni opposi-
tion figures reiterate the problem. “The real democracy is in the capital,” one argued. “In
the outlying regions there’s no competition: everyone works for the state, so they vote
[for the ruling party].”16 Another explained that “the cities are a true reflection of polit-
ical currents in the country, since in the cities it’s actually [issue] politics.” People “come
to the cities, and whereas their villages are voting [for the ruling party],” they do so due
to pressure from the local community. In the cities, however, people “don’t face those
pressures to nearly the same degree, so they can vote their conscience.”17 Cash-​strapped
opposition parties lack the resources to project themselves into the countryside, and, ac-
cordingly, they concentrate their efforts in the high population density areas of the cities
and cede the rural areas to the ruling party—​a huge concession when a large proportion
of the population is rural.
Unsurprisingly, opposition activists complain about the partisan nature of public ser-
vice provision, and there is a widespread belief that access depends on loyalty to the sit-
ting government. One Yemeni opposition party leader explained:

In the existing political system, everything depends on loyalty to the regime. Those
regions that are loyal to the regime get more than their share. Those regions that vote
for a different party—​they’re done for, it’s over. People are just stuck taking patronage
from the [ruling party] if they can get it.18

Statements such as this one capture the common belief that, notwithstanding chronic
resource limitations, the government can turn services on or off at will—​presumably
due to political loyalties. The capacity to exploit state resources for partisan ends,
in the opposition view, is what allows the party in power to win elections despite its
 Clientelism, Constituency Services, and Elections    325

mismanagement. As one opposition-​aligned editor explained, “[W]‌hen we contest

elections here, we’re not facing the ruling party, we’re facing the state . . . . If we face the
state, they win. If we face the party, we win.”19
What, then, can the opposition do to compete with the parties in power? Some reit-
erate their focus on principles in the hopes of overcoming the latter’s advantage in ser-
vice delivery, although at the risk of degenerating into a niche party in the process. Yet
holding firm on principled positions is not the only option for the opposition seeking
to cultivate its constituencies. In addition to relying on nongovernmental funding
sources to pay for services, as many Islamist parties appear to do, opposition parties
have another avenue to provide the services that people expect. In particular, we
might expect opposition MPs to engage in prosaic, on-​the-​books politics: to make use
of their constitutional prerogatives to help people navigate byzantine bureaucracies,
prevent line ministries from overlooking their districts, and ensure that constituents
can access the benefits to which they are formally entitled. Although their practical
capacity to legislate may be minimal, MPs may still be able to engage in real work on
behalf of their constituents. Although this option is not glamorous, it may be surpris-
ingly effective. We take up this possibility in the next section, with novel evidence
from Morocco.

Patronage and Constituency Service

in Morocco

In this section, we explore alternative linkage strategies in Morocco to show that clien-
telism is not the only electorally successful strategy, even in an environment in which
it is highly prevalent. Though the bulk of political power there is held by the unelected
monarch, Morocco holds competitive elections for local, regional, and national political
bodies. The elections themselves are seen as largely free and fair, and a wide range of po-
litical parties participate and win seats. But existing scholarship on Morocco stresses the
overwhelming influence of patronage on political competition in this setting (Denoeux
2007). How do parties lacking access to traditional patronage networks sustain popular
support?
As elsewhere in the Arab world (Lust 2009), Moroccan patronage networks have at
their core the regime—​in this case, the monarchy and its deep pockets. The unofficial
influence of the palace on electoral competition is evidenced by the existence of sev-
eral parties with personal linkages to the regime (Lust-​Okar 2005; Willis 2002). Many
of these “royalist” parties are dominated by local notables descended from the historical
elite, who continue to play an outsize role in politics, especially in rural areas (Liddell
2010). Such actors primarily attract support through clientelistic linkages: in addition
to patronage projects targeted at their supporters, there are widespread perceptions—​
though few direct measures—​of vote-​buying in national elections (Sweet 2001).
326    Daniel Corstange and Erin York

Yet not all parties in this setting use clientelism as their primary linkage strategy. In
general, parties that lack palace connections should have correspondingly reduced ac-
cess to patronage networks. We recap key findings from York 2020 to provide evidence
that such parties can adopt an alternative linkage strategy—​the use of institutionalized
constituency service—​to satisfy constituent expectations and compete for votes. Like
deputies in more democratic contexts, Moroccan legislators are granted a set of insti-
tutional authorities that give them oversight of the bureaucracy. Opposition parties use
these authorities of office to address constituent issues and provide constituency service
without relying on personal networks.
One such power is the ability to submit written requests for information or bureau-
cratic action (queries) to ministry leadership. This is a common feature of parliamen-
tary institutions in both democratic and autocratic states (Martin 2011). In Morocco,
it allows deputies to gather detailed information about their district and pressure
bureaucrats to fix local problems. In interviews, deputies stress the utility of written
queries for addressing local issues and constituent concerns.20 And the subject matter
of actual queries submitted demonstrates that they are regularly used to address public
goods provision within a district. In 2015, for example, a delegate from a nonroyalist
party submitted a request to the Ministry of Health regarding “[p]‌roviding maintenance
to the health center in the municipality of Mrija in Jerada.” Similarly, in 2016, a deputy
from the main Islamist opposition party asked the Ministry of Energy, Water, and the
Environment to address the lack of drinkable water in a remote area of his home dis-
trict. Content analysis reveals that a majority of queries (63 percent) reference district
or subdistrict issues.21 In other words, the evidence (from both deputy characterizations
and the queries themselves) suggests that these queries function as an institutionalized
mechanism for constituency service.
Moroccan deputies submit a huge number of queries each year—​during the legisla-
tive term from 2011 through 2016, they generated 27,196 unique requests submitted to
more than thirty ministers and delegate ministers. But deputies do not make equal use
of this authority. There is substantial variation in the number of queries submitted at
the individual level: 15 percent of deputies did not submit a single query throughout the
five-​year legislative term, while others made hundreds or even thousands of requests.
Notably, nonroyalist parties—​those without regime linkages and corresponding ac-
cess to patronage networks—​use written queries to a much greater degree than their
royalist-​linked counterparts. Figure 15.4 shows the distribution of query submission at
the deputy level, revealing stark variations between queries submitted by the opposi-
tion (top panel) and royalist (bottom panel) parties. No royalist deputy submitted more
than ninety-​seven queries; this was only in the top 40th percentile among opposition
deputies.
This differential pattern across party type is statistically significant. Table 15.3 reports
output from OLS regression at the deputy level of the log number of written queries
submitted on an indicator for whether or not the deputy in question was a member of
30

Nonroyalist
20

10

30

Royalist
20

10

0
4 32 256 2048
No. Written Queries (Log Scale)

Figure 15.4: Written query submission in Morocco: Royalist versus Nonroyalist. Figure shows
distribution on a log scale of number of written queries submitted at the deputy level by affiliation
with royalist (bottom panel) or nonroyalist (top panel) parties.

Table 15.3: Written Query Submission in Morocco: Royalist versus Nonroyalist

Dependent variable:

Log Written Queries

(1) (2)

Royalist -​2.131** -​2.183**

(0.772) (0.773)
Controls ✓
Observations 413 413
R2 0.293 0.400
Adjusted R2 0.291 0.393

* p<0.05;
** p<0.01;
*** p<0.001.
Note: Table 15.3 shows coefficient estimates and standard errors (in parentheses) from OLS regression
of the log number of written queries submitted by Moroccan deputies on their status as members
of royalist or nonroyalist parties. Model 1 is estimated without controls; model 2 includes controls
for incumbency, committee leadership, deputies elected as part of a national list, and participation
in the governing coalition. Standard errors are clustered at the party level and constructed through
bootstrapping.
328    Daniel Corstange and Erin York

a royalist party, with and without controls for a range of deputy characteristics.22 The
results are striking: deputies from royalist parties submitted, on average, more than two
log-​units fewer queries than their nonroyalist counterparts. In other words, the average
opposition deputy submitted a whopping ten times as many queries as the average roy-
alist deputy.
So there is clear evidence that nonroyalist parties make greater use of this form of
legislative activity. But does constituency service provided through institutional
mechanisms constitute a successful linkage strategy? We next look at the relationship
between written queries and electoral outcomes. To do so, we pool written queries
submitted by party deputies within a given district to generate a measure of legislative
activity at the party level.23 We measure electoral outcomes as the change in party vote
share within a district from the 2011 to the 2016 legislative elections (thereby controlling
for past legislative performance) and regress this on the measure of legislative activity
including region fixed effects and controls for other activity not linked to constituency
service.
The results reported in Table 15.4 reveal that legislative activity is closely linked to vote
share. Across the full sample of parties, a log-​unit increase in the number of written
queries submitted is associated with a one percentage point increase in party vote

Table 15.4: Written Queries and Vote Share in Morocco: Royalist versus

Nonroyalist
Dependent variable:

Change in Party Vote Share, 2011–​2016

(1) (2) (3)

Log Written Queries 0.010*** 0.0003 0.017***

(0.003) (0.008) (0.003)
Sample Full Royalist Nonroyalist
Controls ✓ ✓ ✓
Mean Change in Vote Share 0.000 0.001 -​0.001
Observations 263 115 148
R2 0.136 0.164 0.261
Adjusted R2 0.091 0.066 0.189

* p<0.05;
** p<0.01;
*** p<0.001.
Note: Table 15.4 shows coefficient estimates and standard errors (in parentheses) from OLS regression
at the party-​district level of the change in party vote share from 2011–​2016 on the log number of
written queries submitted by deputies from that party. All models include controls for region and other
deputy activity (oral queries submitted).
 Clientelism, Constituency Services, and Elections    329

share in the subsequent elections. But what of the difference in usage by royalist and
nonroyalist parties? We subset the sample to look at effects by party type. For royalist
parties, there is essentially no association between this constituency-​focused legisla-
tive activity and vote share: the coefficient on written queries is insignificant and close
to zero. By contrast, nonroyalist parties see an increase in vote share of 1.7 percentage
points for each log unit increase in the number of queries submitted. This is both sub-
stantively and statistically significant, given that a vote share in the single digits is often
sufficient to win a legislative seat in this context.
The evidence here suggests that even in an environment where patronage is highly
prevalent and programmatic objectives are difficult to implement, clientelism based on
patronage networks is not the only successful linkage strategy. Instead, parties can use
institutionalized constituency service—​the provision of district-​level goods via con-
stitutional legislative authorities—​to satisfy constituents and attract electoral support.
That this activity and its electoral significance vary substantially by party type suggests
that it may offer a functional substitute for clientelism, allowing parties without access
to regime patronage to satisfy constituent expectations.

Conclusion

In this chapter, we have examined linkage strategies in Muslim societies to understand

whether the conventional wisdom on political competition in such settings is consistent
with the reality. We have shown through cross-​national evidence that although there are
disparities in political outcomes between Muslim and non-​Muslim societies, many of
these can be attributed to differences in wealth and governance that generally have little
to do with social or cultural attributes.
And yet clientelistic linkage strategies are quite prevalent in electoral competi-
tion in many Muslim-​majority states, though not necessarily more so than in com-
parable non-​Muslim polities. We have characterized the specifics of such appeals
and demonstrated how they are deployed in a pair of Arab states. This latter discus-
sion reveals that it is largely ruling parties that benefit from this form of competi-
tion, and that the most vulnerable communities—​poor and rural voters—​suffer the
consequences. In particular, the “perverse accountability” relationship created by cli-
entelism often places voters at a disadvantage in dealing with their representatives—​
stuck with politicians that do the bare minimum and unable to transfer their support
lest they be cut out of future goods entirely. That poor communities are most sus-
ceptible to this political strategy suggests that clientelism may exacerbate existing
inequalities in goods provision.
From this perspective, then, the results from the Moroccan context present a more
optimistic view of political accountability. Rather than following a clientelistic logic in
which districts that show strong support for incumbents receive disproportionate at-
tention, the electoral pattern in Morocco suggests a form of accountability that favors
330    Daniel Corstange and Erin York

voters: politicians that devote attention to their constituents tend to receive their sup-
port in future elections.
But while this pattern might imply better distributive outcomes for the electorate, the
findings in Morocco also reveal that it is not the only relationship we observe. Parties
that rely on clientelistic strategies underwritten by regime patronage remain suc-
cessful and do not seem to face the same costs for failing to provide constituency service
through institutional means. These differences suggest that the two linkage strategies
can substitute for one another, perhaps creating within-​society discrepancies in goods
provision based on the category of parties competing within a district. At the very least,
the varied nature of the approaches profiled here reveals a greater complexity to elec-
toral competition in these societies than is obvious on the surface.

Notes
1. Here and elsewhere, we count the following states and territories as Muslim-​majority,
based on 2018 population data: Afghanistan, Albania, Algeria, Azerbaijan, Bahrain,
Bangladesh, Bosnia-​Herzegovina, Brunei, Burkina Faso, Chad, Comoros, Djibouti, Egypt,
Gambia, Guinea, Guinea-​ Bissau, Indonesia, Iran, Iraq, Jordan, Kazakhstan, Kosovo,
Kuwait, Kyrgyzstan, Lebanon, Libya, Malaysia, the Maldives, Mali, Mauritania, Morocco,
Niger, Nigeria, Northern Cyprus, Oman, Pakistan, Palestine, Qatar, Sahrawi Republic,
Saudi Arabia, Senegal, Sierra Leone, Somalia, Syria, Tajikistan, Tunisia, Turkmenistan,
Turkey, the United Arab Emirates, Uzbekistan, and Yemen. Arab states and territories in-
clude Algeria, Bahrain, Comoros, Djibouti, Egypt, Iraq, Jordan, Kuwait, Lebanon, Libya,
Mauritania, Morocco, Oman, Palestine, Qatar, Sahrawi Republic, Saudi Arabia, Somalia,
Sudan, Syria, Tunisia, UAE, and Yemen.
2. The six indicators include control of corruption, political stability and absence of vio-
lence/​terrorism, rule of law, government effectiveness, regulatory quality, and voice and
accountability. These data are available at https://​datacatalog.worldbank.org/​dataset/​
worldwide-​governance-​indicators.
3. Global Party Survey data are accessible at https://​dataverse.harvard.edu/​dataverse/​
GlobalPartySurvey. The data set includes information on party behavior in 161 societies,
including 16 of 22 Arab states and 40 out of 51 Muslim countries. Another project, the
Democracy Accountability and Linkages Project, offers conceptually similar data, but are
from an expert survey in 2008 and covers only 88 societies—​including only 3 Arab coun-
tries, and 13 of 51 Muslim societies in total. Notwithstanding these limitations, we draw the
same broad conclusions from those data as we do from the Global Party Survey data. DALP
data are accessible at https://​sites.duke.edu/​democracylinkage/​data/​.
4. The Global Party Survey data were collected at the party level and include a 10-​point scale
of use of clientelism (“On clientelism, where do parties currently stand on DISTRIBUTING
PUBLIC GOODS, like material benefits, subsidies, construction projects, and jobs. Do
they favor giving universally to all citizens or else primarily to their own supporters?
Where would you place each party on the following scale? 0—​universal distribution; 10—​
distribution only to supporters”). To generate a state-​level clientelism indicator, we weight
party-​level clientelism estimates by party size and standardize this outcome variable for
Table 15.2.
 Clientelism, Constituency Services, and Elections    331

5.Interviews, Beirut and Sanaa, 2005–​2006, 2008–​2009.

6.Interview, opposition party MP, Sanaa, December 2005.
7.Interview, opposition party MP, Sanaa, June 2006.
8.Interview, ruling party official, Sanaa, February 2006.
9.Interview, MP, Beirut, April 2009.
10. Interview, MP, Beirut, July 2008.
11.Interview, NGO director, Beirut, July 2008.
12. Interview, electoral commission member, Beirut, July 2008.
13.Field observations, Dhamar and al-​Bayda provinces, spring 2006.
14. Interview, NGO officials, Beirut, July 2008.
15.Interview, NGO officials, Sanaa, January 2006.
16. Interview, opposition newspaper editor, Sanaa, March 2006.
17.Interview, opposition-​aligned independent, Sanaa, May 2006.
18. Interview, opposition party leader, Sanaa, September 2005.
19. Interviews, opposition newspaper editor, Sanaa, March 2006.
20. Deputy interviews, January 23, 2018, and January 30, 2018.
21. We define district-​level references as those pertaining to the province/​prefecture (the
administrative unit most closely related to legislative districts) and subdistrict as those
addressing communal (municipal) issues.
22. For full multivariate regression results, see York 2020.
23. Because Morocco uses a closed-​list proportional representation electoral system, and
incumbents are dependent on their party to renominate them, this is the most compre-
hensive level at which to test this prediction. Results are consistent when using deputy
level outcomes (i.e., an incumbent winning the subsequent election, conditional on party
nomination).

References
Brooke, Steven. 2019. Winning Hearts and Votes: Social Services and the Islamist Political
Advantage. Ithaca, NY: Cornell University Press.
Cammett, Melani. 2014. Compassionate Communalism: Welfare and Sectarianism in Lebanon.
Ithaca, NY: Cornell University Press.
Cammett, Melani, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17: 187–​206.
Corstange, Daniel. 2007. “Drawing Dissent: Political Cartoons in Yemen.” PS: Political Science
& Politics 40, no. 2: 293–​296.
Corstange, Daniel. 2016. The Price of a Vote in the Middle East. New York: Cambridge
University Press.
Denoeux, Guilain P. 2007. “Corruption in Morocco: Old Forces, New Dynamics and a Way
Forward.” Middle East Policy 14, no. 4: 134–​151.
Keefer, Philip. 2007. “Clientelism, Credibility, and the Policy Choices of Young Democracies.”
American Journal of Political Science 51, no. 4: 804–​821.
Kitschelt, Herbert. 2000. “Linkages between Citizens and Politicians in Democratic Polities.”
Comparative Political Studies 33, no. 6–​7: 845–​879.
Liddell, James. 2010. “Notables, Clientelism and the Politics of Change in Morocco.” Journal of
North African Studies 15, no. 3: 315–​331.
332    Daniel Corstange and Erin York

Lust, Ellen. 2009. “Competitive Clientelism in the Middle East.” Journal of Democracy 20, no.
3: 122–​135.
Lust-​ Okar, Ellen. 2005. Structuring Conflict in the Arab World. New York: Cambridge
University Press.
Martin, Shane. 2011. “Parliamentary Questions, the Behaviour of Legislators, and the Function
of Legislatures.” Journal of Legislative Studies 17, no. 3: 259–​270.
Masoud, Tarek. 2014. Counting Islam. New York: Cambridge University Press.
Stepan, Alfred, and Graeme B. Robertson. 2003. “An “Arab” More Than “Muslim” Electoral
Gap.” Journal of Democracy 14, no. 3: 30–​44.
Stepan, Alfred, and Graeme B. Robertson. 2004. “Arab, Not Muslim, Exceptionalism.” Journal
of Democracy 15, no. 4: 140–​146.
Stokes, Susan C. 2005. “Perverse Accountability: A Formal Model of Machine Politics with
Evidence from Argentina.” American Political Science Review 99, no. 3: 315–​325.
Stokes, Susan C., Thad Dunning, Marcelo Nazareno, and Valeria Brusco. 2013. Brokers, Voters,
and Clientelism. New York: Cambridge University Press.
Sweet, Catherine. 2001. “Democratization without Democracy.” Middle East Report 218: 22–​25.
Willis, Michael J. 2002. “Political Parties in the Maghrib: Ideology and Identification.” Journal
of North African Studies 7, no. 3: 1–​28.
York, Erin A. 2020. “Democratic Institutions under Autocracy.” PhD diss., Columbia
University.
 Pa rt I I I

P OL I T IC A L
AT T I T U DE S
A N D B E HAV IOR
B E YON D VOT I N G
 Chapter 16

Re l igiosit y and P ol i t i c a l
At titudes in T u rk ey
during the A K P E ra

S. Erdem Aytaç

Political attitudes and values held by ordinary citizens play an important role in the
emergence and stability of democratic regimes (Inglehart 2000). Mainwaring (1999,
39), for instance, highlights that the resilience of Latin American democracies in the
late twentieth century can be partly attributed to changes in political attitudes of masses
“towards greater valorization of democracy.” Interpersonal and institutional trust, ap-
preciation of diversity, and ideological moderation are among the attitudes thought
to foster and reinforce democratic governance (Inglehart 2000; Rose, Mishler, and
Haerpfer 1998).
One of the explanations to account for Muslim-​majority countries being democratic
under-​performers focuses on such attitudes in these societies, and especially among
devout Muslims. In recent cross-​national surveys, Muslim-​majority countries tend to
occupy the lowest ranks with respect to average levels of interpersonal trust, perceived
legitimacy of secular officials, and support for freedom of speech and gender equality
(Inglehart et al. 2014; Yuchtman-​Ya’ar and Alkalay 2010). In turn, a voluminous litera-
ture reports a positive relationship between factors associated with modernization (e.g.,
rising levels of education, income, urbanization) and holding pro-​democratic attitudes
in these countries (e.g., Çiftçi 2010, 2012; Jamal 2006; Tessler 2002). These findings sug-
gest that lower levels of socioeconomic development might be the primary culprit for
the relative dearth of pro-​democratic attitudes in Muslim-​majority societies.
Beyond the role of socioeconomic development in the shaping of pro-​democratic
attitudes, we should also note that in many Muslim-​majority countries there have been
severe restrictions on Islamic organizations and communities (e.g., brotherhoods, or-
ders, etc.), as well as on public expressions of religiosity. Examples include crackdowns
on the Muslim Brotherhood in Egypt and Jordan and the severe restrictions around
the wearing of headscarves by women in Turkey and Tunisia in the past until recently.
336   S. Erdem Aytaç

And since in many of these countries there have been regular elections and party
competition—​that is, at least a semblance of democracy—​devout Muslims’ political
attitudes and views on democracy might have been at least partly shaped by the regimes’
distance and even outright hostility toward religiosity. Would there be a change in these
attitudes when the political regime adopts a more favorable stance toward religion? For
instance, would we observe devout Muslims adopting political attitudes more condu-
cive to liberal democracy, such as ideological moderation or higher levels of interper-
sonal trust, if they become more satisfied with the political system?
In this chapter I focus on this question of how policies and approaches toward reli-
gion in public life affect political attitudes among Muslims by leveraging the change in
the Turkish state’s policies during the incumbency of the Justice and Development Party
(Adalet ve Kalkınma Partisi, or AKP) from 2002 to 2018. Turkey is an opportune case
for this task, because when the Islamist AKP came to power in 2002, debates around the
role of religion in public life and politics had peaked, and there were several restrictions
in place regarding the role and visibility of Islam in Turkish society. The AKP gradu-
ally lifted these restrictions, especially after its conservative turn starting around 2011
(Özbudun 2014). Islamic religiosity has become a major defining factor, not only in
public life, but also in education policies and state bureaucracy, while individuals and
organizations with a secular outlook increasingly face discrimination and exclusion. In
the face of these developments, can one observe a change in the political attitudes of de-
vout Muslims?
Drawing on original, nationally representative surveys fielded during the pe-
riod from 2002 to 2018, we are indeed able to diagnose a sense of relief among devout
Muslims in Turkey. In particular, they are much less concerned about restrictions on
religious freedoms, and increasingly think that they are treated fairly by the political
system. In turn, the proportion of devout Muslims favoring the establishment of a
shariʿa-​based religious state in Turkey dropped substantially during this period, and
devout Muslims are more satisfied with the way democracy works in Turkey than are
less devout individuals.
There is no evidence that these changes have been accompanied by a more pluralistic
understanding of democracy, however. We observe no improvement in interpersonal
and institutional trust among devout individuals from 2002 to 2018, and they also have
not moved toward an ideologically more centrist position. Instead, we observe that de-
vout individuals in Turkey hold more populist attitudes than others; in particular, they
are more likely to display a majoritarian view of democracy where the importance of
checks on the executive are downplayed. This affinity of devout individuals for populist
principles is in line with the political platform of the incumbent AKP.
In the following I first present an overview of modern Turkish politics to highlight the
role of religion. This discussion also outlines how religion has become a point of conten-
tion, especially in the period immediately preceding the AKP rule, and touches upon
the developments during the AKP period with respect to the public role and visibility
of Islam. The analyses of religiosity and political attitudes in the AKP era follow this
discussion.
Religiosity and Political Attitudes in Turkey during the AKP Era    337

Religion in Modern Turkish Politics

In a seminal article, Şerif Mardin (1973) presented a framework to understand the major
contours of modern Turkish politics. Mardin argued that a sociocultural cleavage pit-
ting the ruling elites of the “center” against a culturally heterogeneous “periphery” had
been inherited by the Turkish Republic from the Ottoman Empire. In the Ottoman
polity the center was constituted by the imperial house and its ruling apparatus in the
society, and the periphery consisted of various traditionalist ethnic and religious “pri-
mordial groups” that were kept out of power (174). This cleavage was first and foremost
a cultural one; while the center dominated the religious establishment, justice, and ed-
ucation, its claim to legitimacy was bolstered by its self-​ascribed cultural preeminence.
From the nineteenth century, religiosity came to be increasingly identified with the
periphery, as positivism and secularism were the cornerstones of Turkish moderniza-
tion (Göle 1997). Steps taken to modernize the military and administration have led
to the adoption of some form of Western culture, which further alienated peripheral
masses from the rulers and agents of the center. In this process, “a clinging to Islam, to
its cultural patrimony, was the province’s response to the center’s inability to integrate it
into the new cultural framework,” and religiosity became a characteristic marker of the
periphery (Mardin 1973, 179). Atatürk’s reforms that spanned two decades following the
foundation of the Republic in 1923 constituted the zenith of the Turkish modernization.1
As Sunar and Toprak (1983) highlight, Islam had served as a connecting function be-
tween the Ottoman center and periphery, and this link was severely weakened as a result
of these secular reforms, which corresponded to a revolution in cultural values (Mardin
1971). Thus, the center-​periphery cleavage primarily took the form of a “cultural bifur-
cation that crystallized around the secular progressivism of the state and the ascriptive,
religious traditionalism of local communities” (Sunar and Toprak 1983, 426).
Kalaycıoğlu (1994, 403) characterizes the center during the Republican period as the
“estate of a coherent body of nationalist, centralist, laicist elite which holds the view
that it represents and protects the state.” The quasi-​autonomous bureaucracy, espe-
cially that of the judiciary and military, the large, state-​dependent businesses, and the
mainstream intellectual community have been integral parts of the center. In terms of
societal bases of support, the center has traditionally relied on groups with an urban
presence, higher levels of education, lower levels of religiosity, and lower concentration
of ethnic minorities (Çarkoğlu 2012), and the political arm of centrist constituencies was
the founding party of the Republic, the Republican People’s Party (Cumhuriyet Halk
Partisi, CHP).
It was not until the introduction of multiparty elections in 1946 that the peripheral
masses found an opportunity to participate in the political arena. The Democratic Party
(Demokrat Parti, DP) was founded after an elite split within the CHP and emerged as
the representative of the periphery. Mardin (1973, 185) highlights that the DP appealed to
Islam as the culture of the periphery by promising a liberalization of religious practices
338   S. Erdem Aytaç

and “relegitimized Islam and traditional rural values.” The DP won the first free and fair
elections in 1950, and the CHP remained in the opposition throughout the 1950s.
The parties of the periphery, following the DP tradition, have since dominated centrist
parties in elections, despite several interventions by the military, which saw itself as the
guardian of the regime (Cizre-​Sakallıoğlu 1997). These electoral successes have had little
impact on the dominant status and value system of the center, however (Kalaycıoğlu
1994). The Constitution of 1961, drafted in the aftermath of the coup against the DP in
1960, built a regime with strong counter-​majoritarian institutions that were able to in-
sulate themselves from the influence of elected governments through their autonomous
recruitment and promotion practices. Therefore, the elites of the center retained polit-
ical power even though the society’s religious and conservative majority threw support
behind the parties of the periphery at the ballot box.
A number of developments in the 1980s and 1990s brought about the rise of po-
litical Islam in Turkish politics. The military junta (1980–​1983), following the 1980
coup, harshly suppressed leftist movements and organizations, and a combination
of religiosity (Sunni Islam) and Turkish nationalism was promoted with the hope of
undermining left-​wing politics against the perceived threat of communism (Akan
2012; Çarkoğlu and Kalaycıoğlu 2009). This period coincided with the significant
migration of rural masses to big cities, and the formal mechanisms of the state were
simply unable to cope with the accompanying socioeconomic pressures. In the ab-
sence of leftist organizations, which had a strong presence among the urban poor in
the 1970s, Islamist networks and parties became the dominant actors in shantytowns
of big cities like Istanbul and Ankara, inevitably carrying their conservative ideology
with them (Eligür 2010). Around the same time, Islamist intellectuals became more as-
sertive in the public domain through a number of Islamist periodicals and newspapers,
and public expressions of religiosity (e.g., veiled women becoming visible on university
campuses, discouragement of alcohol consumption in public) became more frequent
(Göle 1997).
The “centrist” establishment took two additional blows in the 1990s. First, the inten-
sity of the armed conflict between the Turkish state and the Kurdish separatist organiza-
tion PKK that started in 1984 reached its peak in the first half of 1990s. The clashes had
very negative effects on the daily lives of Kurdish citizens living in the southeastern part
of the country. In this context, many Kurds, already more conservative than the average
Turkish citizen, were attracted to political Islam that emphasized the umma instead of
nationalism and was critical of the Republican institutions (Çarkoğlu and Kalaycıoğlu
2009). Second, the fragmented center-​right and center-​left parties of the day proved to
be incompetent to cope with the economic challenges, leading to a period of unstable
coalition governments and successive economic crises (Öniş 1997).
A consequence of these developments has been a sharp ideological turn toward the
right in the Turkish electorate, and the accompanying rise of political Islam. While
in 1990 about 23 percent of Turkish voters were on the right of the ideological spec-
trum, just six years later, in 1996, this proportion increased to 39 percent (Çarkoğlu and
Kalycıoğlu 2009).2 And in the electoral scene, the Islamist Welfare Party (Refah Partisi,
Religiosity and Political Attitudes in Turkey during the AKP Era    339

RP) first won the metropolitan municipalities of Istanbul and Ankara in 1994, and then
emerged as the leading party in the general election of 1995, obtaining 21 percent of the
votes. In 1996 the leader of the RP, Necmettin Erbakan, became the prime minister in a
coalition government.
This success of political Islam raised alarm bells among the centrist establishment,
especially the military. After some controversial moves by Erbakan and RP executives
that exacerbated the tension, the military declared political Islam as the primary threat
to the Republic in a National Security Council meeting on February 28, 1997, and
demanded action by the government through a directive that they themselves prepared.
The eighteen-​point demands, reluctantly approved by Erbakan, included the exten-
sion of compulsory primary education from five to eight years, which would effectively
close the religious Imam Hatip middle schools, transfer of private dorms and schools
affiliated with Islamic brotherhoods to the Ministry of Education, strict implementa-
tion of laws related to clothing in universities and public offices, and precautions about
Islamist “infiltration” of state institutions. While only one of the directive’s demands was
actually implemented (extension of primary education, see Eligür 2010), this interven-
tion of the military was followed by the mobilization of the secular sectors of society,
leading to the toppling of the RP-​led coalition government and closure of the RP by the
Constitutional Court in 1998. The broader result of this process was a narrower space for
public expressions of religiosity and a retrenchment of Islamist organizations.
Thus, the second half of the 1990s has been a period of showdown between conser-
vative, peripheral actors and the secular establishment. In the political scene, the RP
was replaced by the Virtue Party (Fazilet Partisi, FP), which was subsequently closed
down by the Constitutional Court in 2001. A split within the Islamist movement had
been visible within the FP, and upon its closure the younger generation of Islamists,
headed by Recep Tayyip Erdoğan, formed the AKP (Hale and Özbudun 2010). The coa-
lition governments formed after the toppling of the RP proved to be short-​lived, and in
the aftermath of two consecutive economic crises, in 1999 and 2001, and a devastating
earthquake in 1999, early elections were held in 2002. The AKP won the election with
34 percent of the votes and an unprecedented almost two-​thirds of parliamentary seats
(due to several parties getting votes just below the national threshold of 10 percent),
allowing it to form a single-​party government.

The Post-​2002 AKP Era

The AKP pursued a different strategy than its predecessors by downplaying elements of
Islamist demands and mobilization in its rhetoric and policies until the end of its second
term in government in 2011 (Altınordu 2016). Instead, the party leadership defined
themselves as “conservative democrats” and emphasized a politics of economic com-
petence and democratizing reforms, with the prospect of membership in the European
Union (Öniş 2015).
340   S. Erdem Aytaç

Starting around 2010–​2011, the AKP moved ahead with a discourse and policies
aimed at expanding the public role and visibility of Islam in Turkish society. By this time
the AKP had successfully subdued the elites of the center, especially the military and ju-
diciary, through a series of controversial trials involving hundreds of senior active and
retired military officers, along with constitutional amendments that curbed the power
of secularist groups within the high judiciary. As a result, the era of “military tutelage” in
Turkish politics was effectively over, and the AKP ruled in a setting marked by severely
weakened horizontal accountability (Akkoyunlu and Öktem 2016).
The growing conservative trend in AKP’s policies manifested itself mostly in the
area of education. The middle sections of the religious Imam Hatip schools that were
closed in 1997 reopened in 2012, and many secular high schools were converted to re-
ligious ones. By 2018 the total number of students in these religious schools had risen
fivefold since 2012, and although Imam Hatip schools made up about 11 percent of
the total middle and high school population, they received close to a quarter of the
budget (Butler 2018). At the same time, restrictions on the placement of Imam Hatip
graduates to universities were also lifted. The conservative turn of the AKP has taken
its toll in nominally secular public schools as well. Compulsory religious teaching in
these schools was doubled to two hours per week, and extra optional religious courses
were introduced. While these new courses were optional on paper, in many schools they
were effectively the only available optional courses. References to Islam have steadily
increased in the curriculum, and religious groups outside the education system are now
allowed to conduct special seminars in schools (Gall 2018; Kingsley 2017).
The AKP also engaged in several policy changes with an eye toward pleasing devout
Muslims beyond the realm of education. The restrictions around the religious headscarf
were lifted in this period, which was one of the primary sources of resentment among
the conservative majority of Turkish population (Çarkoğlu 2010). New restrictions on
the sale of alcoholic beverages were introduced, and their advertisement or promo-
tion of any kind were prohibited, together with substantial increases in taxation. These
policies have effectively removed alcohol from public spaces and confined its consump-
tion to mostly private settings. Changes were made to the Turkish civil code in 2017
allowing muftis, religious civil servants, to perform marriages in the same capacity as
municipal officials.
While these developments raised concerns among the more secular-​ oriented
segments of society, anecdotal accounts suggest that they brought a sense of relief to
conservatives. The changes in the education system expanded the opportunities for
parents who want their children to get religious education in a conservative setting
without any limitations upon graduation (Lüküslü 2016). A majority of Turkish adult
women are estimated to wear some kind of a headscarf (Çarkoğlu 2009); the lifting of
restrictions in this area paved the way for them to attend universities and join the bu-
reaucracy without making a change in their appearance. An interviewee for a Reuters
report on Imam Hatip schools said, “Muslims have now reached a point where they can
breathe more easily in their own country . . . in the last 15 years this government has
Religiosity and Political Attitudes in Turkey during the AKP Era    341

shown respect to Muslims” (Butler 2018). In the following section we look at this senti-
ment and its consequences on political attitudes more systematically.

Religiosity and Political Attitudes

in the AKP Era

In this section I track the relationship between religiosity and several political attitudes
during the AKP era in Turkey. My data come from the Turkish Election Study surveys
conducted around the general elections of 2002, 2007, 2011, June 2015, and 2018. These
face-​to-​face surveys are all nationally representative, employing a similar probability
sampling procedure, with sample sizes ranging from 1,642 to 2,156.3 Not every political
attitude question we analyze was asked in every survey, but whenever they were asked,
the question and answer formats were identical. This comparability of the measures
allows me to document the aggregate changes in attitudes over time, and the representa-
tiveness of the samples makes it possible to generalize the results to the adult population
of Turkey.
To analyze whether and how political attitudes vary across individuals with different
levels of religiosity, I separate respondents in the samples into three levels of religiosity
that I label as non-​devout, devout, and very devout. A standard question in all surveys
asked respondents how religious they saw themselves on a 0–​10 scale, where zero
corresponds to “not religious at all” and ten to “very religious,” irrespective of their fre-
quency of prayer. The distributions of answers are heavily skewed toward higher values,
with mean levels ranging from 6.9 in 2002 to 7.4 in 2018. I categorize respondents into
the three religiosity levels such that those who gave answers between zero and six are
labeled as non-​devout, those with answers of seven or eight are labeled as devout, and
those with answers of nine or ten are considered as very devout individuals. These values
are selected such that each category roughly contains about a third of respondents over
the entire period of analysis.
While this measure corresponds to the belief dimension of religiosity (Wald and
Smidt 1993), it also closely reflects its behavioral dimension. In the surveys, respondents
were asked how frequently they prayed (namaz) in the past year, with the answer option
“more than once a week” being the highest indicator of religious practice. Averaging
across all the surveys, only about 26 percent of subjectively defined non-​devout
individuals reported praying more than once a week. This figure increased to 55 per-
cent for devout individuals and to 76 percent for very devout individuals. Thus, there is
a close match between the subjective religiosity measure and reported religious prac-
tice. In the analyses I prefer to use the subjective measure, because it lends itself to a
more equal distribution of individuals into three categories. The substantive results of
the analyses are identical if we use the practice measure.
342   S. Erdem Aytaç

Figure 16.1 presents the distribution of non-​ devout, devout, and very devout
individuals across the surveys. We observe a decrease over time in the proportion
of non-​devout respondents, from 44 percent in 2002 to 28 percent in 2018, while the
proportions of devout and very devout respondents increase from 32 percent to 42 per-
cent and from 24 percent to 30 percent, respectively, during the same period. These
results point to an increase in subjective religiosity during the AKP era. On average,
more devout individuals are more likely to be female, older, reside in rural areas, and
have lower levels of education and income.
We begin our analyses of political attitudes by focusing on respondents’ views on
restrictions that devout Muslims face in Turkey with respect to their religiosity. Two re-
lated questions in this regard probe respondents about whether Muslims in Turkey are able
to worship freely, and whether devout (Muslim) people face oppression. The percentages
of respondents answering these questions affirmatively are presented in panels (a) and
(b) of Figure 16.2, respectively. We see that across all the subgroups considered there
were significant improvements from 2002 to 2018 with respect to perceptions of devout
Muslim’s conditions in Turkey. The change in perceptions among very devout individuals
is especially stark: in 2002 about half of very devout individuals (56 percent) thought that
Muslims in Turkey were able to worship freely, but this figure reached 96 percent in 2018.
Similarly, in 2002 close to half of very devout individuals (47 percent) held the view that
devout people in Turkey faced oppression, whereas in 2018 only 5 percent of them thought
this way. It is remarkable that at the beginning of AKP’s tenure a considerable segment
of even non-​devout individuals acknowledged the hardships devout Muslims faced, and
over time we observe a convergence of views toward a more positive environment in this
regard across individuals with different levels of religiosity.

100%

80%

60%

40%

20%

0%
2002 2007 2011 2015 2018

Non-devout Devout Very devout

Figure 16.1: Distribution of non-​devout, devout, and very devout respondents in the surveys.
 100 100
Percentage of Respondents

Percentage of Respondents
80 80

60 60

40 40

20 20

0 0
2002 2007 2011 2015 2018 2002 2007 2011 2015 2018

Non-devout Devout Very devout Non-devout Devout Very devout

Figure 16.2: Views on restrictions faced by devout Muslims in Turkey.

344   S. Erdem Aytaç

100

Percentage of Respondents
80

60

40

20

0
2007 2011 2015 2018

Non-devout Devout Very devout

Figure 16.3: Are citizens like you treated fairly in our current political system? “Yes” respondents.

These results suggest that devout Muslims in Turkey consider the tenure of AKP gov-
ernment as a period of “decompression of Islam,” similar to the characterization of the
1950s by Sunar and Toprak (1983). This is also reflected in how they view the political
system. Since 2007 our surveys asked respondents whether citizens like them are treated
fairly in the current political system. In 2007 about 52 percent of very devout individuals
responded positively to this question, and this figure steadily increased over time to
63 percent in 2018 (Figure 16.3). While non-​devout individuals consistently had a more
negative view of the fairness of the political system, the percentage of individuals in this
group who think they are treated fairly decreased slightly over this period. Thus, the gap
in positive responses between non-​devout and devout respondents has widened. As of
2018 there is a clear, positive relationship between perceptions of the fairness of the po-
litical system and higher levels of religiosity.
As devout Muslims increasingly consider the political system in Turkey to be fair to-
ward them, have their views about a shariʿa-​based state changed? Figure 16.4 presents
the percentage of respondents from 2002 to 2018 that favor the establishment of a
shariʿa-​based religious state in Turkey. We see a precipitous decline in the proportion
of very devout respondents that support a shariʿa-​based state in Turkey, from about
27 percent in 2002 to 14 percent in 2018. A similar decline, though less dramatic, can
be observed among other groups as well: from 15 percent in 2002 to 12 percent in 2018
among devout respondents, and from 12 percent to 7 percent among non-​devout ones. It
would be wiser to focus on the trend of these figures than on their absolute values, since
the word shariʿa has a rather ambiguous meaning for the Turkish electorate. Çarkoğlu
(2004) presents evidence that when people are asked to choose between the present sec-
ular civil code regulations (about marriage, divorce, inheritance) and arrangements in
accordance with Islamic law, support for Islamic law arrangements is much lower than
the support for shariʿa. He concludes that a certain proportion of the electorate seems
unable to voice objection to a shariʿa-​based rule when asked in a generic manner or es-
tablish its implications to daily life.
Religiosity and Political Attitudes in Turkey during the AKP Era    345

100

Percentage of Respondents
80

60

40

20

0
2002 2007 2015 2018

Non-devout Devout Very devout

Figure 16.4: Would you favor establishment of a shariʿa-​based religious state in Turkey? “Yes”
respondents.

The decrease in preference for the establishment of a shariʿa-​based religious state

among devout individuals has been accompanied by greater levels of satisfaction with
the way democracy works in Turkey. A question in our surveys first included in 2007
asked respondents to indicate how satisfied they were with the way democracy works
in Turkey on a 0–​10 scale, with higher values indicating more satisfaction. In Figure
16.5 we see that the average level of satisfaction among non-​devout respondents has
not changed over the eleven-​year period, and consistently remained lower than among
more devout individuals. In contrast, we do observe an increase in satisfaction among
very devout respondents that is driven by increased levels of satisfaction from 2015 to
2018. The mean level of satisfaction with democracy among very devout respondents is
5.0 in 2007, 4.9 in 2015, and 6.1 in 2018. The period from 2015 to 2018 corresponds to an
increasing centralization of power by Erdoğan, embodied by the transition to a presi-
dential system of government and a sharp conservative turn of the AKP.
The changes in political attitudes among devout Muslims in Turkey from 2002 to
2018 can be summarized as follows. Compared to 2002, devout individuals as of 2018
are much less concerned about restrictions on religious freedoms, and only a small mi-
nority of them think that they face oppression. During this period, devout individuals’
perceptions of being treated fairly by the political system have gradually improved, and
there is a positive association between these perceptions and levels of religiosity. A sim-
ilar positive relationship exists between satisfaction with democracy in Turkey and re-
ligiosity: very devout individuals declare more satisfaction than devout or non-​devout
individuals. In turn, only about one in seven very devout individuals favor the establish-
ment of a shariʿa-​based religious state in Turkey as of 2018, compared to about one in
four in 2002.
Have these changes in political attitudes been accompanied by a more pluralistic un-
derstanding of democracy? There are a couple of ways of approaching this important
346   S. Erdem Aytaç

10

Mean Evaluation Score

6

0
2007 2011 2015 2018

Non-devout Devout Very devout

Figure 16.5: How satisfied are you with the way democracy works in Turkey? 0 corresponds to
“not satisfied at all” and 10 to “very satisfied.” Mean evaluation scores.

question. First, we can look at patterns of interpersonal and institutional trust (Figure
16.6). We see that there has been a slight improvement from 2002 to 2018 in levels of in-
terpersonal trust across all categories of religiosity. During this period, the percentage
of individuals who agree that most people can be trusted increased from 8 percent to
10 percent among non-​devout individuals and from 9 percent to 11 percent among very
devout individuals (panel a). These changes are quite small, and devout individuals do
not differentiate themselves from others, however—​they just follow the general trend.
In panel (b) we consider trust in parliament, and again there is little change from 2002 to
2018. Together, these results suggest that there has not been a substantial improvement
in levels of interpersonal and institutional trust in Turkey from 2002 to 2018.
Another way to explore whether there are signs that the increased satisfaction of
devout individuals led them to adopt a more pluralistic view of democracy is to see
whether they hold more ideologically moderate positions now compared to the past.
Figure 16.7 presents the mean score of respondents over time on the left-​right ide-
ological scale (0 indicating extreme left, and 10 extreme right). Devout and very de-
vout individuals consistently positioned themselves more to the right than non-​devout
individuals, and there has been little change in the mean positions over time. Therefore,
we do not have an indication that devout individuals have shown ideological modera-
tion during the AKP government.
In fact, there is evidence that, as of 2017, religiosity is positively associated with having
more populist attitudes, and thus a majoritarian understanding of democracy. A nation-
ally representative survey (N = 1,954) fielded in 2017 presented respondents with a list
of populist statements to measure the prevalence and correlates of populist attitudes
among Turkish voters.4 The statements tapped into different dimensions of populism: a
Manichean view of politics as moral struggle, anti-​elitism, people-​centrism, and pop-
ular sovereignty (Box 16.1). Respondents were asked to indicate whether they agreed
or disagreed with the statements on a five-​point scale, with options ranging from “I do
 40 10
Percentage of Respondents

Mean Evaluation Score

30
7

20

4
10

0 1
2002 2007 2015 2018 2002 2015 2018
Non-devout Devout Very devout Non-devout Devout Very devout

Figure 16.6: Interpersonal and institutional trust in Turkey.

348   S. Erdem Aytaç

Box 16.1: Statements to Measure Populist Attitudes in Turkey

Manichean outlook
Politics is ultimately a struggle between good and evil.
What people call “compromise” in politics is really just selling out one’s principles.
I would stop talking to a friend who had unacceptable political opinions.
Anti-​elitism
Most politicians do not care about the people.
The power of a few special interests prevents our country from making progress.
The differences between the elite and the people are larger than the differences among the people.
People-​centrism
The people, and not politicians, should make our most important policy decisions.
Referendums are the ultimate measure of the will of the people.
Politicians in the parliament need to follow the will of the people.
Popular sovereignty
Political leaders do not need to be checked by institutions since people make their decision in
elections.
Having a strong leader in government is good for Turkey even if the leader bends the rules to get
things done.
Most of the time parliaments do nothing but preventing the governments to do their jobs.

10

8
Mean Score

0
2006 2007 2011 2015 2018

Non-devout Devout Very devout

Figure 16.7: Self-​positioning on the left-​right ideological position. Mean scores on a 0–​10
scale where higher values indicate more right positions.
Religiosity and Political Attitudes in Turkey during the AKP Era    349

Mean Score 1

0
Overall Manichean People Anti-elitism Popular
populism outlook centrism sovereignty
–1

–2
Non-devout Devout Very devout

Figure 16.8: Support for populism and for its different dimensions across levels of religiosity.

not agree at all” (coded −2) to “I agree completely” (coded 2), the middle position being
“neither agree nor disagree” (coded 0). As agreement with the statements indicates sup-
port for populist attitudes, I created an index that takes the mean value of the answers
given to these statements by a respondent; therefore, the resulting index has a range
from −2 (the most anti-​populist position) to 2 (the most populist position). The mean
score in the sample is 0.36.
Figure 16.8 presents the mean scores of the index of populist attitudes (overall popu-
lism) and of the four different dimensions of populism for non-​devout, devout, and very
devout respondents in the sample. We observe a positive correlation between religiosity
and populist attitudes—​both the overall populism and different dimension scores are
higher for more devout individuals. While support for people-​centrism and anti-​elitism
statements are especially high across all levels of religiosity, the difference between de-
vout and non-​devout individuals is especially stark with respect to the popular sover-
eignty dimension of populism. The mean scores for this dimension of populism stand
at −0.1, 0.27, and 0.34 for non-​devout, devout, and very devout individuals, respectively.
The second-​largest difference in support for populist principles between devout and
non-​devout individuals is on the people-​centrism dimension.
The popular sovereignty and people-​centrism dimensions of populism emphasize the
centrality of people’s will in politics (Mudde and Rovira Kaltwasser 2017). In this view,
politicians should simply respond to people’s demands rather than leading them, which
can be best elucidated through elections and referenda. Institutions of horizontal ac-
countability, such as the judiciary or a strong legislature, are frowned upon by populists,
350   S. Erdem Aytaç

since they can serve as impediments to the exercise of people’s will by the executive.
Our data reveals that devout Turkish individuals are closer to such a view of democracy
than non-​devout ones. For example, about 39 percent of non-​devout respondents in our
sample agree with the statement that “Having a strong leader in government is good for
Turkey even if the leader bends the rules to get things done;” and this figure increases
to about 50 percent among devout and very devout respondents. Similarly, while about
31 percent of non-​devout respondents agree that “Most of the time parliaments do
nothing but preventing the governments to do their jobs,” about 43 percent of devout
and very devout respondents agree with it.
These results suggest that support for populist attitudes are more prevalent among
devout Turkish voters. Rather than adopting a more pluralistic view of democracy
as a result of perceptions of increased religious freedoms and satisfaction with the
political regime, devout individuals are more likely to display a majoritarian under-
standing of democratic governance. Of course, this result cannot be separated from
the fact that devout individuals are more likely to support the Islamist AKP, which
has been in power since 2002. This is a party that caters to the interests of devout
individuals with a heavily populist platform (Aytaç and Öniş 2014), and it seems that
devout individuals have opted to give the executive a blank check to rule, only to be
constrained by electoral performance, in line with the “delegative democracy” charac-
terization of O’Donnell (1994).

Conclusions

In this chapter I examined the relationship between religiosity and political attitudes
in a Muslim-​majority country during a period when the state’s approach and policies
toward religion in public life have fundamentally changed. The AKP in Turkey came to
power in 2002 when there were several restrictions on the public role of Islam, and it not
only gradually lifted these restrictions over time, but also put Islamic religiosity at the
center of state bureaucracy and policies. I asked if one can observe any accompanying
changes in the political attitudes of devout Muslims in response to this more favorable
outlook of the state toward religion, and if yes, in which direction.
The analysis of a series of nationally representative surveys fielded over a period of
sixteen years (2002 to 2018) during which the AKP has been in power reveals two sets
of findings. First, the AKP governments’ positive approach toward Islamic religiosity
in public life seems to have led to a rapprochement of devout Muslims with the polit-
ical regime. Devout Muslims no longer think that they face oppression or limitations
on worship. They also display higher levels of perception of fair treatment by the regime
and satisfaction with the way democracy works compared to less devout individuals.
The second set of findings, however, points to a lack of ideological moderation or
meaningful change in the levels of interpersonal and institutional trust among devout
Religiosity and Political Attitudes in Turkey during the AKP Era    351

Muslims. Rather than a more pluralistic understanding of democracy that might be

expected to develop due to the rapprochement with the political regime, we observe
that devout Muslims display higher levels of populist attitudes than their less devout
compatriots. These populist attitudes emphasize a majoritarian understanding of de-
mocracy by prioritizing vertical accountability through elections and referenda at the
expense of institutions of horizontal accountability.
This picture is compatible with the theories of elite-​driven public opinion (e.g.,
Zaller 1992). As the ruling AKP elites employ a political discourse with frequent pop-
ulist messages, those individuals with affinity to the party and who are satisfied with its
policies, like devout Muslims, are more likely to embrace these messages. The favorable
policies and approaches of the government toward religion in public life seem to have
little effect on political attitudes of devout Muslims, other than making them more com-
fortable with the political regime. This is a rather disappointing finding, as it suggests
that a sizable portion of the Turkish electorate do not feel disturbed by the populist rule
that deliberately weakens institutions of horizontal accountability. This dynamic has
likely played a role in the recent democratic backsliding of the country and its descent
into a competitive authoritarian regime.

Acknowledgments
I thank Melani Cammett and Pauline Jones for their guidance and feedback, and participants
in the Understanding Political and Social Change in Muslim States and Societies Conference
at the University of Michigan in May 2019 for comments and suggestions. I am grateful to Ali
Çarkoğlu for making available the 2002, 2007, and 2011 Turkish Election Study survey data.
Dilale Dönmez and Hevi Gökdemir provided valuable research assistance.

Notes
1. See Eligür (2010) and Sunar and Toprak (1983) for an overview of these reforms.
2. Turkish voters’ self-​placement on the ideological left-​right scale has little economic basis
and is strongly influenced by religiosity and attitudes like preference for the maintenance of
the status quo (Çarkoğlu 2007).
3. The sampling procedure for the surveys starts with the use of Turkish Statistical Institute’s
(TUIK) NUTS-​2 regions. The target samples were distributed according to each region’s share
of urban and rural population in accordance with records of the Address Based Population
Registration System (ADNKS). Next, TUIK randomly selected blocks of four hundred
addresses from each of the NUTS-​2 regions according to probability-​proportionate-​to-​size
(PPS) principle, and twenty addresses were again randomly selected in each block. Selection
of individuals in households is done on the basis of reported target population of eighteen
years or older in each household according to a lottery method. If for any reason that indi-
vidual could not respond to our questions in our first visit, then the same household is visited
up to three times until a successful interview is conducted and no substitution was applied.
4. See Aytaç and Elçi (2019) for details of the survey and associated research.
352   S. Erdem Aytaç

References
Akan, Murat. 2012. “Twin Tolerations or Siamese Twins? Kemalist Laicism and Political Islam
in Turkey.” In Problems Confronting Contemporary Democracies: Essays in Honor of Alfred
Stepan, edited by D. Chalmers and S. Mainwaring, 381–​423. Notre Dame, IN: University of
Notre Dame Press.
Akkoyunlu, Karabekir, and Kerem Öktem. 2016. “Existential Insecurity and the Making of a
Weak Authoritarian Regime in Turkey.” Southeast European and Black Sea Studies 16, no.
4: 505–​527.
Altınordu, Ateş. 2016. “The Political Incorporation of Anti-​system Religious Parties: The Case
of Turkish Political Islam (1994–​2011).” Qualitative Sociology 39, no. 2: 147–​171.
Aytaç, S. Erdem, and Ezgi Elçi. 2019. “Populism in Turkey.” In Populism around the World: A
Comparative Perspective, edited by D. Stockemer, 89–​108. Cham, Switzerland: Springer.
Aytaç, S. Erdem, and Ziya Öniş. 2014. “Varieties of Populism in a Changing Global Context: The
Divergent Paths of Erdoğan and Kirchnerismo.” Comparative Politics 47, no. 1: 41–​59.
Butler, Daren. 2018. “Special Report: With More Islamic Schooling, Erdogan Aims to
Reshape Turkey.” Reuters, January 25. https://​www.reuters.com/​article/​usturkey-​erdogan-​
education/​special-​report-​with-​more-​islamic-​schooling-​erdogan-​aims-​toreshape-​turkey-​
idUSKBN1FE1CD.
Çarkoğlu, Ali. 2004. “Religiosity, Support for Şeriat and Evaluations of Secularist Public
Policies in Turkey.” Middle Eastern Studies 40, no. 2: 111–​136.
Çarkoğlu, Ali. 2007. “The Nature of Left-​Right Ideological Self-​Placement in the Turkish
Context.” Turkish Studies 8, no. 2: 253–​271.
Çarkoğlu, Ali. 2009. “Women’s Choices of Head Cover in Turkey: An Empirical Assessment.”
Comparative Studies of South Asia, Africa, and the Middle East 29, no. 3: 450–​467.
Çarkoğlu, Ali. 2010. “Public Attitudes towards the Turban Ban in Turkey.” Utrecht Law Review
6: 145–​157.
Çarkoğlu, Ali. 2012. “Voting Behavior in Turkey.” In Handbook of Modern Turkey, edited by M.
Heper and S. Sayarı, 160–​170. New York: Routledge.
Çarkoğlu, Ali, and Ersin Kalaycıoğlu. 2009. The Rising Tide of Conservatism in Turkey.
New York: Palgrave.
Çiftçi, Sabri. 2010. “Modernization, Islam, or Social Capital: What Explains Attitudes toward
Democracy in the Muslim World?” Comparative Political Studies 43, no. 11: 1442–​1470.
Çiftçi, Sabri. 2012. “Secular-​Islamist Cleavage, Values, and Support for Democracy and Shari’a
in the Arab World.” Political Research Quarterly 66, no. 4: 781–​793.
Cizre-​Sakallıoğlu, Ümit. 1997. “The Anatomy of the Turkish Military’s Political Autonomy.”
Comparative Politics 29, no. 2: 151–​166.
Eligür, Banu. 2010. The Mobilization of Political Islam in Turkey. New York: Cambridge
University Press.
Gall, Carlotta. 2018. “Erdogan’s Plan to Raise a ‘Pious Generation’ Divides Parents in Turkey.”
New York Times, June 18.
Göle, Nilüfer. 1997. “Secularism and Islamism in Turkey: The Making of Elites and Counter-​
Elites.” Middle East Journal 51, no. 1: 46–​58.
Hale, William, and Ergun Özbudun. 2010. Islamism, Democracy and Liberalism in Turkey: The
Case of the AKP. New York: Routledge.
Religiosity and Political Attitudes in Turkey during the AKP Era    353

Inglehart, Ronald. 2000. “Culture and Democracy.” In Culture Matters: How Values
Shape Human Progress, edited by L. E. Harrison and S. P. Huntington, 80–​ 97.
New York: Basic Books.
Inglehart, R., C. Haerpfer, A. Moreno, C. Welzel, K. Kizilova, J. Diez-​Medrano, M. Lagos, P.
Norris, E. Ponarin, and B. Puranen, eds. 2014. World Values Survey: Round Six—​Country-​
Pooled Datafile Version. Madrid: JD Systems Institute. www.worldvaluessurvey.org/​
WVSDocumentationWV6.jsp.
Jamal, Amaney A. 2006. “Reassessing Support for Islam and Democracy in the Arab
World: Evidence from Egypt and Jordan.” World Affairs 169, no. 2: 51–​63.
Kalaycıoğlu, Ersin. 1994. “Elections and Party Preferences in Turkey: Changes and Continuities
in the 1990s.” Comparative Political Studies 27, no. 3: 402–​424.
Kingsley, Patrick. 2017. “Turkey Drops Evolution from Curriculum, Angering Secularists.”
New York Times, June 23.
Lüküslü, Demet. 2016. “Creating a Pious Generation: Youth and Education Policies of the AKP
in Turkey.” Southeast European and Black Sea Studies 16, no. 4: 637–​649.
Mainwaring, Scott. 1999. “Democratic Survivability in Latin America.” The Helen Kellogg
Institute for International Studies Working Paper no. 267. Notre Dame, IN: Helen Kellogg
Institute for International Studies. https://​kellogg.nd.edu/​documents/​1562.
Mardin, Şerif. 1971. “Ideology and Religion in the Turkish Revolution.” International Journal of
Middle East Studies 2, no. 3: 197–​211.
Mardin, Şerif. 1973. “Center-​Periphery Relations: A Key to Turkish Politics?” Daedalus 102, no.
1: 169–​190.
Mudde, Cas, and Cristobal Rovira Kaltwasser. 2017. Populism: A Very Short Introduction.
New York: Oxford University Press.
O’Donnell, Guillermo. 1994. “Delegative Democracy.” Journal of Democracy 5, no. 1: 55–​69.
Öniş, Ziya. 1997. “The Political Economy of Islamic Resurgence in Turkey: The Rise of the
Welfare Party in Perspective.” Third World Quarterly 18, no. 4: 743–​766.
Öniş, Ziya. 2015. “Monopolising the Centre: The AKP and the Uncertain Path of Turkish
Democracy.” International Spectator 50, no. 2: 22–​41.
Özbudun, Ergun. 2014. “AKP at the Crossroads: Erdoğan’s Majoritarian Drift.” South European
Society and Politics 19, no. 2: 155–​167.
Rose, Richard, William Mishler, and Christian Haerpfer. 1998. Democracy and Its Alternatives:
Understanding Post-​Communist Societies. Baltimore: The Johns Hopkins University Press.
Sunar, İlkay, and Binnaz Toprak. 1983. “Islam in Politics: The Case of Turkey.” Government and
Opposition 18, no. 4: 421–​441.
Tessler, Mark. 2002. “Islam and Democracy in the Middle East: The Impact of Religious
Orientations on Attitudes toward Democracy in Four Arab Countries.” Comparative Politics
34, no. 3: 337–​354.
Wald, Kenneth D., and Corwin E. Smidt. 1993. “Measurement Strategies in the Study of
Religion and Politics” In Rediscovering the Religious Factor in American Politics, edited by D.
C. Leege and L. A. Kellstedt, 26–​49. Armonk, NY: M. E. Sharpe.
Yuchtman-​Ya’ar, Ephraim, and Yasmin Alkalay. 2010. “Political Attitudes in the Muslim
World.” Journal of Democracy 21, no. 3: 122–​134.
Zaller, John R. 1992. The Nature and Origins of Mass Opinion. New York: Cambridge
University Press.
 Chapter 17

Rel igious Prac t i c e a nd

P olitical At t i t u de s
among Shiʿite s i n I ra n
and Iraq

Fotini Christia, Elizabeth Dekeyser, and

Dean Knox

Across Muslim societies, the past decades have seen a resurgence in support for re-
ligion in politics (Grzymala-​Busse 2012). Much work has shown that this support is
often driven by the social or welfare services that religious political movements provide
(Lia 1998; Amuzegar 2007; Mujani and Liddle 2009; Pepinsky et. al 2012; Cammett and
Jones 2014, among others). However, much less is known about whether and how re-
ligious practice itself—​either communal or individual—​shapes support for religion in
politics.
We examine this question among religious Shiʿa in both Iran and Iraq, two coun-
tries that differ significantly in the place of religion within their political system. In
both contexts, the religious individuals surveyed represent important social groups
that provide a base of support in Iran and hold significant sway within Iraq (K. Harris
2017; Mikulaschek, Pant, and Tesfaye 2020). The data for this study is based on a geo-
graphically representative original survey of Shiʿa pilgrims traveling to Karbala, Iraq,
during Arbaʿeen in 2015. This event, the world’s largest annual pilgrimage, draws a dem-
ographically diverse group of religious Shiʿa from across the Muslim world, providing a
unique opportunity to survey an important but understudied group. While Shiʿa make
up roughly 20 percent of the world’s Muslims,1 surveys of the Middle East and North
Africa have been unable to properly cover Iran due to regime restrictions and have had
more limited access in Iraq because of the post-​2003 conflict. The survey covered a wide
range of topics; more details are provided in a separate descriptive report of the survey
(Christia, Dekeyser, and Knox 2019).
356    Fotini Christia, Elizabeth Dekeyser, and Dean Knox

The impact of religion is often considered in terms of three constituent parts—​

identity, belief, and practice. In this work, we focus on how the latter dimension relates
to attitudes about religion in politics in the Iranian and Iraqi contexts, given that
Muslim religious identity remains largely fixed in the two populations, and the inten-
sity of belief is difficult to both quantify and measure. Religious practice, however, can
be further subdivided into two dimensions, individual and communal practice. These
two dimensions would be expected to have different relationships to attitudes about re-
ligion in politics. Individual practice might influence attitudes about religion in politics
if there exists religious legitimacy, or the existence of doctrinal or ideological support for
the concept. Communal practice, on the other hand, could also shape beliefs through
religious socialization, or the non-​ideological transmission of norms through the organ-
izational or social aspects of religious practice (Christia, Dekeyser, and Knox 2019). The
concept of religious socialization builds upon a rich literature, both from the American
(F. Harris 1994; M. Harris 1998; Putnam 2000; Wald, Silverman, and Fridy 2005) and
Middle Eastern (Jamal 2005; Loveland et al. 2005) contexts about the link between re-
ligious practice and political outcomes. Work on religious legitimation, however, has
remained much more limited. These two mechanisms can occur simultaneously and
potentially work in opposite directions; similarly, religious mechanisms operate si-
multaneously with other nonreligious sources of support for religious involvement in
politics.
In addition to (1) an abstract belief about the proper role of religious principles and
leaders in governance, we also examine a number of additional outcome variables that
capture the very different implications of this belief in the Iranian and the Iraqi contexts.
In Iran, the regime claims to be a theocracy—​the ultimate embodiment of religion in
politics—​meaning that support for religious involvement in politics relates directly to
(2) perceptions of regime legitimacy and (3) lower support for democracy, often viewed
as contrasting with the current regime. Our results have important implications for un-
derstanding regime support and regime longevity in Iran. Even work that addresses
the non-​ideological sources of support for the regime highlights how “Shi’a polit-
ical Islam . . . at least creates tacit acceptance of the state as legitimately acting in the
population’s interests” (K. Harris 2017, 26). This work also posits that this link is par-
ticularly strong among certain social groups, including the religious Iranians that we
survey, who have been central to the regime’s longevity. Yet whether there is in fact a link
between religiosity and state support has remained unclear; here, we show that not only
do individuals who engage in individual religious practice support the regime more,
but that communal practice reinforces beliefs about religion in politics that further
reinforces this tendency.
In Iraq, however, the link between the state and religion remains much more ten-
uous. Successive governments have attempted to use religious rhetoric to legitimate
themselves, with limited success. Crucially, however, Iraq claims only to be an “Islamic,
democratic, federal parliamentary republic,” not a theocracy. Therefore, in addition to
abstract beliefs about religion in politics, we examine (3) respondents’ support for de-
mocracy and (4) sectarian political outlook. Understanding how religious individuals
 Religious Practice and Political Attitudes among Shiʿites   357

perceive Iraq’s current government, as well as the broader democratic system, provides
insight into the motivations for shifting bases of support in Iraq.
To measure these concepts, we use an innovative Bayesian principal component
framework to aggregate batteries of questions relating to each topic. The resulting meas-
ures are used in Bayesian linear regressions in a way that captures the uncertainty from
dimension reduction, missingness, and the regression itself. Overall, we find that both
individual and communal religious practice are associated with greater support for
religion in politics, but these aspects of religion play markedly different roles in each
country. In Iraq, communal religious practice is strongly linked to abstract beliefs about
religion’s role in politics, and it shapes opposition to democracy as a system of govern-
ment both directly and through the channel of these abstract beliefs. In Iran, communal
practice is similarly linked to views on religion in politics, though to a lesser extent, and
its influence on regime support is most evident through this channel. Taken together,
these results on the influence of communal religious practice provide support for the
socializing and organizing power of religion to garner support for the state, concurring
with other work on the relationship between religious practice and political outcomes.
Yet the role of religious practice is more complex than mere socialization. In Iran, we
find that individual religious practice matters much more, both for abstract beliefs and
perceptions of the regime—​evidence, we argue, that Iran has succeeded in religiously
legitimating itself among this sector of society and in creating an ideological link be-
tween religious devotion and the regime in this national context. In contrast, the re-
lationship between individual practice and views on religion in politics is three times
weaker in Iraq, and we find no overall connection between individual practice and ei-
ther support for democracy or sectarian political outlooks. These differences highlight
how religious legitimation has encouraged regime longevity in Iran, while its absence
allows for the continued instability of the Iraqi state.
This work presents four important findings. First, by exploring the beliefs of Shiʿites,
rather than Sunnis, it helps move research on religion and politics in the Middle East
beyond majority narratives, dominated largely by Sunni populations, and explore how
these relationships vary within Shiʿite populations. The findings, however, do provide
insight into the potential broader variation across the Middle East in the relationship
between religious practice and political beliefs. Second, different types of religious prac-
tice not only have different effects on political attitudes depending on context, but they
can also reinforce one another. Third, these effects are highly context-​dependent, partic-
ularly on the relationship between the regime and Islam. Fourth, the fact that communal
practice has an effect on political attitudes among Iraqi Shiʿites but individual practice
does not highlights how the various dimensions of religiosity can influence attitudes in
different and surprising ways.
Below, we briefly discuss our data collection and the measurement of individual and
communal practice, as well as outcomes. We then turn to detailed analyses, examining
each outcome in turn and, for concrete questions like regime legitimacy, estimate the
extent to which religion works indirectly through the channel of abstract political views
as opposed to directly shaping beliefs.
358    Fotini Christia, Elizabeth Dekeyser, and Dean Knox

The Survey

The primary data source used in this paper is based on an original survey conducted
during the Arba’een pilgrimage to Karbala in Iraq. This event, the world’s largest an-
nual pilgrimage, attracted a reported 22 million Shiʿa pilgrims during that year, with
Iranians and Iraqis representing the largest groups. Arba’een is a collective mourning
ritual involving a pilgrimage on foot to the shrine of Imam Hussein in Karbala. Walking
is what distinguishes this particular visit, with pilgrims setting out days or weeks ahead
of time, depending on their distance from Karbala, in order to walk to Imam Hussein’s
shrine for the celebrations. The walk involves groups walking from places as far as
Basra (roughly 500 km away) and Baghdad (roughly 120 km away). In the more heavily
frequented portions of the route, roadsides are dotted with clusters of service tents,
known as mawakib, which offer food and shelter to pilgrims. Travelers stop for long
periods of time, not only to dine or drink water and tea, but also to recharge cell phones,
repair shoes, or even get weary feet and legs massaged by local volunteers. Mawakib tend
to have a specific regional identity, hosting people from different governorates and prov-
inces in Iraq and Iran, with separate tents for men and women. Because of their regional
nature, these tents allowed us to target survey respondents by their region of origin.
We surveyed respondents from November 17 to December 12, 2015, a twenty-​five-​
day window broadly associated with the 2015 Arba’een celebrations. Our goal was not to
survey a representative sample of pilgrims, but rather to develop a geographically rep-
resentative sample that would allow us to understand how pilgrims from across Iraq
and Iran perceive a wide variety of issues, and how this differed across the two country
contexts. To define geographic targets for Shiʿa in Iran, we utilized the 2011 population
census, which provides an accurate proxy given that the country is approximately 98 per-
cent Shiʿa. Iraq, however, provides a greater challenge. Not only was its last census in
1997, after which there was significant population displacement, but information on sect
has also remained largely restricted. In the absence of this data, we attempted to sample
pilgrims according to the distribution of Shiʿa politicians in the Iraqi parliament, known
as the Council of Representatives (COR).2 We identified the sect of the parliamentarians
holding these seats, and used the Shiʿa seats as a metric for the Shiʿa proportion of the
population in that area. The nature of the event also allowed us to survey roughly equal
numbers of male and female respondents, a rare opportunity in a region where women
traditionally are excluded from the public sphere. For both genders, respondents were
chosen among pilgrims between eighteen and sixty years old, with roughly half below
the age of thirty-​five. In total, we surveyed 2,410 Iraqis and 1,668 Iranians.
This geographically representative sample of religious individuals from the two coun-
tries provides a central insight into a vital but vastly understudied population. In Iran,
recent work has highlighted the importance of this particular social group to the en-
durance of the revolutionary system (K. Harris 2017). In Iraq, religious Shiʿites have
 Religious Practice and Political Attitudes among Shiʿites   359

provided an important constituency in the ebbs and flows of political power in the post-​
invasion period (Mikulaschek, Pant, and Tesfaye 2020).

The Sample

The Iranians and Iraqis surveyed represented a demographically, socioeconomically,

and ideologically diverse sample. Comparison to the full national population is chal-
lenging in both contexts, particularly for Iraq, given the limited availability of represen-
tative data. Triangulating with alternate data sources, however, allows us to situate our
sample within the range of other surveys conducted in these contexts.
The Iranians surveyed followed the census age distribution closely; were slightly
wealthier and more likely to own a car; and were more likely to have a high school ed-
ucation, but less likely to have a college education. Despite these averages, however, the
survey includes a wide range of individuals from across the educational and economic
spectrum that mirrors the diversity in Iran more broadly.
While data limitations in Iraq do not allow a comparison to the census, we compared
the population to two large-​scale and geographically diverse surveys in the country, the
Arab Barometer and the World Bank’s Living Standards surveys. A weakness of both,
however, is that they do not distinguish Sunni respondents from the Shiʿa group that
we examine here. Our population is poorer and less educated than the Arab Barometer
sample, and wealthier and more educated than the World Bank sample, indicating
that our sample lies within the general range of well-​designed survey pools. More
comparisons and general descriptives can be found in the survey report (Christia,
Dekeyser, and Knox 2016), To Karbala: Surveying Religious Shi‘a from Iran and Iraq.

Analysis

We use a range of detailed survey questions to measure both religious practice and po-
litical outcomes.

Measuring Practice
We measure individual and communal practice using a cluster of questions about religi-
osity from the survey. To measure communal practice, we examine the frequencies with
which respondents reported attending Friday prayer, religious lessons, and mosque.3 To
measure individual practice, we use private religious behaviors. Specifically, we examine
360    Fotini Christia, Elizabeth Dekeyser, and Dean Knox

whether individuals pray daily, watch or listen to religious programs, and read the
Qurʾan or Duʾa.4 These questions can be seen in detail in Tables 17.1 and 17.2.
We then aggregated each cluster of questions using Bayesian principal component
analysis (BPCA), which measures latent constructs by aggregating groups of related
survey questions into overall indices with data-​driven weights. (Unlike traditional prin-
cipal components analysis, BPCA simultaneously addresses missingness on the com-
ponent questions, allowing us to derive more robust results than standard principal
component approaches that drop individuals who do not respond on any component
question.) We find that our measure of communal practice weights each of its constit-
uent religious practices virtually identically. In contrast, the first principal component of
individual practice assigns indistinguishable weights to religious programs and reading
of Qurʾan/​Duʾa, but downweights prayer by roughly 25 percent.5

Measuring Outcomes
We identified one abstract outcome of interest—​views about the role of religion in
politics—​and three concrete political attitudes—​perceptions of democracy, regime
legitimacy (in Iran only), and sectarian politics (Iraq only). Due to differing political
systems, interpretations of these outcomes vary across national contexts, even though
questions are worded identically. In Iran, the perceived legitimacy of a regime that
claims to be theocratic is closely related to an individual’s beliefs about religion in poli-
tics, while support for democracy in Iraq more loosely tied—​while many Iraqi Shiʿites
are dissatisfied with their government because they would prefer a more directly reli-
gious governing system, other sources of discontent also exist. We combined groups of
questions relating to each of these latent concepts using BPCA, focusing on the first prin-
cipal component. (Details about constituent questions are reported in the Appendix.)

Analysis and Findings

We begin by noting basic demographic differences that should be kept in mind when
interpreting results, though we include controls for these characteristics in all analyses.
Communal practicing Iraqis are more likely to be poor, unemployed, over fifty, and
male than nonreligious Iraqis. Those who have higher levels of individual practice are
more likely to be educated, less likely to be wealthy, more likely to be over thirty, and less
likely to be male than their less pious compatriots. Communal practicing Iranians are
more likely than nonreligious Iranians to have a primary or middle school education, be
over fifty, and less likely to be male. Iranians who engage in more individual practice are
less likely to be employed and more likely to be over thirty.
Table 17.1: Variables in Individual Practice Principal Component
Question Responses Iraq Proportion Iran Proportion

Do you pray Yes 0.02 0.04

every day? No 0.98 0.96
How often do you . . . At least once a day 0.45 0.47
Watch or listen to At least once a week 0.34 0.29
religious A few times a month 0.06 0.05
programs/​sermons At most once a month 0.1 0.1
Never 0.05 0.08
How often do you . . . At least once a day 0.64 0.63
Read or listen to At least once a week 0.23 0.21
the Qurʾan or Duʾa A few times a month 0.03 0.04
At most once a month 0.07 0.07
Never 0.04 0.04

Table 17.2: Variables in Communal Practice Principal Component

Question Responses Iraq Proportion Iran Proportion

How often do you . . . At least once a day 0.06 0.25

Attend Friday At least once a week 0.18 0.16
prayers/​women’s prayer A few times a month 0.13 0.11
At most once a month 0.12 0.28
Never 0.5 0.2
How often do you . . . At least once a day 0.09 0.47
Attend religious At least once a week 0.12 0.29
lessons in the A few times a month 0.12 0.05
Mosque At most once a month 0.09 0.1
Never 0.57 0.08
How often do you . . . At least once a day 0.17 0.63
Visit a mosque At least once a week 0.25 0.21
A few times a month 0.17 0.04
At most once a month 0.15 0.07
Never 0.26 0.04
362    Fotini Christia, Elizabeth Dekeyser, and Dean Knox

For each outcome of interest, we then run a linear regression, controlling for years of
education,6 income sufficiency,7 paid work,8 age,9 gender, news interest,10 conservatism
as proxied by gender attitudes,11 civil society involvement,12, and province fixed effects.
In addition to these controls, we also include our variables of interest, individual and
communal practice; uncertainty in the measurement of these BPCA-​based explanatory
variables and outcomes are incorporated by repeatedly sampling, then combined with
the uncertainty of the regression.
Results can be seen in Figure 17.1.
In Iraq, we find that communal practicing individuals are dramatically more likely
to support a strong role for religion in politics (p < 0.001). In practical terms, these
individuals are more opposed to democracy as a system of government (p < 0.001),
and they are more likely to hold sectarian political outlooks (p < 0.038)—​for ex-
ample, believing that out-​group politicians cannot represent their interests. In supple-
mental mediation analyses, we find that a significant portion of communal practice’s
overall contribution to these practical political views flows through the channel of
abstract religion-​in-​politics attitudes (p < 0.001 for both outcomes), but we also
find evidence of a strong direct relationship between communal practice and anti-​
democracy positions (p < 0.008). These results are consistent with the transmission
of nonreligious norms through religious practice. While more individual practicing
individuals are also more supportive of religious involvement in politics, this at-
titude is more weakly linked than communal practice, and we find null—​perhaps
even negative—​associations between individual practice and our practical political
outcomes. Unsurprisingly, we find that neither form of religious practice is signifi-
cantly associated with the legitimacy of the nonreligious government.
In Iran, we find that the relationship between communal practice and an
individual’s support for religion in politics is much weaker. In the Iranian context, in-
dividual practice is far more important. By the same token, more individual practicing
individuals are much more likely to view the regime as legitimate, and also to oppose
democratic systems of governance; a significant portion of individual practice’s con-
tribution to these practical attitudes appears to be mediated through abstract views
about the proper role of religious principles and leaders. Given Iran’s self-​positioning
as a theocracy, these correlations highlight how Iran has succeeded in gaining religious
legitimacy among the highly devout. Communal practice has no discernible connec-
tion to regime support in our sample, though other work on religious legitimation has
highlighted its importance as a conduit for information sharing, organization, and
social pressure that can lead to greater support for a political regime (Ludden 2005,
Wilkenson 2005, Mitchell 2017). (Indeed, we find participation in these group reli-
gious activities still seems to transmit other views, like anti-​democratic beliefs, which
are less directly tied to the regime.) While we do not dispute the importance of these
channels, our findings highlight how religious ideology is a perhaps even more potent
source of support both for religion in politics and for its manifestation in the current
regime.
 Religion in Politics Regime Legitimacy Anti-Democracy Sectarian Politics

Individual Practice

Communal Practice

–0.1 0.0 0.1 0.2 0.3 0.4 0.1 0.0 0.1 0.2 0.3 0.4 0.1 0.0 0.1 0.2 0.3 0.4 0.1 0.0 0.1 0.2 0.3 0.4
Expected Change in Outcome (SD) per SD Change in Predictor

Country Iran Iraq

Figure 17.1: The relationship between religious practice and political outcomes.
Note: Each panel reports the relationship between religious practice and a political attitude. Points (error bars) correspond to posterior means
(95 percent intervals) of Bayesian regressions incorporating uncertainty from BPCA measurement. All models include adjustment for years of edu-
cation, income sufficiency, paid work, age, gender, news interest, conservatism as proxied by gender attitudes, civil society involvement, and province
fixed effects.
364    Fotini Christia, Elizabeth Dekeyser, and Dean Knox

Alternate Explanations

We identify four alternate explanations for correlations between religiosity and political
views: sociability, conservatism, economic interests, and intergroup contact. We do not
argue that these potential confounders and sources of heterogeneity are unimportant;
rather, our results suggest that, even holding these other factors constant, individual and
communal practice can shape political attitudes across both country contexts.
One key concern is that the apparent importance of communal practice is an arti-
fact of confounding—​that “sociable” individuals are both more likely to be involved in
communal activities in general, and also that this type of person has particular political
attitudes. We alleviate this concern by controlling for participation in a variety of other
types of civil society organizations in all reported models; this does not change the rela-
tionship between communal practice and any of our outcomes of interest.
A second concern is that conservativism drives both individual practice and support
for religion in politics. To test for this, we include a control for gender conservativeness—​
based on questions about gender roles and gender stereotypes—​a proxy for overall so-
cially conservative beliefs. Again, we include this variable in all reported specifications,
and results are unchanged by its inclusion.
Third, it is possible that economic hardship is correlated with individual practice,
thus confounding our results. The ability of the Iranian regime (and Islamist parties
more generally) to provide economic benefits is a well-​documented source of support
for both (Mujani and Liddle 2009; Cammett and Jones Luong 2014, among others). Yet
holding this constant, does ideology still play a role? We control for a range of economic
variables in our results, roughly capturing demand for welfare. Yet again, our results
hold, minimizing concerns about this possible link.
Finally, it is possible that the Iran-​Iraq difference is due to the sectarian heterogeneity
in Iraq, which leads to different expectations about what religious involvement in poli-
tics might look like and whose interests it might serve. In particular, Iraqi respondents
might be concerned that democracy advances Sunni interests in Iraq, rather than sec-
ular interests, and this crucial difference might account for the difference in patterns
from Iran. Yet even Shiʿa-​dominated provinces, which mirror the sectarian homoge-
neity of Iran, show similar correlations to the rest of Iraq, indicating that sectarian ho-
mogeneity alone is not the primary determinant of this relationship.
Taken together, these findings provide additional support to the potential for ide-
ology to influence religious legitimacy, and thus the relationship between religiosity and
beliefs about religion in politics.
 Religious Practice and Political Attitudes among Shiʿites   365

Discussion and Conclusion

This research provides a uniquely rich glimpse into the relationship between religiosity
and religion in politics among devout Shiʿites in Iran and Iraq. Using an original survey
of Shiʿite pilgrims during the Arba’een pilgrimage, we find that communal practice, in
both contexts, plays an important role in reinforcing desire for religious principles and
religious leaders in governance. Moreover, this desire manifests in specific and conse-
quential political attitudes, particularly opposition to or support for the government.
We argue that this process occurs largely through religious socialization, or the trans-
mission of potentially nonreligious norms through religious practice. However, in-
dividual religious practice also plays a role in shaping beliefs about politics. Here, we
find that it operates more strongly in Iran, where we argue that the regime’s success in
religious legitimation has led religious individuals to link regime support to religious
devotion.
These findings concur with work that argues communal practice can play an im-
portant role in influencing political outcomes as a result of its organizing potential,
information-​spreading ability, and capacity to socially sanction its congregants (Ludden
2005; Wilkenson 2005; Mitchell 2017). This is what we find in the dynamic Iraqi
environment—​more communal religious practice is associated with greater support for
Islamic systems of governance, but not with support for the (insufficiently religious) re-
gime. Similarly, in Iran, more communal practice is associated with greater support.
Yet in Iran, however, we also see the result of nearly forty years of theocracy, where
even individual practice is associated with greater support for Islamic systems of gov-
ernance, and specifically the Iranian regime, highlighting how Iran has successfully
legitimated its theocracy among devout individuals. We thus find that communal and in-
dividual practice function differently in Iran and Iraq, driven by how their governments
relate to religion. This provides unique insight into the relationship between individual
religiosity and the state in authoritarian and unstable democratic contexts.
We examine factors that could potentially confound the relationship between indi-
vidual and communal practice and support for religion in politics. Our results hold even
outside of self-​selection into communal practice, conservatism, economic interests, and
sectarian heterogeneity. While we do not argue that these cannot be drivers of support
for religion in politics, both hypothetical and applied, we demonstrate that even outside
their influence, piety and practice can play an important role.
More work must be done to better understand the avenues through which religious
legitimation occurs, as in the Iranian context. This religious legitimation could occur
366    Fotini Christia, Elizabeth Dekeyser, and Dean Knox

through a variety of channels, whether religious leaders, political elites, historical

factors, or constitutional factors. For instance, religious leaders could speak about the
regime, thus validating it in the eyes of their members. Alternately, political and gov-
ernment elites could use religious rhetoric, which serves to religiously validate their po-
sition outside the influence of religious leaders. History could also play a role, with the
revolution providing a symbolic background for the integration of religious and polit-
ical life. Finally, constitutional factors—​the fact that Iran is a self-​identified theocracy—​
could also play a role in convincing its citizens of its religious legitimacy.
Our work has certain limitations. First, the nature of the data gathered makes causal
analysis difficult. While we cannot randomly assign individual or communal practice,
one can imagine a design that looks at how the ease of access to communal places of
worship, for instance, influences religious outcomes, to disaggregate whether the rela-
tionship between religious practice and political beliefs is being driven by individuals
who have the desire to practice more in community, or if it is the practice itself. Second,
and relatedly, our work cannot distinguish the direction of the influence. In all likeli-
hood, religious practice influences political beliefs, yet political beliefs also influence
religious practice. Additional work should disaggregate the relative strength of these
different causal directions.
Taken together, these findings provide vital insight into how religion can strengthen
or weaken the state, either through religious legitimation or socialization, with impor-
tant implications for the relationship between religion and the state in Muslim societies.

Acknowledgments
For support in the field we are truly grateful to Abdullah Hammadi, Kufa University president
Dr. Aqeel Abd Yassin Al-​Kufi, and Kufa University professor Hassan Nadhem. Great thanks
also go to our survey supervisors Faris Kamil Hasan, Maytham Hasan Machi, Wael Adnan
Kadhim, Faris Najem Harram, and Nidhal J. Gdhadab, and to our incredible enumerator team.
For advice on surveying in Iraq, we would like to thank Professor Amaney Jamal and Michael
Robbins, who generously shared their experience from the Arab Barometer; Nandini Krishnan,
for sharing the instruments and data from the World Bank household economic survey for
Iraq; and Neha Sahgal, for discussing her experience with the work of the Pew Research Center
in Iraq. Special thanks to Professor Roy Mottahedeh, Dr. Sabrina Mervin, and Geraldine
Chatelard for early input on the project, and to Marsin Alshamary and Ramisa Shaikh for their
research assistance. Fotini Christia carried out the data collection associated with this project
while on an Andrew Carnegie fellowship. She also acknowledges support from ARO MURI
award No. W911NF-​121-​0509. Dean Knox acknowledges financial support from the National
Science Foundation (Graduate Research Fellowship under Grant No. 1122374).

Notes
1. For more, see “Mapping the Global Muslim Population: A Report on the Size and
Distribution of the World’s Muslim Population,” Pew Research Center, October 2009.
 Religious Practice and Political Attitudes among Shiʿites   367

2. The COR has a total of 328 seats across Iraq’s eighteen governorates. The allocation of seats
is based on each governorate’s estimated population size.
3. For religious lessons and mosque, respondents were given the choice of “at least once a
day,” “at least once a week,” “a few times a month,” “at most once a month,” and “never.” For
attending Friday prayer, respondents were divided into those who responded “at least once
a week,” “a few times a month,” “at most once a month,” and “never.”
4. Respondents were given the choice of “Yes” and “No” to the daily prayer, and the choice of
“at least once a day,” “at least once a week,” “a few times a month,” “at most once a month,”
and “never” for religious programs and reading the Qurʾan or Duʾa.
5. The difference in weighting reflects the fact that self-​reported prayer frequency may be
driven by other unobserved factors (such as social desirability) that are captured by the ad-
ditional, discarded principal components.
6. Education is treated as a categorical variable with five levels: none, primary, middle school,
high school, and college and above.
7. Income sufficiency is a categorical variable indicating whether or not individuals felt that
their income was sufficient to meet their daily needs. Using this, rather than an abso-
lute measure of income, allows us to account for differences in purchasing power parity
across the two countries. Possible responses were as follows: “Our household income
did not cover our expenses and we faced significant difficulties meeting our needs,” “Our
household income did not cover our expenses and we faced some difficulties meeting our
needs,” “Our household income covered our expenses without notable difficulties,” and
“Our household income covered our expenses well and we were able to save.”
8. “Paid work” refers to whether a respondent is currently employed. This is an important
control variable, as individuals who do not have a job are more likely to participate in col-
lective religious practice during the work week, and some of the variation we see between
men and women is in fact due to differential employment statuses.
9. We discretize self-​reported age into ranges 18–​30, 30–​50, and 50+.
10. We define an additive index of self-​reported news consumption frequencies from televi-
sion, radio, newspapers or magazines, the Internet, or friends and family.
11. The first principal component extracted from BPCA measure on questions about gender
stereotypes and women’s rights.
12. A binary variable indicating participation in humanitarian or professional organizations,
unions, sports associations, tribal groups, cooperatives, or any other civil society groups.

References
Amuzegar, Jahangir. 2007. “Islamic Social Justice, Iranian Style.” Middle East Policy 14, no.
3: 60–​78.
Cammett, Melani, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17: 187–​206.
Christia, Fotini, Elizabeth Dekeyser, and Dean Knox. 2016. “To Karbala: Surveying Religious
Shi’a from Iran and Iraq.” MIT Political Science Department Research Paper 2016-​39.
Cambridge, MA: MIT.
Christia, Fotini, Elizabeth Dekeyser, and Dean Knox. 2019. “The Nature and Origins of
Sectarian Animosity.” MIT Political Science Department Research Paper 2019-​1. Cambridge,
MA: MIT.
368    Fotini Christia, Elizabeth Dekeyser, and Dean Knox

Grzymala-​Busse, Anna. 2012. “Why Comparative Politics Should Take Religion (More)
Seriously.” Annual Review of Political Science 15: 421–​442.
Harris, Fredrick C. 1994. “Something Within: Religion as a Mobilizer of African-​American
Political Activism.” Journal of Politics 56, no. 1: 42–​68.
Harris, Kevan. 2017. A Social Revolution: Politics and the Welfare State in Iran.
Oakland: University of California Press.
Harris, Margaret. 1998. “A Special Case of Voluntary Associations? Towards a Theory of
Congregational Organization.” British Journal of Sociology 49, no. 4: 602–​618.
Jamal, Amaney. 2005. “The Political Participation and Engagement of Muslim
Americans: Mosque Involvement and Group Consciousness.” American Politics Research 33,
no. 4: 521–​544.
Lia, Brynjar, and Jamal al Banna. 1998. The Society of the Muslim Brothers in Egypt: The Rise of
an Islamic Mass Movement, 1928–​1942. Reading, UK: Ithaca Press.
Loveland, Matthew T., David Sikkink, Daniel J. Myers, and Benjamin Radcliff. 2005. “Private
Prayer and Civic Involvement.” Journal for the Scientific Study of Religion 44, no. 1: 1–​14.
Ludden, David E. 2005. Making India Hindu: Religion, Community, and the Politics of
Democracy in India. New Delhi: Oxford University Press.
Mikulaschek, Christoph, Saurabh Pant, and Beza Tesfaye. 2020. “Winning Hearts and Minds
in Civil Wars: Governance, Leadership Change, and Support for Violent Groups in Iraq.”
American Journal of Political Science (Early View, June 3).
Mitchell, Stephanie Claire. 2017. “The Function of Religion in the Israeli-​Palestinian Conflict.”
MA diss., Portland State University.
Mujani, Saiful, and R. William Liddle. 2009. “Muslim Indonesia’s Secular Democracy.” Asian
Survey 49, no. 4: 575–​590.
Pepinsky, Thomas B., R. William Liddle, and Saiful Mujani. 2012. “Testing Islam’s Political
Advantage: Evidence from Indonesia.” American Journal of Political Science 56, no.
3: 584–​600.
Putnam, Robert D. 2000. Bowling Alone: The Collapse and Revival of American Community.
New York: Simon & Schuster.
Wald, Kenneth D., Adam L. Silverman, and Kevin S. Fridy. 2005. “Making Sense of Religion in
Political Life.” Annual Review of Political Science 8: 121–​143.
Wilkinson, Steven, ed. 2005. Religious Politics and Communal Violence. New Delhi: Oxford
University Press.
 Chapter 18

Repressive Re l i g i ou s
Re gul ation an d P ol i t i c a l
Mobiliz ati on i n
Central Asia
Why Muslims (Don’t) Rebel

Dustin Gamza and Pauline Jones

Much of the literature on political mobilization in the Islamic world has emphasized
religion as a primary factor in explaining why Muslim citizens take to the streets
(Cammett and Jones 2014). Similarly, much of the literature on religious regulation
implies that state regulation of religion indirectly reduces the degree of religiously
motivated political mobilization because it decreases religiosity (e.g., Finke and Stark
1988; Finke 1990; Iannaccone 1991). In this chapter, we bridge these two literatures to
explore the relationship between religious regulation and political mobilization in
Muslim-​majority states (MMS). We argue that, controlling for religiosity, state religious
repression influences whether Muslims rebel against the government, and that the ef-
fect is both direct and positive. More specifically, we argue that religious repression has
a conditional effect on political mobilization: the first condition concerns the degree of
enforcement; the second concerns the type of religious practice that is being repressed.
In sum, Muslims are more likely to support political action against the state when (1) the
degree of enforcement is high1 and (2) the repressive regulation being enforced targets
communal rather than individualistic religious practices.2 When enforcement is high,
Muslims expect state persecution of their religious community to increase. When the
repressive regulation being enforced targets communal religious practice, Muslims ex-
pect state persecution to extract a much greater toll. They are thus more likely to support
taking political action against the state in order to protect their community from this
perceived harm.
370    Dustin Gamza and Pauline Jones

We utilize two novel experiments embedded in a mass survey that we conducted

in Kyrgyzstan (a MMS in Central Asia) to test our hypotheses concerning the direct
and positive relationship between religious repression and political mobilization.
Our results are mixed. First, although we find strong support for our contention that
enforcement of repressive regulation has a significant effect on attitudes toward polit-
ical mobilization, the effect is negative—​that is, in the opposite direction that we ex-
pected. Second, we find support for our argument that it is not only whether repressive
regulations are enforced, but also the degree of enforcement, that matters. The results
of our experiments demonstrate that the threat of arrest and imprisonment has a sig-
nificant effect in a number of cases where fines alone do not have an effect. Yet again,
contrary to our expectations, the effect on attitudes toward political mobilization is
negative. Third, we find support for our contention that the effect of enforcement is
stronger—​and more consistent across different types of political action—​for communal
religious practice (here, Qurʾan study groups in private homes) than it is for individu-
alistic religious practice (here, wearing the hijab in public places). Finally, and perhaps
most importantly, we find that these results are not being driven by the primary alterna-
tive explanation—​the level of religiosity.
In sum, we find that, controlling for level of religiosity, enforcement has a signifi-
cant but negative effect on attitudes toward political mobilization, particularly when it
comes to restrictions on communal religious practice. These findings have important
implications for the literatures on religious regulation, political mobilization, and re-
pression. First, they contribute to our knowledge concerning the effects of religious
regulation on political behavior, which remains limited. Second, they suggest that en-
forcement of religious regulation—​particularly repressive regulation—​has a direct ef-
fect on political behavior. Third, our research provides individual-​level data regarding
the effects of repression that contributes to the divide in the repression literature con-
cerning whether repression has an effect, and, if so, whether it increases or decreases
political mobilization (Davenport and Inman 2012). Contrary to our expectations, it
suggests that, under certain conditions, repression can dampen support for mobiliza-
tion. The negative effect that repression has on attitudes toward political mobilization,
moreover, holds even among those who would be classified as the most devout and
committed to the principles of their faith.

Religious Regulation and

Political Mobilization

Religious market theory posits an indirect causal link between religious regulation and
political behavior via the level of religiosity in a society. Utilizing a political economy
framework, scholars argue that unrestricted religious markets will foster a high level of
religious belief and participation because they promote competition, which encourages
 Repressive Religious Regulation and Political Mobilization    371

religious leaders to innovate to meet the demands of extant and potential adherents
(e.g., Finke and Stark 1988; Finke 1990; Chaves and Cann 1992, Stark and Iannaccone
1994; Stark and Finke 2000). Conversely, restricted religious markets will depress reli-
gious belief and participation because they eliminate competition, which discourages
religious leaders from innovating in order to maintain or attract adherents. In sum, re-
ligious regulation disrupts religious markets, and thus lowers the level of religiosity. The
resulting shifts in religious belief and participation, in turn, have an effect on individual
political behavior (Wald and Martinez 2001; Nicolet and Tresch 2009).
This literature provides limited insight concerning the relationship between repres-
sive religious regulation and political mobilization, particularly in a predominantly
Muslim context, for several reasons. First, the tendency of this literature is to conceptu-
alize regulation as synonymous with state subsidies—​i.e., institutional, fiscal, or political
benefits to particular religious organizations—​and to ignore the role of state repression—​
i.e., restrictions on the selection or profession of certain religious beliefs and practices.
They have thus examined only one of the two dimensions that make up any state effort
to regulate religion (Grzymala-​Busse 2012; Fox 2013). Instead, we reconceptualize reli-
gious regulation as two-​dimensional, consisting of both subsidy, which can be further
disaggregated into general and selective, and repression, which can range from low to
high concerning the severity of restrictions, punishment, and enforcement (Gamza and
Jones 2016). Although our larger project examines the effects of both subsidy and re-
pression, here we focus only on repression—​specifically, the enforcement of regulations
that restrict religious expression.
Second, this literature implies that religiosity is the key factor that mediates the causal
relationship between religious regulation and individual political behavior. We argue
instead that state regulation of religion has a direct effect on individual political beha-
vior; irrespective of the level of religiosity, it influences both whether individuals en-
gage in political action and the form that action takes. This is not because we do not
take religion seriously, but rather because we approach religion as a collective good that
is a fundamental part of one’s daily social life in a predominantly Muslim community,
regardless of how devout one is according to conventional (and unconventional) meas-
ures that only consider individual beliefs and behavior. Thus, one’s willingness to re-
spond to religious repression by mobilizing against the state is at least partly based on
the broader social impact of this repression. In other words, as part of a broader reli-
gious community, an individual is more likely to take into consideration not only the
extent to which repression impedes one’s own freedom of religious expression, but also
the extent to which it impedes the ability of others to freely practice their faith.
Third, the hypothesized relationship between religion and political behavior often
focuses on political participation, such as voting or joining an organization, in a dem-
ocratic context in which there is little risk to such activities (e.g., Esmer and Pettersson
2007; McTague and Layman 2009). We focus instead on various forms of political mo-
bilization in the context of a hybrid regime in which the risk is greater but not fully
known.3 This enables us to investigate both whether citizens are more willing to engage
in some types of political action than others and whether their willingness to mobilize
372    Dustin Gamza and Pauline Jones

despite some risk depends on the degree of enforcement. It also enables us to disen-
tangle political mobilization in MMS from Islamist mobilization, and thus to avoid the
common tendency to conflate them (Cammett and Jones 2014). Moreover, by limiting
the type of repression to restrictions on religious practices, we can better isolate the
effects of religious repression on political behavior.
We argue that religious repression motivates Muslim citizens to rebel under two
conditions. First, they will be more likely to support political action against the state
when the degree of enforcement is high. In other words, what matters is not only that re-
pressive regulations are enforced, but also the severity of the penalty associated with en-
forcement, which ranges from low (e.g., fines) to high (e.g., arrest and imprisonment).
There are two main reasons for this. The first is that average citizens are often unaware
of formal legislation until they have some knowledge of or experience with this legisla-
tion being enforced. In many developing countries, moreover, laws often go unenforced
or are unevenly enforced due to weak state capacity or the central government’s limited
inability to monitor local governments. The second is that enforcement sends a stronger
signal about the state’s intentions than legislation alone. The enforcement of religious
regulations in particular indicates that threats against the freedom of religious expres-
sion are likely to increase. In sum, we expect that Muslims rebel in order to prevent or
mitigate the escalation of state repression—​that is, to protect themselves and, more crit-
ically, their community from a perceived increase in the likelihood of religious persecu-
tion. The adoption of a law that restricts some aspect of belief or practice in and of itself
may not signal such a likelihood, especially given that both the awareness and enforce-
ment of such legislation is likely to be low. When repressive laws are enforced, however,
and as the penalty for violating them becomes more severe, these repressive regulations
can indicate that the state is moving in a draconian direction when it comes to control-
ling religion.
Second, Muslims will be more likely to support political action against the state when
the repressive regulation being enforced targets communal religious practices rather
than individualistic religious practices. As Emile Durkheim recognized over a century
ago, the purpose of exercising religious beliefs and practices is not merely to express
individual piety, but also to “unite in one single community . . . all those who adhere
to them” (Thompson 1982, 129). In other words, religion plays a key social function in
creating and sustaining a moral community among adherents. Religious consump-
tion can play an important role in “attracting individuals to the community, socializing
them to the communal ethos, and drawing symbolic boundaries between community
members and outsiders” (Karatas and Sandikci 2013, 465). Religion can also serve to
bind members of communities together, through the provision of subsidies to their
members and the opportunities that community members sacrifice to maintain their
membership in the religious community (Berman 1999). Individuals within these
communities, therefore, place a higher value on practices that bring members of the
community together and serve to enhance the community (i.e., communal practices)
than on practices that display or are associated with individual piety (i.e., individualistic
practices). They may therefore also be more willing to take risks to prevent or oppose
 Repressive Religious Regulation and Political Mobilization    373

government restrictions on these communal practices in order to prevent or mitigate

policies they perceive to extract a heavier toll on their moral community. As others have
argued in a variety of contexts, communal religious practice plays a prominent role in
shaping political attitudes and behavior (e.g., Lim and Putnam 2010; Khan et al. 2016;
see also the chapter by Christia, Dekeyser, and Knox in this volume). Defending com-
munal practices and the community, moreover, makes mobilization more likely, be-
cause it lowers barriers to collective action via its ability to organize, spread information,
and socially sanction (e.g., Campbell 2004; Hoffman and Nugent 2017; Lussier 2019).
Thus, under each of these conditions, religious repression should have a positive effect
on attitudes toward political mobilization in MMS. Taken together, the effect should be
even stronger. When enforcement is high and the repressive regulation being enforced
targets communal religious practices, Muslims not only expect state persecution of their
religious community to increase, but also that its negative impact on their community
will be greater. They are even more likely, therefore, to support taking political action
against the state in order to protect their community from this perceived harm.

Empirical Strategy and Evidence

To explore the relationship between religious regulation and political mobilization in

predominantly Muslim countries, we conducted a nationally representative survey of
citizens in Kyrgyzstan (N =​2,631, including 2,400 Muslims). As part of the survey, we
randomly assigned roughly 1,800 Muslim respondents to a series of vignettes about re-
pressive regulations that varied the degree of enforcement. In each experimental arm,
after respondents read the vignette, they were asked whether they supported various
forms of political mobilization—​such as writing a letter or participating in a protest—​
as appropriate responses. (See Appendix 18.1 for full text.) Specifically, we designed a
unique survey experiment to test three main hypotheses concerning the direct and pos-
itive relationship that we posit between religious repression and political mobilization,
irrespective of the level of religiosity:

• H1: Support for political mobilization will be greater when repressive legislation is
enforced.
• H2: Support for political mobilization will be greater when the degree of enforcement,
defined as the severity of penalty, increases.
• H3: The effect of repressive regulation on support for political mobilization will be
more significant when the repressive regulation being enforced targets communal
(rather than individualistic) religious practices.

In the control group vignette, respondents are informed that the government has
adopted a law that prohibits the religious practice in question, but that the repressive law
is not yet being enforced. To test H1 and H2, our treatment vignettes both indicate that
374    Dustin Gamza and Pauline Jones

the repressive law is being enforced, and we vary the degree of enforcement from fines
(low) to arrests and imprisonment (high). We estimate a linear probability model to test
each hypothesis.
To test H3, we vary the type of religious practice that is being targeted by repressive
regulations. Experiment 1 concerns repression of an individualistic religious prac-
tice: restrictions on wearing the hijab in public places.4 Experiment 2 concerns repres-
sion of a communal religious practice: restrictions on Qurʾan study groups in private
homes.5
These particular examples of an individualistic and communal practice were selected
for two main reasons. First, they are contextually appropriate examples of individual-
istic and communal religious practices, respectively. Wearing the hijab in Kyrgyzstan,
as in other parts of the Muslim world, is a personal choice that is made by women them-
selves and their families (see, e.g., Carkoglu 2009). Also similar to many other Muslim
contexts, there are state and social pressures not to wear the hijab in Kyrgyzstan. Because
the hijab is not ubiquitous and has only recently grown in popularity among young
women, choosing to wear the hijab may actually alienate the wearer from some other
members of the Muslim community who perceive the hijab as incompatible with being
Kyrgyz (McBrien 2017)—​a discourse that is supported by the state (Suyarkulova 2016,
258).6 Factor analysis also produced two clear dimensions, corresponding to individual-
istic and communal religious practices, with wearing the hijab and Qurʾan study groups
loading onto separate dimensions. Wearing the hijab loaded onto the same factor as ac-
cess to printed religious materials and religious websites, whereas Qurʾan study groups
loaded onto the same factor as attending mosque and celebrating religious holidays
such as Ramadan.7
Our survey results also offer strong evidence to suggest that there is greater opposition
to placing government restrictions on convening Qurʾan study groups in one’s home
than on wearing the hijab in public places—​regardless of an individual’s self-​assessment
of how important religion is in their life (see Table 18.1). In other words, no matter how
religious a respondent considers themself, they are more likely to oppose restrictions on

Table 18.1 Opposition to Restrictions on Individual/​Communal Practices

by Religiosity
Determining Whether the Hijab Prohibiting Qurʾan Group Study
Can Be Worn in Public is a Bad or in Private Homes is a Bad or Very
Degree of Religiosity15 Very Bad Idea Bad Idea

High 17% 55%

Low 26% 36%

Note: Data is weighted. Questions concerning support for religious restrictions also included a neutral
response: “neither a good idea nor a bad idea.”
 Repressive Religious Regulation and Political Mobilization    375

religious practices that are deemed to be communal than to oppose restrictions on reli-
gious practices that are deemed to be individualistic.
The second reason for selecting wearing the hijab and Qurʾan study groups as
examples of individualistic and communal practices, respectively, is that restrictions
on them are very plausible. Although neither wearing the hijab in public or convening
Qurʾan study groups at home was illegal in Kyrgyzstan at the time we fielded the survey,
in recent years both topics have become politically salient issues, with the government
increasingly discouraging wearing the hijab and opposing some forms of informal reli-
gious teaching (Toursunof 2009; Nasritdinov and Esenamanova 2017).8 Additionally, in
neighboring Central Asian states, such as Tajikistan and Uzbekistan, there have been se-
vere crackdowns on both these practices, suggesting that Kyrgyzstani respondents will
be more likely to interpret the hypothetical scenario as a real possibility.9
The use of a survey experiment enabled us to reduce the politically sensitive nature of
the questions concerning political mobilization and to mitigate social desirability bias.
We did not ask respondents whether they would engage in political action, but whether
they would support the decision of others to engage in political action10—​specifically,
those who believed that the religious practice being repressed was “a manifestation
of piety.” The design of the survey experiment also enabled us to avoid the assump-
tion that citizens have complete knowledge of repressive regulations without directly
experiencing repression. It thus controls for an important alternative explanation: that
awareness of repression reduces the likelihood that an individual will engage in dissent
due to greater perception of personal risk (Honari 2018, 957).

Why Kyrgyzstan?
Kyrgyzstan is one of the five Muslim countries in Central Asia that became an inde-
pendent state when the Soviet Union collapsed in 1991. As in the other four MMS that
comprise the region, Kyrgyzstan underwent what many scholars have characterized as
an “Islamic revival” analogous to what occurred throughout the Islamic world in the
1970s and 1980s (Haghayeghi 1994; Olcott 1994; Lipovsky, 1996). In actuality, Islam in
Central Asia has undergone a transformation (Jones 2016). In other words, what we have
witnessed in Kyrgyzstan and throughout the region is not simply the reemergence of
beliefs and practices that were suppressed under Soviet rule, but rather their alteration
in form, nature, and appearance, as individuals and communities have gained direct
access to ideas and information about Islam via a variety of new and newly accessible
sources. Re-​characterizing Central Asia’s Islamic revival as a transformation is not only
more accurate but also more consistent with the notion that Islamic beliefs and practices
throughout the region should be viewed on a continuum from scriptural/​textual to
mystical/​traditional, with most people not located firmly at either end (Tucker 2013).
Although these “competing discourses” have existed in Central Asia for centuries, they
manifest themselves in different ways both at the individual and the community level,
376    Dustin Gamza and Pauline Jones

and reflect what appears to be a “pluralization of Islam” in both beliefs and practices
since independence (Khalid 2007, 123; see also Ro’i and Wainer 2009).
Kyrgyzstan has also experienced a political transformation since independent state-
hood. Once heralded for being the only democratic regime in Central Asia (Anderson
1999), it is now widely considered to be a hybrid regime.11 This transformation has
happened gradually and has paralleled the change over time in the state’s approach to
religious regulation. All five Central Asian states inherited a similarly restrictive reg-
ulatory framework over religion due to their common experience under Soviet rule
(for details, see Tasar 2017). After independence, however, the governments of each
country responded differently to what seemed to be Islam’s growing political potential
(for details, see Gamza and Jones 2020). Kyrgyzstan proceeded gradually to establish
a state infrastructure to regulate religion and initially limited the scope of these regu-
latory bodies, consistent with the principles of individual religious freedom and state
secularism codified in its first (1993) constitution. In contrast, neighboring Uzbekistan
moved quickly to criminalize religion and to expand the authority of its regulatory
bodies to enforce expansive restrictions on religious belief and practice. By the late
2000s, however, Kyrgyzstan’s regulatory strategy appeared to be gradually converging
toward Uzbekistan’s, as the government began to endorse a singular interpretation of
Islam and to expand restrictions on unsanctioned religious beliefs and practices. Yet,
unlike its neighbor, although religious repression has increased, it has not yet become
normalized.
Finally, Kyrgyzstan shares several features with MMS around the world. It is
considered to be highly religious according to standard measures. The level of religiosity
has been growing, moreover, since independence, according to these same commonly
used measures.12 Kyrgyzstan is also similar to other MMS in that its citizens have an ex-
tremely low baseline level of participating in and supporting political mobilization as a
catalyst for political change (Livny, this volume). Based on our survey, less than 1 per-
cent claim to have participated in a protest in the last three years, only 7 percent believe
that protest is very effective in getting the attention of politicians, and 82 percent believe
participating in a protest is somewhat or very risky behavior. And yet, also similar to
their counterparts in MMS, in recent years Kyrgyzstanis have mounted successful mass
protests against unpopular or corrupt regimes, suggesting they are willing and capable
of overcoming obstacles to collective action.13

Summary of Findings

In sum, we find that, controlling for level of religiosity, enforcement of repressive

regulations has a significant but negative effect on attitudes toward political mobiliza-
tion, particularly when it comes to restrictions on communal religious practice. (See
Appendix 18.2 for figures summarizing these results.)
 Repressive Religious Regulation and Political Mobilization    377

H1: Support for political mobilization will be greater when repressive legislation is
enforced

We find that enforcement of repressive regulation has a significant effect, but in the
opposite direction. Respondents are less, not more, likely to have positive attitudes
toward political mobilization when repressive regulations are enforced. This suggests
that the enforcement of repressive regulations decreases, rather than increases, the
likelihood that Muslim citizens will support political mobilization in response. It may
be that fear of repercussions, or a sense of law-​abidingness, overrides any political
will to mobilize; however, we do not find support for this proposition. Alternatively, it
may be that political mobilization in the face of potential fines or arrests is perceived
as ineffective or costly in the Kyrgyz context, resulting in a cost-​benefit analysis that
discourages mobilization in the face of repression. As discussed above, there is a low
baseline level of positive feelings toward political mobilization in Kyrgyzstan in ge-
neral, and it appears that religious repression does not elevate this baseline. At the
same time, however, in the context of the experiments, Kyrgyzstanis were broadly
supportive of mobilization in the face of religious repression; even in the treatment
groups, consistently more than 32 percent of respondents considered each form of
mobilization to be appropriate.14

H2: Support for political mobilization will be greater when the degree of enforcement,
defined as the severity of penalty, increases

We also find some support for our hypothesis that as the degree of enforcement
increases, the effect on attitudes toward political mobilization will also increase.
Contrary to our expectations, however, the effect is negative rather than positive. Point
estimates are lower for the more severe penalty (high enforcement) in Experiment 1
(individual practice), and in some cases (protest action, contacting local authorities to
change policy, and signing or writing a public letter), the degree of enforcement only has
a significant effect on attitudes toward political mobilization when it is high. However,
this pattern does not hold in Experiment 2 (communal practice), where the effect of
both low and high degrees of enforcement on support for mobilization is nearly iden-
tical. These findings suggest that, in the case of communal practices (here: Qurʾan study
groups), respondents do not differentiate among types or levels of penalty for violating
the repressive regulation. Instead, they interpret any penalty as effectively marking the
religious adherents who received this punishment as criminals. This would have longer-​
term repercussions beyond the initial penalty. In other words, a fine today may lead to
greater scrutiny and potential punishments tomorrow.

H3: The effect of repressive regulation on support for political mobilization will be more
significant when the repressive regulation being enforced targets communal (rather
than individualistic) religious practices
378    Dustin Gamza and Pauline Jones

Finally, we find that restrictions targeting communal religious practices do have a

stronger effect on attitudes toward political mobilization than restrictions targeting in-
dividual practices. Yet, once again, this effect is in the opposite direction than expected.
In Experiment 1 (individualistic), enforcement decreases support for political action,
but less significantly than in Experiment 2 (communal). In Experiment 1 (individual-
istic), a low degree of enforcement is significant only for two types of political action
(continuing to wear the hijab and criticizing the policy online), while a high degree of
enforcement is significant for all forms of political mobilization, with the exception of
moving abroad. In Experiment 2 (communal), both low and high enforcement are sig-
nificant and have a negative effect on attitudes toward all forms of political mobilization,
with the exception of moving abroad and participating in protest.
Although our results suggest that the relationship between religious repression and
political mobilization is inverse to our hypothesis, they nonetheless support our claim
that Muslims react differently to restrictions that target communal religious practices
than they do to those that target individualistic ones. Instead of an increased desire to
protect the religious community’s right to engage in the communal practice, however,
what may be driving the results is the desire to avoid communal punishment. In other
words, Muslims may be deterred from mobilizing against the enforcement of repres-
sion targeting communal practices because it increases the risk of collective punish-
ment by bringing the entire community into the state’s spotlight for potential retaliation
or crackdowns. In contrast, enforcement of restrictions on an individualistic practice
would focus the state’s wrath on a subset of individuals within the community who
choose to engage in or mobilize in defense of this practice.
In sum, we find strong support for the contention that enforcement of repressive reg-
ulation has a greater effect than repressive regulation alone; however, the effect is neg-
ative—​that is, in the opposite direction that we expected. Moreover, in the experiment
on individualistic practices, a high degree of enforcement is more broadly effective for
reducing support for political mobilization than a low degree of enforcement, which has
no significant effect in most cases. This suggests that the bar may be lower for supporting
mobilization against restrictions on individualistic practices, because the potential costs
to the community as a whole are lower. Recall that both low and high enforcement are
effective in reducing support when it comes to restrictions on communal practices. We
also find support for our contention that the effect of enforcement is stronger—​and
more consistent across different types of political action—​for communal religious prac-
tice than it is for individualistic religious practice. Finally, and perhaps most impor-
tantly, we find that these results are not being driven by the level of religiosity.

Subgroup Analysis

To better understand these results, we also conducted several subgroup analyses to ob-
serve heterogeneous effects, including the following subgroups: (1) women, (2) the most
 Repressive Religious Regulation and Political Mobilization    379

faith-​abiding (i.e., respondents who agree with the sentiment that faith must take prec-
edence over law), (3) the most devout, (4) the most opposed to regulation (i.e., those
who responded that government regulation was a bad or very bad idea for each type
of restriction), and (5) respondents from the Osh region, where ethnic Uzbek citizens
of Kyrgyzstan are heavily concentrated. Overall, these results provide more evidence to
support our conclusions.
Two subgroups are of particular interest: the faith-​abiding and the most devout.
(See Appendix 18.3 for figures summarizing these results.) The faith-​abiding sub-
group is based on a question asked earlier in the survey: Which of the following three
statements do you agree with most: (1) A person must follow the principles of his/​her
faith, regardless of any legal consequences; (2) A person must take into account the
legal consequences while following the principles of his faith; (3) A person should al-
ways follow the law, even if it contradicts their religious principles. Considering only
members of the subgroup that believes a person must follow the principles of their
faith regardless of legal consequences, in Experiment 1 (individualistic), only fines
in the case of wearing the hijab in public places has a significant negative effect. In
Experiment 2 (communal), results are similar to the full sample, with a high degree
of enforcement (i.e., arrests and imprisonment) having a significant and negative ef-
fect on support for all forms of political action, except for moving abroad and protest.
However, a low degree of enforcement (i.e., fines) does not have a significant effect
on support for political mobilization. Importantly, this suggests that in the case of
communal practices, individuals who believe they must follow religious principles
even when they oppose the law may have a higher threshold for repression before it
discourages a political response.
The most devout subgroup is also based on a question asked earlier in the
survey: How religious do you consider yourself to be? The response options included: Very
religious; Somewhat religious; Not very religious; and Not at all religious. Considering only
members of the subgroup that consider themselves to be “very religious,” in Experiment
1 (individualistic), enforcement only has a significant effect, which is again negative
rather than positive, on whether respondents support one political action: continuing
to wear the hijab in public. In Experiment 2 (communal), the results are again similar to
the full sample, with at least one form of enforcement having a significant negative effect
on all forms of mobilization, except moving abroad and participating in a protest.

Why Muslim’s (Don’t)

Rebel: Implications of Our Findings

Our findings have important implications for the literatures on religious regulation, po-
litical mobilization, and repression. Not only do we find that the enforcement of repres-
sive religious regulation has a direct effect on attitudes toward political mobilization, but
380    Dustin Gamza and Pauline Jones

also that the effect is both significant and negative across various forms of political ac-
tion. This effect is present even when controlling for religiosity and commitment to reli-
gious principles, suggesting that repression reduces support for political action against
the state even among the most devout citizens. Further, the reductive effect is partic-
ularly pronounced when the repression is targeted at communal religious practices as
opposed to individualistic ones.
Why does enforcement of repressive regulations—​particularly those that target com-
munal practices—​appear to reduce rather than increase support for taking political ac-
tion against the state? Several alternative explanations are possible, including the social
desirability and perceived effectiveness of mobilization, as well as the inability to over-
come collective action problems. Yet we find no support for any of these in the context
of our experiments. Instead, we argue that the fear of collective punishment increases as
the degree of enforcement increases—​and that this fear is greatest when the restrictions
being enforced target communal practices. In other words, our initial theory may be
wrong for the right reasons: Muslims are motivated to protect their community from
harm, but the certainty of financial and physical harm outweighs the expectation of
increasing religious persecution. Religion, not religiosity, thus plays an important role
in shaping attitudes toward political mobilization in that it forms the sense of commu-
nity that needs protection. Contra the existing literature linking the defense of com-
munal practices to increased mobilization, our findings suggest that when enforcement
is high, defending communal practices makes mobilization less desirable because it
raises the perceived cost (financial and physical) to the community itself.

Acknowledgments
This chapter is part of a larger multiyear, multicountry research project on Religious Repression
and Political Mobilization in Central Asia. Funding was provided by the National Science
Foundation (no. SES-​1658336). The opinions, findings, conclusions or recommendations
expressed are those of the authors and do not necessarily reflect the views of the National
Science Foundation or any other department or agency. For research assistance, we are grateful
to Anil Menon. We also would like to thank the participants in the 2019 book workshop at the
University of Michigan, especially Megan Ryan and Ragnhild Nordås, for thoughtful feedback
on an earlier version.

Notes
1. We measure the degree of enforcement by the harshness of penalties that are imposed on
those who violate the regulation in question. These penalties range from low (fines) to high
(arrest and imprisonment).
2. We define communal religious practices as those that bring together and serve to enhance
the local community, such as attending mosque or celebrating Eid al-​Fitr, and individual-
istic religious practices as those that display or are associated with individual piety, such as
wearing a hijab or performing salat.
 Repressive Religious Regulation and Political Mobilization    381

3. The risk is also much lower than, for example, in a highly repressive autocratic regime,
where the expectation of citizen compliance is much greater (see, e.g., Blaydes 2018).
4. Experiment 1 included a total of 1,787 respondents; 597 individuals were assigned the con-
trol group, 616 were assigned the first treatment (fines), and 574 were assigned the second
treatment (arrest and imprisonment). Nonresponses were dropped for the analysis.
5. Experiment 2 included a total of 1,780 respondents; 583 individuals were assigned the con-
trol group, 605 were assigned the first treatment (fines), and 592 individuals were assigned
the second treatment (arrest and imprisonment). Nonresponses were dropped for the
analysis.
6. Nonetheless, recent survey research suggests that almost half the population has a positive
opinion toward women who wear hijab, while less than a quarter has a negative opinion
(Jalil 2016).
7. In a separate question on the survey, we ask respondents about their level of support for
restrictions on multiple types of religious practices that correspond to communal and in-
dividualistic. Factor loadings were 0.7 or greater.
8. When conducted under the auspices of state authorities, however, Qurʾan study groups at
home are encouraged as a bulwark against radicalization and a way to empower women,
particularly in the South (Djanibekova 2017).
9. Authorities in Tajikistan have been cracking down on wearing the hijab in public for years,
and in 2017 the state passed an amendment to the law on traditions restricting the hijab.
In August 2017 the authorities approached more than eight thousand women wearing
the hijab in Dushanbe and ordered them to wear their headscarves in the Tajik style
(Agerholm 2017). In Uzbekistan, private religious education is punished, and holding un-
authorized religious meetings can carry lengthy prison sentences (Bayram 2010).
10. The use of proxy subjects is one of many common approaches for mitigating social desira-
bility bias (for details, see, e.g., Nederhof 1985).
11. For example, Freedom House’s political rights and civil liberties scores, which rank coun-
tries from 1-​7 where 7 means the smallest degree of freedom, rated Kyrgyzstan 7 in po-
litical rights and 5 in civil liberties in 2020. Compare this to 1992, the first full year of
independence, for which Kyrgyzstan received a 4 in political rights and 2 in civil liberties.
For details, see: https://​freed​omho​use.org/​rep​ort/​free​dom-​world.
12. In 2003, 74% of Kyrgyzstanis described themselves as a religious person, while in 2011,
92% described themselves as religious. When asked, “How important is God in your life?,”
where 1 represents not at all important and 10 represents very important, 63% selected 8, 9,
or 10 in 2003, compared to 73% in 2011 (Inglehart et al. 2014).
13. Analogous to the wave of mass protests throughout the Middle East in 2010–​2011, com-
monly known as the Arab Spring, Kyrgyzstan is the only Central Asian state to have
experienced mass uprisings that toppled incumbent regimes. For details, see, e.g.,
Temirkulov 2010.
14. Approval for various types of political mobilization across experiments ranges from 33%
to 68%, while disapproval ranges from 28% to 59%. Where each experimental group falls
within this range is driven, in part, by the treatment it received—​that is, the presence and
degree of enforcement.
15. Here we use one of the standard survey questions (from the World Values Survey) to measure
religiosity: How important is religion in your life? A high degree of religiosity includes only
those who answered, “Religion is very important in my life.” A low degree of religiosity
includes those who answered, “Religion is fairly/​not very/​not at all important in my life.”
382    Dustin Gamza and Pauline Jones

References
Agerholm, Harriet. 2017. “Tajikistan Passes Law ‘to Stop Muslim Women Wearing Hijabs.’” The
Independent, September 1. http://​www.inde​pend​ent.co.uk/​news/​world/​asia/​tajiks​tan-​mus​
lim-​hij​abs-​stop-​women-​law-​hea​dsca​rfs-​cent​ral-​asia-​islam-​a7923​886.html.
Anderson, John. 1999. Kyrgyzstan: Central Asia’s Island of Democracy? Amsterdam: Harwood
Academic.
Bayram, Mushfig. 2010. “UZBEKISTAN: Muslims Jailed Long-​Term, Short-​Term Christian
Jailings Re-​start.” Forum 18, April 26. http://​www.foru​m18.org/​arch​ive.php?art​icle​_​id=​1436.
Berman, Eli. 1999. “Sect, Subsidy, and Sacrifice: An Economist’s View of Ultra-​Orthodox Jews.”
Quarterly Journal of Economics 115, no. 3: 905–​953.
Blaydes, Lisa. 2018. State of Repression: Iraq under Saddam Hussein. Princeton, NJ: Princeton
University Press.
Cammett, Melani, and Pauline Jones. 2014. “Is There an Islamist Political Advantage?” Annual
Review of Political Science 17: 187–​206.
Campbell, David. 2004. “‘Acts of Faith: Churches and Political Engagement.’” Political Behavior
26, no. 2: 155–​180.
Carkoglu, Ali. 2009. “Women’s Choices of Head Cover in Turkey: An Empirical Assessment.”
Comparative Studies of South Asia, Africa and the Middle East 29, no. 3: 450–​467.
Chaves, Mark, and David E. Cann. 1992. “Regulation, Pluralism, and Religious Market
Structure: Explaining Religion’s Vitality.” Rationality and Society 4, no. 3: 272–​290.
Davenport, Christian, and Molly Inman. 2012. “The State of State Repression Research since
the 1990s.” Terrorism and Political Violence 24, no. 4: 619–​634.
Djanibekova, Nurjamal. 2017. “Kyrgyzstan: The Quiet Success of an Islamic Woman’s Group.”
Eurasianet, December 14. https://​eur​asia​net.org/​kyr​gyzs​tan-​the-​quiet-​succ​ess-​of-​an-​isla​
mic-​wom​ans-​group.
Esmer, Yilmaz, and Thorleif Pettersson. 2007. “The Effects of Religion and Religiosity on
Voting Behavior.” In The Oxford Handbook of Political Behavior, edited by Russell J. Dalton
and Hans-​Dieter Klingemann, 481–​503. New York: Oxford University Press.
Finke, R. 1990. “Religious Deregulation: Origins and Consequences.” Journal of Church & State
32, no. 3: 609–​626.
Finke, Roger, and Rodney Stark. 1988. “Religious Economies and Sacred Canopies: Religious
Mobilization in American Cities, 1906.” American Sociological Review 53, no. 1: 41–​49.
Fox, Jonathan. 2013. “State Religion and State Repression.” In The Routledge Handbook of
Religion and Security, edited by Chris Seiple, Dennis R. Hoover, and Pauletta Otis, 182–​192.
New York: Routledge.
Gamza, Dustin, and Pauline Jones. 2016. “Reluctant Monopolists: Religious Regulation,
Politicization, and Identity in Central Asia.” Paper presented at the MPSA Annual
Conference.
Gamza, Dustin, and Pauline Jones. 2020. “The Evolution of Religious Regulation in Central
Asia, 1991–​2018.” Central Asian Survey 40, no. 2: 197–​221.
Grzymala-​Busse, Anna. 2012. “Why Comparative Politics Should Take Religion (More)
Seriously.” Annual Review of Political Science 15: 421–​442.
Haghayeghi, Mehrdad. 1994. “Islamic Revival in the Central Asian Republics.” Central Asian
Survey 13, no. 2: 249–​266.
Hoffman, Michael T., and Elizabeth R. Nugent. 2017. “Communal Religious Practice and
Support for Armed Parties: Evidence from Lebanon.” Journal of Conflict Resolution 61, no.
4: 869–​902.
 Repressive Religious Regulation and Political Mobilization    383

Honari, Ali. 2018. “From ‘the Effect of Repression’ toward ‘the Response to Repression.’”
Current Sociology Review 66, no. 6: 950–​973.
Iannaccone, Lawrence. 1991. “The Consequences of Religious Market Structure: Adam Smith
and the Economics of Religion.” Rationality and Society 3, no. 2: 156–​177.
Inglehart, R., C. Haerpfer, A. Moreno, C. Welzel, K. Kizilova, J. Diez-​Medrano, M. Lagos, et al.,
eds. 2014. World Values Survey: All Rounds—​Country-​Pooled Datafile Version. Madrid: JD
Systems Institute. http://​www.worldv​alue​ssur​vey.org/​WVSD​ocum​enta​tion​WVL.jsp.
Jalil, Rouslan. 2016. “The Social Significance of Islam in Post-​Soviet Central Asia: The Case
of Kyrgyzstan.” In Islam, Society, and Politics in Central Asia, edited by Pauline Jones, 3–​28.
Pittsburgh: University of Pittsburgh Press.
Jones, Pauline. 2016. “Reassessing the ‘Islamic Revival’ in Central Asia.” In Islam, Society,
and Politics in Central Asia, edited by Pauline Jones, xi–​xviii. Pittsburgh: University of
Pittsburgh Press.
Karatas, Mustafa, and Özlęm Sandikci. 2013. “Religious Communities and the
Marketplace: Learning and Performing Consumption in an Islamic Network.” Marketing
Theory 13, no. 4: 465–​484.
Khalid, Adeeb. 2007. Islam after Communism: Religion and Politics in Central Asia.
Berkeley: University of California Press.
Khan, Sammyh S., Nick Hopkins, Stephen Reicher, Shruti Tewari, Narayanan Srinivasan,
and Clifford Stevenson. 2016. “How Collective Participation Impacts Social Identity: A
Longitudinal Study from India.” Political Psychology 37, no. 3: 309–​325.
Lim, Chaeyoon, and Robert D. Putnam. 2010. “Religion, Social Networks, and Life
Satisfaction.” American Sociological Review 75, no. 6 (December): 914–​933.
Lipovsky, Igor. 1996. “The Awakening of Central Asian Islam.” Middle Eastern Studies 32,
no. 3: 1–​21.
Lussier, Danielle N. 2019. “Mosques, Churches, and Civic Skill Opportunities in Indonesia.”
Journal for the Scientific Study of Religion 58, no. 2: 415–​438.
McBrien, Julie. 2017. “Mukadas’s Struggle.” In From Belonging to Belief: Modern Secularisms
and the Construction of Religion in Kyrgyzstan, 108–​ 130. Pittsburgh: University of
Pittsburgh Press.
McTague, John Michael, and Geoffrey C. Layman. 2009. “Religion, Parties, and Voting
Behavior: A Political Explanation of Religious Influence.” In The Oxford Handbook of
Religion and American Politics, edited by James L. Guth, Lyman A. Kellstedt, and Corwin E.
Smidt, 330–​370. New York: Oxford University Press.
Nasritdinov, Emil, and Nurgul Esenamanova. 2017. “The War of Billboards: Hijāb, Secularism,
and Public Space in Bishkek.” Central Asian Affairs 4, no. 2: 217–​242.
Nederhof, Anton. J. 1985. “Methods of Coping with Social Desirability Bias: A Review.”
European Journal of Social Psychology 15: 263–​280.
Nicolet, S., and Anke Tresch. 2009. “Changing Religiosity, Changing Politics? The Influence of
Belonging and Believing on Political Attitudes in Switzerland.” Politics and Religion 2: 76–​99.
Olcott, Martha Brill. 1994. “Central Asia’s Islamic Awakening.” Current History 93
(April): 150–​154.
Ro’i, Yaacov, and Alon Wainer. 2009. “Muslim Identity and Islamic Practice in Post-​Soviet
Central Asia.” Central Asian Survey 28, no. 3: 303–​322.
Stark, Rodney and Laurence R. Iannaccone. 1994. “A Supply-​Side Reinterpretation of the
‘Secularization’ of Europe.” Journal for the Scientific Study of Religion 33, no. 3: 230–​252.
Stark, Rodney, and Roger Finke. 2000. Acts of Faith: Explaining the Human Side of Religion.
Berkeley, CA: University of California Press.
384    Dustin Gamza and Pauline Jones

Suyarkulova, Mohira. 2016. “Fashioning the Nation: Gender and Politics of Dress in
Contemporary Kyrgyzstan.” Nationalities Papers 44, no. 2: 247–​265.
Tasar, Eren. 2017. Soviet and Muslim: The Institutionalization of Islam in Central Asia. New York
and Oxford: Oxford University Press.
Temirkulov, Azamat. 2010. “Kyrgyz ‘Revolutions’ in 2005 and 2010: Comparative Analysis of
Mass Mobilization.” Nationalities Papers 38, no. 5: 589–​600.
Thompson, Kenneth. 1982. Emile Durkheim: On Morality and Society, Selected Writings.
[Excerpt from The Elementary Forms of the Religious Life.] Chicago: University of
Chicago Press.
Toursunof, Hamid. 2009. “Kyrgyzstan: Islamic Schools in the Spotlight.” Chalkboard, January
2. http://​cha​lkbo​ard.tol.org/​kyr​gyzs​tan-​isla​mic-​scho​ols-​in-​the-​spotli​ght.
Tucker, N. 2013. “Domestic Shapers of Eurasia’s Islamic Future: Sheikhs, Scholars, Society
and the State.” Presentation at the “Islam in Eurasia” Policy Conference, Kennan Institute,
Washington, DC, June 6–​7.
Wald, K., and Martinez, M. 2001. “Jewish Religiosity and Political Attitudes in the United States
and Israel.” Political Behavior 23, no. 4: 377–​397.

Appendix 18.1 Survey Experiments

Full Text

Survey Experiment no. 1: Wearing Hijab in Public

Control: No Enforcement
Please consider the following situation: In another country where Muslims consti-
tute the majority of the population, the government recently passed a law prohibiting
women from wearing the hijab outside their homes in public places, but this law is not
being enforced. What do you think are appropriate actions for Muslims who believe
that wearing the hijab is a manifestation of piety for women to take in response?

Treatment #1: Regulation with Low Degree of Enforcement

Please consider the following situation: In another country where Muslims consti-
tute the majority of the population, the government recently passed a law prohibiting
women from wearing the hijab outside their homes in public places. Authorities are is-
suing fines to women wearing the hijab in public places. What do you think are appro-
priate actions for Muslims who believe that wearing the hijab is a manifestation of piety
for women to take in response?

Treatment #2: Regulation with High Degree of Enforcement

Please consider the following situation: In another country where Muslims make up the
majority of the population, the government recently passed a law prohibiting women
from wearing the hijab outside their homes in public places. Authorities are arresting
and imprisoning women wearing the hijab in public places. What do you think are
 Repressive Religious Regulation and Political Mobilization    385

Table 18.2: Hijab Repression Experiment Response Options

Yes, this is No, this is not I don’t know Refuse to
appropriate (1) appropriate ( 0) (98) answer (99)

Keep wearing hijab in 1 0 98 99

public places
Criticize this policy on 1 0 98 99
the Internet or through
social networks
Sign or write a public 1 0 98 99
letter to the newspaper
to draw attention to this
policy
Contact local authorities 1 0 98 99
to change this policy
Participate in a protest 1 0 98 99
action to demand
changes to this policy
Move abroad to a 1 0 98 99
country where you can
freely practice your faith

appropriate actions for Muslims who believe that wearing the hijab is a manifestation of
piety for women to take in response?

Survey Experiment no. 2: Gathering in Private Home

for Qurʾan Study

Control: No Enforcement
Please consider the following situation: In another country where Muslims constitute
the majority of the population, the government recently passed a law prohibiting people
from gathering in groups at home to study the Qurʾan together, but this law is not being
enforced. What do you think are appropriate actions for Muslims who believe that co-​
studying the Qurʾan at home is a manifestation of piety to take in response?

Treatment #1: Regulation with Low Degree of Enforcement

Please consider the following situation: In another country where Muslims constitute
the majority of the population, the government recently passed a law prohibiting people
from gathering in groups at home to study the Qurʾan together. Authorities are issuing
fines to people gathering in groups at home to study the Qurʾan. What do you think
386    Dustin Gamza and Pauline Jones

are appropriate actions for Muslims who believe that co-​studying the Qurʾan at home is
a manifestation of piety to take in response?

Treatment #2: Regulation with High Degree of Enforcement

Please consider the following situation: In another country where Muslims make up the
majority of the population, the government recently passed a law prohibiting people
from gathering in groups at home to study the Qurʾan together. Authorities are ar-
resting and imprisoning people gathering in groups at home to study the Qurʾan.
What do you think are appropriate actions for Muslims who believe that co-​studying the
Qurʾan at home is a manifestation of piety to take in response?

Table 18.3: Qurʾan Repression Experiment Response Options

Yes, this is No, this is not I don’t know Refuse to answer
appropriate (1) appropriate ( 0) (98) (99)

Keep gathering at 1 0 98 99
home to study the
Qurʾan
Criticize this policy 1 0 98 99
on the Internet
or through social
networks
Sign or write a 1 0 98 99
public letter to the
newspaper to draw
attention to this
policy
Contact local 1 0 98 99
authorities to
change this policy
Participate in a 1 0 98 99
protest action to
demand changes to
this policy
Move abroad to a 1 0 98 99
country where you
can freely practice
your faith
 Repressive Religious Regulation and Political Mobilization    387

Appendix 18.2 Survey Experiments

Results

Move abroad to a country where Islamic faith can be freely practiced-

Participate in protest action demanding policy change -

Contact local authorities to change policy -

Sign or write a public letter -

Criticize this policy on the Internet -

Keep wearing hijab in public places -

–0.3 –0.2 –0.1 0.0 0.1

Arrest and imprisonment for practice Fined for practice

Figure 18.1. Experiment 1, Wearing Hijab in Public (full sample, N=1787). Bold line indicates
95% confidence interval, thin line indicates 99% confidence interval. The hollow dot represents
the point estimate.

Move abroad to a country where Islamic faith can be freely practiced-

Participate in protest action demanding policy change -

Contact local authorities to change policy -

Sign or write a public letter -

Criticize this policy on the Internet -

Keep gathering at home to study the Qur’an-

–0.3 –0.2 –0.1 0.0 0.1

Arrest and imprisonment for practice Fined for practice

Figure 18.2. Experiment 2, Gathering in Private Home for Qurʾan Study (full sample,
N=1780). Bold line indicates 95% confidence interval, thin line indicates 99% confidence interval.
The hollow dot represents the point estimate.
388    Dustin Gamza and Pauline Jones

Appendix 18. 3 Selected Subgroup

Analysis

Faith-​Abiding Subgroup

Move abroad to a country where Islamic faith can be freely practiced-

Participate in protest action demanding policy change -

Contact local authorities to change policy -

Sign or write a public letter -

Criticize this policy on the Internet -

Keep wearing hijab in public places -

–0.3 –0.2 –0.1 0.0 0.1 0.2

Arrest and imprisonment for practice Fined for practice

Figure 18.3. Experiment 1, Wearing Hijab in Public (sample includes only those who think
that a person must follow the principles of his or her faith, regardless of any legal consequences,
N=362). Bold line indicates 95% confidence interval, thin line indicates 99% confidence interval.
The hollow dot represents the point estimate.

Move abroad to a country where Islamic faith can be freely practiced-

Participate in protest action demanding policy change -

Contact local authorities to change policy -

Sign or write a public letter -

Criticize this policy on the Internet -

Keep gathering at home to study the Qur’ar-

–0.4 –0.2 0.0 0.2

Arrest and imprisonment for practice Fined for practice

Figure 18.4. Experiment 2, Gathering in Private Home for Qurʾan Study (sample includes
only those who think that a person must follow the principles of his/​her faith, regardless of any
legal consequences, N=344). Bold line indicates 95% confidence interval, thin line indicates 99%
confidence interval. The hollow dot represents the point estimate.
 Repressive Religious Regulation and Political Mobilization    389

Most Devout Subgroup

Move abroad to a country where Islamic faith can be freely practiced-

Participate in protest action demanding policy change -

Contact local authorities to change policy -

Sign or write a public letter -

Criticize this policy on the Internet -

Keep wearing hijab in public places -

–0.3 –0.2 –0.1 0.0 0.1 0.2

Arrest and imprisonment for practice Fined for practice

Figure 18.5. Experiment 1, Wearing Hijab in Public (sample includes only those who
identified themselves as “very religious”, N=476). Bold line indicates 95% confidence interval,
thin line indicates 99% confidence interval. The hollow dot represents the point estimate.

Move abroad to a country where Islamic faith can be freely practiced-

Participate in protest action demanding policy change -

Contact local authorities to change policy -

Sign or write a public letter -

Criticize this policy on the Internet -

Keep gathering at home to study the Qur’an-

–0.3 –0.2 –0.1 0.0 0.1

Arrest and imprisonment for practice Fined for practice

Figure 18.6. Experiment 2, Gathering in Private Home for Qurʾan Study (sample includes only
those who identified themselves as “very religious”, N=459). Bold line indicates 95% confidence
interval, thin line indicates 99% confidence interval. The hollow dot represents the point estimate.
 Chapter 19

How Extraordi na ry Was

the Arab Spri ng ?
Examining “Protest Potential” in the Muslim World

Avital Livny

In the months following Mohamed Bouazizi’s self-​immolation in Tunisia, the mass

demonstrations of the Arab Spring captivated public attention. In the years since,
scholars have sought to understand their immediate causes and consequences,
identifying who participated and why, with a special focus on the role of economic
grievances (Campante and Chor 2012; Costello, Jenkins, and Aly 2015), social media
(Clarke and Kocak 2020; Lynch 2011), and Islam (Hoffman and Jamal 2014; Kurzman
and Türkoğlu 2015). But the Arab Spring has rarely been placed in a broader, com-
parative perspective. As such, we do not yet have a clear understanding of whether
these mass political mobilizations were really as extraordinary as they appeared to be.
Understanding how the Arab Spring fits into patterns of political participation, more
broadly, can help us assess whether this type of large-​scale political movement is likely
to happen again.
Before 2010, the Middle East and North Africa (MENA) region was singled out for
low levels of mass political mobilization. In explaining the absence of democracy in the
region, Eva Bellin noted that it had ben devoid of “mammoth, cross-​class coalitions
mobiliz[ing] on the streets to push for reform, as in South Korea” (2004, 150). In this
view, the mass demonstrations of the Arab Spring were doubly remarkable—​not just
for successfully overthrowing two autocratic regimes, but also for effectively reversing
a long history of political passivity. As Bellin herself has admitted, “it was an enormous
empirical surprise to see tens if not hundreds of thousands of ordinary citizens of di-
verse social original joining in the protests for political change across the Arab world” in
392   Avital Livny

2010–​2011 (2012, 135). Alternatively, the events may have simply proven the earlier view
incorrect: the peoples of the Muslim Middle East may have been ripe for political mobi-
lization all along,—​or large-​scale protest may have been taking place without capturing
international attention.
To disentangle these interpretations and contextualize the Arab Spring
appropriately—​vis-​à-​vis the past, as well as the future—​it helps to explore the na-
ture of “protest potential” in MENA, and in the Muslim world more broadly. By “pro-
test potential” I mean the willingness, among average citizens, to participate in mass
politics, shoulder-​to-​shoulder with their compatriots, especially in activities that are
considered to be less “conventional” than voting and joining a political party. These
activities can range from signing petitions and organizing boycotts to joining strikes
and demonstrations. Political scientists have invested considerable energy in assessing
variation in these types of political activities, both within countries and across them.
But none have yet examined how MENA and other Muslim countries compare to the
non-​Muslim world.
Comprehensive studies of Muslim politics (e.g., Fish 2011) have yet to examine polit-
ical participation, and most regional studies of activism (e.g., Lust-​Okar and Zerhouni
2008; Stockemer and Khazaeli 2014) focus on conventional politics only. Otherwise, we
are left with studies of single cases (e.g., Chrona and Capelos 2017; Lussier 2019; Paler,
Marshall, and Atallah 2018) or particular organizations (e.g., El-​Said and Rauch 2015);
ones confined to the Muslim world alone (e.g., Sarkissian 2012); or broad examinations
of religion and politics that assess the role of Islam, among other faiths (e.g., Arikan
and Bloom 2018), or participation of Muslims in exclusively non-​Muslim contexts (e.g.,
Peucker 2018).
In contrast to these extant studies, my goal here is to undertake a comparative inquiry,
asking how citizens in the Middle East and other Muslim countries fare in terms of their
potential to engage in large-​scale political activism. Across a number of measures and
specifications, I find a dramatic participation gap in the Muslim world, similar to the
one hypothesized by Bellin (2004). The deficit I find is statistically and substantively
significant, and it holds across demographic groups and in a wide variety of political
and economic contexts. As such, it transcends differences in opportunities, risks, and
resources between Muslim and non-​Muslim countries. Moreover, it does not appear to
be driven by bias in the ways that Muslim respondents report their willingness to partic-
ipate, whether prospectively or retrospectively.
Although the protest gap is not driven by Islam itself, the relationship be-
tween participation and religiosity among Muslims offers an important clue as to
its source: while personal religious beliefs and private practices tend to depress
individuals’ willingness to become politically active, investments in one’s religious
community—​including regular attendance at mosque and membership in religious
associations—​spur activism. I suggest that the latter, but not the former, speaks to a
salient Islamic identity, which supports political participation by addressing concerns
about social trust among individual activists. While generalized interpersonal trust
tends to be low in Muslim countries, and while this distrust tends to undermine
 How Extraordinary Was the Arab Spring?    393

protest potential, the effect is entirely ameliorated for those with salient Islamic
identities, who tend to trust other Muslims. As such, social trust may be the root cause
of the region’s participation gap as well as the source of a mobilization advantage
among Islamic-​based political groups.

Defining and Measuring

Protest Potential

Beyond voter turnout and party preferences, political scientists have examined partici-
pation in so-​called “unconventional” or “unorthodox” politics (Barnes and Kaase 1979;
Inglehart 1977; Inglehart 1990; Marsh 1974). These varied forms of political activism are
common to democracies and autocracies alike. In the former, grass-​roots movements
regularly hold elected officials accountable, acting as a counterbalance to the moneyed
interests of the elite; in the latter, large-​scale protests are capable of bringing about re-
gime change. In both contexts, these types of activities are a “mechanism by which cit-
izens can communicate information about their interests, preferences, and needs and
generate pressure to respond” (Verba, Schlozman, and Brady 1995, 1). Traditionally,
scholarly emphasis was placed on conventional politics because it was thought to be
the only form of activity that engaged large numbers of citizens (Norris 2007). But as
the Arab Spring proved, mass politics can also take unconventional forms. Similarly,
while unconventional politics were once dismissed as anti-​systemic and violent, most
now recognize the important role they play in accountability and representation
(Stockemer 2014).
In addition to caring about participation in these types of activity for their own sake,
social scientists are invested in understanding protest potential because the willingness
of individuals to become politically active speaks to a collective quality that is valuable
for other outcomes. Because most mainstream political movements are massive in scale
and largely anonymous, their success does not depend on the specific contribution of
any one participant; and since their success results in policy changes that are enjoyed
by all citizens, individuals benefit from these activities irrespective of their actual con-
tribution. As a result, large-​scale political participation represents a quintessential col-
lective action problem: even if all citizens support a movement and its objectives, there
is an incentive for each to opt out and simply free-​ride off of others’ efforts; and without
sufficient numbers, a political movement cannot succeed, leaving everyone worse off
(Chong 1991; Hardin 1971; Olson 2002).
Like the successful provision of public goods, mass politics requires that this free-​
rider problem be effectively addressed. The solution rests in the heterogeneity of indi-
vidual preferences: while some (~20 percent) seem to have rational egoist preferences,
including the incentive to free-​ride, the majority (~40–​60 percent) are “conditional
cooperators,” who prefer to do their part as long as others do the same (Ahn, Ostrom,
394   Avital Livny

and Walker 2003; Ostrom 2000). Collective action—​whether political or otherwise—​

critically depends on these cooperators identifying one another and working together.
Ultimately, it hinges on whether cooperators expect to be working with others like
them, who want to do right by those who do right by them. This expectation of recip-
rocal cooperation, despite incentives to defect, is a common definition of a trust (Hardin
2002; Nannestad 2008). For this reason, protest potential is closely connected to trust
and trustworthiness among average citizens, as well as their ability to work collectively
to solve problems—​the same features that bolster economic development (Knack and
Keefer 1997), health (Kawachi, Subramanian, and Kim 2008), and happiness (Helliwell
and Putnam 2004).
To assess variation in protest potential, scholars have typically relied on self-​reports
from survey respondents who describe whether they have taken part or would take part
in a number of activities, from signing petitions and attending lawful demonstrations
to joining boycotts and strikes and occupying buildings.1 These activities vary in their
degree of “acceptability,” which parallels their perceived costliness, both of which exac-
erbate the underlying collective action problem. And by asking which of these actions
respondents “have done,” which they “would do,” and which they would “never, under
any circumstances do,” both retrospective and prospective activism can be assessed.
Prospective activism fits squarely into the concept of “protest potential” and, like actual
participation, it indicates a capacity to solve the collective action problem.
Using responses from a number of cross-​national survey projects, I am able to es-
timate protest potential among 871,674 respondents from 141 countries,2 including 31
where Muslims constitute a majority of the population.3 In these contexts, Muslims
constitute the largest subgroup and are therefore most likely to lead (or at least pop-
ulate) any mass mobilization effort. Of course, not every Muslim-​majority country
is covered by the surveys, including some particularly repressive regimes, but any
sampling bias should produce a lower bound on the gap between the Muslim and
non-​Muslim world.
As with all large-​scale surveys, there are important questions to ask about the rep-
resentativeness of the data. While each of the surveys is designed to be nationally rep-
resentative, since the number of activists in each country might be quite small, even
small biases in respondent selection or response could impact how we assess protest
potential. Luckily, when combining retrospective and prospective participation, the
number of would-​be activists is not all that tiny. For the activity with the lowest cost—​
signing a petition—​protest potential is well over 60 percent, on average, ranging
from 100 percent to just under 5 percent (µ=0.962, σ=0.213); and nearly 15 percent of
respondents, on average, say that they have done or would consider doing the highest-​
cost activity, occupying a building. Other measurement concerns are likely to be
ameliorated by the number of surveys per country (4.8 on average). While estimates of
protest potential from surveys within each country tend to cluster together, averaging
across these observations will help to address sampling or response biases in any par-
ticular survey.
 How Extraordinary Was the Arab Spring?    395

Assessing Protest Potential in the

Muslim World

Across all respondents, patterns follow existing expectations: the largest share have signed
a petition (28.8 percent) or might consider doing so (32.7 percent), followed by the share
that have taken part in a demonstration (17.6 percent) or might do so (31.9 percent). Smaller
numbers have boycotted (8.2 percent) or would (29.9 percent), joined a strike (5.7 percent,
21.4 percent), or occupied a building (2.2 percent, 12.7 percent). Self-​reported participation
across activities is significantly intercorrelated, particularly when distinguishing those who
have or would participate in each from those who would never consider doing so. As such,
it seems plausible to collapse protest potential across multiple activities into a single index,
reflecting whether respondents have done or would do any of them.
To assess protest potential in the Muslim world, I begin with the raw data, calcu-
lating rates of retrospective and prospective participation in the five separate activities
within each country. Arranging these according to the Muslim population-​share, a dra-
matic pattern emerges (Figure 19.1)—​in each case, protest potential declines with the

Petition Demonstrate Boycott Strike Occupy

100
Have Participated

50

0
100
Would Participate

50

0
0 50 100 0 50 100 0 50 100 0 50 100 0 50 100
Percent Muslim

Figure 19.1: Protest Potential, by Muslim Population-​Share

Notes: Average rates of retrospective (“have participated”) and prospective (“would partici-
pate”) participation across surveys within 141 countries, for each of five separate political activ-
ities: signing a petition, attending a lawful demonstration, participating in a boycott, joining a
strike, and occupying a building. Retrospective participation indicated in the light gray and pro-
spective participation in the medium gray. Across the share of the country’s population that is
Muslim, with lines indicating the lowess trends for both types of participation.
396   Avital Livny

share of Muslims in the country, with the strongest effects for prospective participation
(whether respondents would consider participating). Dividing the set of countries into
two groups—​Muslim-​majority versus non-​Muslim—​I find that the differences are al-
most always statistically significant: respondents in Muslim countries have and would
petition, demonstrate, boycott, and occupy less; and while they are no less likely to have
joined a strike, if they have not done so before, they are significantly less likely to con-
sider doing so in the future.
Of course, there are important differences between Muslim and non-​Muslim coun-
tries that could be driving this pattern. To begin, a range of factors at the individual-​
level, known to correlate with political participation, could be different (or operate
differently) in Muslim-​majority settings. For example, women tend to be less politically
active, on average (Verba, Burns, and Schlozman 1997), and this could be even more pro-
nounced in Muslim societies, where some gender inequities are higher, on average (Fish
2011). Similarly, participation tends to increase with age (Quintelier and Blais 2016), and
Muslim countries have tended to have younger populations (Groth and Sousa-​Poza
2012). And while education tends to bolster participation (Verba, Schlozman, and Brady
1995), Muslim countries have historically had lower graduation rates (Papagapitos and
Riley 2009).
Obviously, political activism also requires some baseline interest in politics (Verba,
Schlozman, and Brady 1995), and one possible explanation for the absence of popular
mobilization in MENA (and the broader Muslim world) could be widespread apathy
(Mednicoff 1998). Alongside interest, expectations about the responsiveness of the
political system to citizen demands also tends to explain variation in participation
(Hooghe and Marien 2013), and although political trust has not been found to be con-
sistently lower in Muslim countries, it may operate quite differently there, strengthening
support for autocratic regimes (Jamal 2007). Individual income is also often taken as
an indicator of both grievances—​pushing low-​income citizens to become political—​
and resources—​boosting participation among the wealthy (Gallego 2007). Income in-
equality could also be higher in the Muslim world (El-​Ghonemy 1998), increasing the
relative share of both political classes.
Beyond differences across individuals, variation at the national level could also ex-
plain why protest potential is lower in Muslim countries. State institutions create both
opportunities and risks for would-​be activists. On the one hand, democracy creates
openings for citizen demands (Dalton, van Sickle, and Weldon 2009), and the Muslim
world (especially MENA) is known for its democratic deficit (Fish 2002). On the other
hand, state repression closes off opportunities for activism, raising the costs of pushing
for change, and some Muslim countries are known to limit political competition (Lust-​
Okar 2004). Meanwhile, natural resources are thought to restrict participation even
further, both in the Muslim world and beyond (Bellin 2004; Smith 2004). Similarly, na-
tional income can both support and depress protest (Gurr 1968; Inglehart 1990), and
many Muslim countries are said to be historically underdeveloped (Kuran 2004).
To understand how these variables might impact the participation gap in the
Muslim world, I begin by creating an index of protest potential, coded one (1) if a
 How Extraordinary Was the Arab Spring?    397

respondent has done or would do any activity, and zero ( 0) otherwise. To maximize
coverage, I focus on participation in petitions, demonstrations, and boycotts alone.
Using this index, I calculate the percent of protest potential across different groups of
respondents, comparing those living in Muslim countries to those living elsewhere.
For example, to assess whether the difference between the two regions is driven by
gender dynamics, I compare levels of protest potential among men and among
women, inside and outside of the Muslim world, checking to see whether the gap is
driven by only one of the two genders.
Gender roughly bifurcates the sample, and I use age categories that divide the
sample into four roughly equal groups: 25 and under, 26–​40, 41–​55, and 56 and older.
For education, I distinguish those who have at least a university degree from those who
completed high school and those who completed their primary education, coding the
rest as having no formal education. To capture variation in apathy, I use a survey ques-
tion about respondents’ interest in politics, creating a binary indicator of being either
“very” or “somewhat” interested in politics. Political trust is another binary indicator
of having either “a lot” or “some” confidence in one’s national government. To create
income groups of roughly equal size, I combine those in the bottom two quintiles and
those in the top two.
Across countries, I define democracies as those with a Polity score of six or above
in the survey-​year, with the remaining country-​years classified as autocracies.4 To as-
sess state control of political opposition, I use the Freedom of Assembly and Association
index from the Cingranelli-​Richards (CIRI) Human Rights Dataset in the country-​year
(see Cingranelli and Richards 2010). Countries are coded as resource-​dependent if they
are above the global average in terms of the percentage of their GDP that comes from re-
source rents (roughly 6 percent). Similarly, in terms of overall income, I code countries
as high-​income if they are above the global average in terms of logged GDP per capita
(roughly 8.75).
The results of these comparisons are displayed in Figure 19.2, showing the averages
in each group within Muslim and non-​Muslim countries as well as the gap between the
two. Across all demographic groups—​men and women, within every age and educa-
tion level—​a statistically significant gap remains. While levels of retrospective and pro-
spective participation are significantly lower in Muslim countries for each group, the
patterns across groups on each dimension are generally the same in Muslim and non-​
Muslim countries: women participate less than men, participation is decreasing in age
and increasing in education and income, and it is lower for those who are apathetic and
who trust their government.
Still, some noteworthy patterns are worth mentioning. The participation gap is
slightly larger for women than it is for men, perhaps reflecting religious-​based norms
against Muslim women’s participation in politics. Meanwhile, the gap is considerably
smaller for the undereducated, which could indicate that a lack of education creates
grievances that propel participation in the Muslim world, while signifying a lack of re-
sources that depress participation elsewhere. A similar story can be told about the re-
lationship between income and participation at the individual level, with a weaker
398   Avital Livny

Individual-Level:
Gender:
Male
Female
Age:
25 and Under
26 to 40
41 to 55
56 and Older
Education:
None
Primary
Secondary
University
Political Interest:
Little
Some
Political Trust:
Little
Some
Income:
Low Income
Middle
High Income
Country-Level:
Polity:
Autocratic
Democratic
Assembly:
Severe
Limited
Unrestricted
Resources:
Non-Dependent
Dependent
GDP per Capita:
Low Income
High Income
–50 0 50 100
Protest Potential

Figure 19.2: Participation Gap in Muslim World, across Covariates

Notes: Retrospective and prospective participation in any of three political activities—​petitions,
demonstrations, boycotts—​in Muslim-​plurality countries (indicated by darker gray) and non-​
Muslim countries (in light gray), with the gap indicated by dashed lines.

correlation between the two in Muslim countries. I also find a smaller gap among those
who trust the current government, which would seem to indicate that Jamal’s (2009)
observations about Palestine hold more broadly.
Similarly, examining variation at the national level, I find the gap remains among
democracies and autocracies alike, although participation rates are relatively higher
in both Muslim and non-​Muslim democracies. The gap remains when Freedom of
Assembly and Association is severely restricted and limited, although it is substan-
tively smaller for those Muslim countries where the political opposition is largely un-
restricted. This pattern—​combined with a larger gap among democracies—​is difficult
to interpret—​it could point to an interaction between regime and restrictions, whereby
participation is limited by either in the Muslim world.
 How Extraordinary Was the Arab Spring?    399

I also find that protest potential is significantly lower among Muslim countries that
are resource-​dependent and among those that have few natural resources, with a slightly
smaller gap among resource-​rich countries. This would seem to indicate that rentierism
cannot explain the participation gap in the Muslim world. The gap holds among both
high-​and low-​income countries, with a larger gap among higher-​income countries.
Interestingly, national income has no effect on protest potential in the Muslim world.
While national income bolsters participation in non-​Muslim countries—​ostensibly due
to citizen resources—​it may generate acquiescence and apathy in Muslim countries.

Can Retrospective and Prospective

Responses Be Trusted?

Further work should explore some of the patterns I have identified between the Muslim
and non-​Muslim worlds, but for my purposes here, it is sufficient to note that there is
a persistent gap in protest potential between Muslim and non-​Muslim countries, one
that is both statistically significant and substantive in magnitude. That retrospective and
prospective participation follows similar patterns across groups, inside and outside of
the Muslim world, attenuates some concerns that there is something different about the
nature or quality of self-​reports from the region. But given that, on average, Muslim-​
majority countries tend to be less democratic, with more severe restrictions on political
activities, respondents living there may feel less comfortable freely admitting that they
are politically active. So, while self-​reports of retrospective and prospective participa-
tion are generally valid (Quintelier and Blais 2016), they may be systematically less so in
the Muslim world.
That the gap in protest potential persists in both democracies and autocracies
challenges this suggestion. So does the fact that the participation gap between Muslim
and non-​Muslim countries remains in self-​reported voter turnout, an activity where
there is pressure to overreport rather than underreport (Selb and Munzert 2013).5 But
neither of these directly assess the validity of self-​reports among Muslim respondents.
To do so, we need to be able compare retrospective and prospective self-​reports to actual
participation, and it is rare that we know anything systematic about participants in a
real-​world political event.
Luckily, a unique survey was conducted among participants in the protest activities
in and around Istanbul’s Gezi Park in June 2013. In addition to demographic indicators,
the survey asks participants whether this was their first time participating in a large-​
scale political activity. In combination with two representative surveys I conducted
in Turkey—​ one in 2012, which happened to ask respondents whether they had
participated or would participate in a mass demonstration, and another in 2015, which
asked about their participation in the Gezi Park events—​I am able to assess the validity
of both prospective and retrospective self-​reports from the Gezi survey.6 Although
400   Avital Livny

Turkey was—​until recently—​one of the most liberal democracies in the Muslim world,
its recent autocratic turn can be said to have started with the Gezi Park events. As such,
even if prospective participation in 2012 would have been self-​reported in a relatively
liberal environment, retrospective participation in 2015 would not have been, making
the Turkish case more similar to other Muslim settings.
Figure 19.3 displays the demographic characteristics of eight different groups of
respondents: those who said, prospectively, that they would never participate in a dem-
onstration, and those who said, retrospectively, that they did not participate in the
Gezi events; those who said, prospectively, that they had never participated but would
consider doing so, and those who said, retrospectively, that they had not taken part in
the Gezi events, but know those who did; those who said, prospectively, that they had
participated in a demonstration before, and those who said, retrospectively, that they
were at Gezi; and those who were surveyed at Gezi Park—​both those who said it was
their first time participating in a large-​scale demonstration and those who self-​reported
participating before. For each group, I assess the proportion that is female, the distri-
bution of ages, and the average level of education (whether respondents had completed
primary, secondary, or tertiary education).

Ex ante:
Would never Participate

Ex post:
Did not Participate

Ex ante:
Haven’t, but would

Ex post:
Know those who did

Ex ante:
Participated before

Ex post:
Participated at Gezi

At Gezi:
First participation

At Gezi:
Participated before
0 .5 1 20 40 60 80 100 0 .5 1
Proporation Female Age Education Share

Figure 19.3: Retrospective and Prospective Participants in the Gezi Park Protests
Notes: Groups of respondents from three representative surveys are compared in terms of gender
balance, age distribution, and completion of primary (dark gray), secondary (medium gray), and
tertiary education (light gray). Ex ante participation—​“have” vs. “would” vs. “would never”—​in
any political demonstration in 2012 compared to participants surveyed during the 2013 Gezi Park
protests.
 How Extraordinary Was the Arab Spring?    401

I find some evidence that women may be less likely to honestly report their partici-
pation, especially ex post: while those surveyed at Gezi were slightly more female than
male, the group that later admitted participating in the Gezi events was overwhelmingly
male. Otherwise, the results are roughly consistent: participants at Gezi were younger
and better educated, and that can be seen in the retrospective self-​reports—​those that
said they took part are the youngest and best educated of the prospective and retrospec-
tive respondents, including those who admitted to being politically active in the 2012
survey, well before the Gezi Park events took place. It is also interesting to see just how
similar the Gezi Park participants were, as a group: it is difficult to distinguish demo-
graphically those who were participating in their first demonstration from those who
had participated before, indicating that it might be right to combine survey respondents
who have participated and those who would into a single set of potential protesters.

The Role of Islam in the Muslim

Participation Gap

Given the participation gap in the Muslim world, it is important to understand the role
that Islam itself might play in undermining protest potential in the region. Generally
speaking, religion can enter into political decision-​making in a number of ways: nom-
inal denomination can effect activism if a particular religious doctrine encourages or
discourages political involvement; religious institutions and their relationship to the
state can foment or suppress political activity; and, across members of the same faith,
the level of religiosity can influence levels of political engagement. In other words, par-
ticipation could be lower in the Muslim world because Islam discourages activism, be-
cause institutions in Muslim-​majority countries do, and/​or because religiosity among
individual Muslims discourages participation.
Political science has a long history of examining the role of religion and religiosity
in political activity, although the great majority of existing research has focused on
Christians in the Western world (Wald and Wilcox 2006). Those studies that have fo-
cused on Islam, in particular, have found that the impact of religion in politics is roughly
similar for Christians and Muslims, although a direct comparison has not yet been
made (Jamal 2005; Sarkissian 2012). The few that have assessed the impact of denomi-
nation and institutions examine Islam as one of the world’s many faiths, so we do not yet
know if it is particularly distinctive (Arikan and Bloom 2018).
To assess the impact of both denomination and institutions, I can leverage variation
at the individual-​level—​between self-​identified Muslims and non-​Muslims—​and at the
country-​level—​between Muslim-​majority and non-​Muslim contexts. If Islam, as a de-
nomination, is the source of the participation gap, then Muslims should have lower rates
of protest potential across both sets of countries; if context is more to blame, then the
participation gap might extend to non-​Muslims living in the Muslim world.
402   Avital Livny

The results of this two-​by-​two comparison are illustrated in Figure 19.4. They show
that Islam, as a doctrine, is unlikely to be depressing protest potential, as Muslims in
non-​Muslim countries participate at rates equal to or greater than their compatriots.7
Meanwhile, context does seem to matter, as participation rates are lower for both
Christians and Muslims living in the Muslim world. That rates are comparatively lower
for Muslims in the region would seem to indicate that faith—​in some form, beyond
doctrine—​is playing an additional role in suppressing protest potential there.
If religious doctrine cannot explain the participation gap for Muslims in the Muslim
world, how else might religion suppress protest potential? Increasingly, scholars of reli-
gion and politics have distinguished between two dimensions of religiosity, each with a
different impact on political engagement. The first—​private religious belief—​seems to
suppress political activism, while the latter—​collective religious engagement—​seems to
support it (Arikan and Bloom 2018; Livny 2020; Wilcox, Wald, and Jelen 2008). Could
the same findings hold among Muslims in the Muslim world?
To directly compare the two dimensions, and their impact on protest potential,
I gathered multiple markers of each from the cross-​national survey dataset, focusing
on self-​identified Muslims in Muslim-​majority countries. Private religious beliefs
can be captured with responses to the following questions: whether religion is impor-
tant in respondents’ lives, whether God is important in their lives, and whether they
self-​identify as a religious person. Collective engagement can be assessed using the
frequency of mosque attendance and membership in a religious association. Sitting

80
Non-Muslim
Countries
Non-Muslim
Countries
Muslim
60 Non-Muslim Countries
Muslim Countries Muslim
Countries Countries
Protest Potential

Non-Muslim
Countries
40 Muslim
Countries

20

0
Petition Demonstrate Boycott Index
Muslims Non-Muslims

Figure 19.4: Islam and Protest Potential, Inside and Outside of the Muslim World
Notes: Rates of retrospective and prospective participation among Muslims and non-​Muslims,
distinguished by whether they live in Muslim-​plurality or non-​Muslim countries.
 How Extraordinary Was the Arab Spring?    403

between the two—​capturing private religious engagement—​is an indicator of whether

respondents self-​report private moments of prayer and reflection. Using principal com-
ponent factor analysis, I confirm the difference between these indicators: while all six
load onto a single “religiosity” factor, only mosque attendance and associational mem-
bership load onto a second factor, one that picks up religion’s more collective aspects.
Rates of protest potential across each measure of religiosity are displayed in Figure
19.5. As elsewhere, among other faiths, a familiar pattern emerges: rates of retrospec-
tive and prospective participation are declining across levels of personal religiosity but
increasing across levels of collective religious engagement. In regression analysis with
demographic controls and country fixed-​effects, I find that each of these effects is statis-
tically significant.
But why the difference? Clearly, religious beliefs—​as preferences—​and religious
practices—​as behavior—​could have different impacts on political engagement, but to
understand the crux of the difference, it is useful to focus on the impact of private prayer
on participation. The negative effect highlights that what is distinctive about mosque
attendance and associational membership is not that they are behaviors but that they
are collective. If we remember that political participation is a quintessential collective ac-
tion, then the connection between the two starts to sharpen. Perhaps those who engage
in religious group-​activities are better able to address the underlying collective action
problem, supporting participation in both religious and political activities.

Religion God Religious Private Mosque Religious

Important Important Person Prayer Attendance Association
100

75

50

25

0
Level of Religiosity

Figure 19.5: Religiosity and Protest Potential among Muslims in the Muslim World
Notes: Retrospective (“have done,” in light gray) and prospective (“would do,” in medium gray)
participation in any of three political activities, across levels of religiosity: importance of religion
and God in respondents’ lives, self-​identified religiosity, a binary indicator of private prayer, fre-
quency of mosque attendance, and a binary indicator of membership in a religious association.
404   Avital Livny

I can capture this dynamic empirically by focusing on a key obstacle to collective

action—​interpersonal distrust. Recall that participation in group activities, whether reli-
gious or political, depends on trustworthy “conditional cooperators” successfully locating
each other and working together. Even if they are individually motivated to participate,
if these cooperators believe that most people in their communities cannot be trusted to
do right by them, they will opt out of group activities. As such, generalized interpersonal
distrust—​the belief that most people cannot be trusted—​limits protest potential (Benson
and Rochon 2004). But if religious group-​activities help to overcome this collective action
problem, either by building social capital that can be applied elsewhere (Kwak, Shah, and
Holbert 2004) or by developing group-​based trust that can substitute for more generalized
trust-​expectations (Livny 2020), then the link between distrust and participation would
be severed for those who are invested in their religious communities.
To assess this possibility, I estimate a model of protest potential where the direct
effects of generalized distrust and religiosity are assessed, along with their interaction.
Combining the direct and interaction terms, I am able to define the effect of distrust
on participation, conditional on the level of religiosity. To compare the two dimensions
of religiosity, I use the markers with the broadest coverage: the importance of religion
in respondents’ lives, as a measure of private religiosity, and frequency of mosque at-
tendance, to capture collective religious engagement. The key results of this model are
displayed in Figure 19.6. While generalized distrust tends to depress protest potential,

Collective Religion

Private Religion

Generalized Distrust

Distrust × Collective

Distrust × Private

–.15 –.1 –.05 0 .05 .1 .15

Effect on Protest Potential

Figure 19.6: Personal and Collective Religion and the Trust Problem
Notes: OLS models of “protest potential,” combining retrospective and prospective participation
in any of three activities. Collective religion measured using frequency of mosque attendance;
private religion measured as self-​reported importance of religion, in the first model (dark gray),
and private prayer, in the second (light gray). Distrust is a binary indicator of the feeling that most
people cannot be trusted. All models include demographic controls (age, gender, education) and
country fixed-​effects with standard errors clustered within countries. Bars indicate statistical sig-
nificance at 90 percent, 95 percent, and 99 percent levels.
 How Extraordinary Was the Arab Spring?    405

the effect is entirely ameliorated for those who attend mosque with some frequency—​at
least once per month. Meanwhile, the interaction between distrust and private religi-
osity is statistically indistinguishable from zero, meaning that it does not address the
underlying distrust problem.
I find the same patterns when I use more comparable measures of private and collec-
tive religiosity—​private prayer compared to prayer at mosque. Even with a more lim-
ited sample size, I continue to find that collective engagement solves the trust problem
while private practices do not. Of course, there are many plausible interpretations of
these results. Those who attend mosque may simply be more trusting than average, or
they may gain practical or psychic resources through attending that make them more
likely to engage in politics. But I find little evidence of either of these. On the one hand,
frequent mosque attendees are statistically indistinguishable from others in terms of
their level of trust. And the direct effect of mosque attendance on political participa-
tion is null, if not negative, indicating that the impact of religious engagement runs only
through those who generally distrust.

Conclusion

In this survey of protest potential across the Muslim world, I have found a substantive
and significant participation gap: individuals living in Muslim-​majority countries are
less likely to engage in politics. This is especially true for those who are Muslims them-
selves, and who engage in religious practices privately or hold strong personal religious
beliefs. The participation gap between Muslim and non-​Muslim countries holds across
demographic categories and in all types of countries. Within the Muslim world, only
engagement in collective religious activities bolsters political participation. While I have
focused my analysis here on communal prayer and formal associations, participation in
Qurʾanic reading circles and enrollment in religious schools are likely to have a similar
positive effect on political activity.
I have argued that the distinguishing feature of these all religious group-​activities
is their ability to address the collective action problem underlying the decision to be-
come politically active. Elsewhere, I have made a more specific claim that different
forms of collective religious engagement help to develop a sense of Islamic group
identity among those who take part, building trust among them (Livny 2020). This
group-​based trust can effectively substitute for more generalized trust-​expectations,
supporting cooperation and coordination in both politics and economics. Ultimately,
this also helps to address the initial puzzle of the participation gap in the Muslim
world: because collective political action rests so heavily on trust, and because
generalized interpersonal trust tends to be lower in Muslim-​majority countries,
individuals opt out of most any political activity that does not tap into their Muslim
identity and the group-​based trust it fosters. The result is lower levels of political (and
economic) participation overall.
406   Avital Livny

Given this, how are we to understand the remarkable events of the Arab Spring, in-
cluding the mass demonstrations that followed in and around Istanbul’s Gezi Park in
2012? I believe we should indeed define them as extraordinary. They were indicative of
overwhelming frustration among the masses which, when mixed with their courage
and tenacity, was able to overcome a pervasive trust problem and a lack of underlying
“protest potential” in the region. Unfortunately, these types of spontaneous (and ex-
traordinary) events are often short-​lived. When it comes to keeping the momentum
going—​ coordinating and cooperating over the longer term to build long-​ lasting
change—​the trust problem will undoubtedly creep back in. Islamic-​based identity poli-
tics may help, in this respect, as could other forces that evoke group-​based trust. Without
a solution to the collective action problem, extraordinary demonstrations, powered by
anger and hope, may punctuate history but are unlikely to change the course of events.

Notes
1. The set of now-​standard questions first appeared in Marsh (1974) and were developed fur-
ther by Kaase and Marsh (1979).
2. Between the World Values Survey 2018, European Values Study 2011, Afrobarometer 2000,
Afrobarometer 2007, Afrobarometer 2014, Asian Barometer 2009, East Asian Barometer
2017, and Latinobarometer 2018, I have the most information about individuals’ propensity
to demonstrate (N = 848,332), but also their willingness to sign a petition (N = 661,455) and
join a boycott (N = 479,702).
3. As defined by Pew Research Center (2015), the Muslim countries in the dataset are Albania,
Algeria, Azerbaijan, Bahrain, Bangladesh, Burkina Faso, Egypt, Guinea, Indonesia,
Iraq, Jordan, Kazakhstan, Kosovo, Kuwait, Kyrgyzstan, Lebanon, Libya, Malaysia, Mali,
Morocco, Niger, Nigeria, Northern Cyprus, Pakistan, Palestine, Senegal, Sierra Leone,
Sudan, Tunisia, Turkey, and Yemen.
4. I prefer to measure all country-​level variables in the survey-​year—​as opposed to the year or
years before—​to capture current dynamics, more likely to impact prospective participation.
5. In Muslim-​plurality countries, just over 60 percent of respondents self-​report voting in the last
election, while the self-​reported turnout rate in non-​Muslim countries is more than 72 percent.
6. To maximize my chances of comparing like with like, I focus only on those respondents
from Istanbul, with easy access to Gezi Park.
7. That there is a difference in the participation index, but not any of the three individual activ-
ities, indicates that Muslims in non-​Muslim countries tend to participate in more than one
activity, while non-​Muslims will specialize in one or two, increasing the number of non-​
Muslims who participate in any of the three.

References
Ahn, T. K., Elinor Ostrom, and James M. Walker. 2003. “Heterogeneous Preferences and
Collective Action.” Public Choice 117, no. 3–​4: 295–​314.
Arikan, Gizem, and Pazit Ben-​Nun Bloom. 2018. “Religion and Political Protest: A Cross-​
Country Analysis.” Comparative Political Studies 52, no. 2: 246–​276.
 How Extraordinary Was the Arab Spring?    407

Asian Barometer Project. 2009. “Asian Barometer Survey Data: 2003, 2004, 2005, 2006, 2007,
2008.” http://​www.asianbarometer.org.
Asian Barometer. 2017. “Asian Barometer Survey Data Release: Wave IV, 2014–​2016.” http://​
www.asianbarometer.org.
Barnes, Samuel H., and Max Kaase, eds. 1979. Political Action: Mass Participation in Five
Western Democracies. Beverly Hills, CA: SAGE.
Bellin, Eva. 2004. “The Robustness of Authoritarianism in the Middle East: Exceptionalism in
Comparative Perspective.” Comparative Politics 36, no. 2: 139–​157.
Bellin, Eva. 2012. “Reconsidering the Robustness of Authoritarianism in the Middle
East: Lessons from the Arab Spring.” Comparative Politics 44, no. 2: 127–​149.
Benson, Michelle, and Thomas R. Rochon. 2004. “Interpersonal Trust and the Magnitude
of Protest: A Micro and Macro Level Approach.” Comparative Political Studies 37, no.
435–​457.
Bratton, Michael, E. Gyimah-​Boadi, and Robert Mattes. 2000. “Compilation of Afrobarometer
Round 1 Survey in 12 Countries: Botswana, Ghana, Lesotho, Malawi, Mali, Namibia,
Nigeria, South Africa, Tanzania, Uganda, Zambia, and Zimbabwe, 1999–​2001.” https://​www.
afrobarometer.org/​data/​merged-​round-​1-​data-​12-​countries-​1999-​2001.
Bratton, Michael, E. Gyimah-​Boadi, and Robert Mattes. 2007. “Afrobarometer Round 3: The
Quality of Democracy and Governance in 18 African Countries, 2005–​2006.” https://​www.
afrobarometer.org/​data/​merged-​round-​3-​data-​18-​countries-​2005.
Campante, Filipe R., and Davin Chor. 2012. “Why Was the Arab World Poised for Revolution?
Schooling, Economic Opportunities, and the Arab Spring.” Journal of Economic Perspectives
26, no. 2: 167–​187.
Chong, Dennis. 1991. Collective Action and the Civil Rights Movement. Chicago: University of
Chicago Press.
Chrona, Stavroula, and Tereza Capelos. 2017. “The Political Psychology of Participation in
Turkey: Civic Engagement, Basic Values, Political Sophistication and the Young.” Southeast
European and Black Sea Studies 17, no. 1: 77–​95.
Cingranelli, David L., and David L. Richards. 2010. “The Cingranelli-​Richards (CIRI) Human
Rights Dataset, v. 2010.08.15.” http://​www.humanrightsdata.com/​p/​data-​documentation.
html.
Clarke, Killian, and Korhan Kocak. 2020. “Launching Revolution: Social Media and the
Egyptian Uprising’s First Movers.” British Journal of Political Science 5, no. 3: 1025–​1045.
Costello, Matthew, J. Craig Jenkins, and Hassan Aly. 2015. “Bread, Justice, or Opportunity? The
Determinants of the Arab Awakening Protests.” World Development 67: 90–​100.
Dalton, Russell J., Alix van Sickle, and Steven Weldon. 2009. “The Individual-​Institutional
Nexus of Protest Behaviour.” British Journal of Political Science 40: 51–​73.
European Values Project. 2011. “European Values Study 1981–​2008, Longitudinal Data File.”
https://​dbk.gesis.org/​dbksearch/​sdesc2.asp?no=4804&db=e&doi=10.4232/​1.12253
Fish, M. Steven. 2002. “Islam and Authoritarianism.” World Politics 55.1: 4–​37.
Fish, M. Steven. 2011. Are Muslims Distinctive? A Look at the Evidence. Oxford: Oxford
University Press.
Gallego, Aina. 2007. “Unequal Political Participation in Europe.” International Journal of
Sociology 37.4: 10–​25.
El-​Ghonemy, M. Riad. 1998. Affluence and Poverty in the Middle East. London: Routledge.
Groth, Hans, and Alfonso Sousa-​ Poza, eds. 2012. Population Dynamics in Muslim
Countries: Assembling the Jigsaw. Berlin: Springer.
408   Avital Livny

Gurr, Ted Robert. 1968. “A Causal Model of Civil Strife: A Comparative Analysis Using New
Indices.” American Political Science Review 62.4: 1104–​1124.
Hardin, Russell. 1971. “Collective Action as an Agreeable n-​Prisoners’ Dilemma.” Behavioral
Sciences 16, no. 5: 472–​481.
Hardin, Russell. 2002. Trust and Trustworthiness. New York: Russell Sage Foundation.
Helliwell, John F., and Robert D. Putnam. 2004. “The Social Context of Well-​ Being.”
Philosophical Transactions of the Royal Society B: Biological Sciences 359, no. 1449: 1435–​1446.
Hoffman, Michael T., and Amaney Jamal. 2014. “Religion in the Arab Spring: Between Two
Competing Narratives.” Journal of Politics 76, no. 3: 593–​606.
Hooghe, Marc, and Sofie Marien. 2013. “A Comparative Analysis of the Relation between
Political Trust and Forms of Political Participation in Europe.” European Societies 15, no.
1: 131–​152.
Inglehart, Ronald. 1977. The Silent Revolution: Changing Values and Political Styles among
Western Publics. Princeton, NJ: Princeton University Press.
Inglehart, Ronald. 1990. Culture Shift in Advanced Industrial Society. Princeton, NJ: Princeton
University Press.
Jamal, Amaney. 2005. “The Political Participation and Engagement of Muslim Americans.”
American Politics Research 33: 521–​544.
Jamal, Amaney. 2007. “When Is Social Trust a Desirable Outcome? Examining Levels of Trust
in the Arab World.” Comparative Political Studies 40: 1328–​1349.
Jamal, Amaney. 2009. Barriers to Democracy: The Other Side of Social Capital in Palestine and
the Arab World. Princeton, NJ: Princeton University Press.
Kaase, Max, and Alan Marsh. 1979. “Political Action: A Theoretical Perspective.” In Barnes and
Kaase 1979, 27–​56.
Kawachi, Ichiro, S. V. Subramanian, and Daniel Kim. 2008. “Social Capital and Health: A
Decade of Progress and Beyond.” In Social Capital and Health, 1–​28. New York: Springer.
Knack, Stephen, and Philip Keefer. 1997. “Does Social Capital Have an Economic Payoff?
A Cross-​Country Investigation.” Quarterly Journal of Economics 112: 1251–​1288.
Kuran, Timur. 2004. “Why the Middle East is Economically Underdeveloped: Historical
Mechanisms of Institutional Stagnation.” Journal of Economic Perspectives 18, no. 3: 71–​90.
Kurzman, Charles, and Didem Türkoğlu. 2015. “After the Arab Spring: Do Muslims Vote
Islamic Now?” Journal of Democracy 26, no. 4: 100–​109.
Kwak, Nojin, Dhavan V. Shah, and R. Lance Holbert. 2004. “Connecting, Trusting, and
Participating: The Direct and Interactive Effects of Social Associations.” American Economic
Review 87, no. 2: 643–​662.
Latinobarómetro Corporation. 2018. “Latinobarómetro, 1995–​ 2018.” http://​www.
latinobarometro.org/​lat.jsp.
Livny, Avital. 2020. Trust and the Islamic Advantage: Religious-​Based Movements in Turkey and
the Muslim World. Cambridge: Cambridge University Press.
Lussier, Danielle N. 2019. “Mosques, Churches, and Civic Skill Opportunities in Indonesia.”
Journal for the Scientific Study of Religion 58, no. 2: 415–​438.
Lust-​Okar, Ellen. 2004. “Divided They Rule: The Management and Manipulation of Political
Opposition.” Comparative Politics 36, no. 2: 159–​179.
Lust-​Okar, Ellen, and Saloua Zerhouni, eds. 2008. Political Participation in the Middle East.
Boulder, CO: Lynne Rienner.
Lynch, Marc. 2011. “After Egypt: The Limits and Promise of Online Challenges to the
Authoritarian Arab State.” Perspectives on Politics 9, no. 2: 301–​310.
 How Extraordinary Was the Arab Spring?    409

Marsh, Alan. 1974. “Explorations in Unorthodox Political Behaviour: A Scale to Measure

‘Protest Potential.’” European Journal of Political Research 2: 107–​129.
Mattes, Robert, E. Gyimah-​ Boadi, and Michael Bratton. 2014. “Afrobarometer Round
5: The Quality of Democracy and Governance in 34 African Countries, 2011–​2013.” http://​
afrobarometer.org/​data/​merged-​round-​5-​data-​34-​countries-​2011-​2013-​last-​update-​july-​2015
Mednicoff, David M. 1998. “Civic Apathy in the Service of Stability? Cultural Politics in
Monarchist Morocco.” Journal of North African Studies 3, no. 4: 1–​27.
Nannestad, Peter. 2008. “What Have We Learned About Generalized Trust, If Anything?”
Annual Review of Political Science 11: 413–​436.
Norris, Pippa. 2007. “Political Activism: New Challenges, New Opportunities.” In The Oxford
Handbook of Comparative Politics, edited by Carles Boix and Susan C. Stokes, 628–​649.
Oxford: Oxford University Press.
Olson, Mancur. 2002. The Logic of Collective Action: Public Goods and the Theory of Groups.
Cambridge, MA: Harvard University Press.
Ostrom, Elinor. 2000. “Collective Action and the Evolution of Social Norms.” Journal of
Economic Perspectives 14, no. 3: 137–​158.
Paler, Laura, Leslie Marshall, and Sami Atallah. 2018. “The Social Costs of Public Political
Participation: Evidence from a Petition Experiment in Lebanon.” Journal of Politics 80, no.
4: 1405–​1410.
Papagapitos, Agapitos, and Robert Riley. 2009. “Social Trust and Human Capital Formation.”
Economics Letters 102: 158–​160.
Peucker, Mario. 2018. “On the (In)compatibility of Islamic Religiosity and Citizenship in
Western Democracies: The Role of Religion for Muslims’ Civic and Political Engagement.”
Politics and Religion 11: 553–​575.
Pew Research Center. 2015. “The Future of World Religions: Population Growth Projections,
2010–​2050.” Washington, DC: Pew Research Center. https://​www.pewforum.org/​2015/​04/​
02/​religious-​projections-​2010-​2050/​.
Quintelier, Ellen, and André Blais. 2016. “Intended and Reported Political Participation.”
International Journal of Public Opinion Research 28.1: 117–​128.
El-​Said, Hamed, and James E. Rauch. 2015. “Education, Political Participation, and Islamist
Parties: The Case of Jordan’s Islamic Action Front.” Middle East Journal 69.1: 51–​73.
Sarkissian, Ani. 2012. “Religion and Civic Engagement in Muslim Countries.” Journal for the
Scientific Study of Religion 51, no. 4: 607–​622.
Selb, Peter, and Simon Munzert. 2013. “Voter Overrepresentation, Vote Misreporting, and
Turnout Bias in Postelection Surveys.” Electoral Studies 32: 186–​196.
Smith, Benjamin. 2004. “Oil Wealth and Regime Survival in the Developing World, 1960–​
1999.” American Journal of Political Science 48, no. 2: 232–​246.
Stockemer, Daniel. 2014. “What Drives Unconventional Political Participation? A Two Level
Study.” Social Science Journal 51: 201–​211.
Stockemer, Daniel, and Susan Khazaeli. 2014. “Electoral Turnout in Muslim-​Majority States: A
Macro-​Level Panel Analysis.” Politics and Religion 7: 79–​99.
Verba, Sidney, Nancy Burns, and Kay Lehman Schlozman. 1997. “Knowing and Caring about
Politics: Gender and Political Engagement.” Journal of Politics 59, no. 4: 1051–​1072.
Verba, Sidney, Kay Lehman Schlozman, and Henry E. Brady. 1995. Voice and Equality: Civic
Voluntarism in American Politics. Cambridge, MA: Harvard University Press.
Wald, Kenneth D., and Clyde Wilcox. 2006. “Getting Religion: Has Political Science
Rediscovered the Faith Factor?” American Political Science Review 100, no. 4: 523–​529.
410   Avital Livny

Wilcox, Clyde, Kenneth D. Wald, and Ted G. Jelen. 2008. “Religious Preferences and Social
Science: A Second Look.” Journal of Politics 70, no. 3: 874–​879.
World Values Survey Association. 2018. “World Values Survey 1981–​ 2014 Longitudinal
Aggregate, v. 2015-​04-​18.” Vienna: World Values Survey Association. http://​www.
worldvaluessurvey.org/​WVSDocumentationWVL.jsp.
 Chapter 20

Il l icit Ec onomi e s a nd
P olitical Vi ol e nc e i n
Central Asia

Lawrence P. Markowitz and

Mariya Y. Omelicheva

The spread of terrorism as a mode of violence in the late twentieth and early twenty-​first
centuries has generated a rich literature on this form of violent political behavior. Yet
post-​9/​11 studies of terrorism have tended to place Islam within a “securitization” par-
adigm (Buzan, Wæver, and de Wilde 1998), viewing the relationship between religion
and violence through the narrow lens of security, and thereby overpredicting the extent
of terrorist violence across societies. This trend has led empirical studies of terrorism
to center on countries in the Middle East and North Africa, while the relationship be-
tween religion and violence in Muslim-​majority countries in other parts of the world
remains understudied. There are also few studies of “negative” cases, where significant
episodes of violence are possible but have failed to materialize (Mahoney and Goertz
2004). What accounts for these cases with such low levels of terrorism?
Long viewed as harboring an underlying potential for large-​scale violence, Central
Asia (Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, and Uzbekistan) has largely
defied this depiction. Despite features of underdevelopment—​ crumbling infra-
structure, state repression, remittance economies, organized crime, and widespread
corruption—​that should make the region vulnerable to nonstate violence, Central Asia
has experienced surprisingly low levels of terrorism since the late-​1990s. In fact, out of
112,252 terrorist incidents registered by the Global Terrorism Database (GTD; https://​
www.start.umd.edu/​gtd/​) between 1994 and 2016 worldwide, only 211 terrorist attacks
took place in the five Central Asian republics, which constitutes about 0.19 percent
of the global total. This is surprising, given that most of the extremist and violent ter-
rorist incidents identified in the GTD during this period occurred in Muslim-​majority
countries (Cordesman 2017). Between 2013 and 2017, a significant number of Central
412    Lawrence P. Markowitz and Mariya Y. Omelicheva

Asians were recruited into the Islamic State in Iraq and Syria (ISIS), but the prepon-
derance of recruits came not from the region but from Central Asian labor migrants in
Russia, where isolation from family and local communities, frequent discrimination,
and ghettoization within Russian society have been important drivers of extremism
(Tucker 2015). As such, Central Asia constitutes a useful site to examine (a) how theories
of terrorism extend to understudied parts of the Muslim world, and (b) what underlying
conditions explain why some of these regions are characterized by low levels of terrorist
violence.
After reviewing the various explanations of terrorist violence, and applying them
to Central Asia, this chapter explores the conditions under which a state’s involve-
ment in illicit economies—​specifically its collusion in the drug trade—​can dampen
levels of terrorist violence. While the state often enters into the comparative study of
terrorism, our approach emphasizes the complex political economy of security that
defines infrastructurally weak states, where political and security apparatuses are often
immersed in informal and illicit economies (Omelicheva and Markowitz 2019). Since
studying politics (of violence) in nondemocratic contexts, where relevant data is often
unavailable or inaccurate, brings particular challenges, we combine quantitative anal-
ysis (including GIS-​enabled tools) with a series of in-​depth expert interviews conducted
in Central Asia. This approach helps us uncover the complex links between religion
and organized violence, where state apparatuses are often drawn into collaborative
relationships with nonstate actors.

Existing Explanations
of Terrorist Violence

Defined as a form of political violence intended to promote intimidation or incite fear

in a larger audience beyond the immediate victims (Hoffman 2006, 3), terrorism usually
involves small, secretive, nonstate actors operating outside of state control (Crenshaw
2011). Despite the small number of perpetrators, however, terrorism remains one of
the most politically significant forms of violent behavior in Muslim-​majority societies,
in part due to the potentially broad-​based appeal of ideologies and incentives driving
this premeditated use of violence. As a result, the comparative study of terrorism often
centers on accounts of political behavior, such as motive and incentive. Motive refers
to a range of individual-​level considerations—​grievances or psychological needs—​that
causes a person to act. Incentive describes a range of strategic considerations that make
violence possible when an individual concludes that violence promises greater reward
compared to the risks and costs. Such incentives are commonly promised by terrorist
organizations, and are often seen as a necessary link in the mobilization process for
terrorist behavior. Scholars have increasingly sought to bridge the social movement
and political violence literatures in their study of terrorism. A focus on opportunity
 Illicit Economies and Political Violence in Central Asia    413

structure and mobilization, for instance, examines how terrorism arises when certain
political, social, and even geographic conditions are ripe for successful violent polit-
ical contention. Indeed, while some early accounts linking religion to terrorist violence
fastened upon its internal drivers or opportunistic leaders (Juergensmeyer 2000; Stern
2004; Pape 2005), it is now well established that this relationship is best understood in
a broader social, historical, economic, and political context. To this end, this chapter
examines the political behaviors underlying terrorism (and the public fear and gov-
ernment response it elicits) as a form of political violence more broadly, considering
an array of potential causal factors. Though some have surveyed these by analytical ap-
proach (e.g., Chenoweth and Moore 2018), a more limited range of relevant propellants
of terrorist violence can be identified at the micro (individual), meso (group or organi-
zation), and macro (societal or national) levels.
At the individual level, a variety of personality characteristics, psychological factors,
and grievances may lead a person to carry out or participate in terrorist attacks. One
micro-​level constraint is familiarity with victims, as most perpetrators do not wish to
target their own communities. While exceptions exist, extremists in Western Europe
have tended to carry out attacks in foreign theaters, not domestically (Hegghammer
2013). Another individual-​level cause is the absence (or loss) of personal ties, social
networks, or family relationships. Most of the ETA’s early recruits into its terrorist or-
ganizations in Spain in the 1970s and 1980s, for instance, were young (teenagers), single,
and urban (Reinares 2004). Similar patterns have been found among those inspired
by ISIS to carry out attacks in Western Europe (Clarke 2019). Psychological well-​being
and health is also a factor: some individuals’ exposure to violence or personal trauma
may make involvement in violent political acts more likely, while others’ low ego or
dependent personality can make them more amenable to group or leader influence
(Zhirkov, Verkuyten, and Weesie 2014). Local drivers of violence (e.g., rivalry, denun-
ciation, revenge) that attend high-​conflict contexts may also influence an individuals’
willingness carry out violence (Kalyvas 2006; Balcells 2017).
Recent empirical findings from studies of Central Asian societies have found some
support for many of these arguments. Persons surveyed who were young, male, unmar-
ried, had experienced state repression or discrimination, and had links to crime, were
more likely to look favorably toward extremist views and the use of violence for religious
ends (Search for Common Ground 2016; Nasritdinov et al. 2019). Lingering psycholog-
ical effects of exposure to violence during Tajikistan’s civil war, however, are indirect,
and often shaped by the contested narratives between the government and opposition
groups over how the conflict is remembered (Epkenhans 2018). In short, micro-​level
factors point to important fine-​grained differences within and across countries, but they
are strongly influenced by an individual’s interaction with the state.
At the meso level, groups and organizations in society also have important influences
on terrorist violence. As in other forms of violence, those groups with greater organiza-
tional resources are more lethal and effective (Asal and Rethemeyer 2008; Kilberg 2012;
Horowitz and Potter 2014) and tend to last longer (Gaibulloev and Sandler 2013; Phillips
2014). Additionally, institutional supports in society for moderation, compromise,
414    Lawrence P. Markowitz and Mariya Y. Omelicheva

and nonviolence, such as religious institutions (mosques, temples, churches), union

or workplace structures, party organizations, and nongovernmental organizations, are
seen as influential in mitigating violence (Varshney 2002; Schwedler 2007).
These meso-​level features help explain generally lower levels of violence across
Central Asia, where groups struggle to overcome the challenges of a weak organiza-
tional base, insufficient funding, and lack of leadership. The two groups designated as
terrorist organizations by the US government with a definite Central Asia connection—​
the Islamic Movement of Uzbekistan (IMU) and the Islamic Jihad Union—​have had to
repeatedly relocate and redefine themselves, engage in illicit entrepreneurial activity
to stay afloat, and remain focused on local targets in Afghanistan or Pakistan. Central
Asia has not seen the emergence of well-​organized and deeply rooted violent Islamist
organizations that exist in other Muslim-​majority regions. Evidence does not support,
however, those arguments pointing to the moderating influence of political and so-
cial institutions in Central Asia. Civil society has been historically weak in post-​Soviet
Eurasia generally (Howard 2003), as have political parties and formal religious organ-
izations. In their stead, informal mechanisms of dispute resolution have emerged in
Central Asian societies that are localized and rooted in the independent authority of
community religious leaders (Khamidov 2017). Moreover, many of these groups and
organizations are either monitored or controlled by government authorities, raising
questions about their degree of autonomy and their ability to reduce levels of terrorist
violence.
At the macro level, particular socioeconomic characteristics are often seen as drivers
of terrorism. One set of explanations focuses on geography as a predictor of terrorism—​
the size of the territory; the length, remoteness, and topography of its shared borders
with other countries; and mountainous topography. The expectation is that physically
large, mountainous areas with poor transportation infrastructures experience higher
levels of terrorism (Asal, Brinton, and Schoon 2015). Another set of arguments holds
that territories with higher levels of socioeconomic disparity, poverty, and illiteracy will
experience more terrorist activity. Socioeconomic inequality, injustice, and unmet ex-
pectations can push individuals toward taking drastic action (e.g., joining an insurgency
or committing a crime) to alter the sources of their discontent (Ehrlich 2002; Frevtag
et al. 2011), although some studies have shown no direct relationship between poverty,
unemployment, and terrorist activity (Krueger 2008; Piazza 2009).
These broader conditions add another layer of analysis, but they also have mixed
results when applied to Central Asia. To examine those in Central Asia, we ran two sets
of statistical tests. First, we analyzed spatial (time-​invariant) determinants of terrorist
violence, and then regressed it on a number of socioeconomic predictors. All analyses
were performed at the provincial level separately on three Central Asian republics
(Kazakhstan, Kyrgyzstan, and Tajikistan), from 2008 to 2016. To measure terrorist vi-
olence, we used an index of terrorism that captures its intensity by assigning weights
to the counts of domestic terrorist incidents (each is multiplied by 1), the total number
of fatalities caused by terrorist incidents (each fatality is multiplied by 3), and the total
number of injuries caused by terrorism (each injury is multiplied by 0.5) in a province/​
 Illicit Economies and Political Violence in Central Asia    415

year (Terror Index) (Hyslop and Morgan 2014). The data on terrorist events come
from the Global Terrorism Database (GTD), maintained by the National Consortium
for the Study of Terrorism and Responses to Terrorism at the University of Maryland.
Topographic predictors include urban land coverage, distance to the nearest contiguous
country, distance to the capital, mountainous area coverage, and the number of border
crossings in a province (Blyth et al. 2002; Weidmann et al. 2010; Meiyappan et al. 2012).
We also separately examined socioeconomic, demographic, and political variables
that were suggested as determinants of terrorist and criminal activity, including meas-
ures of economic performance, education, unemployment, poverty, and popula-
tion. Because of the differences in the data available across the subregions used in the
study, we relied on different empirical indicators as proxies for some of the conceptual
variables. Our expectation was that more populous and economically under-​developed
territories with higher mortality and unemployment rates will experience higher rates
of terrorism. Territories with educational opportunities will have lesser rate of terrorist
incidents. Our assumption is that the greater the number of people employed in these
sectors of economy, the more people in the province will be de facto above the official
levels of poverty. This, in turn, will be associated with the lower levels of terrorism.
The results of a series of negative binomial regressions with errors clustered over the
province are reported in Table 20.1, which displays coefficients (and clustered errors)
for several socioeconomic and demographic determinants (Model 1) and topographic
variables (Model 2). As with explanations at the micro and meso levels, the results are
mixed (particularly in analyses of socioeconomic factors). On the one hand, all the geo-
graphic predictors turned out in the expected direction and were statistically significant.
Urban areas and less mountainous provinces are more likely to experience terrorism.
Distance to border and distance to capital are inversely related to terrorism. Put differ-
ently, territories that are more urban, less mountainous, and in proximity to the cap-
ital and national borders are more likely to experience terrorist incidents. On the other
hand, only some socioeconomic variables provide support for existing explanations.
Contrary to the expectation, infant mortality, a proxy for poverty, returned statistically
significant coefficients with a negative impact on terrorism in Central Asia (i.e., poorer
territories suffer fewer terrorist incidents). However, higher levels of employment con-
tribute to lesser terrorist violence, as well as the volume of cargo turnover (a proxy for
transportation infrastructure, but also signifying the level of economic development).
Interpreted this way, regions with higher levels of economic development are less likely
to experience terrorism, a finding consistent with the outcomes on the “poverty” var-
iable. Other socioeconomic factors, such as crime, population density, and gender,
however, were not statistically significant predictors of terrorist violence. Finally, both
models found that higher volumes of opioid seizures are associated with a higher level of
terrorism, a relationship we explore in the next section.
Existing explanations provide important insights into the factors that promote
increased incidence of terrorist attacks in Central Asia, but the results are uneven.
Common causes such as individuals’ social isolation, groups’ organizational resources,
and certain topographical and socioeconomic conditions certainly apply to the region.
416    Lawrence P. Markowitz and Mariya Y. Omelicheva

Table 20.1: Socioeconomic and Topographic Determinants of Terrorism in Eurasia

Central Asia

Model 1 Model 2

Opioid Seizures 0.012** (0.006) 0.0018*** (0.00045)

Infant Mortality -​0.23** (0.09)
Employment -​29.79*** (10.19)
Cargo Turnover by Automobile Transport -​0.002** (0.001)
Education (Students) 0.042 (0.04)
Population density 1.21 (3.39)
Sex Ratio -​0.126 (0.132)
Crime Rate -​0.00006 (0.001)
Urban Coverage 0.21* (0.282)
Distance to Border -​0.0007*** (0.002)
Distance to Capital -​0.009*** (0.003)
Mountain Coverage -​0.067*** (0.024)
Constant 16.04 (14.6) 0.282 (0.757)
N 239 36
Wald Chi2 23.06 400.08
Prob >chi2 0.0033 (0.00)
Pseudo R-​sq 0.06 0.55

***p<0.01
**p<0.05
*p<0.01

However, others, such as the moderating effect of civil society or exposure to crime, do
not. These arguments, moreover, fall short of explaining the surprisingly low levels of
terrorist violence more broadly across the region. To address this, we turn to the state.

How Illicit Economies Reduce

Terrorist Violence

Uniquely situated at the interstices of security, politics, and the market, many
infrastructurally weak states are often deeply immersed in illicit economic
relationships. Whether it is human trafficking, arms trafficking, the illegal wildlife trade,
or counterfeiting or money laundering, the proceeds of illicit trade are highly lucrative
 Illicit Economies and Political Violence in Central Asia    417

and often linked to organized crime. In states such as those in Central Asia, the same
political and security apparatuses that are mandated to combat these economies are
instead often drawn into them, forging complex ties with a range of nonstate violent
actors—​including insurgents, erstwhile regime opponents, local criminal bosses, and
transnational groups. These relationships erode state capacity, but they also can have the
surprising effect of diminishing terrorist activity by redirecting these actors into other
spheres and forms of violence (such as center-​periphery conflicts, disputes over territo-
rial control, or criminal violence).
This dynamic varies, however, depending on the state’s involvement in illicit trade,
and specifically on the nature of the revenue bargain struck between the state and
leaders in the illicit economy (Easter 2012). Such bargains are critical in Central Asia, a
region through which most of the world’s opiates transit from Afghanistan to the global
market. Drawing on thirty-​five in-​depth expert interviews conducted in Central Asia
in 2016 and 2017, we identify four state-​trafficker revenue bargains that direct how the
state addresses nonstate organized violence generally (and threats of terrorism in par-
ticular). These were individual, semi-​structured interviews with a wide range of security
experts—​including government officials, members of civil society, independent secu-
rity analysts, journalists, and local scholars—​selected for their specialized knowledge
of these topics. While not used here, an additional sixty expert interviews conducted
in Georgia (in 2017), Armenia (in 2018), and Russia’s North Caucasus (in 2017) provide
comparative perspective.
First, a state can play an integrative role when its security services’ deep involvement
in drug trafficking supports and incentivizes the rapid buildup of coercive capabilities
and absorbs potential terrorist actors into this complex illicit political economy. In states
with predatory capacity, administrative and security institutions become corrupt to the
point of their active participation in and facilitation of organized crime, rather than
simply allowing it to occur. In short, the government’s preoccupation with rent-​seeking,
including taxing the drug trade and amassing personal wealth through other forms
of corruption and crime, turns the rest of society into prey. Under these conditions,
state involvement in drug trafficking can dampen levels of terrorism in several ways.
For one, corrupt security institutions can provide protection for organized crime and
trafficking that enables the proliferation of illegal activities but keeps them “in-​house”
and limits potential militants’ access to funding or criminal support. Also, state control
over drug trafficking spheres can be used as an instrument of co-​optation and control of
local elites, thereby limiting unruly elites from allying with or aiding emergent militant
groups. Finally, the state’s integrative role shifts points of conflict between state and non-
state actors to center-​periphery disputes linked to controlling territory, illicit trafficking
routes, and access to markets. For all these reasons, an integrative role of the state tends
to absorb potential militant activity into larger pyramids of drug trafficking under na-
tional and regional elites, thereby diminishing levels of terrorist violence.
Second, a state can play a mitigating role when its limited involvement in the drug
trade undermines its institutional coherence and coercive capacity, but it retains suf-
ficient capabilities to mitigate terrorism while permitting criminal violence to grow.
418    Lawrence P. Markowitz and Mariya Y. Omelicheva

A state’s weakened security capabilities encourages criminal organizations to pursue

lucrative sources of revenue in illicit economic spheres (such as drug trafficking), but
this does not spill over into heightened levels of terrorism. Far from spawning breeding
grounds for extremist recruitment, the appropriation of criminal activities by extremist
groups, or the shift to political violence by criminal groups, is rare and sporadic. In
fact, while criminals may become ideologically radicalized, the opportunities afforded
by a fragmented criminal economy (including the drug trade) with weakened state
capabilities diverts potential recruits into militant groups. Moreover, without guidance
from leaders, criminal organizations’ rank-​and-​file members do not peel off and attach
themselves to insurgent militants, even if led by influential local religious leaders. While
pockets of autonomy from the state open the door to potential terrorists’ violent opposi-
tion to the state or their vulnerability to extremist groups, we expect that advantages will
tilt toward drug traffickers. Criminal violence, not terrorist violence, tends to predomi-
nate in such contexts.
Third, a state can play a hegemonic role when the removal of its security agencies from
the region’s drug trafficking economy enables the independent growth of its coercive
capabilities and the broad use of repression to substantially reduce levels of terrorism.
While the high demand for illicit goods and services may fuel organized crime, coun-
tries with robust law-​enforcement and intelligence institutions and the political resolve
to fight crime will experience less drug trafficking. Comprehensive counterterrorism
policies will have greater effect in denying potential militant groups access to sources
of funding and will harden the targets that are vulnerable to terrorist attack. Low-​level
corruption may cause a breach in the state’s security capability to prevent and respond to
trafficking operations, but it hardly constitutes a threat to state security. Consequently,
the state’s ability to keep trafficking and terrorist networks separate will enable it to pre-
serve a monopoly of the legitimate use of violence and diminish the frequency and in-
tensity of terrorist attacks.
Finally, we introduce an exception—​an outcome in which state failure and civil
conflict provide the context in which state offices’ unrestrained involvement in drug
trafficking open the door to a fleeting convergence of traffickers and militants and
briefly increase levels of terrorist violence. In such states that are characterized by ex-
tremely weak—​or even failing—​administrative, judicial, law-​enforcement, and security
institutions, organized criminal actors can gain control over state institutions. While
wholesale state capture (termed “narcostatization,” when drug trafficking penetrates
state institutions) is extremely rare, it is more common for politicians to cede con-
trol over states and set up informal power-​sharing deals with the leadership of organ-
ized crime. Under these conditions, state involvement in drug trafficking can lead to
increased terrorism in several ways. State officials become overtaken and subsumed by
organized crime and trafficking groups, which enables the explosive proliferation of a
wide range of terrorism and drug trafficking activities that morph into a single web of ac-
tivity. Additionally, the state’s loss of control over local elites leads those elites to become
immersed in the fully intersecting spheres of terrorism and trafficking. And the state’s
 Illicit Economies and Political Violence in Central Asia    419

loss of a monopoly of violence entails the fragmentation of its state security agencies,
many of which separate from the state and join terrorist and criminal organizations. In
this context of a breakdown of state authority, the convergence of terrorist and criminal
activity, and its fusion with coercive agents of the state, will escalate terrorist violence.
We explore the empirical implications of these revenue bargains in our cases below.

Illicit Economies and Low Levels

of Terrorism in Central Asia

Since the late 2000s, Tajikistan has exemplified an integrative role of the state, in which
terrorist violence is absorbed into the country’s illicit political economy. Following civil
war in the 1990s, Tajikistan’s central government struggled to reestablish control over
many parts of the country. One regime strategy for stabilizing postwar Tajikistan was to
cede control over key institutions (including parts of the security apparatus) to former
commanders and prominent politicians and allow them to establish ties to organized
criminal groups (Heathershaw 2009; Markowitz 2013; Driscoll 2015). One of the most
significant features of Tajikistan’s post-​conflict state-​building has been the regime’s rev-
enue bargain with traffickers, which has enabled it to consolidate and exploit its control
over the drug trade. The drug trade in Tajikistan in the 1990s was somewhat fractured,
constituted by several competing medium-​sized groups that were residual formations
of warlord militias from the civil war period. The heads of drug trafficking operations
were also former commanders of militias during the civil war (Paoli et al. 2007). As part
of the post-​conflict power-​sharing and reintegration process, these commanders were
appointed to senior positions in government, which enabled them to conduct the drug
trade from within state structures. While these political elites used their positions to in-
fluence counter-​trafficking efforts, eliminate rivals, and centralize the drug trade, this
also enabled President Emomali Rahmon to concentrate the central leadership’s con-
trol over the drug trade by replacing senior officials with persons beholden to him (De
Danieli 2014).
As a consequence, the drug trade was effectively centralized under the control of
the ruling elite (with ties extending into the presidential administration). By the early
2000s, it was widely believed that drug trafficking was supported and protected by a
range of officials, including border officers, customs officials, and those in the Drug
Control Agency and Ministry of Internal Affairs. As one security official noted, there
are many “hidden hands” aiding the drug trade that serve as intermediaries, enabling
former warlords and others in this illegal activity. It is generally believed that politi-
cally connected traffickers are protected, and in return smaller dealers and sellers are
turned over to keep up an image of regular seizures and arrests. Thus, larger organized
syndicates remained untouched, allowing the central government to continue to benefit
from large profits.
420    Lawrence P. Markowitz and Mariya Y. Omelicheva

The central government’s use of co-​optation and coercion against former opposi-
tion politicians and senior officials (nationally and regionally) has led to several center-​
periphery conflicts. Many who might have moved toward insurgent or terrorist groups
were drawn into these arrangements through their wartime ties to local and regional
leaders. In eastern Tajikistan (in the Rasht Valley and Gorno-​Badakhshan Autonomous
Oblast), many local officials were believed to be heavily involved in the drug trade, and
collaborations between drug traffickers, criminal groups, and militants evolved under
the protection and patronage of those figures. As these were dismissed, or as they pre-
emptively broke from the regime, they mobilized networks of supporters (often former
militants). These conflicts predominated in Tajikistan, absorbing most of the nonstate
militant activity and minimizing the frequency and intensity of religiously inspired ter-
rorist violence. As a result, very few terrorist incidents have emerged in Tajikistan over
the past two decades, and most of these have been disconnected, small-​scale attacks.
Two attacks (in 2008 and 2009) in northern Tajikistan, attributed to the IMU, were in-
dividual shootings of police officers, while a third (in 2010), claimed by a previously
unknown group (Jamaat-​e Ansurallah), involved a car bombing of a regional police
headquarters.
The case of Kyrgyzstan exemplifies a mitigating role of the state, in which relatively
weak government influence over illicit economic activity partially reduces threats of ter-
rorist violence while allowing a rise in criminality. Marked by intra-​elite divisions, state
paralysis, and twice overwhelmed by elite-​led protests (2005 and 2010), Kyrgyzstan has
emerged as an intersection of organized criminal organizations and drug trafficking.
Having initiated political liberalization in the early 1990s, Kyrgyzstan’s national and re-
gional elite remained influential actors, particularly in parliament, where they were in
frequent contestation with President Askar Akaev. Moreover, elites’ frequent rotations
in and out of political positions have undermined the state’s overall capacity during
much of the 1990s (McGlinchey 2011).
Seeking to implement political and economic reform at the time, Kyrgyzstan’s govern-
ment struck a bargain over illicit revenues with traffickers that kept the latter excluded
from positions within the central government apparatus. However, the trafficking of
opium and heroin through southern Kyrgyzstan has been fragmented and difficult to con-
trol since the early 1990s, when much of it was divided between Uzbek and Tajik criminal
groups. During this period, a number of criminal groups who were initially working under
“ruling family representatives” linked to President Akaev became more independent of
their patrons—​a development that further fragmented the drug trade. Even though “there
was open approval for drug trafficking” at the highest levels of the presidential admin-
istration under Akaev’s successor, the state’s revenue bargain with traffickers continued
to exclude them from key posts in government. Between 2005 and 2010 there were re-
portedly thirty-​one different criminal groups (relatively small, between five and fifteen
members) in the country, many of which were operating under the patronage and pro-
tection of their own regional and local elites. Many of these criminal groups were able to
operate without seeking political protection from weakened law enforcement authorities
(UNODC 2007). While Tajikistan is deemed to be closer to a “narco-​state,” in which large
 Illicit Economies and Political Violence in Central Asia    421

portions of the state apparatus are involved in the drug trade, there are only “key persons”
within Kyrgyzstan’s state—​mostly within its law enforcement and security agencies—​that
are believed to provide protection over disparate parts of this economy.
Despite their institutional weaknesses, Kyrgyzstan’s state offices remained free of traf-
ficker influence, and criminal groups did not acquire influence over coercive institutions
of state. Although some individual political elites provided patronage and protection for
criminal activity in exchange for payments, criminal organizations—​including those
with operational linkages to extremists—​could not access coercive resources (arms,
ammunition, personnel) as they could in Tajikistan. At the same time, the space for
criminal organizations in the informal economy grew, pulling in those who might have
gravitated to insurgent or terrorist activities. This has limited nonstate violence in the
country to a relatively few, small-​scale attacks. Despite the many arrests related to ter-
rorism and religious extremism in Kyrgyzstan, especially in the south of the country
(Osh, Batken, and Jalal-​Abad regions), much of the violence has been low-​intensity and
involved criminal groups (settling scores or competing for resources with their rivals),
politicians, or individuals spanning the two spheres.
There have been several small incidents alleged to have been carried out by extremist
groups in Kyrgyzstan. However, most of these were either perpetrated by groups coming
from outside the country, limited to individual attacks (such as political assassinations),
or larger attacks inaccurately attributed to extremist groups. In 1999 and 2000, two sep-
arate incidents of hostage-​taking in rural areas of the country were carried out by IMU
incursions from Tajikistan. Approximately two or three bombings of police stations
and other government buildings were carried out each year over the 2000s, though in
most cases these were not claimed by any extremist groups. In 2010 an explosion took
place outside the Bishkek Sports Palace, which was the venue for a trial of those ac-
cused of ordering and carrying out violence during the April 2010 anti-​Bakiev protests.
While the government attributed the attack (and another attack on a police station)
to a newly created group, Jaysh al-​Mahdi (Army of the Redeemer), ultimately none of
those arrested were convicted, suggesting political motives rather than religious ex-
tremism (Zenn and Kuehnast 2014). Lastly, the Eastern Turkestan Islamic Movement
was believed to have carried out a bombing at the front gate of the Chinese embassy
in Bishkek in August 2016, although the group has not claimed responsibility (Roberts
2012, 20–​21). In short, there have been only a few, scattered, small-​scale terrorist attacks.
Instead, the majority of violent incidents in Kyrgyzstan have involved attacks involving
organized criminal organizations, politicians (often linked to criminal activities), and
law enforcement officials/​offices.
Finally, the case of Uzbekistan exemplifies the hegemonic role of the state in reducing
terrorist violence. Uzbekistan embarked on a program of state-​building in the 1990s
designed to rein in regional power centers and construct a large and robust coercive
apparatus. While its immersion in other economic sectors fostered patterns of coer-
cive rent-​seeking (Markowitz 2013), the regime removed its newly empowered security
apparatus from the drug trade and applied extensive repression against any organized
actors—​criminal or extremist—​operating independent of the state.
422    Lawrence P. Markowitz and Mariya Y. Omelicheva

Uzbekistan’s presidential administration, confronting disturbances in the late 1980s,

invested early on in strengthening its security apparatus, devoting resources to its
National Security Service (SNB), police, and the military. Initially, the central govern-
ment had determined that it needed a well-​trained and well-​equipped security appa-
ratus capable of exercising public control. Over time, however, the leadership under
President Islam Karimov utilized this security apparatus to extend its reach into the dif-
ferent regions and check the growing power of regional elites who controlled important
industries, ran the agricultural sector, and delivered public services to the broader pop-
ulation. In addition, post-​independence Uzbekistan inherited an ambition for regional
dominance and used its security and military forces to carry out limited interventions
in the sovereign affairs of its neighbors (Omelicheva 2011). Consequently, Uzbekistan
embarked on a state-​building trajectory that centered on political control and social sta-
bility, underpinned by powerful security services.
At the same time, extensive state intervention in the domestic economy drew on
these coercive offices to retain government control. Concessions granted to police,
prosecutors, the SNB, and the tax inspectorate (among others)—​in exchange for their
loyalty to Karimov—​gave them license to exploit rent-​seeking opportunities in these
local economies and commodity exports. Using (formal and informal) taxes, customs,
and border controls, law enforcement and security offices became deeply involved in the
country’s lucrative agricultural, industrial, and resource extractive sectors. The govern-
ment of Uzbekistan, therefore, was able to strike a revenue bargain that minimized the
security apparatus’ involvement in drug trafficking, compartmentalized rents from the
drug trade, and forged hegemonic capacities that separated and effectively combated
trafficking and terrorism.
As a result, there have been some terrorist activities in Uzbekistan, but these have
remained largely independent, sporadic attacks. On February 16, 1999, six bombs
exploded in the center of Tashkent city, targeting several government buildings, but
there is no evidence of these attacks being linked to a terrorist-​trafficking connec-
tion. In 2004 there was a short wave of violence, starting with a series of explosions
and shootouts over three days in late March and early April, and followed by three
suicide bombings outside the US and Israeli embassies and the national prosecutor-​
generals’ office in June. While these attacks were attributed to Hizb ut-​Tahrir and the
IMU, these attacks were perpetrated by ordinary citizens organized into local semi-​
autonomous cells of resistance to the country’s increasingly authoritarian regime. The
attackers lacked coordination and training, used primitive explosive devices (some of
which malfunctioned) made from materials widely available in Uzbekistan, and mostly
targeted government and security officials (likely in retaliation for abuses by police and
security services). A third event occurred in May 2009, when a suicide bomber killed a
police officer, and, the same day, several assailants fired rocket-​propelled grenades and
small arms at a police station. Both attacks were later claimed by Islamic Jihad Group (a
group with alleged ties to the al-​Qaʿida hierarchy), though there is little evidence that
transnational groups had any substantial presence in the heavily monitored Andijan
Province where the attacks took place.
 Illicit Economies and Political Violence in Central Asia    423

Overall, Uzbekistan has experienced little conflict between its political elites (as in
Tajikistan) or nonstate criminal violence (as in Kyrgyzstan). Instead, Uzbekistan’s tra-
jectory, and the separate coexistence of trafficking and terrorism within its borders,
has led to sporadic terrorist violence and enabled state repression to predominate
far more than any other forms of nonstate violence. Indeed, repression is consist-
ently perpetrated by the regime for political purposes against independent opposi-
tion figures, civil society groups, and occasional local protests. This reached a peak
on May 13–​14, 2005, when Uzbekistan’s state security forces suppressed an uprising in
Andijan city, resulting in the deaths of hundreds—​possibly seven hundred—​civilians
(including many women and children). Its use of repression against society notwith-
standing, Uzbekistan’s authorities have remained sufficiently strong to suppress most
of activities of the IMU and other potential extremists, and stamp out other forms
of dissent. The government’s counterterrorism and anti-​drug efforts, in short, have
succeeded in removing its security apparatus from the drug trade and keeping levels
of terrorism low.
The sole exception to these patterns of low terrorist violence in Central Asia is the
case of Tajikistan during the 1990s. As this case shows, when the state immerses itself in
the drug trade, it generates higher rates of nonstate violence and enables—​for relatively
short periods of time—​significantly increased terrorist violence as well. This exception
is important to note, since it demonstrates that low levels of terrorism in Central Asia
overall is not the result of broad-​based, region-​wide developments. Cultural explana-
tory accounts of Islam in Central Asia that suggest it is less amenable to radicalization or
path-​dependent arguments contending that Soviet legacies had diminished the capacity
of Islamist mobilization, therefore, provide inadequate accounts of this variation within
the region.
In the wake of state failure in 1992, Tajikistan’s five-​year civil war greatly fragmented
its security apparatus, destroyed many foundations of its economy, and left its popula-
tion traumatized by a conflict that had led to 50,000 deaths and over a million people
displaced. For much of the mid-​1990s, its postwar government remained internally
divided, with the central government’s ministries (as well as Tajikistan’s provinces)
run as the personal fiefdoms of powerful political figures who operated independent
of a figurehead president. The ministerial appointees of the central government often
competed with the local formal and informal authorities for control of the regions.
Under these conditions, these figures ran Tajikistan’s provinces yet operated in a vortex
of organized criminal activity, drug trafficking, and corruption that occasionally inte-
grated members of militant groups with ties to terrorist organizations.
As the fledgling Rahmon government appointed former commanders from pro-​
government and opposition forces to establish its nominal control over the rest of
the country, many of these leaders had already become involved in the drug trade.
According to a witness and a former member of the United Tajik Opposition (UTO), a
fatwa issued by Islamist religious leadership permitted the fighters to smoke hemp and
partake in the trafficking of drugs. Subsequently, the majority of field commanders
and fighters engaged in the drug trade with Afghan smugglers (Izzatullo 2017).
424    Lawrence P. Markowitz and Mariya Y. Omelicheva

Foremost among them was Yaqub Salimov, Tajikistan’s first postwar minister of the
interior. As head of the national police force, Salimov allegedly controlled much of
the drug trade in the early 1990s (reportedly utilizing trains and military airplanes to
transport it to Russia). Another pro-​government commander, Gaffor Mirzoev, was
believed to run an extensive drug-​trafficking operation while commanding Tajikistan’s
elite Presidential Guard (and later as the head of the Drug Control Agency). Some
observers speculate that opposition commanders, who had neither Uzbekistani nor
Russian support, had turned to drug trafficking to finance their war efforts. During
and after the war, Mirzo Ziyoev, military chief of the UTO, was believed to be deeply
engaged in heroin smuggling. Likewise, an opposition leader during the war, Habib
Sanginov, who had risen to be deputy minister of the interior in the postwar govern-
ment, was reportedly killed in a drug operation gone awry. As one in-​depth analysis
has concluded, these examples of central political figures are not merely a few “bad
apples,” but instead represent an emerging “narco-​state” in which high-​ranking gov-
ernment officials used their position to conduct a majority of the country’s drug trade
(Paoli et al. 2007). According to a senior law enforcement official in Dushanbe, these
warlords-​turned-​officials were often directly engaged in the drug trade and other il-
licit activities, in contrast to the regime’s more careful use of intermediaries after the
2000s, such as officials in the military or border control officers. The profits of unre-
strained drug trafficking, moreover, fueled the large patronage networks and patri-
mony of these senior officials, who commanded the support of lower-​level officials,
regional politicians, and local strongmen who controlled key resources in their
jamoats (rural municipalities). In many cases, these patronage networks enabled gov-
ernment officials to straddle licit and illicit economic spheres.
The convergence of traffickers and militants in Tajikistan had several features. While
the state’s pervasive involvement in the drug trade was the key driver for this nexus, the
convergence, although widespread, was relatively short-​lived, occurring only at certain
phases during the 1990s. These groups, however, led to increased terrorist violence. In
Tajikistan, 44 terrorist incidents registered by the GTD took place between June 1997
(when the peace agreement was signed) and the end of 2001. More than 70 people were
killed and over 130 were injured in these terrorist incidents. By contrast, the following
decade (2002–​2016) saw only 14 terrorist attacks that took lives of 22 people.
The most prominent group, the Islamic Movement of Uzbekistan, operated in
Tajikistan in the late 1990s and early 2000s as militant and criminal actors that had
the capacity to subordinate or take advantage of a fragmented state apparatus to their
ends. After belatedly and reluctantly acceding to the 1997 peace agreement that ended
Tajikistan’s civil war, the IMU spent the following five years oscillating between bases of
operations in northern Afghanistan and Tajikistan, strengthened by the Uzbek recruits.
During this period, the military formations that preceded the IMU continued their in-
volvement in the drug trade, and the organization grew to become extensively involved
in the opiate transport business under the leadership of Juma Namangani (Fredholm
2003; Madi 2004). By leveraging its involvement in the drug trade, and the strategic
 Illicit Economies and Political Violence in Central Asia    425

location of its camps along trafficking routes, the IMU established ties to both the
Taliban in Afghanistan and key traffickers Tajikistan (with some, like Ziyoev, holding
key positions in the Tajik government).
In short, the IMU was well situated to take advantage of its cross-​border contacts with
Taliban suppliers, its camps and operations in the same mountain passes used to trans-
port opiates, and its good relations with former UTO comrades-​in-​arms (who were
now commanding border posts, police units, and regional governments) (Levi-​Sanchez
2017). While these relationships interweaving criminal and terrorist networks were rela-
tively short-​lived, they proved useful in organizing the IMU’s 1999 and 2000 incursions
into Kyrgyzstan. Under Namangani, this faction of the IMU had become habituated
to criminal activities and reportedly began pursuing drug trafficking as an end in it-
self, even as it continued its role as one of the region’s transnational terrorist organiza-
tions. At the same time, the IMU was at the center of one of the most sustained episodes
of terrorist violence in Central Asia, having participated in Tajikistan’s civil conflict
(from 1992 to 1997), armed incursions into Kyrgyzstan (1999 and 2000) and Uzbekistan
(2000), and several suicide bombings in Uzbekistan in 2004 (which were carried out
by the IMU’s splinter group, Islamic Jihad Union). Ultimately, however, the IMU was
driven out of Central Asia as the state reasserted its monopoly of violence in postwar
Tajikistan and exerted increasing control over Tajikistan’s drug trafficking and organ-
ized crime (effectively absorbing these activities into the criminal spheres of key govern-
ment and security elites.

Areas of Future Research

Three areas of future research arise from this survey of Central Asia. First, some of the
causal factors associated with varying levels of political violence (and of terrorism in
particular) in the literature are borne out in the Central Asia context. Yet more empirical
work needs to be done to accumulate reliable findings on the political, economic, and
societal factors that explain patterns of terrorist violence in this and other understudied
Muslim-​majority regions. Second, our inquiry suggests that the lack of terrorist vio-
lence in the region as a whole may be overly attributed to the use of state repression,
which is not surprising given the political and financial benefits that flow from this per-
ception. Future research also needs to examine the political economy of security and
its implications for understanding political violence, especially in low-​and middle-​in-
come countries where states lack an effective institutional presence. Third, it suggests
that keeping terrorist violence at low levels has come at the cost of facilitating higher
levels of other forms of violent political behavior (such as enhanced state violence or
rising criminal violence). While much attention is paid to the link between Islam and
terrorism, more studies of the relationship between terrorism and other forms of vio-
lence will better situate this in a broader political context.
426    Lawrence P. Markowitz and Mariya Y. Omelicheva

Acknowledgments
Research conducted for this chapter was supported by research grant ONR N00014-​15-​1-​2788
“Trafficking/​Terrorism Nexus in Eurasia” under the auspices of the Office of Naval Research.

References
Asal, Victor, Milward H. Brinton, and Eric W. Schoon. 2015. “When Terrorists Go
Bad: Analyzing Terrorist Organizations’ Involvement in Drug Smuggling.” International
Studies Quarterly 59, no. 1: 112–​123.
Asal, Victor, and Karl Rethemeyer. 2008. “The Nature of the Beast: Organizational Structures
and the Lethality of Terrorist Attacks.” Journal of Politics 70, no. 2: 437–​449.
Balcells, Laia. 2017. Rivalry and Revenge: The Politics of Violence during Civil War.
Cambridge: Cambridge University Press.
Blyth, Simon, Brian Groombridge, Igor Lysenko, Lera Miles, and Adrian Newton. 2002.
Mountain Watch: Environmental Change & Sustainable Development in Mountains. Series
12. Cambridge, MA: World Conservation Press.
Buzan, Barry, Ole Wæver, and Jaap de Wilde. 1998. Security: A New Framework for Analysis.
Boulder, CO: Lynne Rienner.
Chenoweth, Erica, and Pauline L. Moore. 2018. The Politics of Terror. Oxford: Oxford
University Press.
Clarke, Colin P. 2019. After the Caliphate: The Islamic State and the Future Terrorist Diaspora.
Cambridge: Polity Press.
Cordesman, Anthony H. 2017. “Islam and the Patterns in Terrorism and Violent Extremism.”
Center for Strategic & International Studies Working Paper. Washington, DC: Center
for Strategic & International Studies, October 17. https://​www.csis.org/​analysis/​
islam-​and-​patterns-​terrorism-​and-​violent-​extremism.
Crenshaw, Martha. 2011. Explaining Terrorism: Causes, Processes, Consequences. London and
New York: Routledge.
De Danieli, Filippo. 2014. “Beyond the Drug-​Terror Nexus: Drug Trafficking and State-​Crime
Relations in Central Asia.” International Journal of Drug Policy 25, no. 6: 1235–​1240.
Driscoll, Jesse. 2015. Warlords and Coalition Politics in Post-​ Soviet States.
Cambridge: Cambridge University Press.
Easter, Gerald M. 2012. Capital, Coercion and Post-​Communist States. Ithaca, NY: Cornell
University Press.
Ehrilich, Paul R. 2002. “Some Roots of Terrorism.” Population and Environment 24: 183–​192.
Epkenhans, Tim. 2018. “Oblivion, Ambivalence and Historical Erasure: Remembering the Civil
War in Tajikistan.” In Tajikistan on the Move: Statebuilding and Societal Transformations, ed-
ited by Marlene Laruelle, 195–​222. Lanham, MD: Lexington Books.
Fredholm, Michael. 2003. Uzbekistan and the Threat from Islamic Extremism. Conflict Studies
Research Centre Report K39. Sandhurst, UK: Royal Military Academy, 2003.
Frevtag, Andreas, Jens J. Kruger, Daniel Meierrieks, and Friedrich Schneider. 2011. “The
Origins of Terrorism: Cross-​ Country Estimates of Socio-​ Economic Determinants of
Terrorism.” European Journal of Political Economy 27 (Supplement 1): S5–​S16.
Gaibulloev, Khusrav, and Todd Sandler. 2013. “Determinants of the Demise of Terrorist
Organizations.” Southern Economic Journal 79, no. 4: 774–​792.
 Illicit Economies and Political Violence in Central Asia    427

Heathershaw, John 2009. Post-​ Conflict Tajikistan: The Politics of Peacebuilding and the
Emergence of Legitimate Order. London: Routledge.
Hegghamer, Thomas. 2013. “Should I Stay or Should I Go? Explaining Variation in Western
Jihadists’ Choice between Domestic and Foreign Fighting.” American Political Science
Review 107, no. 1: 1–​15.
Hoffman, Bruce. 2006. Inside Terrorism. New York: Columbia University Press.
Horowitz, Michael C., and Philip Potter. 2014. “Allying to Kill: Terrorist Intergroup Cooperation
and the Consequences for Lethality.” Journal of Conflict Resolution 58, no. 2: 199–​225.
Howard, Marc Morje. 2003. The Weakness of Civil Society in Post-​ Communist Europe.
Cambridge: Cambridge University Press.
Hyslop, D., and T. Morgan. 2014. “Measuring Terrorism with the Global Terrorism Index.”
In Understanding Terrorism: A Socio-​Economic Perspective, edited by R. Caruso and A.
Locatelli, 97–​114. Bingley: Emerald Publishing.
Izzatullo, Sadullom. 2017. “IRPT and ISIL: The Practical Evidence of the Same Teachings,
Activities, and Aims of These Two Extremist Organizations.” Embassy of the Republic of
Tajikistan in Malaysia, November 16. http://​tajemb-​my.org/​en/​embassy/​addresses-​2/​45-​
irpt-​and-​isil-​the-​practical-​evidence-​of-​the-​same-​teachings-​activities-​and-​aims-​of-​these-​
two-​extremist-​organizations.
Juergensmeyer, Mark. 2000. Terror in the Mind of God: The Global Rise of Religious Violence.
Berkeley: University of California Press, 2000.
Kalyvas, Stathis. 2006. The Logic of Violence in Civil War. Cambridge: Cambridge
University Press.
Khamidov, Alisher. 2017. “When Religion Resorts to Violence: Explaining the Spatial Variation
in Religious-​Based Mobilization in Kyrgyzstan.” In Islam, Society and Politics in Central Asia,
edited by Pauline Jones, 199–​220. Pittsburgh: University of Pittsburgh Press.
Kilberg, Joshua. 2012. “A Basic Model of Explaining Terrorist Group Organizational Structure.”
Studies in Conflict and Terrorism 35, no. 11: 810–​830.
Krueger, Alan. 2008. What Makes a Terrorist: Economics and the Roots of Terrorism. Princeton,
NJ: Princeton University Press.
Levi-​Sanchez, Suzanne. 2017. The Afghan-​Central Asia Borderlands: The State and Local
Leaders. New York: Routledge.
Madi, Maral. 2004. “Drug Trade in Kyrgyzstan, Structure, Implications and Countermeasures.”
Central Asian Survey 23, no. 3–​4: 249–​273.
Mahoney, James, and Gary Goertz, 2004. “The Possibility Principle: Choosing Negative Cases
in Comparative Research.” American Political Science Review 98, no. 4: 653–​669.
Markowitz, Lawrence P. 2013. State Erosion: Unlootable Resources and Unruly Elites in Central
Asia. Ithaca, NY: Cornell University Press.
McGlinchey, Eric M. 2011. Chaos, Violence, Dynasty: Politics and Islam in Central Asia.
Pittsburgh: University of Pittsburgh Press.
Meiyappan, Prasanth, and Atul K. Jain. 2012. “Three Distinct Global Estimates of Historical
Land-​cover Change and Land-​use Conversions for over 200 Years.” Frontiers of Earth
Science 6, no. 2: 122–​139.
Nasritdinov, Emil, Zarina Urmanbetova, Kanatbek Murzakhalilov, and Mametbek Myrzabaev.
2019. Vulnerability and Resilience of Young People in Kyrgyzstan to Radicalization, Violence
and Extremism: Analysis across Five Domains. CAP paper no. 213. Washington, DC: Central
Asia Program, January 2019. https://​centralasiaprogram.org/​archives/​12825.
428    Lawrence P. Markowitz and Mariya Y. Omelicheva

Omelicheva, Mariya Y. 2011. Counterterrorism Policies in Central Asia. New York and
London: Routledge.
Omeicheva, Mariya Y., and Lawrence P. Markowitz. 2019. Webs of Corruption: Trafficking and
Terrorism in Central Asia. New York: Columbia University Press.
Paoli, Letizia, Irina Rabkov, Victoria A. Greenfield, and Peter Reuter. 2007. “Tajikistan: The
Rise of a Narco-​State.” Journal of Drug Issues 37, no. 4: 969–​971.
Pape, Robert A. 2005. Dying to Win: The Strategic Logic of Suicide Terrorism.
New York: Random House.
Phillips, Brian. 2014. “Terrorist Group Cooperation and Longevity.” International Studies
Quarterly 58, no. 2: 336–​347.
Piazza, James A. 2009. “Does Poverty Serve as a Root Cause of Terrorism?: No, Poverty Is a
Weak Causal Link.” In Debating Terrorism and Counterterrorism: Conflicting Perspectives
on Causes, Contexts and Responses, edited by Stuart Gottlieb, 35–​ 51. Washington,
DC: Congressional Quarterly Press.
Reinares, Fernando. 2004. “Who Are the Terrorists? Analyzing Changes in Sociological Profile
among Members of ETA.” Studies in Conflict and Terrorism 27, no. 6: 465–​488.
Roberts, Sean R. 2012. “Imaginary Terrorism: The Global War on Terror and the Narrative of
the Uyghur Terrorist Threat.” PONARS Eurasia Working Paper. Washington, DC: PONARS
Eurasia, March 2012. https://​www.ponarseurasia.org/​sites/​default/​files/​Roberts_​
WorkingPaper_​March2012.pdf.
Schwedler, Jillian. 2007. Faith in Moderation: Islamist Parties in Jordan and Yemen.
New York: Cambridge University Press.
Search for Common Ground. 2016. Radicalization of the Population in Osh, Jalal-​Abad and
Batken Oblasts: Factors, Types, and Risk Groups. Internal report. Washington, DC: Search for
Common Ground.
Stern, Jessica. 2004. Terror in the Mind of God: Why Religious Militants Kill.
New York: HarperCollins.
Tucker, Noah 2015. Central Asian Involvement in the Conflict in Syria and Iraq: Drivers and
Responses. United States Agency for International Development Report, May 4. Washington,
DC: ASAID.
UNODC (United Nations Office on Drugs and Crime). 2007. An Assessment of Organized
Crime in Central Asia. Vienna: UNODC.
Varshney, Ashutosh. 2002. Ethnic Conflict and Civic Life: Hindus and Muslims in India. New
Haven, CT: Yale University Press.
Weidmann, Nils B., Doreen Kuse, and Kristian Skrede Gleditsch. 2010. “The Geography of the
International System: The CShapes Dataset.” International Interactions 36, no. 1: 86–​106.
Zenn, Jacob, and Kathleen Kuehnast. 2014. Preventing Violent Extremism in Kyrgyzstan. USIP
Special Report 355. Washington, DC: United States Institute of Peace, October 2014. https://​
www.usip.org/​sites/​default/​files/​SR355_​Preventing-​Violent-​Extremism-​in-​Kyrgyzstan.pdf.
Zhirkov, Kirill, Maykel Verkuyten, and Jereon Weesie. 2014. “Perceptions of World Politics
and Support for Terrorism among Muslims: Evidence from Muslim Countries and Western
Europe.” Conflict Management and Peace Science 31, no. 5: 481–​501.
 Chapter 21

Piet y, Devoti on, a nd

Supp ort for Sha ri ʿa
Examining the Link between Religiosity and Political
Attitudes in Pakistan

Niloufer A. Siddiqui

As scholarly and policy attention has shifted to global Islam and its various
manifestations in recent years, there has been a renewed focus on the link between
individual levels of religiosity and piety, on the one hand, and political behavior and
attitudes, on the other. A wide range of literature, spanning a number of disciplines, has
sought to assess the impact of Islamic religiosity on a host of dependent variables—​from
voting patterns (Esmer and Pettersson 2007; Barreto and Bozonelos 2009; Kalin and
Siddiqui 2016) and political participation (Sobolewska et al. 2015) to foreign policy, in-
ternational conflict (Tessler and Nachtwey 1998), and support for terrorism and mili-
tancy (Ginges, Hansen, and Norenzayan 2009; Kaltenthaler et al. 2010; Fair, Malhotra,
and Shapiro 2012; Fair 2015). Findings have varied, depending on the particular context
of the study, the specific outcome of interest, and, perhaps most importantly, how religi-
osity has been defined and measured. While religiosity is a difficult concept to measure
in any context—​both because the term encompasses many different dimensions and be-
cause it tends to rely on self-​reports that may be distorted by social desirability bias—​
Islam poses particular analytical challenges because, as Cammett and Jones Luong put
it, “there is a tendency to conflate both religiosity with the rise of Islamism . . . and the
emergence of Islamists with popular support for Islamism” (2014, p. 10.4).
In this chapter, I present a conceptualization of religiosity that distinguishes between
three distinct but related components: “piety,” measured as participation in religious
ritual or behavior; “devotion,” measured as the self-​professed importance of religion in
one’s life; and “dogmatism,” measured as belief in laws being made in accordance with
430   Niloufer A. Siddiqui

Islam (i.e., support for shariʿa).1 I explore the link between each of these dimensions of
religiosity and various forms of political behavior in Pakistan, including support for
democracy, concern with religious extremism, support for government restrictions on
blasphemous speech, and attitudes toward neighboring non-​Muslim states. Existing
literature on religion and politics in Muslim-​majority countries has theorized a link
between religiosity, broadly defined, and these key political behaviors, but has found
inconsistent and often contradictory results. This chapter shows that much of this lit-
erature suffers from two distinct problems: (1) inconsistent definitions and measures
of religiosity, including a conflation of personal piety or religious practice with sup-
port for shariʿa law and, in some cases, even Islamic political movements and polit-
ical Islam; and (2) conceptual overlap between support for political Islam and various
political outcomes of interest. This chapter uses the case of Pakistan to refine existing
measures of religiosity to improve our analysis of the relationship between religion and
politics. It does so by offering original empirical analysis that demonstrates that re-
ligiosity does matter in shaping opinions in Pakistan, but not in the way that much
of the literature expects. Using data collected among 1,200 Pakistanis by the Pew
Research Center in 2015, the most recent such survey carried out by Pew, this chapter
finds that religiosity is not a predictor of increasing conservativism in all aspects of
politics. To the extent that religiosity matters, it appears to be positively correlated with
pro-​democracy sentiment. Religiosity measured as devotion and belief in Islamic law
is also negatively correlated with the belief that violence can be justified on religious
grounds; however, personally pious individuals are neither more nor less likely than
the less pious to justify religious violence. Similarly, for outcomes such as attitudes to-
ward India, Pakistan’s Hindu-​majority neighbor and historic enemy, and preferences
for government restrictions on blasphemous speech, the precise measure of religiosity
being employed matters in determining the strength and direction of the relationship.
While the more pious and devoted are more likely to support restrictions on blasphe-
mous speech, support for shariʿa is—​surprisingly—​not a predictive factor.
As the second-​largest Muslim-​majority country in the world, and one that was
founded on the basis of religious identity following its violent partition from India,
Pakistan is an important case study through which to explore the relationship between
the various dimensions of religiosity and political attitudes. Over the last two decades,
Pakistan has seen the increasing influence of “piety movements” and faith-​based
institutions that aim to transform their followers—​often middle-​class women—​into
more “pious” and “pure” subjects (Ahmed 2009; Zia 2009). At the same time, hostilities
along religious and sectarian lines have increased in the country (Ispahani 2015; South
Asia Terrorism Portal 2018). Numerous incidents of mass violence motivated by the ar-
chaic blasphemy laws have also resulted in a spate of killings, with angry mobs mobilized
through rumors spread by word of mouth or on social media platforms (Amnesty
International 2016), and sometimes organized by religious parties or movements.2
Despite the demonstrated ability of Islamist political parties to amass considerable street
 Piety, Devotion, and Support for Shariʿa  431

power, however, they have historically fared poorly in national elections, leading some
to believe that there is little support for the mixing of politics and religion in the country.
These overlapping trends have led scholars, journalists, and analysts alike to question
the exact role played by religiosity in society.
The chapter proceeds as follows. First, I provide an overview of literature examining
the link between religiosity and political attitudes in various Islamic settings, focusing
in particular on the ways in which these studies conceptualize and measure religiosity.
I then outline my hypotheses, motivated both by this literature and the particular case
of Pakistan. Next, I use Pew data to test these hypotheses, using a broad definition of re-
ligiosity that encompasses three distinct dimensions: piety, devotion, and dogmatism.
My final sections outline my primary findings, focusing on support for democracy, re-
ligious extremism, relations with India, and restrictions on blasphemous content as the
key outcome measures, and outline future avenues for research. The results discussed in
this paper are intended to be descriptive and exploratory. Further work remains to be
done to assess the causality behind any relationships we observe.

Religiosity and Political

Attitudes: Problems of Conceptual
Clarity and Measurement

A rich literature has sought to uncover the links between religiosity and political
attitudes on a wide array of outcomes across many different settings. In Islamic contexts,
however, individual religiosity has been far from a clean predictor of partisanship and
other political outcomes of interest. Two factors can help explain these inconsistent
findings that make it difficult to draw many overarching conclusions about these
relationships. First, existing studies use distinct definitions and measures of religiosity.
Second, where studies distinguish between personal and political dimensions of religi-
osity, the outcome measures are often too close to measures of political Islam to be ana-
lytically useful.
The absence of a clear or consistent conceptualization of religiosity means that studies
that speak of “religiosity” are actually comparing different concepts. Pepinsky, Liddle,
and Mujani (2018) rightly point out, for example, that much existing literature has relied
on a single indicator to capture religiosity, and that this indicator varies from one study
to the next. Scholars have conceptualized religiosity as the level of attendance of reli-
gious services (Jamal 2006), support for shariʿa law (Benstead 2014), or participation in
religious ritual such as praying five times a day (Tessler and Nachtwey 1998), to name just
a few examples. More recent studies have sought to incorporate different dimensions
of religiosity into one composite variable or index, arguing that single indicators are
432   Niloufer A. Siddiqui

unable to capture the diversity of behaviors inherent to religiosity (Pepinsky, Liddle, and
Mujani 2018).
Yet different dimensions of religiosity likely affect political attitudes through dif-
ferent pathways. For example, some scholars have pointed to the direct political mo-
bilization opportunities and coordination points provided by frequenting a place
of worship as one potential mechanism (Djupe and Gilbert 2006). Others have
highlighted the role of churches (or mosques) in providing political information and
participatory opportunities to the religious citizenry (Greenberg 2013; Butt 2016).
Through sermons in mosques and churches, or in religious schools, religious clerics
have unique opportunities and regular platforms from which to propagate particular
narratives (Kalin and Siddiqui 2014). Omelicheva and Ahmed (2018) further find that,
on its own, religiosity actually serves as a deterrent to political engagement; rather, it is
membership in religious organizations that make individuals more likely to engage in
political activity. Similarly, Chhibber and Shastri (2014) find that the social equivalence
accorded by attending religious festivals or places of worship makes individuals more
likely to be trusting of democracy. Individual religiosity may also have an impact on
political attitudes or behavior though other mechanisms. Religious ideals may provide,
for example, the “criteria by which to judge governmental authority and . . . to criticize
decisions or policies deemed inconsistent with divine purposes” (Tessler and Nachtwey
1998, 620), or they may provide answers and ethical prescriptions in unpredictable envi-
rons (Wald 1992; Goldstein and Keohane 1993). Thus, there is a case to be made that the
link between political attitudes and individual dimensions and measures of religiosity
should be assessed independently.
Studies have also differed in how they analyze support for political Islam (such
as support for shariʿa or Islamic political parties), with some treating it primarily as
an outcome variable and others including it in their conceptualization of religiosity.
Where support for Islamic law has been identified as a dimension of religiosity, most
studies have nonetheless—​rightly—​sought to distinguish it from personal religiosity.
Indeed, an individual can choose to abide by all tenants of Islam, including, for ex-
ample, praying daily or fasting during the month of Ramadan, but not necessarily be-
lieve in the implementation of facets of shariʿa law. Similarly, individuals may support
Islamic political parties for any number of reasons (see, for example, Cammett and
Jones Luong 2014; Masoud 2014), but not partake in the daily rituals of Islam. Tessler
and Nachtwey find, for example, that those individuals who report high levels of reli-
gious devotion in their personal lives do not have different opinions on the Arab-​Israeli
conflict than those that don’t, but that political aspects of Islam are correlated nega-
tively with support for Arab-​Israeli peace. They conclude, “Attachment to Islam, de-
fined in terms of piety, observance, and an inclination to seek guidance from religious
sources, bears no relationship to attitudes about the most important interstate conflict
in the Middle East” (1998, 631) but find that respondents who believe in having reli-
gion play a larger role in political life and support Islamic political movements are less
likely to support a peaceful resolution of the Arab-​Israeli conflict. Hadad (2003), too,
finds that personal religiosity among Lebanese survey respondents is not a predictor of
 Piety, Devotion, and Support for Shariʿa  433

attitudes toward US policy in the Middle East, but support for political Islam is associ-
ated with unfavorable attitudes.
While important to distinguish between these two dimensions of religiosity, studies
may suffer from conceptual overlap if their measures of political religiosity are too close
to their outcomes of interest. For example, Fair, Littman, and Nugent have suggested
that the ways in which people interpret shariʿa is key to whether they support its im-
plementation, and indeed, whether it determines their political attitudes. They write,
“More than 95 percent of respondents believed a shariʿa government is one that provides
services, justice, personal security, and is free of corruption. In contrast, 55 percent
believed that a shariʿa government is one that uses physical punishments” (2017, 5). They
find that Pakistani respondents who understand shariʿa to be consistent with transpar-
ency and justice will be more likely to support democracy, while those who perceive
shariʿa as restricting the role of women or implementing harsh, physical punishments
will be more likely to support Islamist militant groups. While there is certainly a need
to unpack the various components that make up shari’a support, describing shariʿa in
such illiberal terms as restricting women’s rights when the dependent variable is “sup-
port for democratic values” appears to be an example of the independent and dependent
variables being too close together to offer much analytical traction. Similarly, Haddad
(2003) creates a composite indicator of political Islam that relies on such survey meas-
ures as support for militant violence, support for religious leaders in public office, and
perception of an Islamic state as the most viable political model. The author finds that
support for political Islam is correlated with anti-​American attitudes, an unsurprising
finding insofar as militant violence is often targeted at American audiences or groups
perceived as pro-​American. That is, because most Islamic militants are overtly anti-​
American in their ideology and targeting, we should expect that support for militant
violence and anti-​American attitudes be correlated, almost by definition. Contrasting
personal and political religiosity, in this way, is not particularly helpful, as the latter is too
close to the outcome measures to be informative. The authors may simply be capturing
something rather intuitive, and effectively tautological (see Vreeland 2008 for another
example of such findings).
In this chapter, I posit three distinct dimensions of religiosity that capture different
aspects of the concept. These are “piety,” by which I mean participation in religious
ritual; self-​professed “devotion,” or an assessment of the role of religion in one’s life; and,
“dogmatism,” or support for Islamic law in governance. While these may be related, they
ultimately affect political behavior and attitudes through different mechanisms and
must be assessed separately. Using survey data from Balochistan, Pakistan, Kalin and
Siddiqui (2016), for example, find that respondents are generally skeptical of the mixing
of religion and politics even with very high levels of respondent piety as measured by
adherence to daily rituals such as prayer. Specifically, the majority of respondents—​
irrespective of self-​reported levels of individual piety—​felt that Islamic leaders lose
credibility when they join politics, and that Islamist parties are worse alternatives
than non-​Islamist parties. While some work (e.g., Pepinsky, Liddle, and Mujani 2018)
chooses to exclude any measures of political Islam from a conceptualization of piety,
434   Niloufer A. Siddiqui

I include support for shariʿa as one of my dimensions because it captures an important

aspect of belief in Islamic scripture and dogma.
The next section provides an overview of the role of religion in Pakistan, outlining a
series of hypotheses involving these three measures of religiosity.

Context: The Role of Religion

in Pakistan’s Politics

The relationship between religion and politics was ambiguous at the very founding
of Pakistan.3 Although formed as a state for Muslims, Pakistan’s founder and leader,
Mohammad Ali Jinnah, was avowedly secular and sought to establish a secular, multi-​
confessional state.4 Since those early days, however, the role of Islam in the country’s
legal and sociopolitical life has steadily expanded (Haqqani 2005). The late 1970s and
early 1980s under the military dictator Mohammad Zia-​ul-​Haq, in particular, saw
greater space being accorded to religious political parties and the promulgation of laws
privileging particular schools of Sunni thought. International developments also played
a central role. Pakistan became one of several arenas involved in a broader geopolit-
ical competition between the governments of Saudi Arabia and Iran. This proxy war
brought into Pakistan a funding stream that strengthened and armed Sunni groups in
the country and helped establish a network of madaris (religious schools) following the
Deobandi, Wahhabi, and Ahle Hadith subsects of Sunni Islam. In turn, Iran provided
funding to Pakistani Shiʿa clerics in order to establish their own religious seminaries
in Punjab (International Crisis Group 2011). The anti-​Soviet Afghan jihad further
contributed to the spread of Islamic militancy in Pakistan, with arms and militant
groups proliferating. Some groups—​supported by the Pakistan state in their proxy wars
against their neighbors—​focused outward and directed their ire toward Indian-​occu-
pied Kashmir or targeted the United States and its allies in their war in Afghanistan.
Others, most notably the Tehreek-​e-​Taliban Pakistan (TTP), targeted the Pakistan state
itself, while still others, such as the Lashkar-​e-​Jhangvi, attacked religious minorities in
the country.
Is religiosity a predictor of political attitudes and behaviors in Pakistan? And if so,
which ones? I identify four outcomes of interest: attitudes toward democracy, religious
extremism, foreign relations, and restrictions on blasphemous speech. I outline my the-
oretical priors for each relationship in turn.
First, a rich literature has theorized a link between religiosity, broadly defined, and
support for democracy. Some scholars (Huntington 1996; Fish 2002) have argued that
there is an intrinsic link between Islam and disdain for such liberal ideals as democracy
and pluralism. As such, it would follow that at the individual level, we should expect that
those Muslims who have greater ties to Islam and are personally religious feel differently
about such varied outcomes as support for violence or democracy than those who are
 Piety, Devotion, and Support for Shariʿa  435

more “liberal” or “Western” in their approach. However, despite theoretical predictions,

empirical data on the link between religiosity and support for democracy tends to
show that the two are either unrelated or positively correlated. Survey data from Egypt,
Palestine, and Lebanon, for example, demonstrates a weak relationship between religi-
osity and political preferences, in general, and support for Islamic political movements,
in particular (Harik 1996; Tessler 1997; Tessler and Nachtwey 1998), while survey work
in Bangladesh shows that respondents who want more shari’a prefer more democracy
(Fair and Patel 2019).
In Pakistan, which has historically oscillated between military rule and democracy
since its independence in 1947, and where the military continues to hold influence over
many aspects of foreign policy and national security (Fair 2014; Shah 2014), this rela-
tionship is all the more significant. Military leaders in Pakistan have long used Islam as
a rallying cry to justify their foreign policies and to promote national unity in an eth-
nically heterogeneous society (Cohen 2004; Fair 2014). The military has also engaged
in strategic alliances with Islamist actors over the years. For example, under General
Musharraf ’s tenure, a coalition of religious political parties, the Muttahida Majlis-​e-​
Amal, won power in Khyber Pakhtunkhwa Province (2002–​2008), in what were widely
believed to be rigged elections. Haqqani writes, “Islamist groups have been sponsored
and supported by the state machinery at different times to influence politics and support
the military’s political dominance” (2005, 3). As such, we may expect that there is an in-
verse relationship between support for political Islam and support for democracy.
Hypothesis 1a: Individuals demonstrating greater levels of support for shariʿa will
be less likely to support democracy.
Is there a relationship between religious practice and pro-​democracy sentiment?
One possible mechanism explaining such a relationship, if it exists, is the role played
by religious places of worship as a forum through which religious leaders can propa-
gate certain narratives and viewpoints. In this reading, personally religious people
frequent the mosque or attend dars (Qurʾanic study group) more often or otherwise in-
teract frequently with religious individuals and religious leaders. They may therefore be
mobilized by these leaders or provided information by the similarly religiously inclined,
which shapes their attitudes toward democracy. However, it is unclear whether clerics
or places of worship in Pakistan would necessarily be anti-​democratic, especially given
that local religious leaders are likely to themselves support particular political parties
or politicians at the polls, and indeed a number of religious political parties themselves
contest elections. Additionally, following Chhibber and Shastri (2014), it is possible that
the social aspect of religious practice makes individuals more pro-​democracy because it
serves as an “equalizer,” bringing people into contact with individuals belonging to other
social classes, or indeed, even political figures.
Hypothesis 1b: Individuals demonstrating greater levels of personal piety and de-
votion will be no more or less likely to support democracy than less pious or less de-
voted individuals.
Religious individuals may also have distinct approaches to the issue of extremist vi-
olence, which, in the Pakistani context, has been predominantly religiously motivated.
436   Niloufer A. Siddiqui

Such individuals may be more sympathetic to the groups orchestrating the attacks, be-
cause they do so in the name of jihad and ostensibly for Islamist goals—​whether the
creation of an Islamic shariʿa state or to privilege particular Sunni interpretations.
Or they may attend mosques or religious gatherings that are sympathetic to militant
groups. However, even if one were to accept this reasoning, the proposed relationship
remains too simplistic. First, a Shiʿa individual may be pious but, because he or she is
more likely to be a target of extremist attacks in the country, may still be more likely to
be concerned with extremism in the country than either pious or less pious Sunnis. If
we assume that religious Shiʿas attend the mosque or other religious gatherings more
frequently, these individuals may be privy to more anti-​extremist rhetoric than less
pious Shiʿas. In this case, then, an individual’s sectarian affiliation should be a greater
predictor of concern with extremism than adherence to particular religious rituals.
Indeed, studies have produced inconclusive results about the link between religiosity
and support for militancy. Wiktorowicz (2005), in his qualitative study of the militant
British Islamist group al-​Muhajiroun, demonstrated that religious individuals were less
susceptible to the group’s messages than the less religiously inclined. Fair (2015) finds, in
contrast, that when piety is measured as intensity of religious practice, pious individuals
in Pakistan are more likely to support sectarian violence. However, this result is much
less significant than a respondent’s sectarian identity (or maslak). Ginges, Hansen, and
Norenzayan (2009) find a positive link between attendance at religious services and
support for militant attacks.
Hypothesis 2: Religious Sunni Pakistanis (across all three measures) will be less
likely to be concerned with religious extremism than less religious Sunnis, but reli-
gious Shiʿas will be more likely to be concerned than less religious Shiʿas.
Given the role of religion in the acrimonious history between Pakistan and neigh-
boring India, we may also expect that individual religiosity affects attitudes toward
India. In particular, we may expect that individuals who are more pious or devoted
would be more suspicious of India and less likely to want to pursue peaceful relations
with it, because they frequent mosques/​religious gatherings and interact more with
local clerics who are likely to espouse anti-​Indian rhetoric, or because they value the
role of Islam more highly in their lives. In contrast, those individuals who believe that
Islam should have a greater say in the way the country is run may be no more or less
likely to view India as a threat. On the one hand, such dogmatic individuals may harbor
resentment against India for the violent events of Partition; on the other hand, they may
be more likely to believe in the concept of a Muslim umma (community of believers),
which suggests that Muslims everywhere, regardless of ethnicity or nationality, are part
of the same community.
Hypothesis 3: Pious and devoted individuals will be more likely to consider India a
threat, but more dogmatic individuals will be no more or less likely to do so.
Finally, we would expect personally religious individuals to be more likely to be-
lieve in strict restrictions on blasphemous speech, as this may be considered antithet-
ical to their religious values. Politically religious individuals may similarly believe
in such restrictions, as they believe that Islamic law should play a more central role in
 Piety, Devotion, and Support for Shariʿa  437

government. To be clear, this relationship doesn’t necessarily suggest that individuals

would want blasphemers to be punished, merely that their speech be regulated by the
government.
Hypothesis 4: Religious individuals (across all measures) will be more likely to sup-
port restrictions on blasphemous speech than the less religious.

Measuring Religiosity in Pakistan

Following Marshall’s (2002) theory of ritual practice, which emphasizes beha-

vior, belonging, and belief, I conceive of religiosity as comprising three different
dimensions: piety, devotion, and dogmatism. The first is measured by how often people
offer their daily prayers—​a fundamental tenant of Islam. The second, devotion, is meas-
ured by how important an individual considers religion in their life. The third, dogma-
tism, is measured by the extent to which an individual believes a country’s laws should
follow the teachings of the Qurʾan. All three of these measures are correlated with one
another in the 2015 Pew data used here: each is a significant predictor of the other, in the
expected direction. As with all survey questions, however, we should be aware of the
possibility of social desirability bias, which is perhaps all the more pronounced when it
comes to matters related to religion. Given that Pakistan is generally perceived as a reli-
gious society, we should expect each of the measures to be biased upward.
Table 21.1 depicts the descriptive statistics for each of these three measures. Daily
prayer sees a considerable level of variation, with about 43 percent of Pakistanis saying
that they pray the mandated five times a day, and another 12 percent praying once a
week. Only 11 percent admitted to praying “hardly ever” or only on religious holidays.
In contrast, self-​reports about religion show that the overwhelming majority of Muslim
respondents (92.5 percent) consider themselves to be very religious. Belief in how shariʿa
should be implemented, meanwhile, showed some degree of variation, but with a clear
majority: 76.5 percent of the respondents believe in the importance of shariʿa as a guide
for political and legal life, agreeing with the statement that “laws should strictly follow
the teachings of the Qurʾan.” Only approximately 3 percent believe that the country’s
laws should not be influenced by the teachings of Islam at all.
These results suggest that Pakistanis are generally a religious people. How do they
compare to Muslims in other Islamic countries? With the exception of Turkey, the
other eight countries with sizeable Muslim populations included in this round of Pew
surveys appeared more pious when looking solely at daily prayer. The average number
of Muslim respondents who said that they prayed five times a day across all nine coun-
tries was 69 percent, compared to 43 percent in Pakistani (see Figure 21.1).
How important respondents considered religion in their lives was generally high
across the nine countries, ranging from a low of approximately 58 percent in Turkey to
a high of 97 percent in Senegal, for an average of 83 percent (see Figure 21.2). However,
the overwhelming Pakistani support for shariʿa, or specifically for laws to follow the
 438   Niloufer A. Siddiqui

Table 21.1: Descriptive Statistics

Measures of Individual Religiosity among Muslims in Pakistan

How often, if at all, do you pray? Hardly ever: 9.13%

Only during religious holidays: 2.24%
Only on Fridays: 11.71%
Only on Fridays and religious holidays: 3.10%
More than once a week: 9.39%
Every day at least once: 20.41%
Every day five times: 43.07%
How important is religion in your life? Very important: 92.50%
Somewhat important: 5.92%
Not too important: 1.08%
Not at all important: 0.17%
Which of the following three statements Laws should strictly follow the teachings of the
comes closer to your view? Qurʾan: 76.5%
Laws should follow the values and principles of
Islam but not strictly follow the teachings of the
Qurʾan: 15.17%
Laws should not be influenced by the teachings of the
Qurʾan: 2.92%

Source: Pew Data, 2015.

Religiosity as Frequency of Prayer (Piety)

100%
Percentage of Respondents Who Report

80%
Praying Five Times a Day

60%

40%

20%

0
an

so

ia

an

on

ia

ia

rr.

y
ga

ke
es

ys

Te
Fa
ist

rd

an

ge

ne

r
ala
n

Tu
Jo

st.
k

Ni
a

do

Se
Pa

Le
in

le
In
rk

Pa
Bu

Figure 21.1: Percentage of Muslim respondents who report that they pray five times a day
across Pakistan and nine other Muslim-​majority countries.
 Piety, Devotion, and Support for Shariʿa  439

Religiosity as Importance in Life (Devotion)

100%
Religion is Very Important in Their Life
Percentage of Respondents Reporting

80%

60%

40%

20%

0
an

so

sia

an

ia

ria

r.

y
ga

ke
r
no

ys

Te
Fa
ist

rd

ge
ne

ne

r
ba

ala

Tu
Jo

st.
k

Ni
a

do

Se
Pa

Le
in

le
In
rk

Pa
Bu

Figure 21.2: Percentage of respondents who report that religion is very important in their lives
across Pakistan and nine other Muslim-​majority countries.

teachings of Islam, appears all the more stark when compared to the data from other
countries. Indeed, compared to nearly 77 percent in Pakistan, the average number of
Muslim respondents who expressed their agreement that laws should strictly follow the
teachings of the Qurʾan was only 42 percent in the other nine countries. This ranged
from a low of 13.4 percent in Turkey to 77.4 percent in Malaysia (see Figure 21.3). We
should be careful when assessing these results, because social desirability concerns with
questions about shariʿa are high. Respondents may also generally respond that they be-
lieve laws need to be in accordance with Islam, but that does not imply that they neces-
sarily believe in specific scriptures being invoked. Nonetheless, these comparisons are
interesting insofar as they show that Pakistani personal piety, as measured through daily
practice, is much lower than that observed in other Muslim countries, but that support
for the mixing of politics and religion is much higher.

Results

Do religious respondents in Pakistan hold different attitudes toward democracy, re-

ligious extremism, foreign policy, and blasphemy compared to the less religiously
inclined? This section analyzes data from a Pew survey from 2015 to address these
questions.
 440   Niloufer A. Siddiqui

Religiosity as Support for Religious Law (Dogmatism)

100%
Law Should Strictly Follow Teachings of Quran
Percentage of Respondents Reporting

80%

60%

40%

20%

0
an

so

sia

an

ia

ria

r.

y
ga

ke
r
no

ys

Te
Fa
ist

rd

ge
ne

ne

r
ba

ala

Tu
Jo

st.
k

Ni
a

do

Se
Pa

Le
in

le
In
rk

Pa
Bu

Figure 21.3: Percentage of respondents who report that laws should strictly follow the teaching
of the Qurʾan across Pakistan and nine other Muslim-​majority countries.

Attitudes toward Democracy

Respondents in Pakistan were asked which of the following three statements were
closest to their view: first, democracy is preferable to any other kind of government;
second, in some circumstances, a nondemocratic government can be preferable; or, third,
for someone like me, it doesn’t matter what kind of government we have. Approximately
46 percent of the 1,200 respondents chose the first option, what I refer to as a pro-​
democracy sentiment.
The Pew data show that, across all three religiosity measures and consistent with
findings from other parts of the Muslim world, religious individuals were more likely to
be pro-​democracy. Those individuals who believed in strict enforcement of shariʿa were
significantly more likely to hold pro-​democracy sentiments (p < 0.001). Similarly, those
individuals who prayed more often were also significantly more likely to believe that
democracy was preferable to other forms of government (p < 0.001), as were those for
whom religion was very important in their lives (p < 0.1). When controlling for the other
religiosity measures, as well as such demographic variables as gender, age, whether they
belonged to the Sunni sect, and schooling, support for shariʿa appeared as the strongest
predictor of pro-​democracy sentiment (p < 0.001), but piety and devotion remained sig-
nificantly correlated with higher support for democracy along all three measures (see
Table 21.2).
 Piety, Devotion, and Support for Shariʿa  441

Deciphering why these relationships exist requires further study. Chhibber and
Shastri’s (2014) account of religious practice in India offers some useful insight. They
argue that, in the relatively hierarchy-​free realm of religious practice in India, those
who partake in its social aspects, such as visiting places of worship or attending reli-
gious festivals, may engage in social contact with fellow citizens with whom they other-
wise have little in common. Such religious social space may also provide individuals the
chance to interact with their political representatives, in the process reducing the social
distance between them and making the former more likely to view democracy favor-
ably. Why support for Islamic law is related to pro-​democracy sentiment, however, re-
mains an open question, one which could be addressed at least partly by unpacking what
respondents believe encompasses Islamic law (as in Fair, Littman, and Nugent 2017).

Attitudes toward India

Respondents were asked which of three possible options they considered the greatest
threat to their country: India, the Taliban, or al-​Qaʿida. The wording of this question
was such that the respondent would have to choose India instead of an Islamic militant
group as the most significant threat to Pakistan. While the question does not specify
whether it is referring to the Afghan Taliban or the Tehreek-​e-​Taliban Pakistan (TTP)—​
the latter a more obvious threat to Pakistanis—​it nonetheless pits India against Islamist
groups. Among the three religiosity measures, increased prayer is the only measure
that is statistically significantly related to believing that India is the greatest threat faced
by Pakistan. This relationship remains significant (p < 0.03) even with the inclusion of
demographic controls (see Table 21.2). Given that the other religiosity measures remain
insignificant, this relationship could perhaps best be explained by the role of mosques
and religious gatherings in disseminating information or narratives by religious leaders,
where anti-​India rhetoric would be common.

Attitudes toward Religious Extremism

Religious extremism in Pakistan remains an issue of great import. Tens of thousands
of Pakistanis have died in attacks carried out by militant outfits, many purporting to
be acting in the name of Islam. The Pew data asks two questions that can be used to
measure support for religious extremism. First, respondents were asked, “Some people
think that suicide bombing and other forms of violence against civilian targets are jus-
tified in order to defend Islam from its enemies. Other people believe that, no matter
what the reason, this kind of violence is never justified. Do you personally feel that this
kind of violence is often justified to defend Islam, sometimes justified, rarely justified
or never justified?” Respondents were also asked, “How concerned, if at all, are you
about Islamic extremism in our country these days? Are you very concerned, somewhat
442   Niloufer A. Siddiqui

Table 21.2: Religiosity and Political Attitudes in Pakistan (Multivariate OLS

Regressions)
Belief that India Concerned Belief that Support
Is Greatest with Violence Is Restrictions
Pro-​Democracy Threat Extremism Justified On Blasphemy

Prayer 0.013+ 0.015+ 0.006 -​0.004 0.038***

(0.008) (0.008) (0.012) (0.016) (0.007)
Self-​Report 0.083+ 0.044 0.219** -​0.437*** 0.162***
Religiosity (0.047) (0.047) (0.070) (0.098) (0.044)
Follow Shariʿa 0.153*** 0.046 0.049 -​0.276*** -​0.014
(0.031) (0.032) (0.047) (0.064) (0.029)
Age 0.001 0.002* 0.003+ -​0.001 0.001
(0.001) (0.001) (0.002) (0.002) (0.001)
Female -​0.135*** -​0.134*** -​0.012 -​0.138* -​0.147***
(0.031) (0.031) (0.047) (0.064) (0.029)
Schooling 0.012*** 0.006* 0.022*** -​0.007 0.007*
(0.003) (0.003) (0.004) (0.006) (0.003)
Shiʿa -​0.023 -​0.131+ -​0.162 -​0.099 0.045
(0.070) (0.071) (0.113) (0.153) (0.066)
Constant -​0.248 0.191 2.188*** 0.6362*** 0.017
(0.200) (0.202) (0.299) (0.412) (0.186)

Observations 1,067 1,067 1,009 971 1,067

R-​squared 0.069 0.042 0.041 0.054 0.083

Standard errors in parentheses

*** p<0.001,
** p<0.01,
* p<0.05, + p<0.1

concerned, not too concerned or not at all concerned about Islamic extremism in our
country these days?”
In bivariate regressions on each measure of religiosity, we find that pious, devoted,
and dogmatic respondents were all significantly more likely to be concerned with ex-
tremism, with those self-​identifying as religious being the strongest predictor. When
controlling for the other measures of religiosity and demographic variables, however,
only self-​professed devotion remains significant. Because such a large percentage of
Pakistani respondents self-​reported as religious, we should be cautious about reading
too much into this finding, particularly coupled with the lack of findings associated with
the two other measures of religiosity.
 Piety, Devotion, and Support for Shariʿa  443

When examining whether respondents believed that Islamist violence could ever be
justified, we find that those individuals who are more likely to express that the laws be
governed by Islam and who self-​identify as religious are also significantly less likely to
believe that violence is justified, while those who pray more often are neither more or
less likely to do so. How do these results hold up when taking into account the sectarian
affiliation of the respondents? Here we are faced with a challenge of a small sample
size, with only 54 of the 1,200 respondents identifying as Shiʿa. For Shiʿas, we find that
those individuals who prayed frequently were less likely to believe that violence was
ever justified—​a distinct finding from Sunnis in our sample. There was no significant
finding for any of the other relationships, but this is likely related to small sample size.
Nonetheless, collectively, these findings depict some noteworthy trends. Both religious
Sunnis and religious Shiʿas, to the extent that they were different from the less religious,
were more likely to be concerned with violent extremism. That Shiʿas who prayed more
were less likely than similar Sunnis to believe that violence was justified may indeed be
a function of attending mosques or religious gatherings, which were more likely to con-
demn religious violence targeted against the Shiʿa community than Sunni mosques.

Restrictions on Blasphemous Speech

The Pew survey asked a question gauging support for the government preventing
“statements that are offensive to your religion or beliefs.”5 Approximately 65 percent
agreed with this statement, while 23 percent said that people should be able to say these
sorts of things publicly. Of our three measures of religiosity, prayer and self-​reported re-
ligiosity are positive predictors of such restrictions on blasphemous speech. Along both
measures, as expected, greater levels of religiosity are linked with beliefs that the gov-
ernment should restrict offensive statements. Whether the respondents support strict
enforcement of shariʿa, however, does not appear independently correlated with this
variable—​an interesting finding given that one would expect government restrictions
on blasphemy to be linked to support for Islamic law.

Discussion and Conclusion

This chapter makes two contributions to a burgeoning literature on the link between
Islamic religiosity and various political attitudes. First, it provides an overview of ex-
isting studies seeking to measure religiosity in the context of Islam, all of which use
distinct measures of religiosity, and some of which have sought to distinguish between
personal and political religiosity. It suggests that many of these studies suffer from con-
ceptual overlap between dependent and independent variables. In particular, in their
attempt at providing nuance to a conceptually gray topic such as religiosity, many
measurements of political religiosity have been stretched conceptually to the point that
444   Niloufer A. Siddiqui

they appear to be describing something quite distinct from the concept of religiosity
as commonly understood. Any relationships that are uncovered between political re-
ligiosity and political outcomes such as support for democratic values or support for
militancy may be fundamentally tautological. That is, the manner in which we choose
to measure support for political Islam poses analytical challenges, in large part because
there is often insufficient conceptual separation between the dependent and inde-
pendent variables being studied.
Second, this chapter uses survey data collected by the Pew Research Center to ex-
amine the relationship between religiosity in Pakistan and attitudes toward democracy,
religious extremism, relations with India, and blasphemous speech. Using three dis-
tinct, but related, measures of religiosity, the data shows that higher levels of religiosity
are linked to greater support for democracy across the board. While personal piety was
found to have no relationship with support for violent extremism, devotion and dog-
matism were negatively correlated. These relationships are noteworthy because they
go against many theoretical expectations, even as they find empirical support in other
contexts. More frequent prayer was associated with the belief that India is the greatest
threat faced by Pakistan, consistent with the mechanism that mosques serve an impor-
tant function in propagating particular ideologies and perspectives. While increased
piety and devotion was associated with belief in restricting blasphemous speech,
increased dogmatism was not. More research remains to be done on uncovering the
mechanisms underlying these results.
Future research should continue to focus on the best ways to define and measure
Islamic religiosity, particularly to take into account the importance of support for po-
litical Islam and shariʿa, expansive concepts that can capture quite different political
visions. Similarly, while it remains interesting and necessary to assess the linkages be-
tween religiosity and such variables as militancy and support for democracy, research
agendas should be expanded to include such topics as how changing political beliefs can
lead to changing religious ones (as Margolis 2018 does so well in her study in the United
States), whether liberal religious leaders can lead to more liberal political beliefs (as in
Tuñón 2018), and the ways in which individual religiosity interact with other cleavages,
such as socioeconomic status and partisan affiliation.

Notes
1. If unspecified, the terms religiosity or religious will be used in this chapter as general terms
to refer to all three of these dimensions.
2. Under the law, individuals are prohibited from saying anything that may hurt the
sensibilities of a recognized religion (in practice, Sunni Islam), with punishment ranging
from fines to death. In recent years, this law has been subject to abuse, particularly against
religious minorities who are frequently the targets of false accusations.
3. The vast majority (96.4 percent) of Pakistan’s nearly 200 million people identify as
Muslims, with approximately 85–​90 percent of these identifying as Sunni and 10–​15 percent
 Piety, Devotion, and Support for Shariʿa  445

identifying as Shiʿa. Exact percentages are unknown because the census does not collect in-
formation about sectarian affiliation.
4. For more on the creation of Pakistan, see Jalal 1985; Jalal 1990; and Cohen 2004.
5. The complete wording of the question was: “Do you think that people should be able to
say these types of things publicly OR that the government should be able to prevent people
from saying these types of things in some circumstances.”

References
Ahmed, Sadaf. 2009. Transforming Faith: The Story of Al-​Huda and Islamic Revivalism among
Urban Pakistani Women. Syracuse, NY: Syracuse University Press.
Amnesty International. 2016. “Pakistan: How the Blasphemy Laws Enable Abuse.”
London: Amnesty International, December 21. https://​www.amnesty.org/​en/​latest/​news/​
2016/​12/​pakistan-​how-​the-​blasphemy-​laws-​enable-​abuse/​.
Barreto, Matt, and Dino Bozonelos. 2009. “Democrat, Republican, or None of the Above?
The Role of Religiosity in Muslim American Party Identification.” In “Muslims in
America,” edited by Ted G. Jelen and Sabrina P. Ramet, special issue, Politics and Religion 2
(August): 209–​229.
Benstead, Lindsay J. 2014. “Why Do Some Arab Citizens see Democracy as Unsuitable for
Their Country?” Democratization 22, no. 7: 1183–​1208.
Butt, Ahsan. 2016. “Street Power: Friday Prayers, Islamist Protests, and Islamization in
Pakistan.” Politics and Religion 9, no. 1: 1–​28.
Cammett, Melani, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17: 187–​206.
Chhibber, Pradeer, with Sandeep Shastri. 2014. Religious Practice and Democracy in India .
New York: Cambridge University Press.
Cohen, Stephen P. 2004. The Idea of Pakistan. Washington, DC: Brookings Institution Press.
Djupe, Paul A., and Christopher P. Gilbert. 2006. “The Resourceful Believer: Generating Civic
Skills in Church.” Journal of Politics 68, no. 1: 116–​127.
Esmer, Yilmaz, and Thorleif Pettersson. 2007. “The Effects of Religion and Religiosity on
Voting Behavior.” In The Oxford Handbook on Political Behavior, edited by Russell J. Dalton
and Hans-​Dieter Klingemann, 481–​503. New York and Oxford: Oxford University Press.
Fair, C. Christine. 2014. Fighting to the End: The Pakistan Army’s Way of War. New York: Oxford
University Press.
Fair, C. Christine. 2015. “Explaining Support for Sectarian Terrorism in Pakistan: Piety, Maslak,
and Sharia.” Religions 6, no. 4: 1137–​1167.
Fair, Christine, Rebecca Littman, and Elizabeth R. Nugent. 2017. “Conceptions of Shari’a and
Support for Militancy and Democratic Values: Evidence from Pakistan.” Political Science
Research and Methods 6, no. 3: 429–​448.
Fair, C. Christine, Neil Malhotra, and Jacob N. Shapiro. 2012. “Faith or Doctrine? Religion and
Support for Political Violence in Pakistan.” Public Opinion Quarterly 76: 688–​720.
Fair, C. Christine, and Parina Patel. 2019. “Bangladesh’ Rational Islamists: A Study of Islamism
and Regime Preferences in An Under-​studied South Asian Country.” Working paper, avail-
able on Social Science Research Network.
Fish, M. Steven. 2002. “Islam and Authoritarianism.” World Politics 55: 4–​37.
446   Niloufer A. Siddiqui

Ginges, Jeremy, Ian Hansen, and Ara Norenzayan. 2009. “Religion and Support for Suicide
Attacks.” Psychological Science 20: 224–​30.
Goldstein, J., and R. O. Keohane. 1993. Ideas and Foreign Policy: Beliefs, Institutions, and
Political Change. Ithaca, NY: Cornell University Press.
Greenberg, Anna. 2013. “The Church and the Revitalization of Politics and Community.”
Political Science Quarterly 115, no. 3: 377–​394.
Haddad, Simon. 2003. “Islam and Attitudes toward U.S. Policy in the Middle East: Evidence
from Survey Research in Lebanon.” Studies in Conflict & Terrorism 26, no. 2: 135–​154.
Harik, J. P. 1996. “Between Islam and the System: Sources and Implication of Popular Support
for Lebanon’s Hizballah.” Journal of Conflict Resolution 40: 41–​67.
Haqqani, Hussain. 2005. Pakistan: Between Mosque and Military. Washington DC: Carnegie
Endowment for International Peace.
Huntington, Samuel. 1996. The Clash of Civilizations and the Remaking of World Order.
New York: Simon & Schuster Paperbacks.
International Crisis Group. 2011. “Islamic Parties in Pakistan.” Asia report np. 216.
Ispahani, Farahnaz. 2015. Purifying the Land of the Pure: A History of Pakistan’s Religious
Minorities. New Delhi: Harper Collins.
Jalal, Ayesha. 1985. The Sole Spokesman: Jinnah, the Muslim League and the Demand for
Pakistan. Cambridge: Cambridge University Press.
Jalal, Ayesha, 1990. The State of Martial Law. Cambridge: Cambridge University Press.
Jamal, Amaney A. 2006. “Reassessing Support for Islam and Democracy in the Arab World?
Evidence from Egypt and Jordan.” World Affairs 169, no. 2: 51–​63.
Kalin, Michael, and Niloufer Siddiqui. 2014. Religious Authority and the Promotion of Sectarian
Tolerance in Pakistan. United States Institute of Peace Special Report no. 354. Washington
DC: United States Institute of Peace, October.
Kalin, Michael and Niloufer Siddiqui. 2016. “Islam’s Political Disadvantage: Corruption and
Religiosity in Quetta, Pakistan,” Politics & Religion 9, no. 3: 456–​480.
Kaltenthaler, Karl, William J. Miller, Stephen Ceccoli, and Ron Gelleny. 2010. “The Sources of
Pakistani Attitudes toward Religiously Motivated Terrorism.” Studies in Conflict & Terrorism
33, no. 9: 815–​835.
Margolis, Michele. 2018. From Politics to the Pews: How Partisanship and the Political
Environment Shape Religious Identity. Chicago: University of Chicago Press.
Marshall, Douglas. 2002. “Behavior, Belonging, and Belief: A Theory of Ritual Practice.”
Sociological Theory 20, no. 3: 360–​380.
Masoud, Tarek. 2014. Counting Islam: Religion, Class, and Elections in Egypt. Cambridge:
Cambridge University Press.
Omelicheva, Mariya Y., and Ranya Ahmed. 2018. “Religion and Politics: Examining the Impact
of Faith on Political Participation.” Religion, State and Society 46, no. 1: 4–​25.
Pepinsky, Thomas, R. William Liddle, and Saiful Mujani. 2018. Piety and Public Opinion:
Understanding Indonesian Islam. New York: Oxford University Press.
Shah, Aqil. 2014. The Army and Democracy: Military Politics in Pakistan. Cambridge,
MA: Harvard University Press.
Sobolewska, Maria, Stephen D. Fisher, Anthony F. Heath, and David Sanders. 2015.
“Understanding the Effects of Religious Attendance on Political Participation among Ethnic
Minorities of Different Religions.” European Journal of Political Research 54, no. 2: 271–​287.
South Asia Terrorism Portal. 2018. “Fatalities in Terrorist Violence in Pakistan 2000–​2018.”
http://​www.satp.org/​satporgtp/​countries/​pakistan/​database/​casualties.htm.
 Piety, Devotion, and Support for Shariʿa  447

Tessler, Mark. 1997. “The Origins of Popular Support for Islamist Movements: A Political
Economy Analysis.” In Islam, Democracy, and the State in North Africa, edited by J. Entelis,
93–​126. Bloomington: Indiana University Press.
Tessler, Mark, and Jodi Nachtwey. 1998. “Islam and Attitudes toward International
Conflict: Evidence from Survey Research in the Arab World.” Journal of Conflict Resolution
42, no. 5: 619–​636.
Tuñón, Guadalupe. 2018. “When the Church Votes Left: How Progressive Religion Hurts
Gender Equality.” Working Paper. Berkeley: University of California.
Vreeland, James Raymond. 2008. “The Effect of Political Regime on Civil War: Unpacking
Anocracy.” Journal of Conflict Resolution 52, no. 3: 401–​425.
Wald, K. 1992. Religion and Politics in the United States. Washington, DC: Congressional
Quarterly Press.
Wiktorowicz, Quintan. Radical Islam Rising: Muslim Extremism in the West. Lanham,
MD: Rowman & Littlefield, 2005.
Zia, Afiya Shehrbano. 2009. “Faith-​based Politics, Enlightened Moderation and the Pakistani
Women’s Movement.” Journal of International Women’s Studies 11, no. 1: 225–​245.
 Chapter 22

Mapping and E x pl a i ni ng
Ar ab At titude s towa rd
the Isl ami c Stat e
Findings from an Arab Barometer Survey and
Embedded Experiment

Mark Tessler, Michael Robbins, and

Amaney Jamal

To what extent do ordinary Sunni Muslim citizens in Arab countries sympathize with
the Islamic State? And among those who do support the organization, albeit passively,
are they attracted primarily by the organization’s Islamist message and objectives, or are
they motivated largely by material and political considerations? This chapter uses data
from an Arab Barometer survey and embedded experiment to address these questions.
It considers the explanatory power of both religious and nonreligious factors in ac-
counting for variance in attitudes toward the goals and tactics of the Islamic State. It
also maps and seeks to explain the views of ordinary citizens about whether the
organization’s actions are compatible with the teachings of Islam.

Introduction

The recruitment of tens of thousands of foreign fighters who have joined the forces of
the Islamic State might suggest that its support is wide-​ranging. This is not the case.
Daesh—​a transliteration of the acronym of the Islamic State’s name in Arabic by which
the organization is frequently known—​has brutally killed or executed thousands of
450    Mark Tessler, Michael Robbins, and Amaney Jamal

civilians, both non-​Muslim and Muslim, throughout its territory and beyond. Yet
public opinion surveys report that the vast majority of Muslims reject the use of vio-
lence against civilians. Even more important, surveys also indicate that the vast majority
of Muslims in the Arab world and in other Muslim-​majority countries are opposed to
the organization (ACRPS 2015, 2018; Lipka 2017; Poushter 2015). Nevertheless, as the
number of foreign fighters who have joined the organization in Syria and elsewhere
implies, there is at least a small minority of ordinary citizens across the Arab world who
sympathize with Daesh.
The Islamic State is an insurgent group that is characterized not only by its use of
violence in pursuit of political objectives, but also by its espousal of an extremist ide-
ology in the name of Islam that at least some ordinary citizens apparently find persua-
sive. The existing literature offers relatively few insights into why Daesh has been able
to achieve greater relative success than most other extremist groups. In recent years,
the literature on insurgent groups has downplayed the importance of ideological attach-
ment compared to the provision of tangible material benefits to the population within
its territory (Beardsley and McQuinn 2009; Berman 2003; Berman and Laitin 2008;
Bhattacharya 2013; Davis and Cragin 2009; Flanigan 2006; Flanigan and O’Brien 2015;
Simon and Martini 2004; Weinstein 2007). But whatever its relevance in other cases,
the provision of material benefits cannot help to explain why Daesh activities in Syria
and Iraq, and perhaps now to some extent in Libya and Yemen, lead some individuals in
other Arab countries to passively support the organization.
With respect to support for insurgent groups that operate in an external environment,
many existing accounts have limited explanatory power in the case of the Islamic State.
Some studies emphasize converging political interests (see Byman et al. 2001). This
finding most often applies to diaspora or refugee populations, which are groups that
are likely to perceive a specific link between their own preferences and the actions of
the insurgent group (Barber 1997; Collier and Hoeffler 2004; Cordovez and Harrison
1995; Ranetunge 2000; Shain and Sherman 1998; Sheffer, 1994; Urban 1987). Among the
small percentage of Muslim Arab populations that sympathize with the Islamic State,
there is no indication that support is concentrated among either refugees or a diaspora
population.
A more plausible explanation is either that the organization has been successful in
converting some ordinary Muslims to its extremist worldview, or that it has activated
support among previously radicalized individuals. In either case, a critical factor may
be the content of the Islamic State’s message and ideology and the way that these are
received and understood by Muslim men and women living elsewhere.1 However, if ide-
ological outreach is a key variable leading external populations to support the Islamic
State, questions remain about the kind and content of the outreach efforts that are most
effective—​and most in need of being countered by those who oppose the Islamic State;
and then about the sectors of the population in which these messages are most likely to
have the impact that Daesh intends.
Mapping and Explaining Arab Attitudes toward the Islamic State     451

Against this background, we first develop and test with Arab Barometer survey data
a series of hypotheses about the attributes of those ordinary citizens who express sup-
port for the Islamic State, or who at least decline to express opposition to the organiza-
tion. These hypotheses and the causal stories they represent consider the explanatory
power of personal religiosity, demographic profile, and political trust. Second, we de-
scribe and consider the results of a survey experiment designed to assess the impact
of different messages about Daesh’s objectives and actions. Finally, we explore the in-
teraction between the independent variables in our hypotheses and the treatments in
the survey experiment. More specifically, we consider whether the personal characteris-
tics considered in the hypotheses are also conditionalities that specify whether, and for
individuals characterized by what attributes, the information conveyed in the experi-
mental treatments has a significant impact on attitudes toward the Islamic State.

Arab Barometer Survey Data

Data from the fourth wave of Arab Barometer surveys permit a deeper look at the views
about the Islamic State, or Daesh, held by ordinary citizens across the Arab world.2 These
surveys were carried out in 2016–​2017 in eight countries: Morocco, Algeria, Tunisia,
Egypt, Palestine, Lebanon, Jordan, and Qatar. A total of 9,518 men and women were
interviewed. Egypt is not included in the present analysis, however, because questions
about the Islamic State were not asked in the Egyptian survey, and only Sunni Muslims
and citizens in the remaining countries are included. This reduces the N to 7323.3
Fourth-​wave surveys also contained an embedded survey experiment pertaining to
the Islamic State. The experiment, described and discussed later in this chapter, involved
random assignment to either a control group or one of four treatment groups. Each
treatment provided respondents with information about the Islamic State, with the kind
and amount of information varying across the four treatments.
The following three questions from the fourth-​ wave interview schedule are
considered in the present analysis. Responses to these questions are the dependent
variables in our analyses.
To what extent do you agree with the goals of the Islamic State?
1: Agree to a large extent; 2: Somewhat agree; 3: Somewhat disagree; 4: Disagree to a
large extent.
To what extent do you agree with the Islamic State’s use of violence?
1: Agree to a large extent; 2: Somewhat agree; 3: Somewhat disagree; 4: Disagree to a
large extent.
To what extent do you believe Daesh’s tactics are compatible with the teachings
of Islam?
1: Definitely compatible; 2: Somewhat compatible; 3: Somewhat not compatible;
4: Definitely not compatible.
452    Mark Tessler, Michael Robbins, and Amaney Jamal

Mapping Arab Attitudes toward the

Islamic State

Table 22.1 presents the distribution of responses to each of the three questions that
measure attitudes toward the Islamic State. Since the three questions were asked after the
treatments of the survey experiment, Table 22.1 is based only on respondents assigned
to the control group. Should these treatments affect attitudes, which is what the experi-
ment sought to determine and measure, the attitudes of respondents in any or all of the
treatment groups would not be representative of Sunni Arab attitudes.
A question of potential significance emerges with respect to respondents who did not
express an opinion, either giving a “Don’t know” response or declining to answer. The
number of respondents in this category is not high, but it is high enough to warrant atten-
tion. The question to be answered is whether the possible sensitivity of questions about
Daesh led individuals who do have opinions about the group to choose one of these
nonresponse options. It is impossible to know with certainty, of course, but it is plausible
enough to suggest that these respondents had at least somewhat positive attitudes toward
Daesh that they were reluctant to express, particularly in a face-​to-​face survey. We con-
sider it instructive, therefore, that these respondents refrained from criticizing Daesh,
and for this reason we include them as a separate category in Table 22.1. They are not oth-
erwise included in the present analysis.
Based on the response distributions in Table 22.1, a general takeaway is that very few
respondents have positive attitudes toward the Islamic State, and that support for Daesh
is extremely low in some countries.4

Hypotheses and Causal Stories

The move from description to explanation in the present section will assess the explan-
atory power of three possible drivers: personal religiosity, demographic category, and
political trust.

Personal Religiosity
Turning first to personal religiosity, the following three survey items were used to cap-
ture the variance and rate respondents along a continuum ranging from very religious
to not religious at all. The validity and reliability of these items were assessed using con-
firmatory factor analysis, the logic being that high loadings on a single factor establishes
unidimensionality, which in turn offers evidence of reliability and provides a strong
basis for inferring validity. The loadings, shown along with the items, range from .779 to
Table 22.1: Attitudes toward the Islamic State in Seven Arab Countries in 2016–​2017

To what extent do you agree with the goals of the Islamic State?

Agree to a Somewhat Somewhat Disagree to a Don’t Decline to

large extent agree disagree large extent know answer

Algeria 1.7 5.2 9.6 83.5

Algeria 2.1 2.1
Jordan 0.0 0.0 .04 99.6
Jordan 1.1 0.0
Lebanon 0.9 0.9 14.7 83.5
Lebanon 0.9 0.0
Morocco 0.0 0.4 2.4 97.2
Morocco 2.3 0.4
Palestine 1.7 6.5 12.1 79.7
Palestine 2.1 0.0
Qatar 0.0 2.1 7.2 90.7
Qatar 2.8 6.5
Tunisia 1.3 0.9 0.0 97.8
Tunisia 1.3 0.4
To what extent do you agree with the Islamic State’s use of violence?

Agree to a Somewhat Somewhat Disagree to a Don’t Decline to

large extent agree disagree large extent know answer

Algeria 3.9 3.0 7.8 85.2

Algeria 2.9 1.3
Jordan 0.4 0.0 0.0 99.6
Jordan 0.7 0.0
Lebanon 0.0 0.0 5.5 94.5
Lebanon 0.0 0.0
Morocco 0.0 0.0 1.2 98.8
Morocco 0.0 0.0
Palestine 3.5 3.5 10.6 82.4
Palestine 3.4 0.4
Qatar 2.1 1.0 4.2 92.7
Qatar 3.7 6.5
Tunisia 1.8 0.4 0.9 96.9
Tunisia 0.9 0.9

(continued)
454    Mark Tessler, Michael Robbins, and Amaney Jamal

Table 22.1: Continued

To what extent do you believe Daesh’s tactics are compatible with the teachings of Islam?

Certainly
Certainly Does not does not
represents Represents true represent represent Don’t Decline to
true Islam Islam true Islam true Islam know answer

Algeria 3.9 7.9 7.5 80.7

Algeria 2.9 2.1
Jordan 0.4 0.4 0.7 98.5
Jordan 1.4 0.0
Lebanon 0.9 2.8 13.9 82.4
Lebanon 1.8 0.0
Morocco 0.4 0.4 4.9 94.3
Morocco 3.9 0.4
Palestine 3.1 6.6 15.4 75.0
Palestine 3.4 0.0
Qatar 2.1 2.1 7.3 85.5
Qatar 3.7 6.5
Tunisia 4.1 2.3 1.8 91.8
Tunisia 4.7 0.4

.806.5 With high loadings on a single factor, and hence unidimensionality, the analysis
then combined the items, weighted on the basis of their factor loadings, and generated
an index based on factor scores. The factor scores range from −1.21454 to 3.23202, with
higher values indicating a lower level of personal religiosity.
In general, would you describe yourself as religious, somewhat religious, or not
religious?

• Factor Loading: .794

Do you pray daily?

• Always, Most of the time, Sometimes, Rarely, Never

• Factor Loading: .806

Do you listen to or read the Qurʾan or other religious programs?

• Always, Most of the time, Sometimes, Rarely, Never

• Factor Loading: .779
Mapping and Explaining Arab Attitudes toward the Islamic State     455

Hypotheses about the explanatory power of personal religiosity are suggested

by the centrality and salience of Islam with respect to both the dependent variable
and the independent variable. More specifically, a plausible and perhaps persuasive
causal story posits that more religious men and women are more likely to hold posi-
tive views about Daesh, or, conversely, less likely to hold negative views about Daesh,
because their stronger attachment to Islam makes them more receptive to Daesh’s
claims and declared objectives than their less religious counterparts. This includes,
presumably, Daesh’s insistence that it is leading a struggle to establish a Muslim cal-
iphate, a community of believers in which life will be governed by Islamic law, prop-
erly understood. Accordingly, the caliphate would not only unify Muslims and make
them better able to oppose their enemies, but it would also, and just as importantly,
enable Muslims to purge from their ranks the deviant or secular elements that are
present in many Muslim-​majority countries. The latter is given tangible expression
in Daesh’s denunciation of those, particularly Shiʿi Muslims, whom it claims have
deviated from the true and proper path of the religion. The possibility that greater
personal religiosity increases the likelihood that Sunni Muslim Arabs will look more
favorably on the claims and actions of the Islamic State is expressed in Hypothesis
1a (H1a).
H1a postulates and tests the assumption that a shared connection to Islam, albeit a
very different kind of connection, predisposes Sunni Muslim Arabs to believe that
Daesh’s actions are in the interest of the religion. But there is also another hypothesis,
H1b, to consider. H1b postulates that Sunni Muslims in the Middle East and North
Africa who are more religious are disproportionately likely to have negative views about
the Islamic State.
It should be evident that H1b differs from H0, the null hypothesis. H0 posits that
personal religiosity has no impact on attitudes toward Daesh. But H0 is not the only
alternative to H1a. Rather, it is possible that personal religiosity does have an impact
on views about the Islamic State, but that this relationship is not the one represented
by H1a. H1b, in this case, advances the possibility that personal religiosity pushes
toward an unfavorable view of Daesh among Sunni Muslims in the Arab world. The
underlying causal story to which H1b calls for consideration is that more religious
individuals are more knowledgeable about Islam, and this in turn leads them to rec-
ognize, or at least believe, that the interpretations of Islam that animate the Islamic
State deviate very significantly from the proper and true understanding of what
Islam teaches and what it enjoins upon believers. Indeed, for these men and women,
these understandings may well be seen as grotesque distortions. Thus, whereas H1a
says that personal religiosity and a deeper connection to the religion pushes toward
the view that the actions of the Islamic State are in the interest of Islam, H1b states
the opposite, that Daesh’s message and actions are harmful and detrimental to the
religion.

H1a: Men and women who are more religious are less likely than those who are less
religious to have positive views about the Islamic State.
H1b: Men and women who are more religious are more likely than those who are
less religious to have negative views about the Islamic State.
456    Mark Tessler, Michael Robbins, and Amaney Jamal

Key Demographic
Personal attributes, or demographic features, are another factor that may account for
some of the variance in attitudes toward the Islamic State held by Sunni Muslims in
the Arab world. The focus here is not on the potential explanatory power of any single
demographic variable, but rather on a particular concatenation of three personal
attributes: gender, education, and age. More specifically, the possibility to be considered
is that positive attitudes toward the Islamic State are more likely to be held by younger
poorly educated men than by individuals in other subpopulations defined by personal
attributes. For purposes of the present inquiry, poorly educated younger men constitute,
and may be described as, a key demographic.
This proposition is made plausible by Daesh messaging, which targets this key dem-
ographic more than other personal attribute groupings, and by the fact that Sunni
Muslim Arabs who have left their countries to join the Islamic State have dispropor-
tionately been members of this demographic. At the heart of both Daesh’s messaging
strategy and key demographic departures is an assumption of relatively greater frustra-
tion, and possibly anger, among poorly educated younger men than among other dem-
ographic categories.
For men more than women, there is an expectation of employment in order to pro-
vide for their families; but with low levels of education, their employment prospects are
limited. And their circumstance also differs from that of poorly educated older men,
who to at least some degree have left behind the uncertainty of young adulthood and
accepted their situation. They also passed through these years at a time when aggregate
levels of education were lower, such that a secondary school education was less likely
to be considered a relatively low level of education. Faced with this situation, poorly
educated younger men are, presumably, more discontent with the status quo and re-
ceptive to ideologies that claim to offer a different, better future. For the same reason,
the thought that Daesh is engaged in a noble struggle, indeed a godly one, may add a
measure to the appeal of the organization’s message.
Until Hypothesis 2 (H2) is tested and confirmed, this causal story rests on a founda-
tion of theorizing rather than one of data and evidence. Nevertheless, to the extent H2
expresses a causal story that is plausible and very possibly even persuasive, a test of H2
will yield valuable insights even if the hypothesis is not confirmed. Should H2 not be
confirmed, it will still be useful to know that a driver of Sunni Muslim Arab attitudes
toward Daesh that appears to be important does not, in fact, have explanatory power.
For purposes of the analysis to follow, members of the key demographic are men who
are 30 years of age or younger and have had no more than a secondary school education.

H2: Men who are 30 years of age or younger and have had no more than a secondary
school education are more likely than individuals who do not share this combination
of personal attributes to have positive views about the Islamic State.
Mapping and Explaining Arab Attitudes toward the Islamic State     457

Political Trust
Political trust is a third possible driver shaping attitudes toward the Islamic State. More
specifically, it is proposed that having positive attitudes toward the Islamic State is in-
versely related to level of political trust, which is defined as confidence, or high confi-
dence, in the country’s various national political institutions, such as the cabinet, the
parliament, the military, and the courts. The relationship between attitudes toward
Daesh and political trust is expressed in Hypothesis 3.

H3: Individuals who have lower levels of trust in their country’s political institutions
are more likely than individuals who have higher levels of political trust to have posi-
tive attitudes toward the Islamic State.

The causal story that assigns explanatory power to political trust is similar to the one
proposed in connection with the key demographic. It states that discontent, perhaps
deep discontent, with the way their country is governed is disproportionately likely to
lead those who are unhappy with the status quo to find alternative political formulae
appealing. In addition, Daesh not only offers a different, and what it claims is a much
better, political future, it also condemns the present political system of virtually every
Arab country in a way that resonates with those individuals who have little or no trust in
their country’s political institutions. But while discontent with the status quo is central
to the causal stories that underlie both H2 and H3, they differ in a very important way.
H3 is focused on discontent with the country’s political system, the important elements
being not only that it concerns political rather than socioeconomic factors, as does H2,
but also that it pertains to the polity, not to the individual.
Political trust has been measured with a battery of items in the Arab Barometer
interview schedule that begins by telling respondents, “I’m going to name a number
of institutions. For each one, please tell me how much trust you have in them.” It then
names the institution and asks respondents if they have a great deal of trust, quite
a lot of trust, not very much trust, or no trust at all in that institution. For present
purposes, political trust has been measured by the degree of trust in four political
institutions: The government (Council of Ministers), the elected council of repre-
sentatives (Parliament), the political parties, and the courts and the legal system.
Responses to the questions about trust in these items have been factor-​analyzed fol-
lowing the logic and procedures discussed with respect to the measurement of per-
sonal religiosity.
These items and their factor loadings are listed below.6 As in the case of personal re-
ligiosity, high loadings on a single factor establishes unidimensionality, which in turn
demonstrates reliability and provides an empirical foundation for inferring validity.
Factor loadings vary from .659 to .843. The factor scores, to be employed when H3 is
tested, range from −2.64862 to 1.31206,
458    Mark Tessler, Michael Robbins, and Amaney Jamal

How much trust do you have in the government (Council of Ministers)?

• Factor Loading: .838

How much trust do you have in the courts and legal system?

• Factor Loading: .787

How much trust do you have in the elected council of representatives (Parliament)?

• Factor Loading: .843

How much trust do you have in political parties?

• Factor Loading: .659

Analysis and Findings

Table 22.2 presents the result of an OLS regression analysis that tests the three
hypotheses. Each of the three questions pertaining to the Islamic State is a dependent
variable.7 The independent variables are the three possible drivers presented in H1a,
H1b, H2, and H3. These are personal religiosity, membership in a key demographic cate-
gory composed of poorly educated younger men, and political trust.
Table 22.2 shows that personal religiosity has a strong and statistically significant re-
lationship to attitudes toward the Islamic state for only one of the three questions about
Daesh, the question that asks respondents whether they believe Daesh’s tactics are com-
patible with the teachings of Islam. In this instance, interestingly, the findings do not pro-
vide evidence in support of H1a, but instead provide evidence in support of H1b: greater
personal religiosity pushes toward disagreement that the actions of the Islamic State are
compatible with the teachings of Islam. With respect to the key demographic, belonging
to this demographic category is not related to any of the three questions about Daesh,
and this remains the case when three different cutting points related to age are used—​25
and younger, 30 and younger, and 35 and younger. Finally, political trust has a strong
and statistically significant relationship to all three of the items about the Islamic State.
More specifically, contrary to the inverse relationship posited in H3, lower levels of po-
litical trust push toward more negative views about the Islamic State, including stronger
disagreement with both Daesh’s goals and methods and that the group’s actions and reli-
gious interpretations are compatible with the teachings of Islam.
Table 22.2: Results of an OLS Regression Testing Hypotheses about Personal
Religiosity, Demographic Attributes, and Political Trust
Stronger disagreement with Daesh’s goals

Constant 3.890 3.891 3.890 3.890 3.875 3.887

(.005) (.005) (.005) (.005) (.006) (.006)
Weaker level of personal -​.002 -​.002
religiosity (.005) (.005)
Belongs to key -​.022
demographic: 18–​25 (.019)
Belongs to key -​.005 .000a
demographic: 18–​30 (.015) (.017)
Belongs to key -​.005
demographic: 18–​35 (.014)
Lower political .017*** .015***
trust (.006) (.006)

Stronger disagreement with Daesh’s methods

Constant 3.905 3.907 3.905 3.904 3.890 3.900

(.005) (.005) (.005) (.005) (.006) (.006)
Weaker level of personal .000 -​001
religiosity (.005) (.005)
Belongs to key -​.028
demographic: 18–​25 (.018)
Belongs to key -​.004 -​.001a
demographic: 18–​30 (.015) (.017)
Belongs to key .004
demographic: 18–​35 (.013)
Lower political .010* .018***
trust (.006) (.005)

Stronger disagreement that Daesh represents the true Islam

Constant 3.822 3.823 3.821 3.823 3.797 3.801

(.007) (.007) (.007) (.007) (.008) (.009)
Weaker level of personal -​.028*** -​.029***
religiosity (.007) (.008)
Belongs to key .040*
demographic: 18–​25 (.025)
Belongs to key -​.008 -​.021a
demographic: 18–​30 (.020) (.031)

(continued)
460    Mark Tessler, Michael Robbins, and Amaney Jamal

Table 22.1: Continued

Stronger disagreement with Daesh’s goals
Belongs to key -​.015
demographic: 18–​35 (.018)
Lower political .033*** .038***
trust (.009) (.009)

a Models with personal religiosity and political trust were run with each of the two other age groups

defining the key demographic. The results were the same.

* p<.1
** p<.05
*** p<.01
a Models with personal religiosity and political trust were run with each of the two other age groups
defining the key demographic. The results were the same.
* p<.1
** p<.05
*** p<.01
a Models with personal religiosity and political trust were run with each of the two other age groups
defining the key demographic. The results were the same.
* p<.1
** p<.05
*** p<.01

The finding about personal religiosity deserves emphasis. Whereas it might be

assumed that Muslims who are more devout are disproportionately likely to look with
favor on social and political movements that organize under the banner of Islam, our
data show that this is not the case, at least with respect to the Islamic State. Indeed,
the data show that precisely the opposite is true. More religious Sunni Muslims in the
MENA countries for which we have data are disproportionately likely say that the
actions and religious interpretations of the Islamic State are not compatible with the
teachings of Islam.
The causal story that offers an explanation for this is probably that more religious
individuals are more knowledgeable about the religion, and are therefore better able to
have an informed opinion about what is and is not within the range of legitimate Islamic
interpretation. Along the same lines, more religious individuals are probably more
likely to care about who claims to speak for Islam and whether the sayings and actions
of those who do contribute to a correct public perception of the religion. For more reli-
gious individuals, therefore, Daesh’s proclaimed connection to Islam is not only incor-
rect and unjustified, it is also harmful and probably dangerous.
Findings about the explanatory power of political trust also merit elaboration. H3
proposed that individuals who have lower levels of trust in their country’s political
Mapping and Explaining Arab Attitudes toward the Islamic State     461

institutions are more likely than individuals who have higher levels of political trust to
have more positive attitudes toward the Islamic State. This hypothesis is not confirmed.
There are statistically significant relationships between political trust and views about
the Islamic State. Indeed, these are the strongest and most consistent relationships of
those proposed in the three hypotheses under consideration. But the direction of these
relationships is not the one described in H3. Rather, individuals who have lower levels
of trust in their country’s political institutions are more likely to have more negative
attitudes toward the Islamic State.
The logic on which H3 is based is evidently flawed. But political trust does have a very
strong and significant relationship to views about the Islamic State, so there is clearly a
story to be discovered. One possibility is that a broader conceptualization of political
trust is needed. Whereas H3 rested on the assumption that trust was low in the polit-
ical institutions of the respondent’s country, our data suggest that trust is actually low
in political institutions, and probably also political processes, in general. It therefore
does not tell us, primarily, how well and fairly or how poorly and unfairly the country is
being governed in the judgment of the respondent. It instead tells us something about
the respondent’s own perspective and worldview with respect to political affairs. Thus,
as the data show, this person has unfavorable opinions about the Islamic state’s political
project as well as about political institutions closer to home.
This is an intriguing possibility that appears to offer a plausible explanation for
findings that were not expected. It may not be correct, of course, and at the very least
its serious consideration requires further reflection and additional analysis. But it does
offer a line of inquiry with the potential to be instructive. It also offers an example of
the way to respond to findings that are obviously strong and significant but, at the same
time, are also unexpected—​findings, it might be said, that arrive without a foundation
of theory.

The Experiment

The experiment embedded in the fourth wave of Arab Barometer surveys randomly
assigned respondents to one of five groups, a control group and four treatment groups.
This was done in seven of the eight countries included in the fourth wave: Algeria,
Jordan, Lebanon, Morocco, Palestine, Qatar, and Tunisia. Taking these countries to-
gether, it means that each of the experiment’s five groups contained approximately 1,450
respondents.
The four treatments were statements about the Islamic State that were read to
respondents before they were asked the three questions about Daesh’s goals, methods,
and Islamic interpretation. Selection of these treatments was not based on an over-
arching theory or with a view toward testing specific hypotheses in pursuit of general-
izable causal stories. Rather, the primary goal of the experiment was to shed light on a
real-​world and important question confronting the Arab region, and other regions: to
462    Mark Tessler, Michael Robbins, and Amaney Jamal

what extent were Daesh’s actions and messaging building passive support for the
Islamic State among ordinary Sunni Muslim Arabs. Accordingly, the experiment’s
four treatments asked about Daesh’s use of violence and its declared objectives as
communicated in its social media messaging.8
The first treatment, Treatment A, told respondents about Daesh’s claim to be
establishing an Islamic caliphate and then stated that the group had killed many non-​
Muslims, and also Muslims, in pursuit of this goal. The three remaining treatments,
Treatments B, C, and D, included the text of Treatment A and then added additional text
about one of Daesh’s declared objectives: confronting Iran and Shiʿi Islam, confronting
Arab leaders with a secular orientation, and confronting Western imperialism in the
Arab world. The exact language of the four treatments is presented in Appendix A.

Treatment Effects

Table 22.3 shows the degree to which views about the Islamic State are positive, based on
responses to the questions about Daesh’s goals, methods, and Islamic interpretation. All
respondents, whether in the control group or one of the treatment groups, were asked to
answer these questions. The percentages shown in the table are based on two different
dichotomizations of the responses to each treatment question. In the first case, in the
left-​side column in each cell of the table, the comparison is between those who either to
a large extent agree or somewhat agree with Daesh and those who either somewhat disa-
gree or to a large extent disagree with Daesh. The table in each case gives the percentage
of those who either somewhat agree or to a large extent agree with the Islamic State’s
goals, methods, or interpretation of Islam.
In the second case, in the right-​side column in each cell of the table, the comparison is
between those who to a large extent disagree with Daesh and those who do not hold this
view, including in the latter category not only those who either to a large extent agree
or somewhat agree, but also those who somewhat disagree with Daesh. The figures in
these right-​side columns are the percentages of respondents who gave one of the three
responses in the latter category, which is to say those who do not to a large extent disa-
gree with Daesh. Given the very small proportion of respondents with positive attitudes
toward the Islamic State, having two dichotomizations increases our chances to discern
any treatment impact that might otherwise be missed.
The table also presents response distributions for individuals in either treatment
groups B or C or D and for those in either treatment groups A or B or C or D. These
two treatment group combinations are included to facilitate cross-​group comparison.
The BCD combination is included to see whether there is a treatment impact that is
not related to any particular one of Daesh’s stated objectives. The ABCD combination
is included to see whether there is a treatment impact irrespective of the content of the
treatment.
Mapping and Explaining Arab Attitudes toward the Islamic State     463

The impact of the various treatments, or combinations of treatments, can be assessed

by comparing the attitudes of respondents in any treatment group to the attitudes of
respondents in the control group. Since group assignments were random, the control
group differs from each treatment group only with respect to the treatment, and that dif-
ference thus expresses the treatment’s impact.
Table 22.3 shows, somewhat surprisingly, that in very few of the treatment groups
do respondent attitudes toward the Islamic State differ significantly from respondent
attitudes toward the Islamic State in the control group. There is rarely a difference of
more than 1 percent between the control group and a treatment group, and in not one of
the thirty-​six treatment cells in Table 22.3 is there a 2 percent difference. As a broad and
general takeaway, these findings thus suggest that not only are positive attitudes toward
the Islamic State few and far between, but also that for the most part they are not pushed
even lower by the receipt of information about Daesh’s actions and objectives.
Some additions to this broad and general takeaway are needed. There are a few treat-
ment group–​control group differences large enough to merit attention, and large enough
to be statistically significant. A chi square statistic was computed to determine whether
the attitudinal difference between respondents in the control group and respondents in
each individual treatment group, though small, might nonetheless be large enough to
be statistically significant and therefor worthy of attention. Toward this end, chi square
values were computed for each cell in Table 22.3, and the resulting probability values are
given in parentheses. There are eight instances of a treatment group–​control group dif-
ference large enough to be statistically significant at the more tolerant .10 level, and of
these, four are significant at the .05 level of confidence. The figures are in bold in each of
the eight statistically significant instances.
These findings suggest several addenda to the previously reported general
conclusion—​that attitudes toward the Islamic State are not influenced in meaningful
measure by information about the actions and social media messages of the Islamist
group. It is possible, of course, and perhaps likely, that many respondents have little
prior knowledge about the Islamic State. But in this case, the conclusion remains the
same: the acquisition of information and knowledge about Daesh’s actions and goals
does not change people’s attitudes toward the group, at least not in the short run.
The first addendum is a refinement of this general conclusion. More specifically, it
concerns the particular questions and issues to which this general conclusion does and
does not apply. Interestingly, and to some extent surprisingly, none of the experimental
treatments, or combination of treatments, has a significant impact on the question of
whether Daesh’s actions and religious interpretations are compatible with the teachings
of Islam. This is somewhat surprising given the centrality of Islam in the lives of both cit-
izens and societies in the Arab world. The expectation, therefore, was that it would be in
response to this question that treatment primes would be most likely to have an impact.
Nevertheless, this is not the case. The judgments about religious matters of ordinary cit-
izens are, apparently, sufficiently strong and stable to withstand whatever impetus for
change is provided by the experimental treatments.
464    Mark Tessler, Michael Robbins, and Amaney Jamal

Table 22.3: Impact of Experimental Treatments on Attitudes toward the Islamic State

Attitudes toward Attitudes toward Daesh actions compatible
the goals of the the methods of the with the teachings of
Islamic State Islamic State Islam

Percent agree Percent Percent Percent Percent agree Percent

to a large not agree to a not to a large not
extent or disagree large extent disagree extent or disagree
somewhat to a large or somewhat to a large somewhat to a large
agree extent agree extent agree extent

Control 3.1 8.7 2.9 6.9 5.3 11.9

Group
Treatment 2.1(.081) 6.8(.044) 2.3 (.340) 6.3 (.533) 4.3 (.206) 11.2 (.558)
A: Caliphate and
killings
Treatment 3.1 (.893) 8.6 (.806) 2.0 (.138) 6.3 (.561) 5.1 (.749) 12.2 (.853)
B: Oppose Iran
and Shiʿa
Treatment 2.3 (.129) 7.1(.078) 1.8 (.047) 5.7 (.170) 4.8 (.504) 11.1 (.465)
C: Oppose secular
Arab leaders
Treatment 2.6 (.363) 7.4 (.185) 1.9 (.085) 6.3 (.493) 4.3 (.211) 10.1 (.139)
D: Oppose Western
interference
Treatment 2.6 (.293) 7.6 (.170) 1.9 (.026) 6.1 (.275) 4.8 (.359) 11.1 (.414)
BCD
Treatment 2.5 (.167) 7.4(.086) 2.0 (.041) 6.1 (.295) 4.6 (.263) 11.1 (.412)
ABCD
N = 7,149 N = 7,177 N =7,079

Note: Cells of the table contain percentages and, in parentheses, probability values.
Probability values of .10 or lower are in bold.

A second addendum concerns the questions and issues for which the general conclu-
sion advanced above appears not to apply, and this involves, in particular, the question
about Daesh’s methods. Treatment C and Treatment D both impact respondent attitudes
about the Islamic State’s methods, particularly the killing it has done. Treatment C
references Daesh’s opposition to secular Arab leaders, and Treatment D references the
group’s opposition to Western interference in the Arab region. Upon hearing about each
of these two Daesh objectives, respondents in treatment group C and respondents in
treatment group D express attitudes toward the Islamic State’s methods that are signifi-
cantly more critical than respondents in the control group.
Similarities in the content of Treatments C and D and the way that each differs
from the content of Treatment B suggest what might be an underlying mechanism.
Mapping and Explaining Arab Attitudes toward the Islamic State     465

Treatments C and D are both about politics. They reference political actors and focus
on those that represent a particular agenda. Both Arab leaders and the Western powers
with which they often cooperate support and seek to maintain the political status quo.
Treatment B, by contrast, invites reflection on issues of religion, identity, culture, and
perhaps nationalism. It is only possible to speculate about the reason that treatments
with political content increase disagreement with Daesh’s methods while treatments
with cultural content do not. A possible and plausible causal story is that respondents
concerned with security and political stability, and to that extent support the status
quo, whatever its problems, will be more fearful of the violence and destruction that the
Islamic State has wrought and less inclined to agree with Daesh’s opposition to either
secular Arab leaders or the involvement of Western powers in Arab political affairs. Put
differently, Daesh’s actions and methods represent political disorder in the larger Arab
arena, and when primed to think about Daesh’s opposition to political actors that repre-
sent and seek to maintain the existing political order, however unsatisfactory that order
might be in other respects, these respondents become significantly more critical of the
Islamic State.
A final addendum is suggested by views about the Islamic State’s goals. The treatments
that have a statistically significant impact in this instance are, first, Treatment A, which
talks about Daesh’s commitment to the establishment of an Islamic caliphate and the
killing of both non-​Muslims and Muslims in pursuit of this end; and second, Treatment
C, which restates the content of Treatment A and then adds a reference to Daesh’s op-
position to secular Arab leaders. The significant difference between treatment group
attitudes and control group attitudes is most notable when the comparison is among
respondents whose views do and do not fall outside the “disagree to a large extent” cat-
egory, which is reported in the right-​side columns in Table 22.3. The probabilities asso-
ciated with Treatment A are much lower in the right-​side column, and the probabilities
associated with Treatment C are only significant in the right-​side column. The finding,
then, is that Treatments A and C have moved a significant proportion of respondents
from less determined opposition to Daesh’s goals, or perhaps even a measure of support
for these goals, to a more determined opposition to the Islamic State and its objectives.
It is notable that Treatment A had this impact. This suggests that what might be called
“basic” information about Daesh and its goals is enough to move some respondents
from weak disagreement with Daesh’s goals, or perhaps even weak agreement in some
instances, to very strong disagreement with the Islamic State’s objectives. The experi-
ment thus offers evidence that exposure to even limited or basic information about the
Islamic State is sufficient to move some Sunni Muslim Arabs who previously were not
strongly opposed to Daesh’s goals to views, having acquired this information, that are
more strongly negative. A possible policy implication is that leaders seeking to reduce
passive support or even weak opposition to the Islamic State will find that a better-​
informed public will, in the aggregate, have more negative attitudes toward the Islamist
project.
The finding that Treatment C also significantly reduces the proportion of
respondents who do not strongly disagree with Daesh’s goals, and thus increases the
466    Mark Tessler, Michael Robbins, and Amaney Jamal

proportion who do strongly disagree, invites speculation about why being told that
Daesh opposes secular Arab leaders makes a difference, particularly when being told
that Daesh opposes Iran and Shiʿa, or that it opposes Western interference does not.
One possibility is that opposition to Arab leaders is closer to home for most ordinary
citizens than are the oppositions referenced in Treatments B and D. Closer to home in
this case means that major Islamic State victories would be more likely to affect these
respondents directly.
This also suggests, as noted previously with respect to other findings, that a concern
for security and stability may be motivating many individuals and shaping the way they
think about the various Daesh goals that are referenced in Treatments B, C, and D. With
intense fighting in Syria, Yemen, Iraq, and Libya at the time of the Arab Barometer
fourth-​wave survey, and with the failure of Islamist governments in Egypt and to
some extent Tunisia only a few years before, secular Arab leaders may represent polit-
ical order and stability in the minds of ordinary citizens. These individuals may com-
plain about widespread corruption and serious violations of human rights at the hands
of these leaders, but they nonetheless want leaders who will keep them safe and keep
their country secure. To the extent that the Daesh objective referenced in Treatment
C constitutes more of a threat to security and stability than do the Daesh objectives
referenced in Treatment B and, to some extent, Treatment D, the takeaway is that a con-
cern for political order, or a fear of political disorder, influences the way people think
about Daesh’s goals and assess the potential danger associated with each of them.
These various takeaways are informed by our data and analyses. They are also pre-
sumably plausible, and perhaps even persuasive. Nevertheless, they remain informed
speculations that suggest rather than demonstrate the meaning and significance of
observed variable relationships and underlying causal mechanisms.

Treatments and Drivers

The independent variables in the three hypotheses presented and tested earlier—​per-
sonal religiosity, membership in a key demographic, and political trust—​may condition
the impact of some experimental treatments. Findings about conditioning effects are
reported in the four parts of Table 22.B1 in the appendix, one part for each of the four ex-
perimental treatments. A brief summary of the analysis and findings is presented here.
The appendix offers a fuller discussion.
Table 22.B1 presents the results of OLS regression analyses in which the inde-
pendent variable in each of the three hypotheses and also each of the four experimental
treatments are included as nondependent variables, with the dependent variables again
being the three questions about the Islamic State in the Arab Barometer survey. The table
does not reveal a broad array of instances in which the impact of an experimental treat-
ment varies according to whether or not one or more of the three independent variables
is included in the regression analyses.
Mapping and Explaining Arab Attitudes toward the Islamic State     467

The relatively few cases where the statistical significance of an experimental treat-
ment does change in response to the inclusion of an independent variable are consistent
with previous findings and offer additional evidence of where statistically significant
relationships are to be found. Among the independent variables in the hypotheses
offered earlier, it is clearly political trust that has explanatory power. And among ex-
perimental treatments, it is Treatment A and Treatment C that most often both have
explanatory power and have that power diminish when examined in tandem with
political trust.

Discussion and Conclusion

The fourth wave of Arab Barometer surveys was carried out in eight countries in
2016–​2017. The interview schedule in seven of these countries contained questions and
an embedded experiment pertaining to the Islamic State: Algeria, Jordan, Lebanon,
Morocco, Palestine, Qatar, and Tunisia. The questions and experiment pertaining to the
Islamic State were not included in an eighth Wave IV country, Egypt. The descriptive in-
formation presented earlier, in Table 22.1, makes clear that support for the objectives and
actions of the Islamic State is generally very low, as is the proportion of Sunni Muslim
Arabs who believe that these objective and actions are compatible with the teachings of
Islam. There is, nevertheless, a small proportion of individuals who hold positive views
about the Islamic State, there being, therefore, some variance to be explained.
It is worth reflecting briefly on the question of whether the findings reported here
shed any light on the opinions and attitudes of Sunni Muslim Arab populations in other
countries, or perhaps even non-​Arab Muslim populations outside the MENA region.
Strictly speaking, the answer is negative. There is no way to know for sure whether at-
titudinal patterns and distributions observed in some locations describe patterns and
distributions in other locations without replicating the research in these other locations.
At the same time, findings from the research reported here may, with some caution, pro-
ductively inform thinking about the attitudes toward the Islamic State held by Sunni
Muslim Arabs in countries not included in the fourth wave of Arab Barometer surveys.
With Qatar being the only GCC country to be surveyed, generalization to other coun-
tries in the Gulf Cooperation Council is unlikely to be productive. By contrast, with four
Wave IV countries in North Africa, one of which is Egypt, and three Wave IV countries
in the Mashreq, the possibilities for cautious generalization, particularly if attentive to
scope conditions, should not be arbitrarily and completely dismissed.
The case for thinking about the applicability of findings beyond the Wave IV coun-
tries that were surveyed is further strengthened by the fact that the structure of our
analyses, including those pertaining to the survey experiment, do not consider the
country of which the respondent is a citizen. Rather, focusing on personal religiosity,
for example, or on level of political trust, the analyses do not include country as a var-
iable, and findings are thus based on a large multi-​country pool of data drawn from
468    Mark Tessler, Michael Robbins, and Amaney Jamal

careful surveys conducted across the MENA region. Accordingly, it would appear to be
appropriate, while awaiting additional data-​based investigations, of course, to enter-
tain strongly the possibility, if not the likelihood, that Wave IV findings about the ex-
planatory power of personal religiosity, political trust, or other variables are similar to
those that would be found among populations, among Sunni Muslim Arab populations
in this case, in other parts of the Middle East and North Africa—​and perhaps else-
where as well.
A number of the specific findings offered in the present research report were un-
expected. To an extent, this makes these findings more interesting and instruc-
tive, but it also places some limits on the degree to which the objectives of the study
have been achieved. Turning first to the hypotheses, it is surprising that personal re-
ligiosity does not account for significant variance in views about Daesh’s goals and
methods. And although the possibility is expressed in H1b, the fact that more religious
individuals are more likely than less religious individuals to believe that Daesh’s actions
and pronouncements are incompatible with the teachings of Islam is at least some-
what surprising. It is also a finding of considerable significance, and one for which the
possibilities for generalization would appear to deserve serious attention.
It is also somewhat surprising that we did not succeed in identifying a demographic
subset of the population that is disproportionately likely to be receptive to Daesh’s mes-
sage and to have more positive, or less negative, attitudes toward the Islamic State. The
logic being tested in Hypothesis 2 is that respondents who belong to the modal demo-
graphic profile of individuals who have left their home countries and joined the Islamic
State—​less well-​educated younger men, whom we have called the “key demographic,”
would be more likely than their countrymen with different demographic attributes to
have more positive attitudes toward the Islamic State. This logic, however persuasive it
might be in the abstract, does not appear to describe the facts on the ground, and this
is the case even though different cutting points for age were considered. The absence of
positive findings about a key demographic does provide useful information, but it prob-
ably falls short of our goal of identifying and shedding light on factors that predispose
some Sunni Muslim Arabs to hold positive views about the Islamic State.
The relationship between political trust and attitudes toward Daesh is strong and
statistically significant, but the direction of the relationship is not that proposed in
Hypothesis 3. H3 posited an inverse relationship: individuals with lower levels of polit-
ical trust are more likely to have positive attitudes about the Islamic State, presumably
because Daesh criticizes the system of government by which they are ruled and with
which they are discontent, and because Daesh then offers an alternative. But this is not
the case; individuals with lower levels of political trust have negative views about both
their government and about Daesh. Even if the direction of the relationship is not that
expected, this finding does identify and shed some light on factors that play a role in
shaping attitudes toward the Islamic State. Of potentially equal or even greater value
are the questions this finding raises about how political trust should be conceptualized
and about what is actually being measured by questions that ask about specific political
institutions.
Mapping and Explaining Arab Attitudes toward the Islamic State     469

Proposed for consideration is the possibility that this finding says as much about a
person’s worldview and general perspective on politics as it does about his or her judg-
ment of the specific political institutions of the person’s country. As in the case of per-
sonal religiosity and the incompatibility of Daesh actions with the teachings of Islam,
the finding about political trust points to an attitude-​shaping dynamic that had not
been expected but does have significant explanatory power and merits additional
consideration.
Our survey experiment has produced several interesting findings, but, overall, its con-
tribution is not as significant as we had hoped. None of the four experimental treatments
has a significant impact on all three of the questions about Daesh; no treatment has an
impact on the question about the teachings of Islam; one of the treatments, the one
about Daesh’s opposition to Iran and Shiʿi Muslims, has no impact on any of the three
questions about the Islamic State; and only one of the four treatments, the one about
Daesh’s opposition to secular Arab leaders, has a significant impact on both of the two
remaining questions, one about Daesh’s goals and one about its methods. Accordingly,
we learn that knowing more about the Islamic state’s objectives and actions and the con-
tent of its messaging on social media for the most part does not impact attitudes about
the extremist group.
But findings of no relationship are not necessarily without value. They help to rule
out plausible patterns that might otherwise be thought to have explanatory power. And
this being the case, they more clearly identify and call attention to the small number of
relationships that are both statistically significant and substantively and analytically im-
portant. In this category are the impact of Treatments A and C on views about Daesh’s
goals, and of Treatments C and D on views about Daesh’s methods. These findings, in-
cluding findings of no significant relationship, also provide a foundation for considering
whether the impact of the experimental treatments changes when the independent
variables in the hypotheses are included in the analysis. There are very few such changes,
however, and so it may be concluded that findings about the impact of the experimental
treatments is the same, or at least very similar, for both more religious and less religious
individuals, for both less well-​educated younger men and those who do not fit this dem-
ographic profile, and for both individuals with a higher level and individuals with a
lower level of political trust.
In sum, the value of our findings often lies in the doubt they cast on causal stories that
would otherwise be persuasive. Such findings are not without value. But there are also
some unexpected findings about significant and instructive relationships that might
otherwise have been missed. Among these are the relationship between personal religi-
osity and the view that Daesh’s actions and religious interpretations are not compatible
with the teachings of Islam. So, too, is the finding that lower political trust pushes to-
ward negative views of the Islamic State.
When the fourth wave of Arab Barometer surveys was conducted, in 2016–​2017,
and even more in 2015, when the instrument for Wave IV was being constructed and
pretested, the Islamic State was much stronger and more frightening than it is at the time
of this writing. It was only one year earlier, in June 2014, that the group’s leader, Abu Bakr
470    Mark Tessler, Michael Robbins, and Amaney Jamal

al Baghdadi, announced the formation of a caliphate stretching from Aleppo in Syria to

Diyala in Iraq and gave the group he led the name the Islamic State. The US carried out
airstrikes on the Islamic state during this period and had some success. But the group
made gains on the ground in Syria and remained firmly in control of its territory there.
Daesh also demonstrated its ability to hit overseas targets. In June 2015, for example,
Daesh claimed responsibility for a terrorist attack on a tourist resort in Sousse, Tunisia.
Thirty-​eight people were killed, thirty of whom were British tourists. This picture helps
to explain the Arab Barometer’s interest in asking questions about the Islamic State. But
it also raises questions about the possibility of scope conditions that specify when par-
ticular attitudes about Daesh and their determinants do and do not characterize Sunni
Muslim Arab populations. The present study thus provides baseline data for subsequent
investigations—​investigations that will shed light on the prospects for generalization
over time and offer thoughts about the temporal conditionalities that need to be made
part of causal stories about the way the Islamic State is seen and evaluated by ordinary
Sunni Muslim Arab citizens.

Notes
1. This is not to claim that ideological convergence is the only relevant factor, but simply that it
represents an important factor in winning sympathizers in Arab countries outside of those
in which the Islamic State is actively operating.
2. Arab Barometer surveys are based on face-​to-​face interviews with probability-​based and
nationally representative samples of men and women aged 18 or older. To date, there have
been five waves of Barometer surveys, more than 75,000 individuals have been interviewed,
and at least one survey has been conducted in fifteen different Arab countries. Data from all
Arab Barometer surveys are available for downloading from the Barometer’s website. The
website also contains an online analysis tool, through which can be observed the distribu-
tion of responses to any survey question in any country in any wave and across any one of
various demographic categories.
3. Extensive general and country-​specific methodological information is available on the
Arab Barometer website (arabbarometer.org).
4. Attention is not given to the possible reasons that support for Daesh is not as low in some
countries as it is in others. Some ideas about the determinants of cross-​country variance
might suggest themselves, but without controls to significantly reduce over-​determination,
it is impossible to assess the accuracy of these ideas.
5. These factor loadings are very comfortably beyond the generally accepted minimum of
+.60 or −.60 that is needed to establish that an item measures, at least approximately, what
the factor itself measures.
6. Trust in the army and in the police were initially included in the factor analysis. They both
had low loadings on the political trust factor and high loading on a second factor, indicating
that we have indeed measured political trust and not trust in governance more generally.
7. Responses to each question is a 4-​point scale. For the questions about Daesh’s goals and
methods, this ranges from agree to a large extent to disagree to a large extent. For the
Mapping and Explaining Arab Attitudes toward the Islamic State     471

question about the compatibility of Daesh’s actions with the teachings of Islam, the 4-​point
scale ranges from definitely compatible to definitely not compatible.
8. Additionally, as with all Arab Barometer surveys, the fourth-​wave questionnaire containing
the experiment was reviewed by the Barometer’s in-​country partners and if necessary re-
vised before beginning the survey.

References
ACRPS (Arab Center for Research and Policy Studies). 2015. “Arab Opinion Index 2015: In
Brief.” Al Dhaayen, Qatar: ACRPS. https://​www.dohainstitute.org/​en/​lists/​ACRPS-​
PDFDocumentLibrary/ ​ A rab_ ​ O pinion_ ​ I ndex_ ​ 2 015_ ​ R esults_ ​ i n_ ​ B rief_ ​ R EVISED_​
JUNE_​2016.pdf.
ACRPS (Arab Center for Research and Policy Studies). 2018. “The 2017–​2018 Arab Opinion
Index: Main Results in Brief.” Al Dhaayen, Qatar: ACRPS. https://​www.dohainstitute.org/​
en/​News/​Pages/​ACRPS-​Releases-​Arab-​Index-​2017-​2018.aspx.
Barber, Ben. “Feeding Refugees, or War?” Foreign Affairs 76, no. 4 (1997): 8–​14.
Beardsley, Kyle, and Brian McQuinn. “Rebel Groups as Predatory Organizations: The Political
Effects of the 2004 Tsunami in Indonesia and Sri Lanka.” Journal of Conflict Resolution 53,
no. 4 (2009): 624–​645.
Berman, Eli. 2003. “Hamas, Taliban and the Jewish Underground: An Economist’s View of
Radical Religious Militias.” NBER Working Paper 10004. Cambridge, MA: National Bureau
of Economic Research, September 2003.
Berman, Eli, and David D. Laitin. “Religion, Terrorism and Public Goods: Testing the Club
Model.” Journal of Public Economics 92, no. 10 (2008):1942–​1967.
Bhattacharya, Srobana. 2013. Who Supports Terrorism? A Comparative Study of Civilian
Support for Terrorist Groups. Carbondale: Southern Illinois University at Carbondale, 2013.
Bloom, Mia. Dying to Kill: The Allure of Suicide Terror. New York: Columbia University
Press, 2005.
Byman, Daniel, Peter Chalk, Bruce Hoffman, William Rosenau, and David Brannan. Trends in
Outside Support for Insurgent Movements. Santa Monica, CA: Rand Corporation, 2001.
Collier, Paul, and Anke Hoeffler. “Greed and Grievance in Civil War.” Oxford Economic Papers
56, no. 4 (2004): 563–​595.
Cordovez, Diego, and Selig S. Harrison. Out of Afghanistan: The Inside Story of the Soviet
Withdrawal. New York: Oxford University Press, 1995.
Davis, Paul K., and Kim Cragin. Social Science for Counterterrorism: Putting the Pieces Together.
Santa Monica, CA: Rand Corporation, 2009.
Flanigan, Shawn Teresa. “Charity as Resistance: Connections between Charity, Contentious
Politics, and Terror.” Studies in Conflict & Terrorism 29, no. 7 (2006): 641–​655.
Flanigan, Shawn Teresa, and Cheryl O’Brien. 2015. “Service-​Seeking Behavior, Perceptions
of Armed Actors, and Preferences Regarding Governance: Evidence from the Palestinian
Territories.” Studies in Conflict & Terrorism 38, no. 8 (2015): 622–​651.
Lipka, Michael. “Muslims and Islam: Key Findings in the U.S. and Around the World.” Pew
Research Center FacTank. Washington, DC: Pew Research Center, August 2017. https://​
www.pewresearch.org/​fact-​tank/​2017/​08/​09/​muslims-​and-​islam-​key-​findings-​in-​the-​u-​s-​
and-​around-​the-​world/​.
472    Mark Tessler, Michael Robbins, and Amaney Jamal

Poushter, Jacob. “In Nations with Significant Muslim Populations, Much Disdain for ISIS.”
Pew Research Center FacTank. Washington, DC: Pew Research Center, November 2015.
https://​www.pewresearch.org/​fact-​tank/​2015/​11/​17/​in-​nations-​with-​significant-​muslim-​
populations-​much-​disdain-​for-​isis/​.
Ranetunge, Dushy. “British Charities Fund Terrorists.” Society for Peace, Unity and Human
Rights Information Paper, October 2000.
Shain, Yossi, and Martin Sherman. “Dynamics of Disintegration: Diaspora, Secession and the
Paradox of Nation-​States.” Nations and Nationalism 4, no. 3 (1998): 321–​346.
Sheffer, Gabriel. “Ethno-​National Diasporas and Security.” Survival 36, no. 1 (1994): 60–​79.
Simon, Steven, and Jeff Martini. “Terrorism: Denying al-​ Qaeda Its Popular Support.”
Washington Quarterly 28, no. 1 (2004): 129–​145.
Urban, Mark. War in Afghanistan. New York: Palgrave Macmillan, 1987.
Weinstein, Jeremy M. Inside Rebellion: The Politics of Insurgent Violence. New York: Cambridge
University Press, 2007.

Appendix A
Embedded Experiment
The interview schedule used in the fourth wave of Arab Barometer surveys included
an embedded experiment that randomly assigned respondents to one of five groups: a
control group and four treatment groups. The four treatments, shown below, were
statements about the Islamic State that were read to respondents before they were asked
questions about Daesh’s goals, its methods, and the compatibility of its tactics with
the teachings of Islam. As can be seen, Treatment A references Daesh’s claims about
establishing an Islamic caliphate and then states that the group had killed many non-​
Muslims, and also Muslims, in pursuit of this goal. The three remaining treatments
include the text of Treatment A and then add text about one of Daesh’s declared
objectives: confronting Iran and Shiʿi Islam, confronting Arab leaders with a secular ori-
entation, and confronting Western imperialism in the Arab world.
Treatment A: As you may know, Daesh has emerged as a potent force in the region
and the world. In 2014 it declared a caliphate based in Raqqa, Syria. Daesh’s goal is to
extend the caliphate across the Muslim world and it has killed many Muslims and non-​
Muslims in pursuit of this aim.
Treatment B: As you may know, Daesh has emerged as a potent force in the region
and the world. In 2014 it declared a caliphate based in Raqqa, Syria. Daesh’s goal is to
extend the caliphate across the Muslim world and it has killed many Muslims and non-​
Muslims in pursuit of this aim. Another of Daesh’s stated objectives is to limit Shiʿa in-
fluence across the Muslim world as well as opposing Iranian-​led Shiʿa forces in Syria,
Iraq, Yemen, and elsewhere.
Treatment C: As you may know, Daesh has emerged as a potent force in the region
and the world. In 2014 it declared a caliphate based in Raqqa, Syria. Daesh’s goal is
to extend the caliphate across the Muslim world and it has killed many Muslims and
Mapping and Explaining Arab Attitudes toward the Islamic State     473

non-​Muslims in pursuit of this aim. Another of Daesh’s stated objectives is to defend

Islam from attacks by secular leaders and other elites whose goal is to limit the role of
Islam in government and public life.
Treatment D: As you may know, Daesh has emerged as a potent force in the region
and the world. In 2014 it declared a caliphate based in Raqqa, Syria. Daesh’s goal is to
extend the caliphate across the Muslim world and it has killed many Muslims and non-​
Muslims in pursuit of this aim. Another of Daesh’s stated objectives is to counter inter-
vention in the region by the United States and other Western powers who have engaged
in military attacks against it.

Appendix B
Treatments and Drivers
The independent variables in the hypotheses and causal stories presented and tested
earlier may condition the impact of some experimental treatments. The three variables
are personal religiosity, membership in what we have called a key demographic, and po-
litical trust. H1a and H1b are the hypotheses pertaining to personal religiosity. H2 deals
with membership in a particular demographic category: poorly educated younger men.
The independent variable in H3 is political trust. Findings about conditioning effects
are reported in the four parts of Table 22.B1, one part for each of the four experimental
treatments.
To recall earlier findings, hypotheses about the influence of personal religiosity are
confirmed in only one instance, when the dependent variable is the compatibility of
Daesh’s actions with the teachings of Islam. In this case, there is a statistically significant
relationship between greater personal religiosity and the view that the organization’s
tactics, meaning its use of violence, are not compatible with the teachings of the reli-
gion. Hypotheses about the significance of membership in the key demographic cate-
gory are not confirmed. A statistically significant relationship between political trust
and attitudes toward the Islamic State has been observed, but it is not in the inverse di-
rection proposed in H3. Rather, respondents with a lower level of political trust have
more negative attitudes toward the Islamic State.
Tables 22.B1a, 22.B1b, 22.B1c, and 22.B1d present the results of an OLS regression
analysis in which each of these three variables and also each of the four experimental
treatments are included as nondependent variables, with the dependent variables being
the three questions about the Islamic State in the Arab Barometer surveys: Daesh’s
goals, its methods, and whether its tactics are compatible with the teachings of Islam.
The principal objective of the analysis reported in Table 22.B1 is to determine whether
the statistical significance, and hence the explanatory power of each experimental treat-
ment, changes, becoming either stronger or weaker, when any one or all of the three in-
dependent variables is included in the regression. Thus, for example, the first finding in
Table 22.B1a, excluding the constant, shows that the impact of Treatment A on attitudes
Table 22.B1a: Impact of Treatment A with Other Variables in the Models
Attitudes toward the goals of the Islamic State

Constant 3.872 3.871 3.875 3.867 3.869

(.011) (.012) (.012) (.014) (.014)
Treatment A: Caliphate .032** .032** .031** .017 .017
and killings (.016) (.016) (.016) (.019) (.019)
Lower level of religiosity -​.003 -​.014
(.008) (.010)
Key demographic: -​.025 -​.010
Age 30 or younger (.024) (.029)
Lower level of .022** .025**
political trust (.010) (.010)

Attitudes toward the methods of the Islamic State

Constant 3.885 3.884 3.889 3.878 .3877
(.12) (.012) (.013) (.015) (.016)
Treatment A: Caliphate .014 .014 .014 -​.001 -​.001
and killings (.017 (.017 (.017 (.021) (.021)
Lower level of religiosity -​.001 -​.005
(.009) (.011)
Key demographic: -​.026 .006
Age 30 or younger (.025) (.033)
Lower level of .010 .011
political trust (.011) (.01)

Daesh actions compatible with teachings of Islam

Constant 3.806 3.807 3.806 3.788 3.790
(.015) (.015) (.016) (.019) (.020)
Treatment A: Caliphate .020 .019 .020 .007 .006
and killings (.022) (.022) (.022) (.026) (.013)
Lower level of religiosity -​.018* -​.029**
(.011) (.013)
Key demographic: -​.007 .012
Age 30 or younger (.032) (.040)
Lower level of .043*** .047***
political trust (.014) (.014)

a Models with personal religiosity and political trust were run with each of the two other age groups

defining the key demographic. The results were the same.

* p<.1
** p<.05
*** p<.01
Table 22.B1b: Impact of Treatment B with Other Variables in the Models
Attitudes toward the goals of the Islamic State

Constant 3.872 3.871 3.875 3.869 3.871

(.012) (.017) (.012) (.014) (.015)
Treatment B: Oppose Iran .005 .006 .004 -​009 -​.010*
and Shiʿa (.017) (.017) (.017) (0.20) (.020)
Lower level of religiosity -​.010 -​.017
(.008) (.010)
Key demographic: -​.028 -​.003
Age 30 or younger (.026) (.032)
Lower level of .005 .007
political trust (.011) (.011)

Attitudes toward the methods of the Islamic State

Constant 3.885 3.884 3.891 3.881
(.011) (.012) (.012) (.015)
Treatment B: Oppose Iran .025 .026 .024 .016
and Shiʿa (.016) (.016) (.016) (.020)
Lower level of religiosity -​.005 -​.009
(.008) (.010)
Key demographic: -​.043* -​.014
Age 30 or younger (.025) (.031)
Lower level of -​4.41E-​5 .002
political trust (.010) (.011)

Daesh actions compatible with teachings of Islam

Constant 3.806 3.808 3.810 3.792 3.797
(.016) (.016) (.016) (.019) (.019)
Treatment B: Oppose Iran .004 .003 .003 .001 -​.003
and Shiʿa (.022) (.022) (.022) (.026) (.026)
Lower level of religiosity -​.039*** -​.042***
(.011) (.013)
Key demographic: -​.032 -​.003
Age 30 or younger (.033) (.041)
Lower level of .018 .023*
political trust (.014) (.014)

a.Models with personal religiosity and political trust were run with each of the two other age groups
defining the key demographic. The results were the same.
* p<.1
** p<.05
*** p<.01
Table 22.B1c: Impact of Treatment C with Other Variables in the Models
Attitudes toward the goals of the Islamic State

Constant 3.872 3.871 3.875 3.869 3.870

(.011) (.011) (.012) (.014) (.014)
Treatment C: Oppose .031** .033** .030** .011 .014
secular Arab leaders (.016) (.016) (.016) (.019) (.019)
Lower level of religiosity -​.008 -​.012
(.008) (.009)
Key demographic: -​.026 .001
Age 30 or younger (.024) (.029)
Lower level of .005 .007
political trust (.010) (.010)

Attitudes toward the methods of the Islamic State

Constant 3.885 3.885 3.890 3.878 3.879
(.011) (.011) (.012) (.014) (.014)
Treatment C: Oppose .034** .037** .034** .023 .026
secular Arab leaders (.016) (.016) (.016) (.019) (.019)
Lower level of religiosity -​.013* -​.017*
(.008) (.009)
Key demographic: -​.037 -​.001
Age 30 or younger (.024) (.030)
Lower level of .012 .015
political trust (.010) (.010)

Daesh actions compatible with teachings of Islam

Constant 3.806 3.807 3.812 3.789 3.795
(.015) (.015) (.016) (.019) (.019)
Treatment C: Oppose .019 .021 .019 .009 .011
secular Arab leaders (.021) (.021) (.021) (.026) (.026)
Lower level of religiosity -​.024** -​.024*
(.010) (.013)
Key demographic: -​.047 -​.026
Age 30 or younger (.032) (.040)
Lower level of .036*** .042***
political trust (.013) (.014)

a.Models with personal religiosity and political trust were run with each of the two other age groups
defining the key demographic. The results were the same.
* p<.1
** p<.05
*** p<.01
Table B1d: Impact of Treatment D with Other Variables in the Models
Attitudes toward the goals of the Islamic State

Constant 3.872 3.871 3.874 3.868 3.869

(.012) (.012) (.012) (.014) (.014)
Treatment D: Oppose .021 .021 .022 .014 .013
Western interference (.016) (.017) (.016) (.019) (.019)
Lower level of religiosity -​.002 -​.006
(.008) (.009)
Key demographic: -​.020 -​.006
Age 30 or younger (.024) (.030)
Lower level of .013 0.014
political trust (.010) (.010)

Attitudes toward the methods of the Islamic State

Constant 3.885 3.884 3.892 3.879 3.881
(.012) (.012) (.012) (.014) (.015)
Treatment D: Oppose .023 .024 .032 .020 .021
Western interference (.016) (.017) (.016) (.020) (.020)
Lower level of religiosity .002 -​.002
(.008) (.010)
Key demographic: -​.048* -​.017
Age 30 or younger (.024) (.030)
Lower level of .004 .005
political trust (.010) (.011)

Daesh actions compatible with teachings of Islam

Constant 3.806 3.807 3.805 3.789 3.790
(.015) (.015) (.016) (.019) (.019)
Treatment D: Oppose .030 .028 .030 .020 .019
Western interference (.022) (.022) (.022) (.026) (.026)
Lower level of religiosity -​.029*** -​.038***
(.010) (.013)
Key demographic: .002 .026
Age 30 or younger (.032) (.040)
Lower level of .036*** .040***
political trust (.014) (.014)

a.Models with personal religiosity and political trust were run with each of the two other age groups
defining the key demographic. The results were the same.
* p<.1
** p<.05
*** p<.01
478    Mark Tessler, Michael Robbins, and Amaney Jamal

toward the Islamic State’s goals is statistically significant when none of the independent
variables is included in the model, and that it remains significant when either personal
religiosity or membership in the key demographic is included in a model but then
loses statistical significance when political trust is included. Political trust apparently
overrides the impact of the treatment. More specifically, when political trust is low,
meaning among individuals with a lower level of political trust, receiving Treatment
A does not significantly alter a person’s attitudes toward the goals of the Islamic State.
Table 22.B1 shows that the inclusion of the independent variables in most cases
does not affect the relationship between an experimental treatment and attitudes to-
ward the Islamic State. The pattern reported above, however, about the relationship
between Treatment A and attitudes about Daesh’s goals, is also found in two other
instances: when the treatment is Treatment C and the dependent variable is attitudes
toward either Daesh’s goals or the organization’s tactics. As shown in Table 22.B1c, the
impact of Treatment C is statistically significant when political trust is not included in
the analysis, but ceases to be significant when it is included. The only other instance of a
change in the impact of an experimental treatment is shown in Table 22.B1b. Treatment
B does not have a statistically significant impact when there is no independent variable
in the analysis, and this remains the case when any one of the independent variables is
included. But Treatment B becomes significant, though only at the p < .10 level, when
the dependent variable is Daesh’s goals and the model includes all three of the inde-
pendent variables.
Table 22.B1 does not reveal a broad array of instances in which the impact of an ex-
perimental treatment varies according to whether or not one or more of the three inde-
pendent variables is included in the regression analysis. As noted above, the inclusion
of the independent variables in most cases does not affect the relationship between an
experimental treatment and attitudes toward the Islamic State. In other words, the im-
pact of the treatment, whether statistically significant or not, is not different for a subset
of the population based on personal religiosity, demographic profile, or level of political
trust than it is for the population as a whole. There are a few cases where the statistical
significance of an experimental treatment does change in response to the inclusion of
an independent variable, and these are consistent with previous findings and offer ad-
ditional evidence of where statistically significant relationships are to be found. Among
the independent variables in the hypotheses offered earlier, it is clearly political trust
that has explanatory power. And among experimental treatments, it is Treatment A and
Treatment C that most often both have explanatory power and have that power diminish
when examined in tandem with political trust.
 Pa rt I V

S O C IA L
M OB I L I Z AT ION
 Chapter 23

So cial Mov e me nts ,

Parties, and P ol i t i c a l
Cleavages in Moro c c o
A Religious Divide?

Adria Lawrence

How do religious cleavages work in states that are predominantly Muslim? In states that
include adherents to multiple world religions, religion marks off different communities
of belief. In predominantly Muslim societies that are relatively homogenous with re-
spect to religion, the question that sets opposition groups apart is the degree to which
religion should influence politics. Adherents of political Islam seek a greater role for
religion in the political life of the country, while secular groups do not (Ayoob 2004, 1;
Schwedler 2013). Specified in this way, religious and secular ideologies are alternatives
to one another. Organizations orient themselves in the political landscape according
to their views on the appropriate role of religion in politics: they may self-​identify as
Islamist or draw on secular frames such as class, nationalism, and regionalism.
This understanding of political cleavages as dimensions along which political organ-
izations’ ideologies can be plotted originated in the study of political party formation
in democracies.1 These studies focused on the relationship between voters and political
parties, analyzing how parties established programmatic identities that allowed them
to appeal to supporters on the basis of their ideological commitments—​on where they
stood on questions such as what role religion should play in state laws and policies.
Yet the logic of political cleavages in non-​democracies is different. This chapter
focuses on predominantly Muslim states ruled by electoral authoritarian regimes.
Electoral authoritarian regimes are the most common type of authoritarianism in the
contemporary world. They have institutional features of representative democracies,
such as elections and political parties, but these institutions do not provide the op-
position with the opportunity to lead (Schedler 2015). Unlike democracies or hybrid
regimes, political authority under electoral authoritarianism resides in a regime that
does not alternate in power.2
482   Adria Lawrence

This chapter makes three main claims about political cleavages under electoral au-
thoritarianism. First, it argues that programmatic differences such as secular versus reli-
gious ideologies matter less than the stance an opposition group takes toward the ruling
regime. In a democracy, parties are primarily concerned with capturing votes and win-
ning elections, but in an authoritarian regime, opposition groups must contend with the
regime itself. The primary political question in these settings is whether to adopt a more
cooperative or confrontational approach to the regime; this cleavage divides the oppo-
sition, with some actors opting to participate in the pseudo-​democratic institutions in
place, while others criticize from outside those institutions.
This cleavage is only apparent if both political parties and groups engaged in conten-
tious politics are included in the analysis. Thus, the second claim this chapter makes
is that studying both formal and contentious politics is useful for understanding how
opposition works under electoral authoritarianism. A study that includes both political
parties and social movements is better able to show the range of political opposition in
an authoritarian country than one that focuses exclusively on either formal or informal
politics. For instance, a study that examined only political parties might overemphasize
ideological differences among the parties because it omits, by design, opposition groups
that have not formed political parties. In this chapter, I show that social movement or-
ganizations (SMOs) have common goals even when they differ ideologically. Likewise,
an Islamist political party can have more in common with a leftist political party than it
does with an Islamist SMO.
The third, related claim I make in this chapter is that political parties and SMOs do
not have to choose to be either religious or secular; they can be both. In a predominantly
Muslim country, religious and secular frames are not mutually exclusive alternatives—​
they do not exist at opposite ends of a spectrum. The lines between religious and sec-
ular political aims are frequently blurred. Denoeux (2002, 6) defines political Islam as
“a form of instrumentalization of Islam by individuals, groups and organizations that
pursue political objectives.” In a predominantly Muslim country, all political parties and
SMOs may seek to use religion instrumentally, regardless of whether they are explicitly
identify as Islamist or not. Secular opposition groups can draw on Islam to justify their
policy platforms, recruit followers, or criticize the regime. Their main concern is not al-
ways limiting the role of religion in politics. Further, Islamist organizations do not solely
or consistently advocate for deepening the role of religion in the political sphere. They
have secular goals that they may share with some non-​Islamist groups.
The difference between explicitly Islamist-​oriented groups and other groups is not
reducible to their respective positions on the desirable role of religion in politics. For
this reason, as Pepinsky writes in this volume, “voting for an Islamic party is not always a
vote for Islam, and voting for a non-​Islamic party sometimes is.” Similarly, secular goals,
such as reducing unemployment or poverty, can lead a person to support an Islamist
SMO, just as religious reasons can lead to support for SMOs that are considered secular
in orientation.
This essay draws on the case of Morocco to illustrate these claims. Morocco is an
electoral authoritarian state; it is headed by a hereditary monarch and has a parliament
 Social Movements, Parties, and Political Cleavages in Morocco     483

with limited powers to which parties are elected.3 The population is 99 percent Sunni
Muslim, and over 90 percent of Moroccans self-​describe as either religious or somewhat
religious (Arab Barometer 2017). Morocco has had both secular and Islamist political
actors, making it a good exploratory case for investigating the role of religion in political
cleavages.
The next section begins with a historical overview of political opposition in Morocco
that sets up the empirical questions to be addressed. It then discusses political opposi-
tion in three different time periods. I begin with the anti-​colonial movement and show
how activists relied on religion to advance an ostensibly secular nationalist cause. Next
I discuss the postcolonial period, considering the leftists and the two most prominent
Islamist organizations, al-​ʿAdala wa Tanmiya (the Party of Justice and Development, or
PJD) and al-​ʿAdl wal-​Ihsan (the Justice and Charity Group, or JCG). I show that Islamist
organizations do not consistently seek to change the religious orientation of the political
system, and they share some political goals with leftists. I then discuss the 2011 protests,
describing how political actors attempted to align their goals with religious principles.
As Ayoob (2004, 1–​2) writes, “the political manifestations of Islam, like the practice of
Islam itself, are to a great extent context specific, the result of the interpenetration of re-
ligious precepts and local culture, including political culture.” Yet though context shapes
the role of Islam in politics, we can and do observe commonalities across different cases.
The arguments in this chapter are stated in general terms and may be assessed in other
settings. In principle, the arguments apply to any predominantly Muslim state with an
electoral authoritarian regime.4 In my empirical section, I consider whether it is reason-
able to treat Morocco as comparable to other states with electoral authoritarian regimes.
A hereditary monarchy is one type of electoral authoritarianism. If Islamism is different
in monarchies than other types of authoritarian regimes, the argument may not gen-
eralize. I consider whether the Islamist opposition is exceptionally weak in Morocco
because of the king’s religious legitimacy, and I argue that the presence of a religious
head of state has not foreclosed opposition on religious grounds. I argue against the view
that there is an ex ante reason to dismiss the potential wider applicability of the insights
I derive from the Moroccan case, and I draw on new scholarship on other electoral au-
thoritarian regimes that suggests the ability of these claims to travel elsewhere. The con-
clusion situates the theoretical points made here in the growing literature on the rise of
political Islam.

Political Cleavages and Electoral

Authoritarianism in Morocco

The role of Islam in Moroccan politics initially appears to conform well to the ideas that
(a) the main difference between Islamist and non-​Islamist organizations lies in their op-
posing views on the extent to which Islam should inform politics, and (b) secular and
484   Adria Lawrence

religious frames are alternatives to one another, with some frames more prominent in
particular periods than others. Indeed, when I began to study the role of Islamist frames
in Moroccan politics, my initial thought was to conceptualize religious and secular
frames in this way—​as alternatives that resonate more in some periods than others—​
and to trace the appeal of Islamist versus secular ideologies over time.
This approach made sense because a quick overview of political opposition over the
last century suggests that the importance of religion in Morocco has waxed and waned
over time. Secular and Islamist movements and parties have been dominant at dif-
ferent points in time. A loose periodization of opposition activity in Morocco starting
from the French conquest to the present day could easily be described in terms of the
rise, fall, and rise again of Islamist frames for collective action. The French conquest
prompted a salafiyya movement that saw the calamity of conquest as a punishment
for departing from the correct practice of Islam (Eickelman 1985; Munson 1993). This
movement evolved into a reform movement seeking rights within the framework of the
French protectorate; when reform efforts failed, the movement became a nationalist one
seeking independence (Lawrence 2012, 2013). Nationalism dominated starting from the
Second World War, lasting into the postcolonial era, with the Istiqlal (Independence)
Party forming an alliance with the new king. Then, starting in the 1970s, leftist groups
mounted a challenge that culminated in the formation of a political party, the Socialist
Union of Popular Forces (USFP), which eventually won elections in 1998. The shift
from contentious politics to inclusion in formal politics as a political party resulted in
co-​optation (on this process, see Lust 2006). With the waning of the leftist opposition,
Islamist groups gained in popularity. Abdesssalam Yassine founded al-​ʿAdl wal-​Ihsan
(JCG), an Islamist/​Sufi social movement group, in 1987 (Daadaoui 2016). The Islamist
political party al-​ʿAdala wa Tanmiya (PJD) also benefited from the cooptation of the
leftists, increasing its electoral gains starting in 2002. In 2011, protests spread across
Morocco, organized by a coalition of human rights groups, leftists, and Islamists that
resulted in a new constitution and the victory of the PJD in the 2011 parliamentary
elections.
It is possible to recount this history as one in which secular and Islamist frames have
competed with one another over time, with arguments for or against an increased role
for Islam in the political life of the country resonating more in some periods than in
others. This section seeks to complicate this narrative. The ideologies that have been
invoked by Moroccan social movements and political parties—​nationalism, region-
alism, socialism, human rights, and Islamism—​are not a menu of options from which
only one must be chosen. I suggest that religion often plays an important role, regardless
of whether an SMO’s or party’s dominant frame is secular or religious. To some extent,
almost all political activists in Morocco are Islamist, because they invoke Islam to justify
their platforms, and because they rely on mosques and religious networks to mobilize.
Religious arguments and counterarguments are pervasive among political organiza-
tions in Morocco. The differences between them are differences of emphasis and degree,
not a clash between religious and secular positions. The main cleavage in Morocco is not
a religious one, despite the presence of groups that self-​define as Islamist or secular. The
 Social Movements, Parties, and Political Cleavages in Morocco     485

primary question that divides opposition groups in Morocco is whether to challenge the
authoritarian regime or work within the limitations it imposes.

Islam and Anti-​C olonialism

An examination of political opposition during the protectorate era offers the opportu-
nity to look at political cleavages under a different authoritarian system than the one
that was established at independence. The French protectorate was declared in 1912. It
left the sultan in place; he worked with the colonial administration and acted as a figure-
head. During the period, the French created some electoral institutions, holding local
elections in 1947 and 1951 for chambers that served an advisory purpose (Lawrence 2013,
198–​200).5
One initial reaction to the French conquest resembles responses to colonial incursions
elsewhere in the Muslim world: a group of Moroccan elites, influenced by the Egyptian
thinkers Djamal ad-​Din al-​Afghani and Sheikh Mohamed Abdou, saw the conquest as
an indicator that the country had become vulnerable to external domination because
it had departed from the true practice of Islam (Halstead 1964, 437). Accordingly, these
elites, who were largely educated at the Qarawiyyin Mosque in Fes, were inspired by
Salafist thought to promote religious reforms (Ait Mous 2013, 744; Joffé 1985, 290).
They focused their efforts not on the French invaders, but on opposing the worship of
marabouts (saints) and the religious brotherhoods that they argued had distorted and
corrupted the religious faith (Halstead 1964, 440).
Members of this group shifted toward a stance of heightened confrontation with
the French when the “Berber Dahir” crisis erupted in 1930. The dahir (decree) issued
by the French was intended to codify existing customary laws in the Amazigh (called
Berber at the time) areas of Morocco; it was a short document, not intended to alter
the law but to preserve the status quo (Hoisington 1978, 434). A group of urban elite
activists, including former members of the salafiyya movement such as Allal al-​Fasi,
politicized the dahir, arguing that it was an attempt to divide Berbers and Arabs and
limit the sultan’s sovereignty in Amazigh areas. They depicted the decree as a threat to
the umma (Muslim community) because it removed the Berbers from the jurisdiction
of shariʿa. Rumors circulated that the intention was to convert Berbers to Catholicism
(Lawrence 2012, 477; Wyrtzen 2015, 2). The crisis gained international attention when
Chakib Arslan, a Geneva-​based pan-​Islamic activist, took up the cause and began to
publicize it across the Muslim world as a French effort aimed at the de-​Islamization of
the Berber population (Wyrtzen 2015, 2).
Protests erupted in multiple towns, and the prayer of latif (said in times of calamity)
was recited in mosques throughout the country (see Hoisington 1978; LaFuente 1999;
Pennell 2000). The French authorities, in partnership with local leaders, suppressed
the protests and had the sultan send a letter to be read in mosques in Fes, Meknes,
Salé, Rabat, Casablanca, and Marrakech. In the letter, the sultan described the decree
486   Adria Lawrence

as a simple administrative measure, and he forbade the use of mosques for political
discussions (Hoisington 1978, 437).
Though protests subsided, the crisis had two important lasting effects. First, as
Wyrtzen (2015) argues, it served to shape Moroccan national identity in ways that have
persisted into the contemporary era, politicizing categories such as “Arab” and “Berber,”
which produced, over the course of the colonial period, a dominant identity centered on
Arab-​Islamic high culture.
Second, the protests prompted ongoing mobilization, as the (primarily) Arab elites
who had politicized the dahir established a reformist group, the Kutlah al-​Amal al-​
Watani, composed of religious and literary groups seeking reforms from the protec-
torate administration. These religious leaders made secular demands; the reforms they
sought were inspired by leftist platforms in France and by the promises of the protec-
torate treaty itself. Some of their demands included the appointment and promotion
of Moroccans in all branches of the administration, freedom of the press, freedom to
assemble, improvements to the educational system, and the establishment of municipal
councils and a national council elected by the population.6
The movement for reform was not a mass-​based social movement, but an elite effort
with limited popular support. Mass action began during World War II, after the leaders
of the reform movement formed the Hizb al-​Istiqlal (Independence Party) and began
organizing protests in favor of independence. The independence movement was broad-​
based and relied on protests, strikes, boycotts, petitions, elections, and, at the close of the
colonial period, rural insurgency and urban violence.7
The shift from a religious revivalist movement, to a reformist effort, to an explicitly
nationalist mass movement could be interpreted as suggesting a decline in the cen-
trality of religion to politics over the course of the colonial period. Yet neither the re-
formist movement nor the nationalist movement can accurately be coded as secular.
While the reformists advocated for changes that drew upon the democratic principles
that the French claimed to be importing to North Africa, they were not secular in orien-
tation. The majority were learned elites trained at the Qarawiyyin or at the other great
institutions of learning in the Islamic world, such as al-​Azhar. Many were religious
scholars. While some of their goals may be described as secular, they also sought to in-
crease the number of Arabic schools, and they stressed Islam as a core part of Moroccan
identity. This was not a secularist effort.
The nationalist movement initially appears to be a better candidate for coding as sec-
ular. Nationalism is a flexible ideology: it is the straightforward principle that the nation
has the right to exist.8 This flexibility helps explain its widespread appeal; at the time of
the Moroccan independence movement, nationalism was salient across the globe. The
principle of nationalism does not require that the nation have any particular character-
istic; the nation’s identity could be based on religion, language, ethnicity, shared history,
or any experience that generates a sense of solidarity (Renan 1996 [1882]).
Yet in Morocco, both the content of the movement and its strategies reflected the cen-
trality of Islam to political mobilization. Religion was an important part of the claim to
nationhood. The protectorate officials had, since their arrival, coded the population by
 Social Movements, Parties, and Political Cleavages in Morocco     487

religion, crafting their reports around the European, Muslim, and Jewish communities.
Moroccans, too, saw Islam as an important part of national identity; and in their
speeches, nationalist leaders suggested that defending the nation was a religious duty.9
Some scholars have suggested that members understood the national party as a type of
religious brotherhood (Halstead 1969, 252; Rivet 1999, 369–​371). Moroccans also relied
upon religious sites and holidays to mobilize effectively. Nationalist demonstrations
often began at the mosque; after prayer, a group would gather and march through the
streets. Mosques also provided meeting places and places of refuge when repression
occurred. Demonstrations also occurred with greater frequency during religious hol-
idays, such as Aïd es-​Seghir, Aïd al-​Kebir, and Aïd Mouloud (Lawrence 2013, 203–​204).
Moroccan nationalists were not secularists. In important ways, the Moroccan nation-
alist movement was also a religious movement.
Mobilization throughout the colonial period incorporated both religious and secular
appeals and goals. Religious and secular claims were not in competition with one an-
other, but were complementary—​both secular and religious reasons could be marshaled
to oppose French policies and ultimately the French presence. What did change over the
course of the period was the stance that Moroccan leaders took toward the protectorate.
Initially, reformers focused on religious advocacy that did not challenge the French re-
gime.10 During the reform period, they directly criticized French policy, pointing to
discrepancies between colonial aims and their actual practices. During World War II,
the movement changed. Instead of seeking reform within the context of the protec-
torate, the nationalist movement rejected French rule altogether and fought to end it.
Moroccans drew on both religious and secular ideologies to protest colonial rule. This
makes sense, since the French were foreigners who were not co-​religionists. The next
section turns to the post-​independence period, when secular and religious objectives
might be expected to come into conflict more than they did under colonial rule. In
the postcolonial era, the ability of political groups to combine religious and secular
rationales is less intuitive.

Postcolonial Political Cleavages

In the years following independence, social movement actors and political parties in
Morocco primarily identified as either Islamist or leftist, which suggests that the secular-​
religious divide was the main political cleavage. Yet a focus on self-​definitions is mis-
leading. The primary cleavage that has defined Moroccan politics in the postcolonial
period is not a religious one. Instead, it is the divide between royalists and those who
have opposed the monarchy’s move to monopolize political power.
In support of this claim, this section considers the orientation of both social move-
ment organizations and political parties. Both Islamist and leftist groups are divided in
their stances toward the monarchy. Further, conditional on their orientation toward the
monarchy, leftists and Islamists have taken similar positions on questions of political
488   Adria Lawrence

participation and policy. Their goals are not reducible to claims about the extent to
which religion should play a role in politics in the kingdom.

Leftist Opposition in the Postcolonial Era

With the arrival of independence in 1956, King Mohammed V sought to limit the power
of the Hizb al-​Istiqlal and gain control over other groups that had sprung up during the
final years of the anti-​colonial struggle.11 In 1959 a leftist party was created when Mehdi
Ben Barka, Abbderrahim Bouabid, and Abderrahmane Youssoufi split with the Istiqlal
to form the National Union of Popular Forces (Union Nationale des Forces Populaires,
or UNFP). Its platform included agrarian reform and democracy. The UNFP took a
more confrontational stance toward the monarchy than the Istiqlal party, led by Allal
al-​Fasi. With the death of the king in 1961, al-​Fasi supported the new king, Hassan II,
calling for national unity (Miller 2013, 163). Yet in creating the 1962 constitution, the
king sought to ensure that the Istiqlal would not retain its dominance. The constitution
prohibited a single ruling party (Buehler 2018, 46; Waterbury 1970, 256). The king also
built loyalist parties led by palace insiders to offset opposition parties, a strategy that has
continued to the present day (Buehler 2018, 47).
While the Istiqlal accepted the new constitution, the UNFP opposed it, with Ben
Barka declaring that “the primary task of the Moroccan people is to battle this totally
feudal regime” (quoted in Miller 2013, 167). Ben Barka’s (1966) essay, The Revolutionary
Option in Morocco, laid out his confrontational strategy, and led to widespread arrests
of UNFP militants and the declaration of a state of emergency, which lasted from 1965
to 1970. Ben Barka himself was kidnapped on the streets of Paris in 1965 and never seen
again, presumably assassinated.
The UNFP was not the only organization confronting the regime; during the late
1960s and early 1970s, protests and strikes occurred on university campuses, organ-
ized by the Moroccan National Students Union. High school students also carried out
demonstrations (Miller 2013, 168). On the far left, an organization called Ila al-​Amam
(To the Forefront) opposed the monarchy and faced severe repression in the early 1970s.
One of its founders, Abraham Sefarty, was condemned to life imprisonment in 1977, an
event that some suggest marks the beginning of what is known as the “years of lead”
(approximately 1975–​1990) during which the regime imprisoned and tortured political
opponents (Miller 2013, 170).
Not all leftists agreed with these groups’ confrontational stance toward the mon-
archy. Bouabid, formerly of the UNFP, established the Socialist Union of Popular Forces
(Union Socialiste des Forces Populaires, or USFP) in Rabat; it renounced violence and
announced that it would participate in electoral politics (Buehler 2018, 61–​62). The USFP
advocated agrarian reform and state economic planning. Its secularist stance included
refusing to wear djellabas for a royal speech on the grounds that these cloaks were associ-
ated with Islamic symbols of monarchy. This move was not so much a rejection of religion
as a rejection of the monarchy’s appropriation of traditional and religious symbols.
 Social Movements, Parties, and Political Cleavages in Morocco     489

The monarchy prevented the USFP from gaining too much electoral power after the
state of emergency ended and through the 1980s by fielding loyalist candidates running
as “independents” and creating royalist parties. The monarchy also controlled most
ministry appointments. In the 1980s, as Lust (2006, 130) writes, political contestation
remained primarily between the king and the parties, although bread riots and other
contentious strikes and protests occurred during the decade in response to neoliberal
reforms. In 1996, constitutional reform reduced the number of king-​appointed minis-
ters, which motivated the USFP to enter the government. In 1998 the USFP, led by the
former political prisoner and founder of the UNFP Abderrahman al-​Youssoufi, formed
the Alternance government, leading the parliament for the first time. Al-​Youssoufi’s ap-
pointment as prime minister prompted hopes that a genuine democratic transition had
begun. The USFP, however, lacked a majority and had to share power with rural loyalist
parties. It did not control important ministries and could not implement deep reforms.
Ultimately, the USFP came to be seen as a party that had been co-​opted by the mon-
archy, and it faced competition from Islamist opponents.

Islamist Opposition in the Postcolonial Era

In 1969 ʿAbd al-​Karim Moutiʾ, a former UNFP member, founded the al-​Shabiba al-​
Islamiyya (Islamic Youth) organization, which focused on clandestine activities, as
well as public and educational initiatives (Clark and Dalmasso 2015; Miller 2013, 188).
A dozen members of this group withdrew to create the Islamic Group, which fo-
cused solely on social and educational activities, renounced the use of violence, and
recognized the monarchy as legitimate. This group created At-​Tawhid wa al-​Islah (the
Unity and Reform Movement) in the 1980s, which sought official recognition and the
right to form a political party (Willis 2004, 55). They entered electoral politics in 1996, in
partnership with an existing political party. In 1998, they renamed their organization the
Party of Justice and Development (PJD), keeping the Unity and Reform Movement as a
social movement organization (Clark and Dalmasso 2015, 194–​95).
The PJD won nine seats in parliament in 1997, and was invited to join the coalition
government led by the USFP. The party declined, preferring not to associate itself with a
government whose chances of success were uncertain. It adopted a position of “critical
support” for the new government in the hope that this would allow the party to back al-​
Youssoufi’s reformist agenda while maintaining the flexibility to distance itself from the
new government’s failures (Willis 2004, 56).
The PJD’s political aims differed from the USFP’s in several respects. While the
leftists opposed the regime’s neoliberal reforms, resisting budget austerity and privati-
zation, the PJD offered moderate support for the regime’s agenda. PJD leader Abdelilah
Benkirane stated, “We believe in the free market, but we think the state needs to take up
its role to restore economic balance and justice in society. We support pro-​social neolib-
eralism” (quoted in Buehler 2018, 65). The PJD also differed in its stance on reforming
the Moudawana, the law governing the family, opposing proposals to ban polygamy,
490   Adria Lawrence

raise the minimum age for marriage, and alter the right of husbands to repudiate their
wives (Willis 2004, 56). While the PJD’s religious orientation shaped the latter stance, its
economic policies had little to do with religion.
The PJD and the USFP’s Alternance government also converged in important areas.
Both had opposed Morocco’s pro-​Western stance during the 1991 US war in Iraq; their
pro-​Palestinian views were shared by most parties in parliament, and they opposed
the creation of a palace-​sponsored political party (Buehler 2018, 68; Willis 2004, 56).
Further, their overall stance toward the royal palace and the system of government was
the same: while pushing for moderate reforms, neither the USFP nor the PJD has ever
challenged the monarchy’s political dominance. In contrast to both leftist and Islamist
groups that operate outside the party structure, the USFP and the PJD accept the institu-
tional rules set in place by the monarchy.
The other major Islamist force in Moroccan politics is the JCG, an association that
operates outside formal institutional channels and reportedly has more supporters
than the PJD, although its actual membership is unknown. The JCG differs from other
Islamist social movements in the wider Muslim world, largely because it was organized
by a charismatic individual, the former high school teacher and Sufi mystic Abdessalam
Yassine (Willis 2004, 58). Yassine gained notoriety when he sent a public letter to Hassan
II in 1974 in which he questioned the king’s religious legitimacy and called upon him to
repent and “return to God” (Miller 2013, 189). In response, Yassine was sent to a mental
hospital for three years. When he was released, he was placed under house arrest, but he
continued to build his movement, creating the JCG in 1987.
Since its inception, the JCG has operated as a social movement, with a strong base
among university students and in the cities, despite the fact that it was outlawed in 1990.
The association’s success has depended upon its ability to provide a range of social serv-
ices, such as healthcare, basic welfare, blood banks, and literacy courses (Daadaoui 2016;
Lust 2006, 139). It has elected to remain outside institutional party structures. After his
release from house arrest in 2000, Yassine told the press that his association “is a move-
ment focused on spiritual education, not a political party . . . but we are partly inter-
ested in politics” (quoted in Daadaoui 2016). Yassine continued to confront the king
when Mohammad VI took the throne after his father’s death in 1999, criticizing the
monarchy’s accumulation of wealth, and directly attacking the king’s religious legiti-
macy by calling the traditional ceremony in which elites pledge their allegiance, “a sacri-
legious ceremony in which the king is worshipped” (quoted in Daadaoui 2016).
The association’s political stance has had more in common with the banned leftist
parties like the UNFP or Sefarty’s Ila al-​Amam than with the PJD, despite sharing an
explicitly religious orientation.12 Yassine, like Ben Barka and Sefarty, explicitly criticized
the king’s monopolization of wealth and power. His concern, like theirs, was with the
system of rule in Morocco, and he too was repressed for directly confronting the king.
As Lust (2006) has argued, a key feature of the Moroccan opposition lies in the distinc-
tion between groups that operate within the formal political system and those that exist
outside it. Yassine’s association, like the leftist organizations that preceded it, is anti-​
royalist and concerned with social justice and economic redistribution. The difference
 Social Movements, Parties, and Political Cleavages in Morocco     491

between the leftists and the JCG is that Yassine, in his publications and texts, marshals
religious arguments to defend these goals.

Islamists and Leftists in Morocco’s

2011 Uprising

In February 2011, as popular uprisings spread across the Arab world, Moroccans took to
the streets in approximately sixty cities and towns.13 Buehler (2018), seeking to explain
why these protests failed to produce the same unity of purpose that occurred in Tunisia,
stresses the failure of Islamists and leftists to ally in support of the protests. He writes
that “whereas many of Morocco’s Islamists participated in the protests, its leftists did
not” (Buehler 2018, 166). From this perspective, the Islamist-​secular divide was a critical
determinant of participation in this mass wave of protest. Buehler, however, focuses ex-
clusively on the realm of formal politics; by Islamists, he means the PJD, whose members
participated at high levels early on in the protests, and by leftists, he means the USFP,
which did not endorse the protest movement.
Broadening the analysis to include social movement organizations produces a dif-
ferent picture. The February 20th movement was largely made up of human rights
activists and leftist groups that were not associated with the USFP. It also included
activists seeking Amazigh rights and student union groups, and it relied on the trade
union for meeting space. Only one small group, the Mouvement Alternatif pour les
Libertés Individuelles (MALI), was openly anti-​religious.14 The majority focused on
corruption, economic problems, and authoritarianism. Leftists were among the first
movers that organized the initial protests, but they were not aligned with the USFP.
Islamists were a part of a coalition that participated in the 2011 protests. PJD leaders
and members participated in the protests up until March 2011, although Benkirane
refrained from officially endorsing them. A more important role was played by the
JCG, which formed an alliance with the February 20th movement and participated in
protests until December 2011. A coalition that included Islamists, leftists, and a small
number of atheists and republicans was not easy to maintain, though the activists
were committed to a broad tent strategy that they argued would maximize the chances
of gaining real movement toward “a king who reigns but does not rule.”15 When prog-
ress toward that goal appeared stalled, Yassine withdrew his organization from the
protests and resumed a rejectionist stance toward the monarchy until his death one
year later.
In response to the protests, the king authorized changes to the constitution and called
for new legislative elections. In November the PJD came first in the elections, winning
more seats than the royalist Party of Authenticity and Modernity or the USFP, now
widely seen as having failed to implement meaningful change during its years leading
the parliament. Upon the PDJ’s victory, Benkirane immediately reached out to leftists to
492   Adria Lawrence

form a coalition government, but they rejected his invitation, and Benkirane was forced
to cooperate with centrist and royalist parties (Buehler 2018, 170–​172).
The PJD’s electoral triumph mimics the victories that occurred in the wake of the
uprisings in Tunisia and Egypt. Yet the reasons for the PJD’s victory have little to do with
its religious orientation, or the desire to deepen the role of religion in the state. In the
run-​up to the election, Moroccans who voted for the PJD reported that they did so not
because they supported a religious platform, but because the other parties running had
already tried and failed to address the country’s pressing economic woes. By this logic,
it was the PJD’s “turn” to take the reins, and voters hoped that the PJD’s Islamist orienta-
tion, and particularly Benkirane’s reputation as an upright and moral character, meant
that they would be less corrupt and more honest than their predecessors had been.16

Hereditary Monarchy and

the Religious-​Secular Cleavage

Morocco is a religiously homogenous country led by a king who is a descendent of the

Prophet and holds the title “Commander of the Faithful.” The monarchy in Morocco has
a “quasi-​divine” aura (Daadaoui 2011, 6). Religion is a central component of the regime’s
identity; the state’s motto is “God, the King, the Nation,” and these three words are
engraved on hills throughout Morocco. The king’s religious legitimacy sets him apart
from other leaders of electoral authoritarian regimes. Some scholars have suggested
that religious legitimacy explains the regime’s longevity (see Daadaoui 2011; Hammoudi
1997). If these claims are right, we might expect to see less opposition in Morocco than
we do in other kinds of authoritarian regimes. Indeed, the amount of critical opposition
described above may seem surprising if the monarchy’s characteristics truly shelter it
from challenge.
This chapter, however, is not aimed at addressing regime duration or the overall
strength of the opposition.17 For the purpose of this analysis, the relevant question is
whether the monarchy’s religious authority affects the religious-​secular cleavage in such
a way that this chapter’s arguments are unlikely to hold elsewhere. The king’s religious
role may mean that the religious cleavage is less pronounced in Morocco than in coun-
tries where the political leader is not also a religious leader. The fusion of religious and
political authority in the person of the king may leave fewer openings to politicize reli-
gion, or to argue about the role of religion in the state.
Yet while the salience of the cleavage in Morocco may differ from other cases, it is
worth exploring these arguments elsewhere. The religious authority of the monarchy
has not prevented Islamist challenges that resemble those that have occurred in other
places. As Miller (2013, 188) writes, “[t]‌he monarch, despite all of his galvanizing poten-
tial, has never held absolute control over the definition of Islam, leaving an open space
for other contenders.”
 Social Movements, Parties, and Political Cleavages in Morocco     493

During the colonial period, Allal al-​Fasi envisioned an independent state in which
religious authority would be exercised by an educated and essentially apolitical ulama
(Miller 2013, 188). In the 1970s, the Islamic Youth adopted an anti-​royalist stance,
questioning the monarchy’s religious authority. As described above, the JCG has
questioned the king’s religious legitimacy since the late 1980s. Opposition leaders can
and do question royal plans and policies by arguing that they deviate from Islamic prin-
ciples and practices.
Even when they do not question the regime, SMOs and political parties take positions
on religious questions. Heated debates over the rights of women to divorce or inherit,
for instance, have occurred during periods of monarchy-​initiated legal reform. Religion
remains salient for politics. My claim is not that religious questions are not contentious,
but that the religious cleavage does not differentiate SMOs and political parties as clearly
as their stance toward the regime does.
Three other historical episodes illustrate the ways that the king’s religious authority is
not automatic, but requires active work on the part of the palace and is subject to contes-
tation from below. The first is King Hassan II’s reaction to the series of coup attempts that
occurred in the early 1970s. In addition to reorganizing the armed forces, he revitalized
ties to the conservative Islamic establishment in order to create a counterweight to the
military and political parties (Miller 2013, 175). His efforts to build those ties show that
the king is not the only important source of religious authority in Morocco. The support
of the religious leadership did not automatically follow from his religious credentials.
The second episode occurred in 1975, when the king organized the Green March, a
350,000-​person march of volunteers armed only with Qurʾans and Moroccan flags to
take back the Spanish-​occupied region of the Sahara. With this event, the king sought to
shore up flagging popular and military support by acquiring Morocco’s “lost territory,”
not via military action, but by wielding the word of God. This move worked to unify
the country behind the king, but by the 1980s, that unity had again fragmented and the
king faced new domestic challenges in the form of strikes, protests, and the formation of
Islamist and other civil society organizations.
The third episode was the construction of the Hassan II Mosque, completed in
Casablanca in 1993. The king commissioned the mosque to demonstrate his piety, but the
project provoked intense public criticism over the size and cost of the mosque. Critics
argued that the king’s initial plan to build the second-​largest mosque in the world was
ostentatious and impious, an inappropriate use of resources in a highly impoverished
country.18 Cartoons, jokes, and articles pointed to corruption in the collection of funds;
the poor were said to be repeatedly taxed for the project.
The king’s religious role has not foreclosed contestation in Morocco. While few
among the opposition, or the wider public, question his religious role as a descendent
of the Prophet and the Commander of the Faithful, religious and political authority are
not the same. The SMOs that have adopted a more confrontational stance toward the
regime have argued that the king’s religious role does not confer the right to consolidate
political control. Members of the February 20th movement and the JCG have pushed
for a genuine constitutional democracy that retains the king in a symbolic and religious
494   Adria Lawrence

role, echoing the calls for a limited monarchy made in the early twentieth century. These
claims are part of the national conversation, suggesting that vibrant debate exists even in
a hereditary monarchy with religious legitimacy.
Further research on other cases can help to clarify whether the king’s role has deflected
religious opposition in Morocco, making this case markedly different from other places
in the Muslim world. In fact, recent work points to similar dynamics elsewhere. In this
volume, Pepinksky studies the cases of Indonesia and Malaysia from the perspective of
voters, and argues that “voting for Islam is a more complicated and conditional phe-
nomenon than the assumption of Islam as a programmatic identity implies.” Like this
chapter, he suggests the need to look beyond party labels to the policies driving iden-
tification with Islamist platforms. In his study of Tunisia, Grubman (2020) is directly
concerned with the same outcome I address here: the formation of political cleavages.
In electoral authoritarian regimes, Grubman argues, parties have difficulty developing
distinctive ideological platforms. Before the 2011 revolution, parties under Ben Ali were
shaped by their relationship to the regime:

Faced with a highly repressive and exclusive regime cloaking itself in the rhetoric of
liberalism and pragmatism, opposition parties had little opportunity or incentive to
organize around shared platforms for governing—​let alone to communicate them to
the public. Instead, they built their “brands” primarily by navigating their distance
from the regime, earning formal recognition or not, participating in elections or not,
and ganging up with each other or not (Grubman 2020, 86).

Grubman’s focus is on economic cleavages, but it is encouraging that in a non-​

monarchical authoritarian regime, he too found that the key factor shaping opposition
groups was how they handled the regime, a question that parties and SMOs in a democ-
racy do not face.

Conclusion

Opposition groups in Morocco, both institutional and extra-​institutional, identify

with religious or secular frames. But a focus on the ideologies that opposition groups
have used to situate themselves in the political spectrum obscures the primary political
cleavage that has divided Morocco since the colonial era: the division between those who
are critical of the monarchy’s political dominance, and those who accept it. Paraphrasing
Rémy Leveau, Miller (2013, 214) writes that “the crux of the political struggle in Morocco
is not over ideology in all its competing forms, but rather over tactics and accommoda-
tion of interests.” Leveau (2000, 122–​123) states, “Morocco’s most important institutional
problems . . . concern the interaction and power-​sharing between the monarchy and
Morocco’s political parties.” These insights are reflected in work that focuses on the im-
portance of co-​optation for regime stability (see Buehler 2018; Lust 2006).
 Social Movements, Parties, and Political Cleavages in Morocco     495

Scholars writing on Moroccan politics often focus either on political parties or on

social movements, but considering both kinds of opposition is important for under-
standing how attitudes toward the palace affect opposition strategies. Antagonistic
stances toward the monarchy are largely expressed outside the realm of formal politics.
These stances have been taken by both religious and secular actors, including leftists,
civil society organizations such as human rights and women’s rights groups, and Islamist
movements such the JCG, as well smaller underground Islamist organizations. Formal
political parties, whether they are leftist or Islamist in orientation, adopt cooperative
stances toward the monarchy.
Through discussing opposition groups in Morocco, I have sought to advance three
general claims. First, I have stressed that political cleavages should not be expected
to work similarly in democracies and autocracies. The logic of political cleavages in
democracies is that they reflect an underlying distribution of preferences for policies—​
they reflect what people want their government to do. Since people do not all sup-
port the same policies, cleavages mark off programmatic differences. In autocracies,
policy preferences cannot be translated into policy outcomes without approval by a
regime whose power is not constrained by free-​and-​fair competitive elections, as in
functioning democracies. Parties and SMOs in electoral authoritarian regimes have to
decide how to contend with a regime, not just how to win support. It is principally
their stance toward the regime that situates them in the political arena. That stance also
matters for how and whether they are able to gain support from different segments of
the broader public.
Second, this analysis has called attention to the need to consider both formal political
parties and informal political groups when attempting to understand opposition poli-
tics under electoral authoritarianism. The most critical actors are typically not political
parties who have accepted the authoritarian rules of the game, and by focusing solely
on political parties, it can be easy to overstate regime stability, as many who follow the
Middle East learned after the 2011 protests. In this chapter, including both SMOs and
political parties made it possible to see the platforms that are shared by organizations
with different ideological identities. SMOs often shared a similar stance toward the re-
gime and espoused the same economic and political goals, regardless of whether they
were Islamist or leftist. Political parties also differed less than a focus on ideology might
lead us to expect.
Third, I argued that actors do not need to choose between religious and secular
frames, but can incorporate elements of both. In a predominantly Muslim society, sec-
ular organizations may justify and propose policies on religious grounds, and Islamist
groups can support secular aims. Indeed, key issues such as poverty alleviation, edu-
cation, and corruption are not easily classifiable as either secular or religious. The
PJD in recent years has focused on economic problems, prioritizing those over issues
such as prostitution, alcohol consumption, or women’s rights, areas that are often key
components of Islamist platforms. The USFP has supported pan-​Arab and pan-​Muslim
causes such as the plight of Palestinians or Iraqis. Social movement actors also have both
secular and religious goals, and they draw on religious justifications to make claims
496   Adria Lawrence

about social justice, welfare, and development. Opposition groups are more diverse
than a focus on the secular-​religious cleavage would suggest. In seeking to build support
and carve out a position toward the regime, they can adopt a variety of different policy
platforms.
In making these claims, this chapter seeks to contribute to a growing body of work
that has questioned the causal role of religiosity in producing Islamist success at the
ballot box (see Grubman 2020; Masoud 2014; Pepinsky, Liddle, and Mujani 2012). I ex-
tend this insight from parties and voting to social movement organizing, adding to work
that suggests that religiosity may not be the variable that sets opposition groups apart.
This chapter also aimed to contribute to understanding the logic of cleavages in elec-
toral authoritarian settings, advising caution when important insights from the study
of cleavages in democracies to non-​democracies. Much of the existing work on political
Islam has been concerned with the question of whether Islamists, or even Muslims more
generally, would be able to abide by democratic rules once democracy was attained.19
But not all regimes are democratizing, and political cleavages are not the same in places
where there is a regime that seeks to weaken and co-​opt critical voices. In making this
point, I join scholars like Schwedler (2006) and Yadav (2013) who study how Islamist
parties act within authoritarian regimes.
This chapter has set aside some important questions for future study, such as how a
regime’s use of repression affects which political groups survive and even thrive under
authoritarianism, or why some ideologies are more popular and resonant at partic-
ular points in time. These are worthy topics to address. This chapter’s aim was to think
through the nature of the secular and religious cleavage in predominantly Muslim coun-
tries under electoral authoritarianism, and to suggest that cleavages are not always what
they seem to be.

Notes
1. See Boix (2009); Downs (1957); Lipset and Rokkan (1967).
2. For a discussion of hybrid regimes, see Levitsky and Way (2010). On the role of democratic
institutions in authoritarian regimes, see Gandhi (2008).
3. On the monarchy’s power and the limited role provided for parliament by the constitution,
see Maghraoui (2011); Waterbury (1970).
4. Electoral authoritarian regimes include monarchies, military regimes, personalistic
regimes, and single-​party regimes. In single-​party regimes, in which the ruling regime is a
party that participates in the electoral institutions, some modifications to the theory may
be required, though the claim that the stance toward the regime is the opposition’s primary
concern still holds. In single-​party regimes, the cleavage structure may look different than
it does in a personalistic regime or hereditary monarchy because the ruling party’s ide-
ology may shape the ideologies available to the opposition. On these subtypes of authori-
tarianism, see Geddes et al. (2014) and Smith (2005).
5. Questions about how to classify colonial regimes are not settled in the literature; French
colonial rule was clearly undemocratic and sometimes had features of modern electoral
 Social Movements, Parties, and Political Cleavages in Morocco     497

authoritarianism, though inconsistently. This chapter does not assume that French rule is
the same type of rule as the postcolonial monarchy; including it allows for consideration of
cleavages in a different context.
6. For more on the plan of reforms, see Abun-​Nasr (1975); al-​Fasi (1970); Lawrence (2013,
chap. 2); Rézette (1955); Waterbury (1970).
7. On these tactics, see Lawrence (2010, 2013).
8. On the definition, see Gellner (1983); on its flexibility, see Kocher et al. (2018, 124).
9. These claims did not go uncontested; in one instance, a reformist leader was ejected from
the mosque by prayer-​goers who argued that he was using the mosque for political, not re-
ligious purposes (Lawrence 2013, 202).
10. French rule was challenged directly during the early years of the protectorate in the areas
that had not been conquered; in mountainous areas, tribes fought French rule until 1934
(Porch 1982).
11. On the splintering of the nationalist movement into multiple groups, see Lawrence (2010).
On the king’s consolidation of power in the early years of independence, see Waterbury
(1970) and Zartman (1964).
12. In contrast, Daadaoui (2011, 11) sees both the PJD and the JCG as sharing the goal of
mitigating the authoritarian effects of the regime; I would suggest that the JCG is far more
confrontational about this goal than the PJD is, and the former’s participation in the 2011
protests supports this view.
13. On the organization, goals, and mobilization processes of the uprising in Morocco, see
Benchemsi (2014) and Lawrence (2017).
14. This group was known for holding picnics in public during Ramadan in protest of a law
prohibiting the breaking of the Ramadan fast. The group is also pro–​gay rights. See http://​
www.manifeste.org/​rubrique.php3?id_​rubrique=93.
15. Author field notes from the February 20th movement’s General Assembly, November 2011.
16. See Cammett and Jones Luong (2014) for a broader discussion of how Islamists’ reputation
for trustworthiness can be an advantage at the ballot box.
17. On the question of hereditary monarchy and regime longevity, see Brownlee et al. (2013);
Lawrence (2014); Menaldo (2012); Yom and Gause (2012).
18. Building a mosque that would rival the ones in Mecca and Medina was criticized as dis-
respectful of Muslim holy sites. Today, the mosque is the largest in Africa, and the fifth
largest in the world.
19. For a sample of work in this vein, see Grewal (2020); Kalyvas (2000); Langohr (2001);
March (2009).

References
Abun-​Nasr, Jamil M. 1975. A History of the Maghrib. Cambridge: Cambridge University Press.
Ait Mous, Fadma. 2013. “The Moroccan Nationalist Movement: From Local to National
Networks.” Journal of North African Studies 18, no. 5: 737–​752.
Arab Barometer. 2017. “Morocco Five Years after the Arab Uprisings.” http://​www.
arabbarometer.org/​wp-​content/​uploads/​Morocco_​Public_​Opinion_​Survey_​2016.pdf.
Ayoob, Mohammed. 2004. “Political Islam: Image and Reality.” World Policy Journal 21,
no. 3: 1–​1 4.
498   Adria Lawrence

Barka, Mehdi Ben. 1966. Option révolutionnaire au Maroc: Suivi de Écrits politiques, 1960–​1965.
Paris: La Découverte.
Benchemsi, Ahmed. 2014. “Feb20’s Rise and Fall: A Moroccan Story.” In Taking to the
Streets: Activism, Arab Uprisings, and Democratization, edited by Ellen Lust and Lina Khatib,
199–​235. Baltimore: John Hopkins University Press.
Boix, Carles. 2009. “The Emergence of Parties and Party Systems.” In The Oxford Handbook
of Comparative Politics, edited by Carles Boix and Susan C. Stokes. New York: Oxford
University Press.
Brownlee, Jason, Tarek Masoud, and Andrew Reynolds. 2013. “Why the Modest Harvest?”
Journal of Democracy 24, no. 4: 29–​44.
Buehler, Matt. 2018. Why Alliances Fail: Islamist and Leftist Coalitions in North Africa. Syracuse,
NY: Syracuse University Press.
Cammett, Melani, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17, no. 1: 187–​206.
Clark, Janine A., and Emanuela Dalmasso. 2015. “State Actor-​Social Movement Coalitions and
Policy-​Making under Authoritarianism: The Moroccan Party of Justice and Development in
the Urban Municipality of Kenitra.” Middle East Law & Governance 7, no. 2: 185–​211.
Daadaoui, Mohamed. 2011. Moroccan Monarchy and the Islamist Challenge: Maintaining
Makhzen Power. New York: Palgrave Macmillan.
Daadaoui, Mohamed. 2016. “Islamism and the State in Morocco.” Hudson Institute, April 29.
Denoeux, Guilain. 2002. “The Forgotten Swamp: Navigating Political Islam.” Middle East
Policy 9, no. 2: 56–​81.
Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper.
Eickelman, Dale F. 1985. Knowledge and Power in Morocco: The Education of a Twentieth
Century Notable. Princeton: Princeton University Press.
al-​Fasi, Allal. 1970. The Independence Movements in Arab North Africa. New York: Octagon Books.
Gandhi, Jennifer. 2008. Political Institutions under Dictatorship. New York: Cambridge
University Press.
Geddes, Barbara, Joseph Wright, and Erica Frantz. 2014. “Autocratic Breakdown and Regime
Transitions: A New Data Set.” Perspectives on Politics 12, no. 2: 313–​331.
Gellner, Ernest. 1983. Nations and Nationalism. Ithaca, NY: Cornell University Press.
Grewal, Sharan. 2020. “From Islamists to Muslim Democrats: The Case of Tunisia’s Ennahda.”
American Political Science Review 114, no. 2: 519–​535.
Grubman, Nathan. 2020. “Party Systems and Social Cleavages in New Democracies: Skipping
Class in Postuprising Tunisia.” PhD diss., Yale University.
Halstead, John P. 1964. “The Changing Character of Moroccan Reformism, 1921–​1934.” Journal
of African History 5, no. 3: 435–​447.
Halstead, John P. 1969. Rebirth of a Nation: The Origins and Rise of Moroccan Nationalism, 1912–​
1944. Cambridge, MA: Harvard University Press.
Hammoudi, Abdellah. 1997. Master and Disciple: The Cultural Foundations of Moroccan
Authoritarianism. Chicago: University of Chicago Press.
Hoisington, William A. 1978. “Cities in Revolt: The Berber Dahir (1930) and France’s Urban
Strategy in Morocco.” Journal of Contemporary History 13, no. 3: 433–​448.
Joffé, E. G. H. 1985. “The Moroccan Nationalist Movement: Istiqlal, the Sultan, and the
Country.” Journal of African History 26, no. 4: 289–​307.
Kalyvas, Stathis N. 2000. “Commitment Problems in Emerging Democracies: The Case of
Religious Parties.” Comparative Politics 32, no. 4: 379–​398.
 Social Movements, Parties, and Political Cleavages in Morocco     499

Kocher, Matthew Adam, Adria K. Lawrence, and Nuno P. Monteiro. 2018. “Nationalism,
Collaboration, and Resistance: France under Nazi Occupation.” International Security 43,
no. 2: 117–​150.
LaFuente, Gilles. 1999. La Politique Berbère de La France et Le Nationalisme Marocain.
Paris: L’Harmattan.
Langohr, Vickie. 2001. “Of Islamists and Ballot Boxes: Rethinking the Relationship between
Islamisms and Electoral Politics.” International Journal of Middle East Studies 33, no.
4: 591–​610.
Lawrence, Adria K. 2010. “Triggering Nationalist Violence: Competition and Conflict in
Uprisings against Colonial Rule.” International Security 35, no. 2: 88–​122.
Lawrence, Adria K. 2012. “Rethinking Moroccan Nationalism, 1930–​44.” Journal of North
African Studies 17, no. 3: 475–​490.
Lawrence, Adria K. 2013. Imperial Rule and the Politics of Nationalism: Anti-​Colonial Protest in
the French Empire. New York: Cambridge University Press.
Lawrence, Adria K. 2014. “Kings in a Democratic Age: Collective Protest and the Institutional
Promise of Monarchy.” Paper Presented at the Annual Meeting of the American Political
Science Association, September 2014.
Lawrence, Adria K. 2017. “Repression and Activism among the Arab Spring’s First
Movers: Evidence from Morocco’s February 20th Movement.” British Journal of Political
Science 47, no. 3: 699–​7 18.
Leveau, Rémy. 2000. “The Moroccan Monarchy: A Political System in Quest of a New
Equilibrium.” In Middle East Monarchies: The Challenge of Modernity, edited by Joseph
Kostiner, 117–​130. Boulder, CO: Lynne Rienner.
Levitsky, Steven, and Lucan Way. 2010. Competitive Authoritarianism: Hybrid Regimes after the
Cold War. New York: Cambridge University Press.
Lipset, Seymour Martin, and Stein Rokkan. 1967. Party Systems and Voter Alignments: Cross-​
National Perspectives. New York: Free Press.
Lust, Ellen. 2006. Structuring Conflict in the Arab World: Incumbents, Opponents, and
Institutions. Cambridge: Cambridge University Press.
Maghraoui, Driss. 2011. “Constitutional Reforms in Morocco: Between Consensus and
Subaltern Politics.” Journal of North African Studies 16, no. 4: 679–​699.
March, Andrew F. 2009. Islam and Liberal Citizenship: The Search for an Overlapping
Consensus. New York: Oxford University Press.
Masoud, Tarek E. 2014. Counting Islam: Religion, Class, and Elections in Egypt. Cambridge:
Cambridge University Press.
Menaldo, Victor. 2012. “The Middle East and North Africa’s Resilient Monarchs.” Journal of
Politics 74, no. 3: 707–​722.
Miller, Susan Gilson. 2013. A History of Modern Morocco. New York: Cambridge University Press.
Munson, Henry, Jr. 1993. Religion and Power in Morocco. New Haven, CT: Yale University Press.
Pennell, C. R. 2000. Morocco since 1830: A History. New York: New York University Press.
Pepinsky, Thomas B., R. William Liddle, and Saiful Mujani. 2012. “Testing Islam’s Political
Advantage: Evidence from Indonesia.” American Journal of Political Science 56, no.
3: 584–​600.
Porch, Douglas. 1982. The Conquest of Morocco. New York: Alfred A. Knopf.
Renan, Ernest. 1996. “What Is a Nation.” In Becoming National, edited by Geoff Eley and Ronald
Grigor Suny, 41–​55. Oxford: Oxford University Press.
Rézette, Robert. 1955. Les Partis Politiques Marocains. Paris: Librarie Armand Colin.
500   Adria Lawrence

Rivet, Daniel. 1999. Le Maroc de Lyautey à Mohammed V: Le Double Visage Du Protectorat.

Paris: Editions Denoel.
Schedler, Andreas. 2015. “Electoral Authoritarianism.” In Emerging Trends in the Social and
Behavioral Sciences, edited by Robert A. Scott and Stephen M. Kosslyn, 1–​16. Hoboken,
NJ: John Wiley & Sons.
Schwedler, Jillian. 2006. Faith in Moderation: Islamist Parties in Jordan and Yemen. Cambridge:
Cambridge University Press.
Schwedler, Jillian, ed. 2013. Understanding the Contemporary Middle East. 4th ed. Boulder, CO:
Lynne Rienner.
Smith, Benjamin B. 2005. “Life of the Party: The Origins of Regime Breakdown and Persistence
under Single-​Party Rule.” World Politics 57, no. 3: 421–​451.
Waterbury, John. 1970. Commander of the Faithful: The Moroccan Political Elite—​A Study in
Segmented Politics. London: Weidenfeld and Nicolson.
Willis, Michael J. 2004. “Morocco’s Islamists and the Legislative Elections of 2002: The Strange
Case of the Party That Did Not Want to Win.” Mediterranean Politics 9, no. 1: 53–​81.
Wyrtzen, Jonathan. 2015. Making Morocco: Colonial Intervention and the Politics of Identity.
Ithaca, NY: Cornell University Press.
Yadav, Stacey Philbrick. 2013. Islamists and the State: Legitimacy and Institutions in Yemen and
Lebanon. London: I.B. Tauris.
Yom, Sean, and F. Gause. 2012. “Resilient Royals: How Arab Monarchies Hang On.” Journal of
Democracy 23, no. 4: 74–​88.
Zartman, I. William. 1964. Morocco: Problems of New Power. New York: Atherton Press.
 Chapter 24

T he Rise and I mpac t

of M u slim Wome n
Preaching Onl i ne

Richard A. Nielsen

Female religious authority is on the rise in many Muslim societies (Bano and Kalmbach
2012). Women now occupy official positions in the state-​run religious hierarchies in
Morocco (Rausch 2012), Turkey (Hassan 2011), and elsewhere in the Muslim world. This
transformation of Islamic authority is also happening in transnational communities of
Muslims, such as the Salafi social movement. Salafi Islam is a global revivalist movement
that seeks to return Islam to what it believes were the practices and beliefs of the salaf,
the first few generations of Muslims during and after Muhammad’s lifetime. Doctrinally,
the Salafi movement espouses patriarchal values and appears to have very little space for
women’s religious authority. Yet recent research documents the rise of female preachers
among the Salafi movement’s ranks (Al-​Rasheed 2013; Le Renard 2012; Nielsen 2020).
The prevailing view is that religion is the key factor explaining the increase and im-
pact of women’s preaching. According to this view, women’s preaching follows a logic
based on religious norms. Religious ideas about appropriate gender relations, such
as doctrines forbidding public cross-​gender interactions, create a need for female
preachers who preach about “women’s issues” to exclusively female audiences (Al-​
Rasheed, 2013; Le Renard, 2012). In this chapter, I argue instead that the rise and impact
of female preachers in conservative Islamic settings follows a social movement logic.
In this view, female preachers help Islamic social movements reach new audiences of
both women and men, potentially enlarging the movement. Therefore, religious ideas
prohibiting gender mixing are not the cause of women’s preaching, but rather a norma-
tive constraint that female preachers circumvent by preaching online.
In this chapter, I investigate the impact these female preachers are having on their
online audiences, and, by extension, on society and politics. Assessing the impact of fe-
male preachers is complicated by the difficulty of surveying Muslim publics about po-
tentially sensitive gender issues in many of the countries where these female authorities
502   Richard A. Nielsen

are finding audiences. Instead, I look at patterns of voluntary online engagement with
female preachers using digital trace data. Some of the websites where these women pub-
lish their writing allow users to interact with the writing through comments. I collect
and compare these reactions to examine how reader engagement differs depending on
the gender of the author. My statistical tests answer this question: If a reader encounters
online preaching by a woman, do they respond differently than if they had encountered
similar preaching by a man?
First, I examine who is engaging with the online writing of female preachers.
I test whether female preachers get more or less engagement relative to their male
counterparts. I collect data on the gender makeup of those who respond to writing by
female preachers. Previous scholarship has concluded in other Islamic contexts that
“women are confined to female audiences . . . because believers by and large still prefer
male religious authorities” (Kloos and Künkler 2016, 488). If at least some men are
paying attention to online preaching by women, this represents an important shift.
Second, I examine how lay Muslims are engaging with the preaching of female
preachers. Using the content of online comments, I test whether there is any indication
that female preachers are changing readers’ perceptions about the role of women in reli-
gion and society by looking for evidence of admiration and criticism.
My results show that many lay Muslims are responding positively to the online
preaching of female preachers. I find that writing by women garners more comments
from website readers, and that the majority of these commenters are men. Compared
to comments on a similar set of documents by men, the writings of women elicit more
praise, although some commenters do become critical when female preachers challenge
patriarchal norms. The comments on women’s writing also come from more diverse
geographic locations than comments on comparable writing by men. Taken together,
these findings support the argument that social movements might selectively promote
female authorities in contravention of their patriarchal norms because they engage new
audiences.
More broadly, my findings have implications for the acceptance of female authorities
in Muslim societies, though my conclusions on this point are necessarily speculative.
Religious authority in Islam has generally been held by men, so the rise of female Muslim
preachers on the Internet may be fundamentally changing the nature and impact of re-
ligious authority in Muslim societies. Although these women would likely claim they
are not political actors, when women gain authority in society, it is necessarily political.
The trends I document in this chapter have the potential to fundamentally change the
gender politics of many Muslim societies.

Female Authority in Islam

Although there have always been some female authorities in Islam (Bano and Kalmbach
2012; Nadwī 2007), Islamic authority has historically been dominated by men. Yet
 The Rise and Impact of Muslim Women Preaching Online    503

women are gaining authority, both in state-​sponsored “official Islam” and in Islamic so-
cial movements such as Salafism.
Why is this rise in female authority happening? It is likely that there are several root
causes that may vary from one context to another. To explain the existence of female
Salafi preachers, Al-​Rasheed (2013) argues that religiously based gender norms created
demand for gender-​segregated preaching, and rising levels of religious education
among women have produced a supply of capable female preachers. I agree that this is
part of the story, but in my own work, I argue that male Salafi movement leaders are in-
creasingly promoting female authorities because they attract new audiences that men
cannot reach (Nielsen 2020). The spread of the Internet has created the conditions for
this to happen. As Le Renard (2014, 38) points out, the Internet allows women to preach
while technically following conservative social norms prohibiting “gender mixing.” It
also creates conditions where a movement faces a lower cost of rapidly expanding its
audience through mass communication. My contention is that male leaders of the Salafi
movement remain skeptical about women’s authority in principle, but they have ac-
cepted it in practice because it increases the influence of their movement.
How are these women carving out space for their exercise of authority in domains
that are not always permissive? Le Renard (2012, 113–​116) argues that scholarly expertise
and charisma are important features of the preaching she observes in the study circles
of female clerics. If these factors help women gain authority, then perhaps the “how?”
of women’s preaching is not particularly distinctive from men, because men also rely
on expert training and charisma to preach. Other accounts emphasize aspects that are
unique to women’s preaching. Ben Shitrit’s (2016) account emphasizes frames of excep-
tion that women deploy to justify behaviors that would ordinarily be transgressions. By
casting their activism as a temporary necessity to meet the needs of exceptional times,
the women she studies in several religions carve out space to exercise greater authority
than before. My own research shows that female preachers rely on identity authority
when establishing their authority (Nielsen 2020). While male Salafi preachers typically
use citations to establish the authority of their arguments, female preachers use citations
far less, even when writing on the same topics as men. Instead, they are more likely to
support their arguments by appealing to their identities as women, a finding that echoes
Ben Shitrit’s observations about the “motherhood” frames women use to justify certain
forms of activism.
Less is known about the impact of increasing women’s authority in Islam, either for
those exposed to their preaching or to society more generally. Ethnographers describe
a certain awe and respect among the participants in face-​to-​face preaching (Ben Shitrit
2016; Le Renard 2012). However, the reception of online preaching by women could
be very different from the face-​to-​face reception of their devoted followers. As far as
I know, my recent analysis of Twitter reactions to the writing of female Salafi preachers
on the website saaid.net is the only systematic analysis of the effects of women’s online
preaching in Islam (Nielsen 2020). I find that writing by women prompts more engage-
ment from women. It also brings in more individuals who otherwise have never engaged
with the content of the website, suggesting that women reach new audiences. At the
504   Richard A. Nielsen

same time, I find that women are not only preaching to women. In fact, almost 75 per-
cent of the Twitter reactions to female preachers’ writings come from men. However,
these Twitter reactions—​“retweets” and “favorites”—​reveal very little about what im-
pact this preaching is having.

Online Engagement with the Writing

of Female Preachers

In this chapter, I look at a more substantive form of engagement: online comments by

readers of women’s preaching. Many of the major Islamic websites permit users to in-
teract with the content posted on the website in various ways: “likes,” ratings, and, in
some cases, comments. The major benefit of studying comments is that they are gen-
uine, natural, in situ reactions to women’s preaching. There are no concerns about
respondents giving responses that they think will please researchers. Yet although
comments are a very immediate indicator of impact, the data cannot speak directly
to many of the pressing questions around women’s preaching. Does experiencing fe-
male religious authority change people’s attitudes about the appropriateness of women
having religious authority? Might this spill over to affect their attitudes about the suita-
bility of women for political office? Their attitudes about the roles of women in society
writ large? I cannot answer these questions, because I am not able to survey respondents
about their attitudes after exposing them to women’s preaching. Such a survey would be
fascinating, but is challenging in the places where these women’s preaching is primarily
consumed: Saudi Arabia, Egypt, Algeria, and Morocco.
There are other forms of engagement, such as page view counts, that I examine briefly
but do not emphasize. I focus on comments because they are the richest form of engage-
ment I can find. In some cases, I find examples of dialogue between preachers and their
followers in the comments section. Likes and other thin indicators of engagement do
not convey as much information about how the teaching is received.
What do I expect to find? The religious norms logic, exemplified by Al-​Rasheed
(2013), argues that female preachers are writing almost exclusively on women’s issues
to female audiences. Kloos and Künkler (2016, 488) conclude that Muslims by and large
still prefer male authorities. These prevailing arguments in Islamic studies suggest the
expectation that female preachers will be less popular than men and that they will ex-
clusively serve female audiences. This would suggest that documents written by fe-
male preachers will receive fewer page views than similar documents written by men.
Women’s documents should also receive fewer comments if readers are more interested
in male authorities. And when female preachers do receive comments, they will be al-
most exclusively from women. Men will be largely ignore women’s preaching, or per-
haps express hostility.
 The Rise and Impact of Muslim Women Preaching Online    505

The social movement logic I espouse makes contrasting predictions. If the strategic
considerations of social movements explain the rise and impact of female preachers,
I expect to find that women get as many page views as men, or perhaps more on subjects
for which women might be viewed as particularly authoritative because of their
gender identity. I expect that female preachers will get more comments than their male
counterparts, and that these comments will come from a more diverse set of people, be-
cause the role of these female preachers is partly to attract new audiences. I expect that
the writing of female preachers will garner comments from more women than similar
writing by male preachers. Nevertheless, I also expect that female preachers will get a
substantial portion of their comments from male readers and that these comments will
be positive.
To summarize, the key empirical question that distinguishes the social movement
logic from the logic of religious norms is this: Is there is a receptive male audience for
female preachers?

Data and Design

My prior work examined the writing of 43 female Salafi preachers and 172 of their male
counterparts on the Salafi missionary website saaid.net (Nielsen 2020). In that research,
I sought to understand the impact of these specific female preachers because they are
members of a patriarchal social movement in Islam that is suspicious of women’s au-
thority. But saaid.net does not permit the types of engagement I would like to study;
I cannot tell from the website who is engaging with the texts beyond some basic infor-
mation about who interacts with the website’s official Twitter feed. From my earlier anal-
ysis, I know that these female Salafi preachers are reaching new audiences of both men
and women. However, these Twitter users don’t tend to comment as they like or repost
writings from saaid.net, so it is difficult to tell whether these reactions indicate a positive
or negative reception.
To find reactions with more substance, I turn to the website alukah.net, a large Islamic
aggregator site in Arabic that allows users to comment on writings, with moderation
from the website managers. The website gets a lot of traffic: I record over 1.55 billion
page views over the life of the website (since October 2006). Although the website is not
focused exclusively on Salafism, it contains writings by about half of the female Salafi
preachers on saaid.net. It also contains writing by many other female authors, though
not all of these authors are preachers. The process by which writing appears on the web-
site is opaque, but the website welcomes submissions, and I expect that this is the pri-
mary way that articles come to be on the website.1 Submissions must be made by the
author, must be in modern standard Arabic, must be not previously published else-
where online (but this is frequently violated), and must be “sensitive to the teachings of
our true religion and Sunni doctrine.”
506   Richard A. Nielsen

Website visitors can leave comments on any document on the website using a
standard web form at the bottom of each page. This form requires that readers enter an
e-​mail address (though it need not be a valid one), a name, and a title for their comment.
It also permits readers to specify their country (open response) and to write as much
as they want in the text of the comment. After submitting a comment, a message reads
“Thank you. We will publish your comment after a review by the administration. You
may comment again by returning to this page.” The nature of this review is unclear, and
my request for clarification from the website administrators went unanswered. I found
that my comments were generally posted within ten minutes, but I only tested with
short positive, comments reading “Thank you,” because I did not want to disturb this re-
ligious training environment (Nielsen 2015).2 In my reading of thousands of comments
from hundreds of articles, I found critical comments but never crude or openly deroga-
tory comments. It appears that a combination of moderation and social norms keeps the
comments on this website more civil than many Internet forums.
I collected all of the Arabic-​language writings from the website, resulting in an initial
data set of 114,892 documents by 11,677 authors. Along with these texts, I got the number
of times each document had been viewed (median = 3,917 views), the time it was posted,
the text of the document, the number of comments (31,020 total comments), and the
first page of comments for each document.3
The gender of authors is not indicated on the website, except through their name, and
in some cases through a biography as well. I coded my perception of the gender of the
author based on Arabic naming conventions. Of the 11,677 authors, I code 671 (5.7 per-
cent) as women and 7,013 (60 percent) as men. The remaining 3,993 (34 percent) are
ambiguous, difficult to code, or entities that are not individual authors. I discard the
documents by these 3,993 from the data set because the effort to determine the gender
of each far outstrips the usefulness of additional data. The documents for which I can
readily code the gender of the author provide plenty of data to observe how readers re-
spond to men’s and women’s writing, and I do not believe there are serious threats to my
inferences below by omitting the more ambiguous cases. My coding procedures retain
the authors in the data that are most readily classified as men and women based solely
on names, which is ideal for testing my arguments. However, I should note that I can’t
provide evidence about the effect of ambiguous perceptions of author gender.
I exclude documents that must be downloaded because extracting text from them is
difficult and website visitors engage them differently than documents which render in a
web browser. I also exclude any documents that have fewer than thirty characters. These
exclusions leave 49,192 documents by 442 female authors and 4,386 male authors.
My goal is to estimate the how readers’ perceptions of an author’s gender change their
reactions while holding the writing constant. My strategy for doing so is to find female
and male authors who write similar documents, and then compare the comments they
receive. This strategy is inspired by experimental audit studies, in which respondents
react to comparable documents (often resumes), where the name is randomized to
convey information about gender, race, or some other characteristic. Because of the
challenges of surveying the target population for this preaching, it is not feasible to carry
 The Rise and Impact of Muslim Women Preaching Online    507

out an audit study in this context, but my aim is to design an observational study that
that mimics the logic of an audit design. I assume that if I can compare documents in
which male and female authors write very similarly, any differences in the comments are
due to the perceived gender of the author rather than differences in the content of the
writing.
Prior qualitative and quantitative research shows that female preachers generally
write on different topics than men (Al-​Rasheed 2013; Nielsen 2020). Because of this,
I cannot simply compare the comments on all documents by women to the comments
on all documents by men; differences in the content are likely to be the cause of most of
the differences in the comments. Instead, I identify documents by women and men that
have similar content using a combination of matching procedures and regression adjust-
ment. The strength of this design is that any cross-​gender differences in the comments
on these documents are more likely to be the effect of perceived author gender, rather
than spurious correlation from differences in content. But the design has weaknesses,
too. Not all documents by female authors have comparable documents by men, and vice
versa. This means my results cannot answer questions such as how readers would react if
women were to write on topics that are currently only discussed by men.
To identify comparable documents by female and male preachers, I use text matching,
introduced by Roberts, Stewart, and Nielsen (2020). I condition on the text to find
subsets of documents in which men and women write on similar topics. Specifically,
I estimate a structural topic model (Roberts et al, 2014) with the author gender as a
covariate. This produces fifteen estimated topics that I use to match documents with
similar topical content to each other.
In addition to facilitating my efforts to match similar documents to each other, this
topic model gives a sense of the general themes of documents on alukah.net. Figure 24.1
summarizes the fifteen topics using topic numbers randomly assigned by the model,
labels I assign, and key words generated by the model. On the right, I plot the proportion
of women’s and men’s writing devoted to each topic using black and white disks.4 These
estimates show that women heavily favor the three topics at the top, which I have labeled
Family, Lived Experience, and Children. Men favor the remaining topics, especially
those at the bottom relating to Islamic law, doctrine, and sacred texts. To reiterate, the
reason for comparing documents with similar topic proportions is that I anticipate an
article primarily about lived experience and family problems to draw different reactions
from readers than an article primarily about sin and Islamic law.
I also match on other variables that are likely to influence reader engagement, such
as the number of years since a document was posted (measured in fractions) and the
category of the document listed in the website.5 For analyses where some aspect of the
comments are the outcome variable, I also condition on the number of page views be-
cause these strongly predict the number of comments.
I use coarsened exact matching (Iacus, King, and Porro 2012) to construct two sets of
matched documents.6 In the first set, I limit the female authors to the seventeen authors
(with 443 documents) that also appear on the saaid.net website. This matched data
set allows me to analyze the impact of the set of female authors that is most directly
508   Richard A. Nielsen

0 0.1 0.2

13. Family: marriage, young woman, husband, blessings, wife, youth, divorce, sister, problem
4. Lived Experience: is not, sadness, pain, said, I know, hopefully, was, life, moment, suddenly
9. Children: child, negative, therapy, yourself, feelings, children, syndrome, feeling, answer
5. Culture: civilization, politics, country, economic, culture, society, creed, West
10. Teaching: teacher, teaching, goals, administration, students, number, study, methods
6. Holidays: sleep, Ramadan, day, fasting, food, holiday, night, fast, forgiveness, month
15. Authority: innovation, creed, authored by, sciences, reciter, article, sheikh, knowledge
7. Grammar: letter, context, language, evidence, case marking, dialect, Quran, meaning
3. Places and Poetry: Beirut, Damascus, tradition, poetry, dictionary, Cairo, poet, Farsi, Arabic
12. Text Labels: name, seven, five, sources, number, abridgement, cities, ten, three, four
11. Hadith: hadith attribution, al−Safi, al−Nawawi, Fath al−Bari, Hanbali, Shafi'i
8. Sin: forgiveness, punishment, sin, disobedience, creator, air, Lord, damn, path,
2. Sunna: Prophet, polytheist, said, conquest, Aisha, Abu Hurayra, al−Bukhari, al−Tarmidhi
14. Islamic law: jurisprudence, certificate, questions, definition, question, Sharia, attribution
1. Manuscripts and Officialdom: office, al−Azhar, file, professor, manuscript, Dr., photographer

0 0.1 0.2
Estimated topic proportion

Female authors
Male authors

Figure 24.1: Male and female authors on alukah.net focus on different topic.
Note: Estimated topic proportions for female and male authors in the alukah.net corpus. Topics
are ranked by the difference in proportions between women and men. I match documents on
these topics to improve comparability in my analysis below.

comparable to my prior analysis (Nielsen, 2020) and who I am certain are preachers.
The matched data set contains 242 documents by women and 993 matched documents
by men. The vast majority of the 44,000 men’s documents in the full data set are dropped
in the matching procedure, indicating that relatively few men’s documents are similar to
women’s documents on all dimensions. Most of my analysis below concentrates on this
matched data set.
The estimated topics show that men’s and women’s writing is quite different on av-
erage. Approximately 21 percent of writing by the seventeen female preachers from
saaid.net is devoted to a topic on family (keywords: marriage, young woman, husband,
blessings, wife, youth, divorce, sister, problem), while only 3 percent of men’s writing
is devoted to this topic. In the matched sample, I am able to make the portions of this
topic roughly equal: 40 percent for women and 37 percent for men. On the other end,
only 2 percent of these women’s writing is devoted to a topic on hadith, the sayings of the
Prophet Muhammad and his companions, while 8.5 percent of men’s writing is on this
topic (keywords: hadith attribution, al-​Safi, al-​Nawawi, Fath al-​Bari, Hanbali, Shafi’i). In
the matched sample, I am able to find comparison documents by men that devote only
 The Rise and Impact of Muslim Women Preaching Online    509

1.8 percent of words to this hadith topic. No two texts are identical, so this text matching
is only approximate. However, I find that the matching results in pairs of relatively sim-
ilar texts. For example, when I sample pairs of matched texts, I cannot reliably predict
which was authored by a woman without seeing the authors’ names.
In the second matched sample, I attempt to find matches for all of the female authors
appearing on alukah.net. This set of individuals is more varied and includes many
authors who are unlikely to consider themselves preachers. I am able to match 2,264
documents by women with 7,754 suitably similar documents by men. Again, there are
predictable differences between the topics that men and women emphasize. I am able to
reduce those differences for all topics and eliminate the differences for some topics.
Matching improves balance substantially, but significant imbalances remain in both
data sets when I examine the mean differences between men and women in the matched
samples. Because of these remaining imbalances I adjust for each of the matching
covariates using parametric models (Ho et al. 2007).
The matching approach helps with another challenge of using data from alukah.
net: collecting the full set of comments for each article. Comments appear in reverse
chronological order, ten at a time. When there are ten comments or less, they all are
rendered in the html of the website, but if there are more, then they are only acces-
sible through JavaScript and I can only download them by hand. Matching restricts the
sample so that collecting the complete set of comments by hand is feasible.
In addition to collecting some comments by hand, I also have to hand-​code my per-
ception of the commenter gender, their self-​reported location, and whether they iden-
tify themselves as the author of the article. Again, by limiting the sample, a matching
strategy makes this herculean task feasible. However, it commits me to the particular
matches I choose—​I can’t just change the matching strategy to see how the results might
change. This is a strength of my analysis, rather than a weakness. There are fears that
flexibility to select matched samples after looking at the results allows for p-​hacking
(Miller 2013), but I have no such flexibility because I only collect many of the outcome
variables after deciding on matched set.

Findings

Before turning to the findings, I reiterate what I expect to find and how it contrasts
with the predictions of alternative explanations for the rise and roles of female Muslim
preachers. I argue that a social movement logic dominates. Female preachers serve a par-
ticular role in the Salafi movement: to engage new audiences online that male preachers
are unable to reach. This theory predicts that women are likely to garner more engage-
ment from a more diverse audience of both men and women, and that this engagement
will be positive. In contrast, the prevailing view of these female preachers follows a reli-
gious norms logic in which women preach primarily to women, and male lay Muslims
strongly prefer male preachers instead. If this view is correct, then I expect to see less
510   Richard A. Nielsen

Page Views Comments % Female Commenters # of Commenters Countries % Praise Words

30000 1.4 100 3.0 40
25000 1.2 2.5
80 30
1.0 2.0
20000
0.8 60
15000 1.5 20
0.6 40
10000 1.0
0.4 10
5000 20 0.5
0.2
0 0.0 0 0.0 0

Women Men Women Men Women Men Women Men Women Men

Figure 24.2: Summary of findings: Female authors get fewer page views but deeper engagement.
Note: Each panel summarizes the result of a regression model in the following sections. I find that
female authors get fewer page views (panel 1), get more comments (panel 2), get the majority of
their comments from men (panel 3), draw comments from a more geographically diverse audi-
ence (panel 4), and receive more praise in comments (panel 5). Bar heights indicate regression
predictions and dashed lines indicate 95 percent confidence intervals. All result are based on the
first matched sample of women from saaid.net, containing 242 documents by women and 993
matched documents by men, that collectively receive 983 comments.

engagement with women’s writing, and that what engagement there is will be almost ex-
clusively from female readers. If men react to women’s preaching, their reactions may be
negative. However, because the website I examine moderates comments, an absence of
negative comments may simply reflect the choices of website moderators to screen crit-
icism. Even with comment moderation, this alternative theory would not predict that
women should get more positive comments than men.
I summarize the key findings in Figure 24.2; the details of each calculation follow in
the sections below. I find that female authors get fewer page views (panel 1), get more
comments (panel 2), get the majority of their comments from men (panel 3), draw
comments from a more geographically diverse audience (panel 4), and receive more
praise in comments (panel 5). This pattern of results shows that writings by women are
viewed less frequently than comparable writing by men, but women’s writing elicits
more engagement and praise from readers. These results largely align with my expecta-
tions based on a social movement logic.

Women Get Fewer Page Views

I find that articles by women get fewer page views than similar articles by men. This is
true for both the matched sample of saaid.net women and the larger matched sample
of all the women on alukah.net. It is also true in the unmatched samples, so matching
does not make much of a difference in this analysis. The outcome variable in this anal-
ysis, page views, is highly skewed, so I also estimate models using the natural log of
page views as the outcome variable. If I use the untransformed page view variable, the
estimates indicate that the saaid.net women received over 5,000 fewer page views than
comparable documents by men, but this is likely influenced by the extreme skewness.
Using log transformed page views as the outcome variable, I estimate that the saaid.
net women got 23 percent fewer page views than their male counterparts. These same
 The Rise and Impact of Muslim Women Preaching Online    511

estimates are roughly comparable when comparing all of the women on alukah.net to
the matched men: they receive 4,400 fewer page views, or about a 7 percent decrease.
This finding contradicts my expectations from the social movement logic and offers
preliminary evidence in support of the religious norms logic. However, the results below
suggest that once I look at substantive engagement through comments, rather than thin
engagement through page views, the social movement logic finds more support.

Women Get More Comments

Although women get fewer page views, I find that they get more comments. Documents
by the female preachers from saaid.net get on average 0.6 more comments than compa-
rable documents by make preachers; this estimate remains statistically significant if I use
a quasi-​poisson generalized linear model to account for the over-​dispersed counts in the
outcome variable.7 In the full sample of women, the effect is not as large, but it is still
statistically detectable. On the whole, women on alukah.net get 0.3 more comments on
average than men. I find no statistically detectable difference in the length of comments
left on documents by men and women.
These numbers may sound small, but they are large relative to the baseline number of
comments. The modal number of comments in both matched samples is zero, and the
mean number of comments for articles by men is around 0.3. My best estimate is that
women are getting at least 100 percent more comments on average than comparable
documents by men.
What drives this deeper engagement with writing by women? At first, I assumed that
comments would increase proportionally with the number of page views, and indeed
I find that page views are a very strong predictor of the number of comments a docu-
ment will get. This makes sense: the more people who read a document, the more people
who will comment on it. A model fit to the unmatched data suggests that the base rate
of comments is low, and increases approximately linearly. The average document gets
1.6 additional comments for every 100,000 page views. However, this can’t explain why
women get more comments than men, because women are getting fewer page views on
average.
If women get fewer views, but more comments, it suggests that women are more effi-
cient at converting thin engagement (page views) into thick engagement (comments).
How? My investigation of the data suggests that female authors are more likely to en-
gage with visitors that leave comments, and this interaction increase the number of
comments. Both male and female authors can leave comments on their own documents,
but I find that women are more likely to do so. My data only allow me to quantify this
trend for the female preachers from saaid.net and the male authors matched to them,
because this is the only set for which I have extracted the names of commenters to de-
termine which comments are from the authors themselves. Commenting by authors is
common; out of 983 comments in this matched sample, 136 are from the authors. Sixty-​
seven percent of these author comments are by women, even though women authors
512   Richard A. Nielsen

make up only 20 percent of this matched sample. I find that these author comments are
an important predictor of reader engagement. When I account for author comments in
my statistical model, I find that each author comment is associated with 1.7 additional
comments from website visitors, and author gender no longer predicts the number of
comments.
These results suggest that female authors get deeper engagement from readers be-
cause they offer deeper engagement as authors. To get a sense of how these engagements
unfold, I read thirty-​five exchanges between female authors and their readers, and
a handful of exchanges with male authors for comparison. Most of these exchanges
are complimentary to the authors. In many cases, commenters simply offer their
congratulations and praise for publishing the piece, and the authors respond with
short expressions of gratitude. In some cases, commenters seek specific advice from
the authors on topics related to the article, or seek clarification on some aspect of the
writing. In six of the thirty-​five exchanges, the commenter adopts a critical tone, either
criticizing the ideas or writing of the author. In some cases, these critical commenters
claim some authority and offer advice for how the female author might improve her
writing for the future.
A particularly interesting exchange happens in the comment section of Sara
Bint Muhammad Hassan’s article “Diary of an Exemplary Husband and a Grumpy
Wife,” which recounts the feelings of Amal, a Muslim wife whose neglectful husband
announces to her that he has decided to take another wife.8 The social commentary in
this article is biting and could be read as indirect criticism of the institution of polygamy.
There are forty-​six comments on this article, many of them critical. For example, a com-
ment left by Abu Abd Allah from Riyadh, Saudi Arabia, accuses the wife of jealousy and
says, “In order to judge, we need to hear the other side,” to which the author retorts, “I see
that you are very biased against the sister.”9 A more ambiguous commenter thanks her
for her “impartial and objective” treatment of the issue, in a statement that appears com-
plimentary but could be deeply sarcastic.10 Bint Muhammad Hassan replies “Brother
Hisham, may God bless you, I did not think that I was objective, I thought I was biased
towards women!”11

Female Preachers Get Substantial Engagement from Men

I code the gender of commenters by hand using the self-​reported names they list when
leaving comments. I am able to code the gender of 82 percent of the names commenters
list for themselves (170 women and 421 men). The remaining 128 “names” are not
strongly associated with one gender or are not proper names (e.g., “student of life”).
Because this process is time-​consuming, I only do this for the matched sample of women
from saaid.net.
I expected to find that female authors would get more comments from women, be-
cause I found previously that these same female authors got more responses from female
users of the social media platform Twitter (Nielsen 2020). I find only weak evidence of
 The Rise and Impact of Muslim Women Preaching Online    513

a difference. On average, the matched articles by men get approximately 23 percent of

their comments from women, which is similar to the percentage of women’s comments
on men’s documents in the unmatched sample (between 19 percent and 39 per-
cent based on a random sample of fifty articles). Articles by women get an estimated
10 percentage points more comments from women, meaning that roughly 33 percent
of their comments come from women. However, this increase is imprecisely estimated
(p = 0.176), so this difference is not statistically different from zero according to con-
ventional thresholds. I cannot conclude from these results that female authors get more
comments from women.
Substantively, these results show that the majority of commenters on women’s
posts are men, corroborating my finding (in Nielsen 2020) that female preachers are
garnering reactions from male audiences, too. The willingness of female authors to
engage in debates with male commenters undermines the notion that female Muslim
authorities are writing solely for women.

Commenters on Women’s Documents Are More

Geographically Diverse
The comments on women’s writing comes from more diverse locations, even for the fe-
male preachers from saaid.net, who themselves are mostly from Saudi Arabia. To show
this, I match the location listed by each commenter to a country, which is possible for
914 of the 983 comments. Sixty-​eight percent of commenters come from four coun-
tries: 25 percent from Saudi Arabia, 22 percent from Egypt, 13 percent from Algeria, and
8 percent from Morocco. The remaining 31 percent come from thirty-​two other coun-
tries, with Iraq (4.7 percent) and the UAE (3.9 percent) at the top, and Sweden, Senegal,
Russia, Australia, Mauritania, Kenya, and Israel at the bottom, with one comment each.
I measure the diversity of commenter locations in two ways. First, for each document,
I count the total number of countries from which comments come. Second, I count the
number of comments on each document that came from commenters not in the top
four countries. Documents by male authors typically get comments from 1.8 countries,
while documents by female authors get comments from 2.5 countries on average. The
fact that a woman is the author doubles the average number of comments from outside
of Saudi Arabia, Egypt, Algeria, and Morocco from 0.8 to 1.6. These results support the
argument that a major role female preachers play is to reach out to new audiences who
are more diverse.

Women Get More Positive Comments Than Men

I examine the content of comments in two ways. First, I test whether the substance of
the comments on women’s and men’s documents are different, using a structural topic
514   Richard A. Nielsen

0 0.1 0.2 0.3

2. Relationships: husband, man, woman, marriage, woman, sister, house, second, many
1. Praise: May God reward you, bless, benefit, distinguish, good
3. Compliments: article, author, wonderful, thank you, beautiful, topic, style
4. Miscellaneous: Doctor, big, good luck, well, professor, day, special, yes
5. The Prophet: peace be upon him, prophet, Muhammad, almighty, bless, world
6. Doctrine: Islam, Muslim, religion, Arab, hadith, said, Imam, I know

0 0.1 0.2 0.3

Estimated topic proportion
comments for female authors
comments for male authors

Figure 24.3: Comments on documents by male and female authors differ.

Note: Estimated topic proportions for comments left on documents by female and male authors
on the alukah.net website. Topics are ranked by the difference in proportions between women
and men.

model. I find that there are notable differences, shown in Figure 24.3. The comments
on women’s documents discuss issues relating to relationships more, and use terms re-
lated to other aspects of Islamic doctrine and hadith citation less. Recall that all of the
articles in this sample have been matched topically, so there are just as many articles
devoted to family relationships by men. This suggests that religious norms are still influ-
ential. Commenters engage female authors more on “women’s issues,” (family, children,
and relationships), presumably because they are considered to have natural expertise in
these areas by virtue of their gender.
The topic model produces two topics expressing positive sentiment. Topic 3 (Compli­
ments) places high probability on words expressing positive sentiment in secular terms:
“thank you for your article; what wonderful style.” Commenters use these terms equally
when reacting to the writing of men and women. Topic 1 (Praise) emphasizes words that
express positive sentiment religiously, such as “May god reward you.” Women’s writing
receives detectably more of this religious style of praise; 26 percent of words in comments
on women’s writing are praise, compared to only 22 percent for men’s writing.
To further examine the tone of the comments, I use a dictionary-​based approach.
I look at the 179 words in the document-​term matrix for the comments and classify fif-
teen of them as praise words.12 I then count the number of times these praise words
are used in the comments on each article. I find that female preachers get more pos-
itive comments than male preachers, measured either as the number of praise words
per comment or the proportion of each comment that is praise. Intriguingly, when
I analyze comments by men and women separately, I find that this praise for female
authors appears to come especially from men, although the statistical models do not
give enough precision to be confident of this conclusion at traditional statistical signifi-
cance thresholds.
 The Rise and Impact of Muslim Women Preaching Online    515

From my reading of exchanges between commenters and authors, the vast majority of
this praise is sincere, and is accepted sincerely by the authors. The only example of sar-
castic or patronizing praise I have come across is the ambiguous comment on the article
“Diary of an Exemplary Husband and a Grumpy Wife,” discussed above.
I find little evidence of backlash negativity from men (or women) on this platform.
I do find critical comments, but these tend to be substantive rather than insulting or de-
rogatory. This is presumably because comments are moderated for negativity, though
I received no response when I tried to confirm this with the website administrators.

Discussion

Taken together, my findings paint a picture of how lay Muslims are responding to the
rise of online female preachers. Users of a major Islamic aggregator website are some-
what less likely to click on the writings of female authors, suggesting that Arabic-​
speaking Muslim audiences worldwide are less interested in women than men as
religious authorities. Despite this lower level of general interest, it is worth noting the
sheer quantity of people who read the writings of these women: 83.6 million page views
for all of the women on the website. Even if many Muslims still prefer male authorities, it
is hardly plausible that these women have no influence.
Although female preachers have lower readership numbers, once website users click
on the writings of women, they are often more engaged. Readers leave more comments
on documents by women, and those comments contain more praise. This increase in
comments appears to be driven by the propensity of female authors to engage more with
their readers in the comment section. Especially notable is the fact that men differen-
tially praise female authors more in their comments. Presumably, selection effects are at
play—​that is, the type of man who would criticize a female authority in the comments
is opting simply not to click. But this is very different from many open forums on the
Internet where women claiming some authority or expertise are heaped with sexist
abuse. And it suggests that there is a large segment of the male Muslim population
worldwide that views female religious authority in a positive light.
There is no way to be sure that the gender ratio in the comments matches the gender
ratio of page views, but if it does, then the writings of these women have reached upwards
of 40 million men since 2006, and it is truly remarkable that almost none of them have
taken to the comments to complain. Comments on the website are moderated, so I am
not surprised that there is no name-​calling that I can find, but there is certainly crit-
icism, especially when women articulate views that challenge patriarchal norms. The
fact that this criticism remains a minority of comments, and that, on the whole, men are
more likely to praise female authors than male authors, is remarkable. Further research
might investigate reactions to women’s Islamic writing on unmoderated websites and
forums to see if they elicit significant backlash there.
516   Richard A. Nielsen

All of this matters because attitudes about the appropriate roles for women in so-
ciety remain very conservative in the Arab world. The Arab Barometer surveys from
2018–​2019 show ostensible support for the idea that a woman can be a head of state, with
every country surveyed agreeing at 50 percent or more except Algeria.13 But this agree-
ment masks a stark difference in men’s and women’s attitudes. Among men, support for
women’s access to the highest office is reliably 15 points lower than among women, and
falls below 50 percent in about half of countries surveyed. Moreover, these numbers do
not fully capture the depth of patriarchal attitudes, which becomes apparent when Arab
Barometer respondents answer other questions about gender (Benstead 2020).
My findings lead me to wonder whether Muslims interacting with authorita-
tive preaching by women online might begin to change their views about women in
positions of authority, even if the content of the preaching itself supports patriarchal
values. These results show that some men may be interested in women’s religious writing,
hinting that the preconditions may be in place for large shifts in the gender norms of
Muslim societies. My results also suggest that the Internet might contribute to shifting
ideas about the proper roles of women in society, which could have far-​reaching effects.
However, while these trends could presage increasing space for women’s authority in the
Arab Muslim world, they may represent an evolution of patriarchal norms, rather than
the abandonment of those norms. Patriarchy can come in many flavors. As Benstead
(2020) shows, different measures of gender attitudes in the Arab Barometer produce
markedly different pictures of patriarchy in the region. There are “varieties of patri-
archy” across contexts (Blaydes and Platas 2019; Kandiyoti 1988), and these results may
signal a shift from a patriarchy that forbids women from entering the religious sphere to
one that allows certain women to have authority while pedastalizing and constraining
them. Male support for women’s writing may be conditional on whether women pre-
sent “acceptable” messages. Although I find that female authors get significantly more
praise on average, it is revealing that the most critical, negative comments I found are in
response to a female author, Sara Bint Muhammad Hassan, who implicitly criticizes pa-
triarchal practices.14
Yet even if the online environment I study in this chapter is patriarchal in many ways,
it allows significant space for women’s agency. My assessment from reading some of
these articles and the exchanges between readers and authors in the comments is that fe-
male authors generally find publishing Islamic writing to be an empowering experience.
Female authors appear to take pleasure in sharing their writing with an engaged, mixed-​
gender audience. While reading through the comments, I find that female authors often
thank their readers for commenting, which in turn further increases reader engage-
ment. Even when comments become critical, I find that female authors feel empowered
to push back against aggressive disagreement from male readers. The tone is always civil,
with frequent references to shared Islamic values of respect between fellow Muslims, but
these women do not generally acquiesce when a male reader tells them they are wrong.
This corroborates previous findings that women can enjoy increased autonomy, agency,
and pleasure while participating in and supporting patriarchal social organizations
(Ben Shitrit 2016; Deeb 2011; Inge 2017; Kandiyoti 1988; Mahmood 2005; Ulrich 2017).
 The Rise and Impact of Muslim Women Preaching Online    517

It is important to begin to understand the effects of visible female preachers on

Muslim societies. Female preachers are already working in various official capacities
in a number of other countries, including Morocco and Turkey, and Saudi Arabia has
flirted with employing official female muftis, potentially hiring some of the women
I study here. Unfortunately, the evidence in this chapter is not sufficient to say with con-
fidence how these societies might respond when these governments appoint women to
positions of religious authority. All of the preaching I study here is mediated through the
Internet, as are the comments, which are moderated. Would the same reactions prevail
if these women were preaching in person to mixed-​gender audiences? Is moderation
of some sort necessary to induce positive reactions to women’s preaching? Will women
garner positive reactions no matter what topics they preach on, or is audience praise
conditional on women confining their preaching to “women’s issues?” I cannot yet de-
finitively answer any of these questions, and further research is needed. However, my
analysis of reactions to women’s preaching online suggests reasons for optimism. If a
significant number of lay Muslims react positively to female preachers on the Internet,
female Muslim authorities may eventually enjoy a wide and positive reception in other
spheres.

Acknowledgments
Many people have helpfully commented on this piece and my related research published
elsewhere. Thanks to Melani Cammett, Pauline Jones, Lisa Blaydes, Mark Tessler, Peter
Krause, Lindsay Benstead, Sarah Mohamed, Steven Brook, Amaney Jamal, Tarek Masoud,
Marc Lynch, Mirjam Künkler, Ana Weeks, Marsin Alshamary, Rebecca Nielsen, Bruce
Rutherford, Zehra Arat, Henri Lauzière, Aaron Rock-​Singer, Ari Schriber, and Malika
Zhegal, and audiences at the Oxford Handbook workshop, the NYU Center for Data Science,
the University of Connecticut, the Northeast Middle East Politics Working Group, the
2016 AALIMS conference, the 2016 MESA conference, and the Salafiyya Workshop held at
Harvard on March 11, 2016. This publication was made possible by a grant from the Carnegie
Corporation of New York, supporting me as an Andrew Carnegie Fellow. The statements
made and views expressed are solely my responsibility. The data and code to replicate all
analyses in this chapter are available on the Harvard Dataverse Network, at https://​doi.org/​
10.7910/​DVN/​SVTH0W.

Notes
1. Submissions can be made through https://​www.alukah.net/​contribute.aspx or https://​
www.alukah.net/​researches/​contribute.aspx, depending on whether the submission is
“scholarly.”
2. Authors would be rightfully annoyed if I randomly left profane and critical comments on
their articles in my attempts to probe the limits of the comment moderation.
3. Subsequent pages of comments must be collected by hand, a feature of the website that
shapes my analysis strategy.
518   Richard A. Nielsen

4. I also plot 95 percent confidence intervals, but these are so narrow that they are obscured
by the disks showing the point estimates.
5. These categories are sharia (18,701), culture (6,767), library (5,462), literature_​language
(5,089), fatawa_​counsels (3,704), social (3,694), spotlight (2,627), web (2265), world_​
muslims (417), publications_​competitions (73), and translations (27).
6. Coarsened exact matching requires analyst-​defined cut-​points. For the fifteen topics, I use
breakpoints of 10 percent and 50 percent. I use three evenly spaced cut-​points for years
since posting (0 to 3.25 years, 6.5 years, 9.75 to 13 years) and, when relevant, the number of
page views (0 to 2226, 5282, and 13,132 to 4,562,169).
7. I remove comments by the authors themselves from this analysis.
8. https://​www.alukah.net/​social/​0/​9510, accessed April 1, 2020.
9. Comments made at 13-​03-​2010, 06:24 a.m., and 18-​07-​2010, 11:46 p.m., on https://​www.
alukah.net/​social/​0/​9510, accessed April 1, 2020.
10. Comment made at 19-​ 02-​2010, 03:23 a.m. on https://​www.alukah.net/​social/​0/​9510,
accessed April 1, 2020.
11. Comment made at 18-​ 07-​2010, 11:46 p.m., on https://​www.alukah.net/​social/​0/​9510,
accessed April 1, 2020.
12. The fifteen praise words are: thank, thank, good, good luck, wonderful, benefit, good,
good, beautiful, reward, reward, bless, blessings, most beautiful. Duplicates are due to
translation from Arabic to English.
13. https://​www.arabbarometer.org/​wp-​content/​uploads/​AB_​Women_​August2019_​Public-​
Opinion_​Arab-​Barometer.pdf and https://​www.arabbarometer.org/​wp-​content/​uploads/​
AB_​WomenFinal-​version05122018.pdf, accessed October 25, 2019.
14. See my discussion above of Sara Bint Muhammad Hassan’s article “Diary of an Exemplary
Husband and a Grumpy Wife,” available at https://​www.alukah.net/​social/​0/​9510.

References
Al-​Rasheed, Madawi. 2013. A Most Masculine State: Gender, Politics, and Religion in Saudi
Arabia. New York: Cambridge University Press.
Bano, Masooda, and Hilary Kalmbach, eds. 2012. Women, Leadership and Mosques: Changes in
Contemporary Islamic Authority. Leiden: Brill.
Ben Shitrit, Lihi. 2016. Righteous Transgressions: Women’s Activism on the Israeli and Palestinian
Right. Princeton, NJ: Princeton University Press.
Benstead, Lindsay J. 2020. “Conceptualizing and Measuring Patriarchy: The Importance of
Feminist Theory.” Mediterranean Politics (Online First, February 19).
Blaydes, Lisa, and Melina R. Platas. 2019. “Religion, Patriarchy and the Perpetuation of
Harmful Social Conventions: Explaining Women’s Attitudes toward Female Genital Cutting
in Egypt.” Forthcoming in the Journal of Demographic Economics. Available online at https://​
blaydes.people.stanford.edu/​sites/​g/​files/​sbiybj1961/​f/​fgm1.pdf.
Deeb, Lara. 2011. An Enchanted Modern: Gender and Public Piety in Shi’i Lebanon. Princeton,
NJ: Princeton University Press.
Hassan, Mona. 2011. “Women Preaching for the Secular State: Official Female Preachers
(Bayan Vaizler) in Contemporary Turkey.” International Journal of Middle East Studies 43,
no. 3: 451–​473.
 The Rise and Impact of Muslim Women Preaching Online    519

Ho, Daniel E., Kosuke Imai, Gary King, and Elizabeth A. Stuart. 2007. “Matching as
Nonparametric Preprocessing for Reducing Model Dependence in Parametric Causal
Inference.” Political Analysis 15, no. 3: 199–​236.
Iacus, Stefano M., Gary King, and Giuseppe Porro. 2012. “Causal Inference without Balance
Checking: Coarsened Exact Matching.” Political Analysis 20, no. 1: 1–​24.
Inge, Anabel. 2017. The Making of a Salafi Muslim Woman: Paths to Conversion. New York:
Oxford University Press.
Kandiyoti, Deniz. 1988. “Bargaining with Patriarchy.” Gender & Society 2, no. 3: 274–​290.
Kloos, David, and Mirjam Künkler. 2016. “Studying Female Islamic Authority: From Top-​
Down to Bottom-​Up Modes of Certification.” Asian Studies Review 40, no. 4: 479–​490.
Le Renard, Amélie. 2012. “From Qur’anic Circles to the Internet: Gender Segregation and the
Rise of Female Preachers in Saudi Arabia.” In Women, Leadership, and Mosques: Changes in
Contemporary Islamic Authority, edited by Masooda Bano and Hilary Kalmback, 105–​126.
Leiden: Brill.
Le Renard, Amélie. 2014. A Society of Young Women: Opportunities of Place, Power, and Reform
in Saudi Arabia. Stanford, CA: Stanford University Press.
Mahmood, Saba. 2005. Politics of Piety: The Islamic Revival and the Feminist Subject. Princeton,
NJ: Princeton University Press.
Miller, Michael K. 2013. “The Uses and Abuses of Matching in Political Science.” Available at
https://​sites.google.com/​site/​mkmtwo/​Miller-​Matching.pdf, accessed April 8, 2020.
Nadwī, Muammad Akram. 2007. Al-​Muaddithāt: The Women Scholars in Islam. Oxford:
Interface Publications.
Nielsen, Richard A. 2015. “Ethics for Experimental Manipulation of Religion.” In Ethics and
Experiments: Problems and Solutions for Social Scientists and Policy Professionals, edited by
Scott Desposato, 42–​65. London: Routledge.
Nielsen, Richard A. 2020. “Women’s Authority in Patriarchal Social Movements: The Case of
Female Salafi Preachers.” American Journal of Political Science 64, no. 1: 52–​66.
Rausch, Margaret J. 2012. “Women Mosque Preachers and Spiritual Guides: Publicizing
and Negotiating Women’s Religious Authority in Morocco.” In Women, Leadership and
Mosques: Changes in Contemporary Islamic Authority, edited by Masooda Bano and Hilary
Kalmbach, 59–​83. Leiden: Brill.
Roberts, Margaret E., Brandon M. Stewart, and Richard A. Nielsen. 2020. “Adjusting for
Confounding with Text Matching.” American Journal of Political Science 64, no. 4: 887–​903.
Roberts, Margaret E., Brandon M. Stewart, Dustin Tingley, Christopher Lucas, Jetson Leder-​
Luis, Shana Kushner Gadarian, Bethany Albertson and David G. Rand. 2014. “Structural
Topic Models for Open-​Ended Survey Responses.” American Journal of Political Science 58,
no. 4: 1064–​1082.
Ulrich, Laurel Thatcher. 2017. A House Full of Females: Plural Marriage and Women’s Rights in
Early Mormonism, 1835–​1870. New York: Knopf.
 Chapter 25

Religion a nd
Mobiliz ation i n t h e
Syrian Uprising a nd Wa r

Wendy Pearlman

Many popular commentaries on the Syrian uprising and war emphasize the role of re-
ligion, or of animosities related to religious difference, in propelling conflict (Reuters
2012; Council on Foreign Relations 2013; Vallely 2014). Many Syria experts critique this
view, suggesting that violence has fueled Syria’s sectarianization more than sectarianism
has fueled violence (Droz-​Vincent 2014; Yassin-​Kassab and Al-​Shami 2015; Pinto 2017).
This chapter affirms the second view and builds on it in three ways. First, it addresses
these questions with a unique and extensive data source: open-​ended interviews that
I have conducted with more than four hundred displaced Syrians in Jordan, Turkey,
Lebanon, the UAE, Germany, Sweden, Denmark, and the United States from 2012 to
2018. Second, it brings together in one analysis two dimensions of religion: as a set of
beliefs and practices with political implications for ideology, public piety, and state
policies (religion); and as a collective identity, social structure, and marker of differ-
ence within a heterogeneous society (sect). Third, it sharpens the focus on changes over
time by reviewing the role of religion and sect in Syria’s pre-​2011 politics and asking four
questions pertaining to successive stages of the uprising: (1) What was the role of religion
in the motivations driving protest? (2) What was its role in the processes generating col-
lective action? (3) What was its role in the initial escalation of mobilization to militarized
means and radicalized ends? (4) What was its role in transforming mobilization from
uprising to war?
I argue that while religion came to occupy an increasingly prominent place in mo-
bilization over time, its role in the Syrian conflict has been less attributable to religion
per se than to the ways religion is entwined with power, privilege, and the dynamics of
violent conflict itself. Where religion has sometimes appeared significant, such as in the
tendency of demonstrations to begin at mosques, its power lay not in piety but in struc-
tural constraints. Religion and sect became increasingly important in mobilization as
522   Wendy Pearlman

the conflict escalated and transformed into war. Yet this too was largely an outgrowth
of contentious dynamics not particular to Islam, including state repression, regime
strategies of divide and conquer, and the impact of horrific violence in reshaping actors
and understandings.
The chapter illustrates these arguments with a focus on the conflict’s first two years.
The first section offers brief historical background on the nexus between religion and
politics in Syria. The next four sections examine, respectively, how religion relates to
protest motivations, collective action processes, the escalation of contention, and the
transformation of conflict. The final section concludes with a discussion of the value
of viewing religion in political mobilization in Muslim countries primarily as a var-
iable that interacts with other variables, rather than as an independent driver in it is
own right.

Historical Context

On the eve of the Arab Spring, an estimated 69 percent of Syria’s population was Sunni
Muslim, and 14 percent was Christian. About 11.5 percent was Alawite, a historically
persecuted and marginalized offshoot of Shiʿi Islam. The remainder pertained mostly to
other Muslim sects, primarily Druze or Ismaʿili (van Dam, 2011, 1).
While most of modern Syria was distinguished by coexistence, different religious
groups historically had different statuses and rights. Ruling Syria as a Mandate from
1920 to 1946, France encouraged sectarian identities in order to counter the nationalism
of the Arab-​Sunni majority, and also recruited minorities into the military (van Dam
2011, 4). Marginalized communities’ receptiveness to that avenue of social mobility, as
well as the tendency of the urban Sunni bourgeoisie to shun military careers, increased
minorities’ disproportionate representation in the armed forces.
In the decades following independence, military interventions overthrew sev-
eral governments. A faction of the Baʿth Party seized control of the state in 1963. Baʿth
officers’ origins in peripheral communities alienated them from urban folk and the
Sunni majority (Batatu 1999, 171), while the party’s secular pan-​Arabism offended many
pious Muslims. In 1970 General Hafez al-​Assad sidelined competitors and seized power.
Over the next three decades, he established a durable authoritarian regime based on a
well-​organized ruling party, populist welfare policies, alliances across society, and ex-
tensive employment in a bloated public sector (Hinnebusch 1990; Perthes 1997; Batatu
1999). A far-​reaching coercive apparatus monitored the population, restricted speech,
detained people at will, and tortured political prisoners (Middle East Watch 1991).
Against a backdrop of accumulating grievances regarding inflation, corruption, and
security force abuses, civic and professional associations agitated for human rights in
the late 1970s. A breakaway faction of the Muslim Brotherhood, the country’s strongest
opposition movement, began a campaign of violent attacks against the regime. Assad
retaliated against the Brotherhood, spurring the group’s radicalization and 1982
 Religion and Mobilization in the Syrian Uprising and War    523

insurrection in Hama (Lefèvre 2013). The regime responded with a scorched-​earth as-
sault that killed tens of thousands, effectively eradicating the Brotherhood’s domestic
presence.
Assad’s rule crystalized complex, even contradictory, relations between religion, sect,
and politics. With regard to religion as belief, the regime constructed an image of it-
self as the protector of the country’s religious minorities. It demonized the Muslim
Brotherhood, painting it as the treacherous “them” against which the regime was “us”
(Khatib 2018). Apprehensive of any regrouping of Islam-​inspired opposition, regime in-
telligence monitored imams and mosques and suspended the licenses of many religious
charitable associations (Abboud 2018, 45). It banned even nonpolitical expressions of
Islamic practice and piety, such as wearing headscarves in schools (Pierret 2013, 70–​7 1).
Yet the regime was at least as repressive toward secular dissidents (Khatib 2018). It
also courted Islam as an expedient to co-​opt, placate, and maintain control of society.
Saleh (2017, 215–​217) notes that the regime thus preserved Sunni Islam’s dominant po-
sition as Syria’s “public religion,” allowing it to have “nearly universal jurisdiction” over
religious education, civil status, and public festivities. In addition, Assad undertook
symbolic acts such as public prayer and quoting from the Qurʾan (van Dam 2011, p.142).
He built alliances with ʿulamaʾ, granting favors to religious leaders deemed loyal while
marginalizing or eliminating those who were not (Pierret 2013).
The regime’s relationship to religion as a social structure and collective identity was no
less complicated than its relationship to religion as faith. On the one hand, Baʿth Party
ideology was not sectarian. The political dominance of Alawite officers after 1963 was
due primarily to the “ruralization of the armed forces” and splits among Sunni officers
(Batatu 1999, 156–​159). Assad placed members from his own family, region, and sect at
the helm of the military and security services. Nonetheless, many of Assad’s confidants
were not Alawite. Perthes (1997, 182) was adamant: “Confession as such . . . is not the
criterion for membership in this circle; loyalty to the President is.” Furthermore, the re-
gime built a decisive partnership with Sunni Damascene merchants (Ismail 2010, 23).
Throughout the 1970s, Assad created predominantly Alawite military units and of-
ficer corps for the purpose of regime protection and “coup-​proofing” (Quinlivan 1999,
147). These units were part of larger Alawite trust networks, in Tilly’s (2004, 12) sense of
strong ties, “within which people set valued, consequential, long-​term resources and
enterprises at risk to the malfeasance, mistakes, or failures of others.” Nevertheless the
lynchpin of this strategy was not religious identity as much as the structuring of relations
of inclusion and exclusion through “special loyalties” and interests (Quinlivan 1999,
141). Indeed, some of Assad’s fiercest challengers—​the Baʿth Party rival from whom he
seized power in 1970, and his own brother, who attempted a coup in 1983—​were also
Alawite.
Thus, sect alone cannot explain political allegiance. Most Alawites did not directly
benefit from Assad’s rule, and some opposed it. Nevertheless, these policies allowed this
once under-​represented community to profit “over-​proportionally” from public sector
employment (Perthes 1997, 183) and in turn discriminate against others (van Dam 2011,
139–​140). Farouk-​Alli (2014, 217) observes that “the bitter compromise that was struck
524   Wendy Pearlman

by the Alawi community with the Assad regime was the acceptance of dictatorship in
return for social mobility.” At the same time, a totalizing discourse of national unity
prevented frank conversation about interconnections of sect, power, and privilege.
Instead, Saleh (2017, 279) lamented that “to bring up the phenomenon of sectarianism
was to find oneself accused of sectarianism.”
Tensions in the interplay between religion and politics took new forms after 2000,
when Hafez al-​Assad died and his son Bashar assumed the presidency. Tightening its
alliance with Iran, the regime increased the visibility of Shiʿism through such measures
as reconfiguring mosques and permitting Iranian Shiʿi missionary activity (Sindawi
2009; Pinto 2011, 2017). Simultaneously, the regime fused Islamic and national idioms in
new ways, from billboards imploring God to protect Syria to the celebration of Aleppo
as the “Capital of Islamic Culture” (Pinto 2011). Bashar also embarked on what Lister
(2015, chap. 3) called a “flirtation with Jihadism.” Following the 2003 toppling of Saddam
Hussein, the regime facilitated fighters’ passage into Iraq to join the insurgency against
American forces. As some Jihadists returned to Syria in 2006, the government impris-
oned them, lest they target the Assad regime as another “apostate enemy.” That turn
of policy added another layer of complexity to Syria’s complex history of religion and
politics.

Motivations for Mobilization: The

Role of Religion in Driving Protest

After four decades, many Syrians were fed up with authoritarian rule. However, few
dared to voice their criticism publicly until the wave of popular protest known as the
“Arab Spring.” Syrians joined in a few tentative demonstrations in March 2011 that
slowly spread over space, were sustained over time, and launched a nationwide uprising.
The main drivers of protest were the calls for greater political freedoms, accountability,
and good governance, as well as frustration with corruption, everyday indignities, and
deteriorating livelihoods (Yassin-​Kassab and Al-​Shami 2015; Abboud 2018; Baczko,
Dorronsoro, and Quesnay 2018). These grievances and aspirations were overwhelm-
ingly civic, political, liberal, and unconnected to religion. For most Syrians, where reli-
gion did emerge as a motivator for protest, it was not related to religion per se, but rather
to the ways that religious identities were infused with unequal power and the denial of
basic liberties.
In my interviews, ostensibly religious aspects of fundamentally nonreligious
grievances came to light in three ways. First, many invoked sect in their criticism of un-
fairness and discrimination. “You can find people from different sects living together,
and we don’t have any problem with this,” a woman said. “Our dispute is with favoritism”
(Interview in Irbid, Jordan, September 27, 2013). Resentment of this sort often focused
on public-​sector employment, which employed at least a quarter of Syrian workers and
 Religion and Mobilization in the Syrian Uprising and War    525

usually offered greater benefits than private-​sector jobs (Abboud 2013, 166). A Sunni
man from Homs, where Alawite residents worked for the state in disproportionate
numbers (Nakkash 2013, 1), expressed his frustration:

We go into the branches of the government, and 95 percent of the employees are from
the Alawite sect, which is his sect. Here is how you get a job: by having connections
to his sect. After they occupied the country, they gave the country to his sect to rule.
It’s all about his possessions and interests. I studied. But if your father is an officer in
the army from the Alawite sect, there is no need for education. This is what made the
young [rise up]. (Interview in Zarqa, Jordan, September 24, 2012)

Second, some who invoked religion as grounds for protest insisted that they were not
calling upon the state to expand the role of Islam in the public sphere, but rather to with-
draw policies inhibiting Muslims from practicing their faith. A man from the Damascus
suburbs explained such religious grievances as a matter of rights and dignity:

One of my friends is a Muslim and prays five times a day. Then he did his military
service in Syria and wasn’t allowed to do his prayers. The only place that he could
pray was in the bathroom. This is not right in Islam, but if you can’t pray in the
right place, you do it in wherever you can any time. (Interview in Amman, Jordan,
October 16, 2012)

A married couple echoed these views:

Wife: Some people say that our revolution is an Islamic revolution, but that is not
true. Yes, it’s not a religious revolution but we want to be free to practice our religion
because they refuse to let us pray. A Saudi company built Tishreen Hospital. They
built a small room inside the hospital to be a prayer room, because it’s part of any
Saudi design. But the hospital made it into a storage room instead. Because it’s a mili-
tary hospital, we aren’t allowed to have congregational prayer. We accept that, but on
top of that, they don’t allow us to pray individually, either. Also, it’s not allowed for
men to have beards.

Husband: In Aleppo University, our prayer room was in the basement. We didn’t
have a problem with that, but whenever we finished praying, the security people
would write down our names.

Wife: One time, the security officer . . . asked our friend why he was praying during
his shift inside the hospital. He answered that he was working 24 hours a day, which
means he had to pray at the hospital. The officer told him, “You will not pray here. Go
pray in your house as much as you want, and pray for us.” They even mock religion.

Third, some referred to Bashar’s enlargement of space for Shiʿi Islam, which far
outpaced the approximately 1 or 2 percent of the population that was Shiʿi. One of my
526   Wendy Pearlman

interlocutors expressed such discontent, clarifying that it was not related to ideas about
heresy, but rather a sense that the government privileged foreigners at the expense of its
own citizens:

During the rule of Bashar al-​Assad there was a kind of sectarian provocation, and es-
pecially toward Sunni Muslims. He began to promote the Shiʿi school of thought and
build Shiʿi mosques . . . . It contributed to a strong feeling inside people. A feeling that
there was a war in the country on one particular sect. This contributed to a sense of
psychological submission among people. There was a sense that Shiʿa are expanding
in the country, but Sunnis can’t expand. I tell you this to show you that people’s psy-
chological situation was one of being strangulated, in sectarian terms. (Interview in
Amman, Jordan on September 16, 2012)

For many Syrians who supported the uprising, therefore, religion factored into their
motivations not as a primary grievance, but as an outgrowth of their frustration with au-
thoritarianism. They were not moved by a call for a greater role of religion as much as for
a state that approached religion with a fairer, or even a neutral, stance.

Processes of Mobilization: The Role

of Religion in Collective Action

Once protest was under way, initial protest was not propelled or structured by religion,
or even ideology, broadly speaking. The tendency of popular demonstrations to form at
mosques at the conclusion of Friday prayers, often perceived as religiously motivated,
was in fact the outcome of collective action logistics. Due to decades of the stifling of
civil society (Yassin-​Kassab and Al-​Shami 2015; Abboud 2018; Baczko, Dorronsoro,
and Quesnay 2018), Syrian dissenters lacked a preexisting infrastructure of autono-
mous organizations, free social spaces, and robust networks of trust that could serve as
the “mobilizing structures” that social movement scholars find crucial for organizing
challenges to authority (McAdam, McCarthy, and Zald 1996).
In addition, pervasive security forces ready to crush dissent rendered mosques the
only space from which a crowd could get a demonstration off the ground. Although
mosques were state-​ monitored and informants could be clandestine among
worshippers, the ritual of weekly collective prayer provided one of the few intersections
of time and place that permitted a critical mass to form (Pearlman 2020). Repression
severely constrained opportunities to gather in other spaces. For example, Homs’s
attempt at a Tahrir-​like peaceful occupation of a public square lasted only hours be-
fore security forces opened fire, killing at least seventeen protestors (Human Rights
Watch 2011).
The utilitarian, rather than religious, function of mosques came to light in my
interviews with participants in the uprising’s first large-​scale demonstration in the
 Religion and Mobilization in the Syrian Uprising and War    527

southern province of Daraa on March 18, 2011 (Pearlman 2018). Activists explained
that they had planned to hold a demonstration in front of the central court building
on March 15, in conjunction with the call on Facebook for a nationwide “Day of Rage.”
Though they circulated word as secretly as possible, security forces were waiting on the
scene and quashed the demonstration before it began. A small group of activists de-
cided to try again after Friday prayers. “Why Friday prayers?” one Daraa resident asked
rhetorically. “Because, in Syria, it is prohibited for more than five people to assemble
without prior security approval . . . . Prayer inside the mosque is the only opportunity
people have to gather” (Interview in Amman, Jordan, September 16, 2012).
Had these oppositionists selected a mosque on the basis of religious significance, they
might have chosen the city’s main religious institution, al-​Omari. Yet they knew that
this would attract a large security presence, and instead chose to demonstrate from a
new and scarcely known mosque in a peripheral neighborhood. This offered the best
conditions for a demonstration to gather participants before it could be suppressed.
The oppositionists’ plan succeeded; they launched a demonstration that gathered hun-
dreds and became the most sustained street protest to date. Ending when security forces
killed two protestors, it was also a turning-​point in launching the uprising (Pearlman
2020). The following Friday, tens of thousands joined in demonstrations across the
country, most of them likewise gathering inside mosques and pouring forth from there
(Slackman 2011).
Friday protests became a regularized expectation for protestors and security
forces, each taking on a unique name through a weekly process of online deliberation
(Friedman 2011). Many people I interviewed insisted that both non-​Muslims and non-
practicing Muslims came to mosques for the purposes of protesting. One worshipper
recalled a young man at Friday prayers with a necklace tucked inside his shirt:

He lined up and prayed with everyone else. And when he bowed, the necklace fell
out. The pendant was a cross. People said to him, “Either you’re wearing that neck-
lace by mistake, or you came to the mosque by mistake.” And the Christian young
man said, “I came here to go out in the demonstration with all of you.” (Interviewed
in Marj al-​Hamam, Jordan, October 2, 2012)

Similar dynamics took place on Muslim religious holidays that, like Friday prayers,
brought people together with potential for protest. A Damascene described Laylat al-​
Qadr (Night of Power), the holiest night of the Islamic calendar, in August 2011:

Everybody said that the regime would collapse during the month of Ramadan, be-
cause instead of gathering at the mosque for prayer only on Fridays, people gathered
every night. The atmosphere was pumped with energy.
The twenty-​seventh of Ramadan is a holy day, and people stay up all night praying
and reading the Qurʾan. Every year over five thousand people gathered at the mosque
near our house. Volunteers from the neighborhood helped prepare a meal for people
to eat before sunrise. I don’t pray, but I always participated in preparing the meal, be-
cause I thought it was a beautiful social event.
528   Wendy Pearlman

People started arriving. There were a lot of old people, but also guys with body
piercings and strange haircuts. You could see that they had no idea what to do. Some
guys were wearing shorts, which you aren’t supposed to do at a mosque. Out of re-
spect, they were trying to pull their shorts down toward their ankles. But then you
could saw their backsides exposed. It was a beautiful scene of the complex social
fabric that we had in Damascus. (Interview in Copenhagen, Denmark, June 12, 2016)

Like holy places and days, religious language also served mobilizational functions.
While the most common shout was simply the word “freedom,” protests also included
chants such as “God is Great” and “There is no god but God.” Many of my interviewees
explained that these statements of devotion carried a religious power for Muslims specif-
ically because they helped invigorate them with the courage to “break the barrier of fear”
and face the high risk of protest. The language of Islam was thus a “meta-​political” um-
brella that could encompass Islamist and secular leftist discourses (Baczko, Dorronsoro,
and Quesnay 2018, 179), in part because it mirrored the religious nationalism that Bashar
himself had promoted during the prior decade (Pinto 2017, 126). A powerful example of
protestors seizing upon existing frames at the intersection of piety and patriotism was
the replacement of the government slogan “God, Syria, Bashar, and that’s all!” with the
revolutionary “God, Syria, Freedom, and that’s all!”
At the same time, protestors used anti-​sectarian slogans, such as “No to sectarianism,
one, one, the Syrian people is one,” and “We are not (Muslim) Brothers, nor foreign
agents, we are all Syria, whether Muslims, Alawites, Druze, or Christians.” In an expres-
sion of cross-​sectarian outreach, protestors named Fridays to honor Alawite, Christian,
or Kurdish communities, or to emphasize inclusive national unity (Droz-​Vincent 2014,
47, 49).
A final question relevant for assessing the role of religion in collective action concerns
the degree to which religious figures lent leadership to the protest movement. Syrian
preachers based in exile played a different role, as I elaborate below. The positions of
local ʿulamaʾ, however, varied with the communities and networks in which they were
embedded. Muslim clergy in provincial towns where protest was strong and where
the government responded violently tended to endorse the uprising, while those in
Damascus and Aleppo, where the popular response was ambivalent, showed mixed
stances (Pierret 2013, 217). Some establishment figures, like the Grand Mufti personally
appointed by Bashar al-​Assad in 2005, denounced political unrest in terms very similar
to the regime. “Our imams and Muslim scholars let us down,” one oppositionist judged
such personalities. “Their stance was neutral and negative: ‘Oh people, we don’t want
disorder.’ They told us to calm down” (Interview in Zarqa, Jordan, August 11, 2013).
When religious leaders boldly endorsed the uprising, they often did so on the
same nonreligious grounds articulated by demonstrators on the streets. Pierret (2013,
228) notes that, leading clerics released a petition enumerating “distinctly liberal
demands such as the establishment of the rule of law, respect for civil and political rights,
and the abrogation of Article No. 8 of the constitution according to which the Ba‘ath
lead society and state.” The mobilizing impact of local Muslim leaders, as with Islamic
 Religion and Mobilization in the Syrian Uprising and War    529

slogans and places, was thus not primarily due to Islam. Rather, they aided collective ac-
tion to the degree that they strengthened resources and opportunities for a movement in
which religion was ancillary to nonreligious demands and aspirations.

Escalation of Mobilization: The

Role of Religion in Militarization
and Radicalization

It was in the escalation of mobilization that religion, as faith and sectarian identity, came
to play an increasingly prominent role. This was not due to a static or preexisting facet of
protestors’ identities or beliefs. Rather, it illustrates McAdam, Tarrow, and Tilly’s (2001)
insistence that movements evolve with dynamics of contention driven by contingent,
relational mechanisms.
Primarily, in an effort to delegitimize protestors, pro-​regime media explicitly and im-
plicitly accused them of being motivated by sectarianism and religious extremism. In
remarks just nine days after the Daraa protest, government spokeswoman Bouthaina
Shaaban denounced unrest as a sectarian plot (Wimmen 2014, 23). When Assad made
his first public address to parliament three days later, he invoked the word “sectarian” as
many as seventeen times in condemning the uprising (Wimmen 2014, 23). One of my
interviewees from Latakia observed this sectarianization, both in speech and action:

I didn’t even know the difference between Shiʿi and Sunni until this whole thing
started, because nobody cared. Don’t get me wrong—​Shiʿa and Sunnis have been
fighting forever. But nobody was mentioning any of that in Syria at the time. In 2011,
the needs, the goals, and the demands had nothing to do with that. The goals were
political reform, participation, real representation, and some actual active citizen-
ship in the country.

The people in power saw those goals as a fundamental threat to their grip over the
country. The only way they could maintain full control was by reframing the argu-
ment from reform to Shiʿa and Sunnis and radicals. They implemented that policy
through their political speeches and military approach. And they imposed it on the
ground . . . . And eventually things ended up where we are now. (Interviewed in
Copenhagen, Denmark, on June 12, 2016)

In framing the confrontation between government and opposition in religious terms,

the regime sought to “divide and conquer” the uprising. This instrumentalized religious
minorities’ anxieties to scare them into choosing the safety of the status quo they knew
over the uncertainty of a revolution that they did not. In this context, stories circulated
about protestors chanting the phrase “Christians to Beirut, and Alawites in the coffin.”
While reporting on the chant became widespread, most cite it as a rumor rather than
530   Wendy Pearlman

something that witnesses verified having heard (Pinto 2017, 132; New York Times 2011).
Portrayals of the movement as religiously motivated appeared to be deliberate misinfor-
mation, as a writer from the capital observed:

There are certainly Sunni extremists protesting in the streets. But there are also
atheists, communists, and feminists by their side. Although there have been pious
individuals involved in the uprising, and even some who might envision a future re-
ligious state, this has never been a religious movement . . . . Although the fears of
the Christian, Alawite, and Druze minorities are not unfounded . . . they seem to
be exaggerated. The majority of the young revolutionaries and the older opposition
leaders are demanding secular democracy. (New York Times 2011)

In addition to rhetoric, the application of state violence had sectarianizing effects.

A robust literature shows that while repression can paradoxically intensify dissent
(Davenport, Johnston, and Mueller 2005), it often has the effect of splintering nonvio-
lent movements into more “moderate” and “radical” groups (Tarrow 1994), which can
enable the latter to eclipse the former (Schock 2005; Chenoweth and Stephan 2012).
Syria illustrates these dynamics. Observers note that regime forces aimed indiscrimi-
nate and collective repression at Sunni neighborhoods, even in cities where protestors
came from various religious communities (Pinto 2017, 135). They also targeted Sunni sa-
cred sites and objects, such as by desecrating mosques and burning copies of the Qurʾan,
or attacked Sunni rituals, such as shelling communities while people broke the Ramadan
fast (Yassin-​Kassab and Al-​Shami 2015, 110–​111). A sense of being singled out for punish-
ment pushed some Syrians to identify in religious ways that they previously had not.
One activist recalled, “When they started killing Sunni civilians randomly as opposed to
just those protesting . . . my atheist friends began asserting their Sunnism, which is now
more of a social than a religious identity” (Yassin-​Kassab and Al-​Shami 2015, 110).
The identities of the agents of repression intensified the sectarianizing impact of the
targeting of violence. The officer corps of the Syrian army, and especially the elite units
deployed against the uprising, was predominantly Alawite (Droz-​Vincent 2014, 39). That
they took the helm in crushing an uprising in which most participants (like the popula-
tion at large) were Sunni had sectarianizing effects. Military support for the regime from
Iran, Hezbollah, and Iraqi volunteers likewise furthered many Sunni oppositionists’
sense that they were being attacked by a specifically Shiʿi adversary (Pinto 2017, 140).
Compounding sectarianization was the role of Assadist “thugs,” militias, and para-
military units—​different types of armed civilian groups referred to as shabiha (singular,
shabih). Often armed and organized by Syrian intelligence (Lund 2015, 212), shabiha
gained infamy for atrocities against civilians. Videos circulated that showed them forcing
people to say “There is no God but Bashar,” or to prostrate themselves before images of
the president (Saleh 2017, 50). That most, though not all, shabiha were Alawite increased
many Sunnis’ perception that shabiha represented the entire Alawite community.
Here it is important to reiterate that high-​profile members of all religious communities
sided with both the regime and the opposition. Many Alawites felt that their community
 Religion and Mobilization in the Syrian Uprising and War    531

was “taken hostage by the regime” (Droz-​Vincent 2014, 42). Yet others remained si-
lent for fear of retribution or actively stood with the regime for fear of how its collapse
would endanger the community. For many Sunni Muslims, the lack of strongly visible
Alawite participation in the uprising magnified the impact of their strongly visible role
in repression (International Crisis Group 2012, 8). Correspondingly, for many Alawites,
deployment of Alawite-​dominated forces helped the regime project the uprising as anti-​
Alawite, redoubling the belief that the community was fighting for its survival (Rifai
2018, 246).
Against this backdrop, perhaps the greatest factor shaping the role of religion in con-
flict escalation was the enormity of civilian suffering. Yassin-​Kassab and Al-​Shami
(2015, 108–​109) write, “In most cultures the proximity of death will focus minds on the
transcendent—​there are no atheists in foxholes, as the saying goes—​and more so in an
already religious society like Syria’s.” Radicalization in Syria, they argue, is more aptly
understood as traumatization. Echoing this analysis, Saleh (2017, 122–​123) attributed the
rise of a “militant nihilism” to three factors: the regime’s horrifying violence, the ineffi-
cacy of the exile-​based leadership, and the inaction of foreign powers, which left Syrians
feeling abandoned in a world that “is indifferent to them, if not actively conspiring
against them.” Facing loss of trust in everything once believed to be reliable, many found
solace only in religion. A father from Daraa voiced a similar analysis when I spoke with
him in the fall of 2012:

If there is no opportunity, people won’t return to their religion. Only in hard

situations, like what we’re living in now . . . . If the situation continues like this, many
countries will be dragged into a big war. A huge Islamic organization will be created.
It will happen, because no one is responding to us. No one has helped us. No one
has put his hand in our hand and joined with us. Nobody. So what can you do? Just
go back to God. After that, you’ll organize yourself by returning back to the religious
fundamentals . . . . Islam needs an opportunity and this opportunity is ready now
in Syria.
In Syria now, everybody is returning to God. It’s clear nothing works with the re-
gime, not politics or anything else . . . . There is hope only in God. There is no other
solution. What we are saying is that the entire world has left us. Publicly, not in secret.
The world is watching and knows what is happening in Syria. So in this case we have
no one else but God. (Interview in Irbid, Jordan September 17, 2012)

Transformation of Mobilization: The

Role of Religion in Consolidating
the Conflict as War

The protest movement remained overwhelmingly nonviolent through the summer of

2011. That July, military defectors announced the formation of the Free Syrian Army.
532   Wendy Pearlman

While it put forth the exclusively nationalist agenda of toppling the Assad regime, most
members were Sunni Muslims and numerous brigades took Sunni-​inspired names
(Baczko, Dorronsoro, and Quesnay 2018, 180).
Meanwhile, more radical understandings of the conflict gained influence. Adnan
al-​Aroor, a Saudi-​based Syrian Salafi cleric, broadcast “charismatic diatribes” calling
for the uprising to embrace arms as part of a broader struggle of Sunnis against Shiʿa
(International Crisis Group 2012, 29). Aroor’s most controversial statement came in a
June 2011 speech when, in apparent reference to the regimes’ weaponization of rape, he
directed these words to Syria’s Alawite community: “No harm will be done to those who
remained neutral [during the revolution]; as for those who supported the revolution,
they will be with us . . . however those who violated sanctities, . . . we will chop their
flesh and feed them to . . . dogs” (Pierret 2013, 237).
These developments culminated with the establishment of expressly Islamist and even
Salafi-​Jihadi armed groups. Here, the regime’s earlier encouragement and subsequent ar-
rest of participants in the Iraqi insurgency played an important role. In mid-​2011, Assad
pardoned hundreds of the fighters whom he had previously imprisoned (Abouzeid 2014;
Lister 2015). Lister (2015, 53) is one of many who views those amnesties as the regime’s
“devious attempt” to unleash extremists and terrorists into an opposition that it was al-
ready denouncing as extremist and terrorist. Then, in August 2011, eight al-​Qaʿida leaders
crossed from Iraq to Syria with the goal of replacing the Assad regime with an Islamic
state (Abouzeid 2014). They convened with the militants freed from prison and began
establishing cells throughout Syria. These al-​Qaʿida affiliates would officially establish
Jabhat al-​Nusra in January 2012. Other armed groups with different Islamist ideologies
or goals subsequently formed and used religious rhetoric to recruit legions of foreign
fighters. One of those groups, emerging from a rift between Jabhat al-​Nusra and its al-​
Qaʿida parent organization in April 2013, was the Islamic State in Iraq and Syria (ISIS).
It is beyond the scope of this chapter to assess whether members of these groups were
true believers, adopted identities strategically, or were driven by a complex mix of religious
and other impetuses. In mobilizing on Syrian soil, however, they both tapped into religious
ideas preexisting in society and, arguably more importantly, benefited from the context
of extreme violence. Given the latter, they employed religion in claiming to offer civilians
a sense of purity, purpose, and clearly defined enemies; in legitimizing violence as a vir-
tuous response to violence; and in extending some solace in casting death as martyrdom
(International Crisis Group 2012, ii; Saleh 2017, 126). Beyond this, the rise of Salafi-​Jihadism
was facilitated by funding, weapons, organization, discipline, and military experience
gained outside Syria. As Western countries offered scant military aid to the revolt, wealthy
Gulf states and individuals emerged as the best sources of support. Competing to win it,
rival armed groups often portrayed themselves as not only Islamist, but also anti-​Shiʿa
(Baczko, Dorronsoro, and Quesnay 2018, 181). As early as fall 2012, the International Crisis
Group (2012, 10) warned that this escalatory cycle would engulf the entire armed rebellion:

The money flow from conservative donors did more than strengthen Salafi factions
relative to their mainstream counterparts . . . . (I)t also pushed non-​Salafi combatants
 Religion and Mobilization in the Syrian Uprising and War    533

toward joining Salafi units capable of providing them with the requisite weapons and
ammunition. Groups with no ideological affiliation whatsoever began to adopt the
symbols, rhetoric and facial hair associated with Salafism for that purpose. While
such forms of behaviour typically might start as a largely opportunistic phenomenon
and thus lead to exaggerated assessments of a rising Islamist tide, over time they
could well turn into more genuine feelings, as the experience of a religiously-​inspired
struggle permeates a generation of fighters.

Once ascendant, Islamist groups attained and maintained power through approaches
familiar to those of nonreligious political forces, including institution-​building, ter-
ritorial control, external relations, social service provision, propaganda campaigns,
and sheer violence. Their relative weight in the Syrian opposition also profited from
constraints on expressly civic, nonsectarian organizations. Another of my interlocutors
explained:

Our problem is that we don’t have politicians or political parties. We have

personalities, individuals. So when citizens started raising religious slogans, like “We
have no one but God,” they were right. Because they had no political actors or groups.
So all they had was religion. (Interview in Antakya, Turkey, September 5, 2013)

During nearly a decade of conflict, hundreds of citizen initiatives inside and out-
side Syria have worked for a free and civil state in myriad ways (Citizens for Syria
2015). Civic groups have undertaken initiatives from collective gardens to educa-
tional activities designed to counter sectarianism in society (swisspeace, Conflict
Dynamics International, and FarikBeirut.net 2016, 21–​23). Nevertheless, civic groups
faced the enormous challenge of both overcoming the preexisting constraints on
Syrian civil society and building capable leadership and structures while under
military assault. As the conflict escalated and transformed, their ability to offer
nonreligious structures to channel social grievances and aspirations remained hand-
icapped by a lack of funding, as well as attacks by regime and extremist forces alike
(Becker and Stolleis 2016).
Against this backdrop, the Syrian conflict evolved into a multidimensional war with
religious and sectarian dimensions. The summer of 2012 registered ghastly killings
that many observers identified as sectarian massacres, including the slaughter of some
two hundred villagers in Hula and Qubayr at the hands of shabiha and regime troops
(Yassin-​Kassab and Al-​Shami 2015, 112). In the years that followed, armed rebels engaged
in other kinds of gruesome brutality with sectarian overtones (Droz-​Vincent 2014, 56).
Ideological fanaticism contributed to transforming the conflict from a peaceful popular
uprising to a war that has killed more than 500,000 people, disappeared some 100,000,
and forcibly displaced more than half the population of 22 million. “We tried our best to
build something,” a secular activist told me. “We faced a lot, and we faced it alone. And
we faced it with the minimum of hatred toward others. But we lost control” (Interviewed
in Gaziantep, Turkey, on January 11, 2016).
534   Wendy Pearlman

Conclusion

Given the sectarianism and religious extremism that would eventually consume Syria,
it can be tempting to conclude that religion was a primary factor in mobilization from
the start. This would be a mistake. Structures and strategies of rule before and under
Assad rule set the stage for complex connections between politics and religion. In
terms of both belief and collective identity, religion became entwined with issues of
power and privilege, as well as deep feelings of subjugation, resentment, and fear across
society.
Nevertheless, religion occupied only a marginal place in motivating the uprising.
Once protest was underway, the role of Islam in aiding collective action was utilitarian
far more than spiritual. Religion became most salient in mobilization as the conflict
militarized and escalated, consolidating its transformation from popular uprising to a
multidimensional war. Violence intensified the role of religion as both ideology and so-
cial cleavage, as well as the interconnections between the two.
The role of religion in accelerating violence was not inevitable, but in good part inten-
tional. As Droz-​Vincent (2014, 54) writes, the Assad regime sought to hold power at all
costs, and “bet on the fact that precipitating a sectarian breakdown would produce the
most favorable environment for the pursuit of its overwhelming military preponder-
ance.” In condemning peaceful grass-​roots protest as the work of sectarian fanatics, the
government’s rhetoric generated a self-​fulfilling prophecy (Pinto 2017, 136). Extremist
Islamist groups entered the fray and grew in both strength and brutality. Yet even then,
their preeminence was grounded in contentious dynamics not peculiar to Islam, in-
cluding state repression, resources, intra-​opposition competition, and the protraction
of both violence and unmet aspirations.
Analysis of mobilization in the Syrian uprising and war encourages us not to assume
the primacy of religion as a causal variable, even when protestors invoke the name of
God, convene in mosques, or articulate resentment toward people of other faiths.
A deeper examination can reveal that religion can function primarily as an indicator
or symptom of nonreligious problems and processes, or as a variable whose political
significance lays in interactions with variables unrelated to religion. This perspective
generates a host of new questions for further research. In those contexts in which reli-
gion does become an independent driver of mobilization or conflict, what makes it so?
When others perceive religion as being a more significant factor than it is, what drives
those perceptions? What does that teach us not only about politics in Muslim societies,
but also about non-​Muslims’ ways of viewing them?

References
Abboud, Samer N. 2013. “Economic Transformation and Diffusion of Authoritarian Power in
Syria.” In Democratic Transition in the Middle East: Unmaking Power, edited by Larbi Sadiki,
Heiko Wimmen, and Layla Al-​Zubaidi, 159–​177. London and New York: Routledge.
 Religion and Mobilization in the Syrian Uprising and War    535

Abboud, Samer N. 2018. Syria. 2nd ed. Cambridge, MA: Polity.

Abouzeid, Rania 2014. “The Jihad Next Door.” Politico Magazine, June 23, 2014. https://​www.
politico.com/​magazine/​story/​2014/​06/​al-​qaeda-​iraq-​syria-​108214?paginate=false.
Baczko, Adam, Gilles Dorronsoro, and Arthur Quesnay. 2018. Civil War in Syria: Mobilization
and Competing Social Orders. Cambridge: Cambridge University Press.
Batatu, Hanna. 1999. Syria’s Peasantry, the Descendants of Its Lesser Rural Notables, and Their
Politics. Princeton, NJ: Princeton University Press.
Becker, Petra, and Friederike Stolleis. 2016. “The Crushing of Syria’s Civil Actors.” SWP
Comment 2016/​C55. Berlin: Stiftung Wissenschaft und Politik. https://​www.swp-​berlin.org/​
en/​publication/​the-​crushing-​of-​syrias-​civil-​actors/​.
Chenoweth, Erica, and Maria J. Stephan. 2012. Why Civil Resistance Works: The Strategic Logic
of Nonviolent Conflict. New York: Columbia University Press.
Citizens for Syria. 2015. Mapping the Syrian Civil Society Actors—​Phase One. Berlin: Citizens
for Syria. https://​citizensforsyria.org/​OrgLiterature/​CfS-​mapping-​phase1-​EN.pdf.
Council on Foreign Relations. 2013. The Role of Religion in Postconflict Syria. New York: Council
on Foreign Relations, July 2013. Retrieved from https://​www.cfr.org/​interview/​role-​religion-​
postconflict-​syria.
Davenport, Christian, Hank Johnston, and Carol Mueller, eds. 2005. Repression and Mobilization.
Minneapolis: University of Minnesota Press.
Droz-​Vincent, Philippe. 2014. “State of Barbary (Take Two): From the Arab Spring to the
Return of Violence in Syria.” Middle East Journal 68, no. 1: 33–​58.
Farouk-​Alli, Aslam. 2014. “Sectarianism in Alawi Syria: Exploring the Paradoxes of Politics
and Religion.” Journal of Muslim Minority Affairs 34, no. 3: 207–​226. http://​dx.doi.org/​
10.1080/​13602004.2014.946761.
Friedman, Uri. 2011. “How Syrian Activists Name Their Friday Protests.” The Atlantic,
September 16, 2011. https://​www.theatlantic.com/​international/​archive/​2011/​09/​how-​syrian-
​activists-​name-​their-​friday-​protests/​337848/​.
Hinnebusch, Raymond A. 1990. Authoritarian Power and State Formation in Ba‘athist
Syria: Army, Party, and Peasant. Boulder, CO: Westview Press.
Human Rights Watch. 2011. “Syria: Rising Toll in Homs.” New York: Human Rights Watch, July
2. https://​www.hrw.org/​news/​2011/​07/​02/​syria-​rising-​toll-​homs.
International Crisis Group. 2012. Tentative Jihad: Syria’s Fundamentalist Opposition. Middle East
Report 131. Brussels: International Crisis Group. https://​www.crisisgroup.org/​middle-​east-​
north-​africa/​eastern-​mediterranean/​syria/​tentative-​jihad-​syria-​s-​fundamentalist-​opposition.
Ismail, Salwa. 2010. “Changing Social Structure, Shifting Alliances and Authoritarianism in
Syria.” In Demystifying Syria, edited by Fred Lawson, 13–​28. London: Saqi Books.
Khatib, Lina. 2018. “Tutelary Authoritarianism and the Shifts between Secularism and Islam
in Syria.” In The Syrian Uprising: Domestic Origins and Early Trajectory, edited by Raymond
Hinnesbusch and Omar Imady, 92–​105. London and New York: Routledge.
Lefèvre, Raphael. 2013. Ashes of Hama: The Muslim Brotherhood in Syria. Oxford: Oxford
University Press.
Lister, Charles R. 2015. The Syrian Jihad: The Evolution of an Insurgency. New York: Oxford
University Press.
Lund, Aaron. 2015. “Chasing Ghosts: The Shahiba Phenomenon.” In The Alawis of Syria: War,
Faith and Politics in the Levant, edited by Michael Kerr and Craig Larkin. New York: Oxford
University Press.
McAdam, Doug, John D. McCarthy, and Mayer N. Zald, eds. 1996. Comparative Perspectives
on Social Movements: Political Opportunities, Mobilizing Structures, and Cultural Framings.
New York: Cambridge University Press.
536   Wendy Pearlman

McAdam, Doug, Sidney Tarrow, and Charles Tilly. 2001. Dynamics of Contention.
Cambridge: Cambridge University Press.
Middle East Watch. 1991. Syria Unmasked: The Suppression of Human Rights by the Asad
Regime. New Haven, CT: Yale University Press.
Nakkash, Aziz. 2013. The Alawite Dilemma in Homs: Survival, Solidarity and the Making of a
Community. Bonn, Germany: Friedrich Ebert Stiftung. https://​library.fes.de/​pdf-​files/​iez/​
09825.pdf.
New York Times. 2011. “Damascus on the Edge.” New York Times, December 12, 2011. https://​
kristof.blogs.nytimes.com/​2011/​12/​12/​damascus-​on-​the-​edge/​.
Pearlman, Wendy. 2018. “Contingency and Agency in a Turning Point Protest: March 18,
2011 in Daraa, Syria.” In Microfoundations of the Arab Uprisings: Mapping Interactions
between Regimes and Protesters, edited by James Jasper and Frédéric Volpi, 111–​ 134.
Amsterdam: Amsterdam University Press.
Pearlman, Wendy. 2020. “Mobilization from Scratch: Large-​Scale Collective Action Without
Preexisting Organization in the Syrian Uprising.” Journal of Comparative Political Studies
first published online, March 23, 2020.
Perthes, Volker. 1997. The Political Economy of Syria under Asad. London: I.B. Tauris.
Phillips, Christopher. 2015. “Sectarianism and Conflict in Syria.” Third World Quarterly 36, no.
2: 357–​376.
Pierret, Thomas. 2013. Religion and State in Syria: The Sunni Ulama from Coup to Revolution.
New York: Cambridge University Press.
Pinto, Paulo G. H. 2011. “‘Oh Syria, God Protects You’: Islam as Cultural Idiom under Basher
al-​Asad.” Middle East Critique 20, no. 2: 189–​205.
Pinto, Paulo G. H. 2017. “The Shattered Nation: The Sectarianization of the Syrian Conflict.” In
Sectarianization: Mapping the New Politics of the Middle East, edited by Nader Hashemi and
Danny Postel, 123–​142. New York: Oxford University Press.
Quinlivan, James T. 1999. “Coup-​Proofing.” International Security 24, no. 2: 131–​165.
Reuters. 2012. “Syria’s Alawites, a Secretive and Persecuted Sect.” Reuters, January 31, 2012.
https://​www.reuters.com/​article/​us-​syria-​alawites-​sect-​idUSTRE80U1HK20120131.
Rifai, Ola. 2018. “Sunni/​Alawi Identity Clashes during the Syrian Uprising: A Continuous
Reproduction?” In The Syrian Uprising: Domestic Origins and Early Trajectory, edited by
Raymond Hinnesbusch and Omar Imady, 242–​259. London and New York: Routledge.
Saleh, Yassin Al-​Haj. 2017. The Impossible Revolution: Making Sense of the Syrian Tragedy.
Chicago: Haymarket Books.
Schock, Kurt. 2005. Unarmed Insurrections: People Power Movements In Nondemocracies: Social
Movements, Protest and Contention. Minneapolis: University of Minnesota Press.
Sindawi, Khalid. 2009. The Shiite Turn in Syria. Current Trends in Islamist Ideology. Washington,
DC: Hudson Institute. https://​www.hudson.org/​research/​9894-​the-​shiite-​turn-​in-​syria-​.
Slackman, Michael. 2011. “Syrian Troops Open Fire on Protesters in Several Cities.” New York
Times, March 25, 2011. https://​www.nytimes.com/​2011/​03/​26/​world/​middleeast/​26syria.html.
swisspeace, Conflict Dynamics International, and FarikBeirut.net. 2016. Inside Syria: What
Local Actors Are Doing for Peace. Bern, Switzerland: swisspeace, March 2016. http://​www.
swisspeace.ch/​fileadmin/​user_​upload/​pdf/​Mediation/​Inside_​Syria_​en.pdf.
Tarrow, Sidney G. 1994. Power in Movement: Social Movements and Contentious Politics.
New York: Cambridge University Press.
Tilly, Charles. 2004. Trust and Rule. New York: Cambridge University Press.
 Religion and Mobilization in the Syrian Uprising and War    537

Vallely, Paul. 2014. “The Vicious Schism between Sunni and Shia Has Been Poisoning Islam for
1,400 Years—​and It’s Getting Worse.” The Independent, February 19, 2014. https://​www.inde-
pendent.co.uk/​news/​world/​middle-​east/​the-​vicious-​schism-​between-​sunni-​and-​shia-​has-​
been-​poisoning-​islam-​for-​1400-​years-​and-​its-​getting-​9139525.html.
van Dam, Nicholas. 2011. The Struggle for Power in Syria: Politics and Society under Asad and
the Ba’th Party. New York: Palgrave Macmillan.
Wimmen, Heiko. 2014. “Divisive Rule: Sectarianism and Power Maintenance in the Arab
Spring: Bahrain, Iraq, Lebanon and Syria.” SWP Research Paper 4. Berlin: German Institute
for International and Security Affairs. http://​www.swp-​berlin.org/​fileadmin/​contents/​
products/​research_​papers/​2014_​RP04_​wmm.pdf.
Yassin-​Kassab, Robin, and Leila Al-​Shami. 2015. Burning Country: Syrians in Revolution and
War. London: Pluto.
 Chapter 26

C hristian-​Mu sl i m
Rel ations in th e Sha d ow
of C onf l i c t
Insights from Kaduna, Nigeria

Alexandra Scacco and Shana S. Warren

In February 2000, large-​scale Christian-​Muslim riots shook the Nigerian city of

Kaduna, killing at least three thousand people and displacing tens of thousands more.
In the course of four days of the most serious interreligious violence Nigeria had ever
witnessed, dozens of churches, mosques, and entire city blocks were burned to the
ground. Drawing on original survey data from a random sample of three hundred
young men living in particularly conflict-​prone neighborhoods in Kaduna, we analyze
patterns of Christian-​Muslim relations fifteen years after the February 2000 crisis. In
this chapter, we build a case that, contrary to recent literature that points to relatively
quick post-​conflict recovery and documents a positive relationship between violence
exposure and within-​group prosocial behavior (e.g., Bauer et al. 2016; Blattman 2009;
Gilligan et al. 2014; Voors et al. 2012), local exposure to deadly intergroup violence
continues to have profoundly negative social and political effects on intergroup rela-
tions nearly two decades later.
We show that Kaduna residents on either side of the religious divide continue to ex-
hibit high levels of mistrust and prejudice against members of the religious out-​group,
and demonstrate substantial discrimination against members of the religious out-​group
in behavioral games with real material stakes. We highlight three lasting consequences
of exposure to intercommunal violence, each of which complicates post-​ conflict
reconciliation.
First, deadly intercommunal violence erodes intergroup trust and heightens
perceptions of the out-​group as threatening and dangerous. This increases the likeli-
hood of further episodes of violence. Second, intercommunal violence on a large scale
tends to increase residential segregation at the local level, as people become afraid to
540    Alexandra Scacco and Shana S. Warren

live near out-​group communities. At the same time, out-​group communities do not dis-
appear entirely—​they often continue to coexist but occupy separate spaces within the
same locality. We follow others (e.g., Enos 2017; Enos and Celaya 2018; Weiss 2021) in
suggesting that mixed but segregated cities are particularly likely to suffer from high
levels of intergroup prejudice and discrimination, making post-​conflict reconciliation
and recovery more difficult.
Third, exposure to violence can have long-​term psychological consequences. Simply
put, exposure to intercommunal violence traumatizes populations. We draw on sec-
ondary literature documenting high levels of post-​traumatic stress disorder (PTSD) in
Kaduna and Jos—​the cities in Nigeria’s Middle Belt that have experienced the most fre-
quent and deadly episodes of communal violence. Taken together, these consequences of
violence exposure should lead us, on balance, to expect relatively slow, halting progress
toward post-​conflict recovery. We offer evidence on each of these consequences from
the Nigerian context, and caution against expectations that post-​conflict communities
should quickly bounce back from episodes of interreligious violence.
Our data affords us a detailed view of the nature of Christian-​Muslim relations in
Kaduna, fifteen years after deadly intergroup conflict. Although intergroup attitudes
are not uniformly negative, a picture clearly emerges of lingering mistrust, antagonism,
and anxiety in intergroup relations. We are not arguing that Christian-​Muslim violence
is inevitable—​in fact, it remains relatively rare, even in the religiously mixed cities of
Nigeria’s Middle Belt region. Nor are we arguing that post-​conflict reconciliation and
recovery are impossible. We simply urge caution against expectations that communities
can easily and quickly bounce back from violent conflict.

Long-​Term Impacts
of Intercommunal Violence

Reviewing recent empirical work on the consequences of exposure to violent con-

flict, the casual reader might be tempted to conclude that, while initially devastating
in terms of lost lives and property destroyed, the dark side of violence exposure casts a
short shadow. A series of high-​quality (and high-​profile) studies documents a positive
relationship between exposure to wartime violence and within-​community prosocial
behaviors in settings ranging from Uganda (Blattman 2009) and Burundi (Voors et al.
2012) to Nepal (Gilligan et al. 2014). Building on this work, a recent meta-​analysis of
sixteen micro-​level studies of violence exposure during civil and interstate wars reports
a set of surprisingly positive effects of exposure on political and social outcomes, in-
cluding participation in community organizations, participation in local politics, and
prosocial behavior more broadly (Bauer et al. 2016). The study concludes that violence
exposure not only increases individual-​level participation in local civic and political life,
but also leads those who have directly suffered violence to be more trusting of others
 Christian-Muslim Relations in the Shadow of Conflict    541

in their local community, and more altruistic in experimental laboratory games. It fur-
ther suggests that “the increased local cooperation we document might help to explain
why some post-​conflict countries experience almost miraculous economic and social
recoveries” (Bauer et al. 2016, 4–​5).
It would be easy to conclude from this body of research that, if one were to check back
in on a post-​conflict setting like Kaduna, Nigeria, more than a decade later, things should
not look too bad. Unfortunately, as we describe in detail below, that is not borne out in
our survey data. There is little to suggest conditions of reconciliation or rapid recovery.
Our observations from Kaduna are more consistent with a set of studies published in
the past few years highlighting the negative medium-​and long-​term consequences
of violence exposure. These studies suggest that violence exposure deepens negative
intergroup attitudes and may in fact decrease prosocial behavior across social groups as
well as within them (Hager et al. 2019).
In this vein, Beber et al. (2014) find that exposure to communal violence in Khartoum
exacerbated already negative attitudes held by Sudanese Arabs toward South Sudanese
residents of the city. Similarly, Grossman et al. (2015) find that exposure to combat
hardens Israeli attitudes toward Palestinians, reduces support for political compromise,
and increases voting for hawkish parties. Across a range of contexts, including Bosnia,
Kosovo, and Kyrgyzstan, recent studies have found that individual-​and community-​
level experiences with violent conflict solidify ethnic identities, increase distrust to-
ward non-​co-​ethnics, and make ethnic parties more appealing to voters (Hadzic et al.
2020; Hager et al. 2019; Mironova and Whitt 2018). Further, this hardening of negative
attitudes may also be passed down across generations (Lupu and Peisakhin 2017; Nunn
and Wantchekon 2011).
What might explain this disconnect in the literature? One possibility is that the set
of violence exposure studies documenting negative consequences, as well as our own
observations in Kaduna, focus on intercommunal violence in settings where members
of the groups in conflict live in relatively close proximity to one another. This feature
applies to religiously mixed cities across Nigeria’s Middle Belt region. It also describes
present and past settings of Hindu-​Muslim violence in urban India, as well as com-
munal violence in cities in Iraq, Lebanon, Northern Ireland, and the United States,
among others. In contrast, the civil and interstate wars examined by Bauer et al. (2016) in
their meta-​analysis are conflicts in which parties to the violence are less likely to live in
close proximity before, during, or after the fighting is over. Behavioral games designed
to measure prosociality implemented in such settings therefore reflect within-​group
improvements and cannot inform us about intergroup attitudes or behavior in the after-
math of conflict.1 As such, it is reasonable to expect the impacts of violence exposure to
be very different, a possibility that Bauer et al. (2016, 251) themselves acknowledge.
The immediate effects of large-​scale communal violence are easily observed. The case
we explore here illustrates how destructive intercommunal violence can be. Casualty
and displacement figures from the Christian-​Muslim riots that took place in February
2000 in Kaduna are staggering, with conservative estimates of fatalities ranging from
around 2,000 to 5,000 deaths, an estimated 125,000 people temporarily displaced
542    Alexandra Scacco and Shana S. Warren

in the aftermath of the violence, and billions of naira lost in private property damage
(Angerbrandt 2011; Scacco and Warren 2018; Wapwera and Gajere 2017).
Beyond immediate effects, we should expect intercommunal violence on this
scale to have large negative medium-​to long-​term social, political, and psycholog-
ical consequences. We highlight three types of such consequences below. Exposure
to intergroup violence may (1) deepen intergroup mistrust, complicating reconcil-
iation processes, and increasing the likelihood of future violence; (2) result in rad-
ical residential segregation, complicating processes of intergroup reconciliation; and
(3) lead to lasting psychological trauma within affected communities. We discuss each
in detail below.

Intercommunal Violence Erodes Trust

The idea that violent conflict between groups living in close proximity worsens already
fraught intergroup attitudes in divided societies is, of course, not new. Prior episodes
of intercommunal violence make communal identities and social divisions more sa-
lient, hardening boundaries between groups in conflict (Beber et al. 2014; De Waal 2005;
Fearon and Laitin 2000; Hadzic et al. 2020). Histories of conflict can make it more dif-
ficult for people living in divided societies to update negative beliefs when faced with
new, positive information about the out-​group (e.g., Devine et al. 2002; Fiske 1989).
The mistrust generated by communal violence may also decrease prosocial behavior,
both across (Mironova and Whitt 2018) and within (Hager et al. 2019) relevant identity
groups. The impacts of communal violence on intergroup attitudes can be observed far
beyond the months and years following violent events; evidence suggests that negative
out-​group attitudes and high levels of out-​group mistrust can be transmitted across gen-
erations (Lupu and Peisakhin 2017; Nunn and Wantchekon 2011).
Literature in social psychology suggests that deadly communal conflicts can pro-
duce psychological repertoires of fear and grievance toward the out-​group, making it
more difficult to break out of cycles of repeated conflict. Research on such repertoires,
discussed in Bar-​Tal and Avrahamzon (2016) and Sharvit (2014), indicates that settings
of repeated conflict pose what they term a “special challenge” for prejudice reduction,
since even when individuals are able to develop more favorable beliefs toward out-​group
members (in lab experimental settings, for example), such beliefs are difficult to main-
tain in communities that perpetuate a culture of conflict. Further, even when explicit
attitudes or beliefs shift in a more positive direction, recent lab experimental work in the
Israeli context suggests that out-​group animosity may continue to be stored in implicit
beliefs and attitudes (Sharvit 2014).
The notion of a conflict repertoire resonates with field-​based experiences from
Nigeria’s Middle Belt region discussed by Scacco (2020). During fieldwork conducted
in the riot-​prone city of Jos (a few hours’ drive from Kaduna) in 2007, one of the
authors (Alexandra Scacco) experienced what might be best termed a “near-​riot” due
 Christian-Muslim Relations in the Shadow of Conflict    543

to a misunderstanding conceivably amplified by the type of conflict repertoire the psy-

chology literature identifies.
One summer morning, the author woke to cries from neighbors that there was
“trouble in the town” because “the Muslims were mobilizing” for a riot. In Jos, com-
mercial motorcycle drivers (who offer an inexpensive, efficient form of transport that
many local commuters rely on) have historically tended to be Muslim men. In the hours
that followed, rumors spread quickly through Christian neighborhoods that “religious
trouble” was brewing. In reality, the motorcycle drivers had begun marching through
the town and setting up roadblocks as part of a labor strike in protest against a recent
order by the city government that they were no longer allowed to transport passengers
after dark.
A religious riot was prevented through swift military intervention (by chance, due
to the nearby location of a federal military barracks), but the panic and fear generated
by the sight of out-​group members marching together was unmistakable. The town
of Jos had experienced deadly Christian-​Muslim riots in 2001 and 2004, resulting in
thousands of deaths. This history of violence led ordinary people to interpret an innoc-
uous event through the lens of intergroup threat, shedding light on the process by which
violence can beget violence in conflict settings and stubborn prejudices may be diffi-
cult to dislodge. Indeed, the very fact that episodes of communal violence tend to occur
again and again in the same localities (see, for example, Varshney 2003 and Wilkinson
2006 for detailed evidence on the spatial concentration of Hindu-​Muslim riots in post–​
World War II India) suggests how difficult it is for groups to reconcile after violence.

Intercommunal Violence and

Local Segregation

Another feature of life in the wake of violent conflict can further amplify the problem of
intergroup mistrust. Deadly communal conflicts often increase local residential segre-
gation, as people who no longer feel safe living in close proximity to out-​group members
move into in-​group residential enclaves. Intercommunal violence of the form we con-
sider may be particularly likely to produce diverse yet locally segregated environments,
as households relocate within the same locality.2
Although motivated by a desire for safety, local segregation along communal lines
may provide especially fertile ground for intergroup prejudice and animosity. Recent
studies have identified pernicious effects of segregation on racial prejudice and discrim-
ination in American cities (Enos 2017), interreligious mistrust and violence in Jerusalem
(Rokem et al. 2018), cycles of ethnic violence in Iraq (Weidmann and Salehyan 2013) and
Kenya (Kasara 2017), and studies in laboratory settings (Enos and Celaya 2018).
One reason for this is that residential segregation in particular limits opportunities for
meaningful cross-​group social interactions among ordinary people—​and communal
544    Alexandra Scacco and Shana S. Warren

leaders—​that can provide information to quell incendiary rumors and resolve small-​
scale conflicts (Kasara 2013). Although residential segregation may make people feel
more safe where they live, if things do go wrong, it is more difficult to contain communal
violence in its early stages.
Other research links segregation with high levels of prejudice and mistrust more di-
rectly, independent from segregation’s impact on levels of intergroup contact in daily
life. Enos (2017) argues that diverse but segregated local geographies make identities
of groups in tension more salient. Increased salience leads to more biased attitudes and
behavior across the identity cleavage. Enos argues: “Social geography creates space be-
tween us in our heads before it creates a space between us in our relationships” (Enos
2017, 15). If segregation has both direct and indirect (through the standard theorized
path of reduced intergroup contact) negative effects on attitudes toward and perceptions
about the out-​group, we should expect a greater risk of conflict recurrence in diverse but
segregated local settings.

Psychological Consequences
of Violence Exposure

On the surface, it might seem that the populations of cities “bounce back” quickly after
even large-​scale episodes of communal violence. People might resume trading across
lines of cleavage in local markets weeks, or even days, after the riots. However, an
emerging literature on the psychological impact of intercommunal violence suggests
that deeper scars may linger for years in post-​conflict settings.
The effects of violent conflict on mental health may be direct, if they stem from di-
rect exposure to traumatic events, or indirect, if they arise from displacement or any
other consequences of violent conflict, such as disruption of livelihoods. Direct effects
of violence exposure on mental health include post-​traumatic stress disorder (PTSD),
insomnia, anxiety, and depression (Bogic et al. 2015; Gordon et al. 2010; Mpembi et al.
2018). Conflict exposure may affect mental health indirectly, as a result of displacement,
scarcity of basic necessities, such as food and water, and the loss of family and finan-
cial stability (Garry and Checchi 2020). To the extent that violent conflict increases the
number of people living in extreme poverty where it occurs, we can draw on a large body
of literature documenting the negative impacts of poverty on mental health (e.g., Lund
et al. 2010; Patel and Kleinman 2003). Given the scale and intensity of the interreligious
communal violence in religiously heterogeneous contexts, it is reasonable to expect
to observe lingering negative mental health consequences among its survivors many
years later.
In sum, the literature leads us to expect at least three sets of negative consequences of
exposure to intercommunal violence in the medium to long term. That said, we do not
intend to suggest that there is no hope at all for post-​conflict reconciliation. Our goal
 Christian-Muslim Relations in the Shadow of Conflict    545

here is simply to caution against applying optimistic expectations from the empirical
literature on the effects of violence exposure on within-​group prosociality to intergroup
relations. Various dimensions of post-​conflict reconciliation can be difficult and painful.
Rebuilding intergroup trust can take time and can easily experience setbacks. We should
not be surprised if recovery is slow and if gains are reversed.

Christian-​Muslim Relations in Nigeria

We now turn to a discussion of Christian-​Muslim intergroup relations in the Nigerian
context. Decades of conflict across often overlapping ethnic and religious lines within
the Nigerian state shaped the lens through which Kaduna residents could interpret and
respond to the violent riot that shook the city in 2000.
Although social demography has long been, and continues to be, an issue of con-
troversy and contestation, we can accurately claim that Nigeria is a multiethnic state
of approximately 200 million people, with relatively equal proportions of Muslims and
Christians and hundreds of ethnic groups. The Christian-​Muslim political divide dates
to British colonial policy, under which the predominantly Muslim North and largely
Christian South were separately governed as the Northern and Southern Nigerian
Protectorates (until 1914), then unified as the colony of Nigeria (1914–​1954), and finally
re-​divided into three regions by splitting the Southern region into a Christian-​dominant
Eastern region and a religiously mixed Western region. The dominant ethnic group in
each of these regions is one of Nigeria’s three largest ethnic groups: Northern (Hausa,
Muslim), Eastern (Igbo, Christian), and Western (Yoruba, both Christian and Muslim).
Civil and communal violence have occurred frequently across various social cleavage
lines in Nigeria since independence in 1960. Just seven years after Nigeria obtained in-
dependence, a civil war broke out along a reinforcing ethnic and religious divide. The
Eastern region, predominantly Igbo and Christian, attempted to secede as the inde-
pendent nation of Biafra. The Nigerian government and military, under primarily
Northern (Muslim) leadership, was victorious in the three-​year civil war (1967–​1970).
In the process, over two million civilians within Biafra died. Interreligious tensions
in Nigeria continued after the war, and still underpin many contemporary political
grievances within Nigeria (Kendhammer 2014). Throughout the 1980s, many Nigerian
Christians believed that the predominantly Muslim North received a disproportionate
share of both political power and economic resources, fanning resentment across reli-
gious lines (Ibrahim 1991, 135).
During the late 1980s and early 1990s, as Nigeria haltingly inched toward a return to
multiparty politics, violence across reinforcing religious and ethnic cleavages began to
increase (Falola 1998; Ibrahim 1991). Some of this violence occurred within Nigeria’s cul-
tural Middle Belt, the region in which the dominantly Christian South and dominantly
Muslim North meet, including a religious riot in Kafanchan, in Kaduna state, in 1987,
which resulted in twenty-​five deaths, serious property destruction, and spread to other
cities and states in the region. Events such as the Kafanchan riot produced a reference
 546    Alexandra Scacco and Shana S. Warren

point of religious violence through which subsequent communal conflicts in the Middle
Belt and Kaduna could be interpreted. Yet the scale of violence remained small.
With Nigeria’s return to multiparty elections in 1999, communal violence across re-
ligious lines increased more dramatically (Ukiwo 2003). Deadly interreligious violence
recurred regularly in this period. There were only three years between 1990 and 2008 in
which there were no recorded incidents, and at least 4,703 fatalities occurred in the same
period (Basedau et al. 2013). Large-​scale communal violence particularly began to con-
centrate across Nigeria’s cultural Middle Belt, as shown in Figure 26.1, which compares
riot events in two key states in the Middle Belt region—​Kaduna and Plateau—​with riot
events in the rest of Nigeria. The height of each bar represents the number of riot events
that occurred, while the numbers atop each bar represent total casualties from these
events. We observe that violence in Kaduna state in 2000 was an outlier relative to the
rest of the period, with nearly four thousand people killed in three incidents that year.
We will return to the events of 2000 in the next section.
Interreligious tensions in contemporary Nigeria have become self-​reinforcing at the
national level alongside repeated violent incidents. A 2009 Pew Forum on Religion and
Public Life survey of 1,516 Nigerians examined the attitudes of Christians and Muslims
toward religion and each other, situating Nigerian public opinion within a sample of
nineteen sub-​Saharan African countries (Pew Forum 2010). Over half (58%) of Nigerians
said they believe that religious conflict is a very big problem in their country—​the largest
proportion of any of the nineteen sampled countries (tied with Rwanda). Both Christian

4000

3500

3000
Total Deaths per Year

2500

2000

1500

1000

500

0
1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

Year
State
Plateau Kaduna All Other Nigerian States

Figure 26.1: Riot events and fatalities per year in Kaduna and Plateau states.
 Christian-Muslim Relations in the Shadow of Conflict    547

Table 26.1: Religious Beliefs, Religious Practice, and Public Life in Nigeria

Muslim Christian

Conflict between religious groups a very big problem 60% 56%

Conflict between ethnic groups a very big problem 48% 48%
Shariʿa and/​or Bible is literal word of God 90% 87%
Shariʿa and/​or Bible should be official law 70% 71%
Know not very much or nothing at all about other faith 63% 55%
Pray every day 93% 92%

Data from Pew Forum (2010).

and Muslim Nigerians are deeply religious, and large majorities of both Christians and
Muslims pray every day, believe that their holy books are the literal word of God, and
would like the Bible and/​or shariʿa to become the official law of the land, as shown in
Table 26.1. Indeed, Nigerians were more likely than respondents from any other country
in the Pew sample to believe that the Bible and/​or Qurʾan are the literal word of God.
This centrality of religious practice in daily life, combined with an interreligious con-
flict repertoire drawn from recent past experience with violence, reinforces the sali-
ence of religious identity. In short, in Nigerian society, religion is highly politicized, the
Muslim-​Christian divide is highly salient, and Nigerians are deeply concerned about
interreligious conflict.

Christian-​Muslim Relations in the Shadow

of Violence: Kaduna
The city of Kaduna erupted in the worst interreligious violence ever witnessed in
Nigeria’s history over four days in February 2000. The Kaduna state government is-
sued a report documenting 1,295 deaths during the riots, although other sources have
suggested the true number may be far higher (Scacco 2020; Tertsakian 2003). The
fighting was sparked by a public debate over whether to introduce shariʿa law into the
criminal code in Kaduna state. Although shariʿa provisions had long been incorporated
into “personal” or domestic law for Muslims throughout northern Nigeria, the debate
raised concerns in religiously heterogeneous states that shariʿa would be imposed on
Christian communities (Abdu and Umar 2002).
Several months before the violence in Kaduna, the government of Zamfara, an over-
whelmingly Muslim state bordering Niger, decided to incorporate shariʿa laws into its
state criminal code. State legislatures across the northern region soon followed suit,
with the exception of states like Kaduna and Plateau, which contain large Christian
populations. The issue was particularly contentious in Kaduna, and its governor at the
548    Alexandra Scacco and Shana S. Warren

time, Ahmed Makarfi, strongly opposed introducing shariʿa. In January 2000, however,
under pressure from Muslim civil society organizations, the Kaduna State Assembly
opened debate on the suitability of shariʿa criminal law for Muslim residents of the
state. The action divided the assembly along religious lines, and sparked vocal protests
from Christian civil society organizations (Kaduna State Government 2001; Sani 2007;
Tertsakian 2003).
On February 21, the Kaduna branch of the Christian Association of Nigeria organ-
ized an anti-​shariʿa demonstration. Announcements were made in churches across
Kaduna city, urging parishioners to attend the demonstration, and the event attracted
hundreds of demonstrators (Ullah 2002). Accounts of how exactly the fighting began
differ, but small numbers of Christians and Muslims began throwing stones at each
other as demonstrators made their way past Kaduna’s crowded central market (Maier
2002). At this point, the fighting spread quickly, engulfing the town center in violence,
as Christian protesters fought with Muslim traders in the market, using simple weapons
such as stones and wooden planks. The fighting in the largely commercial center of town
spread outward to residential areas across the Kaduna metropolitan area, as rumors
circulated about atrocities committed on either side. Interview respondents from
neighborhoods across Kaduna recalled seeing smoke rising from the center of town,
and hearing stories about churches and mosques being set on fire (Scacco 2020). The
riots lasted for four days, and were only put to rest when the central government ordered
the military to intervene.
Violence in Kaduna in 2000 changed life in the city dramatically, as residents viewed
the riots as evidence that it was no longer safe to live near out-​group members, and
abandoned more integrated neighborhoods and self-​segregated across religious lines
(Angerbrandt 2011). Figure 26.2 presents quantitative evidence of the increased segre-
gation of ten Kaduna neighborhoods, using data from Scacco (2020). The x-​axis meas-
ures the difference in proportions of each religious group living in the neighborhoods
in 2000, and the y-​axis the same data from 2008. If neighborhood composition had
remained unchanged, all points would fall along the dotted line. Points above the line in-
dicate increased segregation, which all ten neighborhoods experienced. While Kaduna
was not a religiously integrated city before the 2000 riot by any means, it had already be-
come much more segregated eight years later.
In the remainder of this chapter, we present a micro-​level analysis of Christian-​
Muslim relations within the city of Kaduna to examine the long shadow cast by se-
vere communal violence fifteen years after a major violent event. We use data from
the control group for an experimental study of intergroup contact conducted in
Kaduna in 2014–​2015 (as outlined in Scacco and Warren 2018). The sampling frame
includes all poor, conflict-​prone neighborhoods within one hour’s travel to central
Kaduna, a city of approximately one million residents in north-​central Nigeria, a few
hours’ drive from the national capital. Sixteen neighborhoods, eight in Kaduna’s ma-
jority Muslim north and eight in Kaduna’s majority Christian south, were selected
from this frame.
 Christian-Muslim Relations in the Shadow of Conflict    549

100 Ungwan Kanawa

Ungwan Television Badarawa
Sabon Tasha Tudun Wada Kaduna

75
I% Christians - % MuslimsI, 2008

Barnawa

Barnawa Gra
50 Hayin Banki
Kawo

25

Ungwan Rimi

0 25 50 75 100
I% Christians - % Muslimsl, 2000
No conflict Conflict

Figure 26.2: Pre-​and post-​riot neighborhood segregation in Kaduna.

The final sample consists of 270 young men, aged 18–​25, who were recruited for a base-
line survey in August 2014, and interviewed about intergroup relations in January and
December 2015. Table 26.2 demonstrates that our sample reflects the segregated con-
text that defines contemporary Kaduna. Two neighborhoods include minority religious
group enclaves: Barnawa and Kakuri. Barnawa is a predominantly Christian neighbor-
hood, but within the market area of the neighborhood there is a small and concentrated
Muslim minority. In both cases, residential segregation at street level is the norm, and
neither are cases of well-​integrated neighborhoods. The residential segregation already
evident in 2008 has continued or deepened, and has led to religious segregation in edu-
cation and other aspects of daily living.
Within the city of Kaduna, as in most of Nigeria, there is also broad overlap be-
tween religious and ethnic identities. Table 26.3 demonstrates the extent of this overlap
within our sample. Our survey asked respondents to name their “home language.”3 The
left-​hand column lists all home languages cited by at least 2 percent of all respondents,
ordered from most to least frequently cited. The remaining two columns present the
share of Muslim or Christian respondents who cited those home languages. All lan-
guages cited by less than 2 percent of either sample are grouped together within the
“Other” category. Significantly, we observe that 84 percent of all Muslims, but only 2 per-
cent of Christians, in the sample are Hausa. Furthermore, among sampled Christians
there is no home language cited by more than 12 percent of all respondents. Similarly,
550    Alexandra Scacco and Shana S. Warren

Table 26.2: Survey Sample: Residential Segregation in Kaduna, January 2015

Muslims Christians Segregation

Neighborhood (N) (N) (% Dominant Group)

Badarawa 21 0 100%
Badiko 13 0 100%
Barnawa 11 4 73%
Hayin Banki 16 1 94%
Kakuri 10 18 64%
Kawo 11 0 100%
Kurmin Mashi 18 3 86%
Malali 8 0 100%
Narayi 0 46 100%
Nassarawa 0 8 100%
Sabon Tasha 0 20 100%
Tudun Nupawa 14 0 100%
Kanawa 8 0 100%
Shanu 15 0 100%
Sunday 0 10 100%
Television 0 15 100%
Total 145 125 -​

Table 26.3: Survey Sample: Religion and

Language (Ethnicity), January 2015
Muslim Christian

Hausa 84% 2%
Kataf -​ 12%
Baju -​ 12%
Igala 2% 10%
Yoruba 3% 6%
Igbo -​ 9%
Idoma -​ 7%
Gabagyi 1% 5%
Jaba -​ 6%
Adara -​ 2%
Ikulu -​ 3%
Ibira 2% -​
Other 8% 26%
100% 100%
 Christian-Muslim Relations in the Shadow of Conflict    551

Table 26.4: Survey Sample: Religious Practice, January 2015

Muslim Christian

Brotherhood or Izala (75%) Catholic (41%)

Denomination Tijaniya (11%) Pentecostal (28%)
Unaffiliated (11%) Evangelical (25%)
Quadriya (3%) Anglican (6%)
House of Worship Multiple Times per Day (99%) Every Day (6%)
Attendance Less than Monthly (1%) Several Times per Week (61%)
Once per Week (30%)
Monthly or Less Often (3%)

there is greater diversity of religious practice among Christian young men than among
Muslim young men in our sample. As shown in Table 26.4, 75 percent of Muslims are
affiliated with the Izala Brotherhood. Catholics, the largest Christian denomination
within our sample, comprise only 41 percent of sampled Christians, and the Pentecostal
and Evangelical categories include many different denominations.
As one would expect, since there is one predominant ethnic group among Kaduna’s
Muslims, and highly fragmented ethnicity among Kaduna’s Christians, intrareligious
ethnic conflict within the city of Kaduna has not emerged as a major fault line. Instead,
the religious divide forms the main social cleavage.
Adherents of both religious faiths also report high levels of religious participation and
engagement, as shown in Table 26.4. Nearly all (99 percent) of Muslim young men indi-
cated that they go to a mosque to pray multiple times a day, and 97 percent of Christians
reported attendance at religious services at least once a week. Despite differences in re-
ligious obligation, both Christian and Muslim young men in our sample reported high
levels of engagement with the observance of rituals of their respective faiths.
In sum, the poor urban neighborhoods we studied in Kaduna replicate Nigeria’s
broad North-​South and Muslim-​Christian divide, along with the religiosity of Nigerians
across the country. This social cleavage within the city of Kaduna is reinforced by a com-
bination of an overlapping ethnic (Hausa–​non-​Hausa) divide and everyday religious
practice. While the vast majority of Muslims are affiliated with the Izala Brotherhood
and go to mosques multiple times per day, among Christians there is extensive variation
in denomination and frequency of church attendance.
In spite of extreme residential segregation, Christian and Muslim residents of
Kaduna do interact in daily life, but in generally superficial ways. Despite some regular
interactions in employment, transportation, and shopping, Christians and Muslims in
Kaduna exhibit high levels of anxiety about interactions and low levels of trust across
group lines, even fifteen years post-​conflict. Within our Kaduna micro sample, 30 per-
cent of Christians and 40 percent of Muslims reported feeling anxious around members
of the religious out-​group. Table 26.5 presents the results of a series of questions about
social trust, as well as estimated “trust deficits,” comparing trust in members of an
552    Alexandra Scacco and Shana S. Warren

Table 26.5: Intergroup Trust, January 2015

All Muslim Christian

Religion and Trust

Own religion 2.08 2.58 1.50
(0.05) (0.05) (0.07)
Other religion 1.75 2.13 1.30
(0.05) (0.06) (0.06)
Religious trust deficit -​0.33 -​0.44 -​0.19
(0.05) (0.06) (0.08)
Ethnicity and Trust
Own tribe 1.90 2.36 1.36
(0.05) (0.06) (0.06)
Other tribes 1.68 2.03 1.26
(0.05) (0.06) (0.06)
Tribal trust deficit -​0.23 -​0.33 -​0.10
(0.08) (0.05) (0.06)
Familial Trust
Relatives 2.35 2.87 1.74
(0.06) (0.04) (0.08)
Observations 270 145 125

Question: “How much do you trust each of the following types of people?”
Responses scaled 0 (Not at all) –​3 (Very much).
Standard errors in parentheses.

individual’s in-​group faith and tribe to trust in members of the out-​group faith and other
tribes. We again show results for the full sample as well as the Muslim and Christian
subsamples. As we would anticipate, respondents trust members of their own families
more than any of the other prompted types. Beyond their own families, respondents re-
port the greatest trust in co-​religionists, and the least in non-​co-​ethnics. Levels of trust
are lower for members of the religious out-​group than those of the same faith, as shown
by the negative estimated religious trust deficits.
Both Muslims and Christians in Kaduna have significantly less social trust than is typ-
ical in Nigeria more broadly. National averages are between “trust somewhat” and “trust
a lot” for relatives, yet are between “trust a little” and “trust not at all” for Kaduna state
residents. Over half of Kaduna residents reported not trusting their relatives at all. Only
residents of the federal capital Abuja reported anywhere near that total, and nationally,
 Christian-Muslim Relations in the Shadow of Conflict    553

Table 26.6: Prejudice Indices, January 2015 and December 2015.

January 2015 December 2015

All Muslim Christian All Muslim Christian

Prejudice index 1.73 2.10 1.31 2.08 2.38 1.70

(0.07) (0.11) (0.08) (0.07) (0.09) (0.08)
Observations 225 119 106 246 138 108

Responses scaled 0 (most prejudice) -​4 (least prejudice). Standard errors in parentheses.

less than 13 percent of Nigerians report a complete lack of trust in their relatives. With
respect to trust in neighbors, Kaduna residents are again an extreme outlier in compar-
ison to the rest of the country (Pew Forum 2010).
Table 26.6 presents summary results by religious group for an index of explicit prej-
udice. Respondents were asked how well each of a series of interspersed positive and
negative attributes described members of the religious out-​group. Negative attributes
consisted of arrogant, ungrateful, unreasonable, and fanatical. Responses are scaled such
that higher values indicate stronger disagreement with negative attributes. Within poor
neighborhoods in Kaduna, both Muslims and Christians expressed overtly prejudiced
attitudes toward members of the religious out-​group. The mean Negative Attributes
index value for all respondents of 1.73 in the January 2015 survey indicates that the av-
erage respondent fell between agreement with the negative attributes and ambivalence.
As has been the case in other surveys in Nigeria and sub-​Saharan Africa, Christians
have more negative perceptions of Muslims than Muslims do of Christians (e.g., Pew
Forum 2010).
Table 26.7 presents summary results by religious group for a behavioral game
implemented during two rounds of our Kaduna survey in January and December 2015.
Respondents played ten rounds of a dictator (“divide the dollar”) game in both January
and December 2015. Within behavioral economics and political science, the dictator
game has been widely used to measure altruism and discrimination. In each round of
the game, survey respondents were asked in a row to choose how to divide 100 Nigerian
naira (about .50 US dollars at the time of game implementation) from a common pool
with another randomly assigned survey respondent. We primed survey respondents
with the first name of the individual to whom they could choose to allocate money, and
asked them to indicate how many 10-​naira notes they would give to the named indi-
vidual. Importantly, within Kaduna, first names clearly and unambiguously signal
religious affiliation. Respondents were informed that game play was nonreciprocal
(for example, being asked to divide 100 naira with Shehu does not necessarily mean
Shehu will be asked to divide 100 naira with you). This design mitigated concerns
about retribution. All behavioral games were administered privately at respondent
554    Alexandra Scacco and Shana S. Warren

Table 26.7: Discrimination
January 2015 Nov-​Dec 2015

All Muslim Christian All Muslim Christian

Play outgroup -​0.19** -​0.34** -​0.02 -​0.26** -​0.29** -​0.21*

(0.06) (0.08) (0.07) (0.06) (0.07) (0.08)
Constant 2.57** 2.59** 2.55** 2.78** 2.62** 2.97**
(0.10) (0.14) (0.15) (0.12) (0.15) (0.19)
Observations 2700 1450 1250 2730 1430 1300

Standard errors in parentheses) clustered by respondent.

* p < 0.5,
** p < 0.01

households; respondents circled the number of notes to give in each round on a separate
response sheet.
We implemented this game to examine how the treatment of individuals from the
out-​group differs from those from the in-​group based solely on group affiliation, the
only cue offered in the game. We found that respondents systematically discriminated
against members of the religious out-​group, with the exception of Christian respondents
in the January 2015 survey. We note that Nigeria held an election in early February 2015,
in which a Christian incumbent president lost the election to a Muslim challenger,
representing Nigeria’s first democratic transfer of power between parties since its inde-
pendence in 1963. While ethnic identity is often heightened in the run-​up to elections
(Eifert et al. 2010), tensions on the eve of the election may have encouraged Christian
respondents to censor their behavior.
Turning to the question of the potentially long-​term consequences of exposure to
communal violence for mental health, although our own survey did not include meas-
ures of respondents’ mental health, we draw on three recent public health studies that
examine large samples of survivors of Christian-​Muslim violence in Kaduna and Jos,
another religiously mixed state capital city in northern Nigeria that experienced large-​
scale Christian-​Muslim violence in 2001. All three studies indicate widespread trauma
in adult populations of survivors. Taken together, the Kaduna studies document high
levels of PTSD (more than 40 percent of respondents) and depression (between 50 and
60 percent) among interview respondents—​who witnessed violence, were temporarily
displaced from their homes, and/​or experienced personal property destruction—​many
years after large-​scale episodes of communal violence in 2000 and 2001 (Obilom and
Thacher 2008; Sheikh et al. 2014, 2015).
 Christian-Muslim Relations in the Shadow of Conflict    555

Conclusion

In this chapter we have demonstrated that large-​scale intercommunal violence across re-
ligious lines may produce severe and long-​lasting consequences. Using micro-​level data
from Kaduna, Nigeria, gathered fifteen years after a major violent event that produced
lasting residential segregation, we find that lack of trust and prejudice across religious
group divides may be particularly tenacious, given the close intragroup ties that are
formed and reinforced through everyday religious practice and residential segrega-
tion. In contrast to optimistic assessments of relatively speedy post-​conflict recoveries
in homogeneous local settings, and positive consequences of violence exposure for
within-​group cooperation and trust, we find that violence exposure may necessitate a
longer recovery in more diverse settings. Localities like Kaduna, Baghdad, Belfast, and
Ahmedabad that have experienced interreligious communal violence should expect
post-​conflict reconciliation to be slow, difficult, and far from guaranteed.
How then can social recovery occur after communal violence across religious lines?
Building and strengthening interreligious economic ties at the local level may increase
trust and decrease riot events in the long run, as has been demonstrated in India (Jha
2013). Social contact theory suggests that efforts to desegregate public life and increase
social contact across group lines may rebuild trust, decrease prejudice, reduce discrim-
ination, and thereby reduce the potential for violent conflict (Allport 1954). Recent
experimental studies of Christian-​Muslim social contact interventions in Kaduna
(Scacco and Warren 2018) and Baghdad (Mousa 2020) have demonstrated that although
interreligious prejudice may be slow to change, micro-​level interventions may be helpful
in reducing discriminatory behaviors, even in settings with particularly difficult conflict
histories.

Acknowledgments
The United States Institute of Peace (USIP) and the New York University Research Challenge
Fund (URCF) provided funding for this study. This research was approved via NYU IRB
Protocol 14-​9985. We are grateful to Chantal Pezold for research assistance. All errors and
omissions are our own.

Notes
1. For instance, the common outcome measure of “trust” used in the studies included in Bauer
et al. (2016) is “how much do you trust members of your village?,” where the villages are
typically homogeneous with respect to the identity cleavages along which the violence was
wrought.
556    Alexandra Scacco and Shana S. Warren

2. For an ethnographic account of “house swapping” by Christians and Muslims within

the Kaduna metropolitan area as part of a systematic process of post-​conflict residential
sorting, see Scacco (2020).
3. “Home language” is equivalent to ethnic affiliation in Nigeria.

References
Abdu, Hussaini, and Lydia Umar. 2002. Hope Betrayed: A Report on Impunity and State-​
Sponsored Violence in Nigeria. OMCT Report. Lagos, Nigeria: World Organization Against
Torture and Center for Law Enforcement Education.
Allport, Gordon Willar. 1954. The Nature of Prejudice. Reading, MA: Addison-​Wesley.
Angerbrandt, Henrik. 2011. “Political Decentralisation and Conflict: The Sharia Crisis in
Kaduna, Nigeria.” Journal of Contemporary African Studies 29, no. 1: 15–​31.
Bar-​Tal, D., and T. Avrahamzon. 2016. “Development of Delegitimization and Animosity
in the Context of Intractable Conflict.” In Cambridge Handbook of the Psychology of
Prejudice, edited by C. G. Sibley and F. K. Barlow, 582–​606. Cambridge: Cambridge
University Press.
Basedau, Matthias, Johannes Vüllers, and Peter Körner. 2013. “What Drives Inter-​religious
Violence? Lessons from Nigeria, Côte d’Ivoire, and Tanzania.” Studies in Conflict &
Terrorism 36, no. 10: 857–​879.
Bauer, Michal, Christopher Blattman, Julie Chytilová, Joseph Henrich, Edward Miguel, and
Tamar Mitts 2016. “Can War Foster Cooperation?” Journal of Economic Perspectives 30, no.
3: 249–​274.
Beber, Bernd, Philip Roessler, and Alexandra Scacco. 2014. “Intergroup Violence and Political
Attitudes: Evidence from a Dividing Sudan.” Journal of Politics 76, no. 3: 649–​665.
Blattman, Christopher. 2009. From Violence to Voting: War and Political Participation in
Uganda. American Political Science Review 103, no. 2: 231–​247.
Bogic, Marija, Anthony Njoku, and Stefan Priebe. 2015. “Long-​Term Mental Health of
War-​Refugees: A Systematic Literature Review.” BMC International Health and Human
Rights 15, no. 1: 1–​4 1.
De Waal, Alex. 2005. “Who Are the Darfurians? Arab and African Identities, Violence and
External Engagement.” African Affairs 104, no. 415: 181–​205.
Devine, Patricia G., E. Ashby Plant, David M. Amodio, Eddie Harmon-​Jones, and Stephanie
L. Vance. 2002. “The Regulation of Explicit and Implicit Race Bias: The Role of Motivations
to Respond without Prejudice.” Journal of Personality and Social Psychology 82, no. 5: 835–​848.
Eifert, Benn, Edward Miguel, and Daniel N Posner. 2010. “Political Competition and Ethnic
Identification in Africa.” American Journal of Political Science 54, no. 2: 494–​510.
Enos, Ryan D. 2017. The Space between Us: Social Geography and Politics. Cambridge: Cambridge
University Press.
Enos, Ryan D., and Christopher Celaya. 2018. “The Effect of Segregation on Intergroup
Relations.” Journal of Experimental Political Science 5, no. 1: 26–​38.
Falola, Toyin. 1998. Violence in Nigeria: The Crisis of Religious Politics and Secular Ideologies.
Rochester, NY: University Rochester Press.
Fearon, James D., and David D. Laitin. 2000. “Violence and the Social Construction of Ethnic
Identity.” International Organization 54, no. 4: 845–​877.
 Christian-Muslim Relations in the Shadow of Conflict    557

Fiske, Susan T. 1989. “Examining the Role of Intent: Toward Understanding Its Role in
Stereotyping and Prejudice.” In Unintended Thought, edited by J. S. Uleman and J. A. Bargh,
253–​283. New York: Guilford Press.
Garry, S., and F. Checchi. 2020. “Armed Conflict and Public Health: into the 21st Century.”
Journal of Public Health 42, no. 3: e287–​e298.
Gilligan, Michael J., Benjamin J. Pasquale, and Cyrus Samii. 2014. “Civil War and Social
Cohesion: Lab-​in-​the-​Field Evidence from Nepal.” American Journal of Political Science 58,
no. 3: 604–​619.
Gordon, Stuart, Amanda Baker, Alexia Duten, and Paul Garner. 2010. Study Exploring the
Evidence Relating Health and Conflict Interventions and Outcomes. Technical report.
London: UK Cross Government Group on Health and Conflict.
Grossman, Guy, Devorah Manekin, and Dan Miodownik. 2015. “The Political Legacies of
Combat: Attitudes toward War and Peace among Israeli Ex-​combatants.” International
Organization 69, no. 4: 981–​1009.
Hadzic, Dino, David Carlson, and Margit Tavits. 2020. “How Exposure to Violence Affects
Ethnic Voting.” British Journal of Political Science 50, no. 1: 345–​362.
Hager, Anselm, Krzysztof Krakowski, and Max Schaub. 2019. “Ethnic Riots and Prosocial
Behavior: Evidence from Kyrgyzstan.” American Political Science Review 113, no.
4: 1029–​1044.
Ibrahim, Jibrin. 1991. “Religion and Political Turbulence in Nigeria.” Journal of Modern African
Studies 29, no. 1: 115–​136.
Jha, Saumitra. 2013. “Trade, Institutions, and Ethnic Tolerance: Evidence from South Asia.”
American Political Science Review 107, no. 4: 806–​832.
Kaduna State Government. 2001, April. “White Paper on the Report of Judicial Commission of
Inquiry into Kaduna State Religious Disturbances of February 2000.”
Kasara, Kimuli. 2013. “Separate and Suspicious: the Social Environment and Inter-​Ethnic Trust
in Kenya.” Journal of Politics 75, no. 4: 921–​936.
Kasara, Kimuli. 2017. “Does Local Ethnic Segregation Lead to Violence?: Evidence from
Kenya.” Quarterly Journal of Political Science 11, no. 4: 441–​470.
Kendhammer, Brandon. 2014. “Citizenship, Federalism and Powersharing: Nigeria’s Federal
Character and the Challenges of Institutional Design.” Ethnopolitics 13, no. 4: 396–​417.
Lund, Crick, Alison Breen, Alan J. Flisher, Ritsuko Kakuma, Joanne Corrigall, John A. Joska,
Leslie Swartz, and Vikram Patel. 2010. “Poverty and Common Mental Disorders in Low
and Middle Income Countries: A Systematic Review.” Social Science & Medicine 71, no.
3: 517–​528.
Lupu, Noam, and Leonid Peisakhin. 2017. “The Legacy of Political Violence across
Generations.” American Journal of Political Science 61, no. 4: 836–​851.
Maier, Karl. 2002. This House Has Fallen: Nigeria in Crisis. Boulder, CO: Westview Press.
Mironova, Vera, and Sam Whitt. 2018. “Social Norms after Conflict Exposure and
Victimization by Violence: Experimental Evidence from Kosovo.” British Journal of Political
Science 48, no. 3: 749–​765.
Mousa, Salma. 2020. “Building Social Cohesion between Christians and Muslims through
Soccer in post-​ISIS Iraq.” Science 369, no. 6505: 866–​870.
Mpembi, Magloire Nkosi, Thierry Lukeba, Damien Mayemba, Victoria Kubuta Massamba,
Ma Nzuzi, Symphorien Gokara, Etienne Vermeiren, Gilbert Manangoa Lelo, and Samuel
M. Ma Miezil. 2018. “Stress Syndrome in Patients Receiving Outpatient Treatment at
558    Alexandra Scacco and Shana S. Warren

the General Hospital, in Bangui, in a Context of Armed Conflict.” Pan African Medical
Journal 29: 26.
Nunn, Nathan, and Leonard Wantchekon. 2011. “The Slave Trade and the Origins of Mistrust
in Africa.” American Economic Review 101, no. 7: 3221–​3252.
Obilom, Rose E., and Tom D. Thacher. 2008. “Posttraumatic Stress Disorder Following
Ethnoreligious Conflict in Jos, Nigeria.” Journal of Interpersonal Violence 23, no. 8: 1108–​1119.
Patel, Vikram, and Arthur Kleinman. 2003. “Poverty and Common Mental Disorders in
Developing Countries.” Bulletin of the World Health Organization 81: 609–​615.
Pew Forum. 2010. Tolerance and Tension: Islam and Christianity in Sub-​Saharan Africa.
Technical report. Washington, DC: Pew Forum on Religion & Public Life. https://​www.
pewforum.org/​interactives/​africa/​.
Rokem, Jonathan, Chagai M. Weiss, and Dan Miodownik. 2018. “Geographies of Violence in
Jerusalem: The Spatial Logic of Urban Intergroup Conflict.” Political Geography 66: 88–​97.
Sani, Shehu. 2007. The Killing Fields: Religious Violence in Northern Nigeria. Ibadan,
Nigeria: Spectrum Books.
Scacco, Alexandra. 2020. Anatomy of a Riot: Participation in Ethnic Violence in Nigeria. Book
Manuscript. Berlin: WZB Berlin Social Science Center.
Scacco, Alexandra, and Shana S. Warren. 2018. Can Social Contact Reduce Prejudice and
Discrimination? Evidence from a Field Experiment in Nigeria.” American Political Science
Review 112, no. 3: 1–​24.
Sharvit, Keren. 2014. “How Conflict Begets Conflict: Activation of the Ethos of Conflict in
Times of Distress in a Society Involved in an Intractable Conflict.” Journal of Experimental
Social Psychology 55: 252–​261.
Sheikh, Taiwo Lateef, Mohammed Abdulaziz, Samuel Agunbiade, Ike Joseph, Bill Ebiti, and
Oluwatosin Adekeye. 2015. “Correlates of Depression among Internally Displaced Persons
after Post-​election Violence in Kaduna, North Western Nigeria.” Journal of Affective
Disorders 170: 46–​51.
Sheikh, Taiwo Lateef, Abdulaziz Mohammed, Samuel Agunbiade, Joseph Ike, William
N. Ebiti, and Oluwatosin Adekeye. 2014. “Psycho-​ Trauma, Psychosocial Adjustment,
and Symptomatic Post-​traumatic Stress Disorder among Internally Displaced Persons in
Kaduna, Northwestern Nigeria.” Frontiers in Psychiatry 5: 127.
Tertsakian, Carina. 2003. Nigeria: The “Miss World Riots”: Continued Impunity for Killings in
Kaduna. New York: Human Rights Watch.
Ukiwo, Ukoha. 2003. “Politics, Ethno-​religious Conflicts and Democratic Consolidation in
Nigeria.” Journal of Modern African Studies 41, no. 1: 115–​138.
Ullah, Haroon. 2002. “Perspectives on Shariah, Federalism and Religious Associations in
Nigeria: An Urban Case Study of Kaduna and the Riots of 2000.” Technical report. Ann
Arbor: University of Michigan, Department of Political Science.
Varshney, Ashutosh. 2003. Ethnic Conflict and Civic Life: Hindus and Muslims in India. New
Haven, CT: Yale University Press.
Voors, Maarten J., Eleonora E. M. Nillesen, Philip Verwimp, Erwin H. Bulte, Robert Lensink,
and Daan P. Van Soest. 2012. “Violent Conflict and Behavior: A Field Experiment in
Burundi.” American Economic Review 102, no. 2: 941–​964.
Wapwera, Samuel Danjuma, and Jiriko Kefas Gajere. 2017. “Ethnoreligious Urban Violence and
Residential Mobility in Nigerian Cities: The Kaduna Experience.” Urban Studies Research
2017: 4624768.
 Christian-Muslim Relations in the Shadow of Conflict    559

Weidmann, Nils B., and Idean Salehyan. 2013. “Violence and Ethnic Segregation: A
Computational Model Applied to Baghdad.” International Studies Quarterly 57, no. 1: 52–​64.
Weiss, Chagai M. 2021. “Diversity in Health Care Institutions Reduces Israeli Patients’ Prejudice
Toward Arabs.” Proceedings of the National Academy of Sciences 118, no. 14: e2022634118.
doi:10.1073/​pnas.2022634118
Wilkinson, Steven I. 2006. Votes and Violence: Electoral Competition and Ethnic Riots in India.
Cambridge: Cambridge University Press.
 Chapter 27

New M edia and I sl a mi st

Mobiliz ation i n E g yp t

Alexandra A. Siegel

In the early days of the Arab Spring, as crowds gathered in Tahrir Square, academics and
journalists hailed social media as a “liberation technology” empowering young liberal
activists to protest and demand democratic change (Diamond and Plattner 2012). They
argued that by facilitating the creation and distribution of user-​generated content, social
media has democratized access to information and communication tools.1 A consensus
formed among techno-​optimists that Web 2.0 tools were tipping the scales, bolstering
the powerless more than the powerful (Castells 2012; Earl et al. 2013; Howard and
Hussain 2013; Tufekci and Wilson 2012). But young liberals were not the only actors to
exploit these tools. As early adopters of new media technologies with a well-​organized
ground game, Islamists were particularly well positioned to use social media to mobilize
when protests erupted in January 2011.
Despite the wealth of journalistic and scholarly attention paid to young activists’
use of social media in the early days of the Egyptian revolution, this chapter suggests
that Islamists were perhaps better equipped to use these tools to mobilize and amplify
their messages. Motivated by resource mobilization theory, I argue that movements
with stronger organizational structures, greater access to resources, and more co-
herent ideologies are able to use new media technologies more successfully than more
fragmented and less-​well-​funded groups.
Drawing on almost a decade of public Egyptian Facebook posts, this chapter
demonstrates that Islamist actors were particularly successful at gaining visibility and
spreading narratives that advance their goals on social media, relative to other political
actors. It also explores the political consequences of this online success, suggesting that
the same social media technologies that facilitated the Muslim Brotherhood’s mobiliza-
tion efforts in the post-​revolution period may have also undermined the organization
by accelerating its fragmentation, amplifying extremist voices, and giving the military
regime a new authoritarian toolkit with which to fight the Brotherhood on-​and offline.
562   Alexandra A. Siegel

New Media and Social Mobilization

While the Arab Spring protests and other recent episodes of collective action organ-
ized online have sparked a growing literature on the impact of new media technologies
on social mobilization,2 “new” media technologies have long shaped how groups mo-
bilize, as well as the consequences of this mobilization (Tarrow 2011; Tilly 1981). Since
the rise of print media and mass literacy, diverse movements have relied on pamphlets,
newspapers, radio, and television to broaden their constituencies, secure legitimacy,
and advance their agendas (Caren, Andrews, and Lu 2020).
The potential of media technologies to facilitate social mobilization can be under-
stood in terms of resource mobilization theory (McCarthy and Zald 1977). Resource mo-
bilization theory posits that centralized, formally structured organizations will be more
effective at mobilizing resources and mounting sustained challenges than decentralized,
informal movements, and their success is largely determined by strategic factors and
political opportunity structures (Jenkins 1983). This suggests that the ability of oppo-
sition elites—​including organizations, community leaders, and activists—​to mobilize
using new media technologies is moderated by organizational structure and access to
resources. Social movements will be able to use these tools more effectively when they
have strong centralized and efficient leadership and sufficient funding.
While social media is often heralded as a great equalizer that democratizes commu-
nication, viewed through the lens of resource mobilization theory, not all actors are
equally well positioned to mobilize on social media. Although social media can be effec-
tive in mobilizing the “powerless,” it has often reproduced or even intensified preexisting
power imbalances. Describing this phenomenon, Schradie (2019) argues that conserva-
tive movements dominate the online sphere due to their comparative advantage in off-
line resources, infrastructure, and ideologies that favor maintaining or returning to a
status quo. Although social media platforms are indeed low-​cost and nonhierarchical,
greater resources can be harnessed to maximize digital production and reach. Because
progressive movements tend to comprise diverse coalitions, they often struggle to prop-
agate a coherent message in the online sphere. As a result, while liberal movements have
received a great deal of attention for their ability to “go viral” and mobilize in the streets,
conservative digital media infrastructures have more quietly emerged, thrived, and
persisted.
In the Egyptian context, this suggests that Islamists—​particularly members of the
conservative and well-​established Muslim Brotherhood organization—​may have been
better positioned to effectively use social media tools than the liberal activists. At the
start of the Egyptian revolution, the Muslim Brotherhood had powerful networks on the
ground, extensive financial resources, and a relatively coherent ideology (Ullah 2017).
Moreover, as I describe in detail below, Islamists have a long history of exploiting media
technologies to mobilize, even under conditions of severe repression. Digital media
 New Media and Islamist Mobilization in Egypt    563

technologies are the latest manifestation of this trend, enabling Islamists to mobilize and
amplify their messages in real time.

Islamists’ Long History

with “New” Media

From audio tapes and photocopy machines to satellite television and social media,
Islamists have historically capitalized on cutting-​edge communication technologies
to mobilize followers. By imbuing their messages with religious meaning and drawing
on networks fostered through participation in religious community and institutions,
Islamists have frequently taken advantage of available media tools to gain mass appeal.
Such strategies have allowed diverse actors to build a shared set of symbolic references
among constituents, shape mass preferences on the role of Islam in politics, and open
new channels of increasingly rapid communication. This has enabled Islamists to more
effectively solve collective action problems—​both by increasing the degree of personal
gain individuals anticipate from mobilizing and by lowering individual costs of mobili-
zation (Mecham 2017).
In the 1950s and 1960s, as many Arab countries nationalized the press and began
to develop state monopolies of radio and television, Islamists faced barriers to
disseminating their messages. In Egypt, state radio and television broadcasted Friday
sermons and hosted programs where state-​sanctioned clerics would discuss religion,
but the Muslim Brotherhood and other Islamist organizations were banned and blocked
from the airwaves (Skovgaard-​Petersen 2011). In this period of almost complete state
control over the media, trust among the Egyptian public began to wane. Notoriously, in
1967, newspapers, radio stations, and television outlets across the Arab world reported
that Arab armies had soundly defeated Israel and that Israeli planes were falling from
the skies “like flies.” Egyptian citizens soon learned the truth of their defeat from foreign
sources, however, exacerbating widespread lack of faith in mass media (Fandy 2000).
In the 1970s and 1980s, amid this atmosphere of distrust of state media, the Islamic re-
vival put pressure on Arab states to allow more religious content on radio and television.
Most Arab countries slowly began airing Islamic programs as a means of demonstrating
that Islam was an integrated and valued part of the official national culture. At the same
time, however, Islamists use of mainstream media remained severely curtailed. While
the Egyptian regime permitted the publication of a few small Islamic magazines that
allowed some Islamist-​leaning clerics to spread their message, overtly Islamist actors
were barred from mass-​circulation newspapers, radio, and television (Skovgaard-​
Petersen 2011). As a result, Islamists developed other creative approaches to dissemi-
nate their messages. They distributed tape cassettes, copies of pamphlets, and later video
tapes to reach a broader audience discretely (Sivan 1990).
564   Alexandra A. Siegel

These inexpensive methods of communication also enabled a wider variety of

Islamist actors to reach potential followers. Indeed, across the Arab World, begin-
ning in the 1960s, audio cassettes of sermons were widely distributed—​primarily
among the lower classes—​mostly for the purpose of promoting individual piety and
correcting religious practice, though also to spread political messages (Echchaibi
2011). In Egypt, underground audiocassettes were particularly popular—​often more
popular than the content produced by state-​sanctioned clerics and formally edu-
cated religious scholars. This was, in part, because the tapes were in Egyptian dialect,
rather than the Modern Standard Arabic of the hadith and Qurʾan that was favored
by preachers at al-​Azhar (Egypt’s premier institution of Islamic learning). This meant
that messages on cassettes could reach less-​educated Egyptians in Cairo and rural
areas alike (Fandy 2000).
On the one hand, physical media like audiocassettes and photocopies were inefficient
means of transmitting a message, as they were highly dependent—​at least initially—​on
personal contact at mosques or rallies to spread the word. But they allowed for multiple
points of production and distribution. This meant that they were nearly untraceable
and irrepressible, making them powerful political messaging tools that authoritarian
regimes found difficult to control (Sreberny-​Mohammadi 1990). On the other hand,
these underground forms of communication did not insulate Islamists from state re-
pression. Daʾia, or Islamic preachers, who produced sermons through pamphlets or
audiocassettes that were considered “insurgent” or anti-​regime were frequently jailed,
tortured, or exiled (Hirschkind 2006).
Islamists also seized on the advent of satellite television in the early 1990s, achieving
new levels of mass-​media penetration (Lynch 2006; Ullah 2017). For example, the
Muslim Brotherhood–​affiliated cleric Yusef al-​Qardawi’s show on Al Jazeera, Sharia and
Life, launched in 1996, boasted 40–​60 million viewers worldwide (Rock 2010). The pro-
gram addressed controversial contemporary issues from the point of view of Islamic
law (Skovgaard-​Petersen 2011). Unlike cassette sermons, which were largely popular
among lower-​class Muslims and served no direct commercial purpose, these television
channels were and remain popular among middle-​and upper-​class Muslims and highly
profitable for media companies (Echchaibi 2011).
Although satellite television greatly expanded the national and transnational reach of
many clerics and Islamist movements, the license application, studio space, equipment,
and start-​up costs of launching a TV channel prohibited many Islamists from entering
this sphere (Mecham 2017). Along these lines, states were (and still are) able to exert
a great deal of control over satellite media by legislating, regulating, and owning sta-
tions, as well as through withholding funding, controlling trade unions, or threatening
journalists or editors. Indeed, in most countries throughout the Arab world, private TV
companies rely so heavily on the “good will” of the state that they express as much po-
litical loyalty as state channels (Sakr 2001). Thus, just as in the early days of radio and
terrestrial television networks, even as satellite television transformed the Arab media
environment, many Islamist actors remained excluded or limited by state control of the
airwaves.
 New Media and Islamist Mobilization in Egypt    565

Continuing their long history of strategically capitalizing on new media technologies,

Islamists were early adopters of the Internet. Even though Internet penetration was rela-
tively slow to expand across the Middle East, it facilitated the formation of new organiza-
tional structures that helped Islamists to mobilize. Operating as “portable headquarters,”
websites offered the Muslim Brotherhood newfound flexibility to expand their reach
both within Egypt and beyond national borders. In the early days of the Internet, some
Islamist organizations began to host their websites abroad to bypass authoritarian sur-
veillance, and frequently moved web addresses after government authorities shut down
their sites (Lerner 2010).
After decades of being excluded from television and radio shows, the Muslim
Brotherhood recognized the advent of the Internet as an opportunity to amplify its
message. In 1998 the organization commissioned a website for its print publication,
al-​Daʿwa (The Call), and launched other websites focusing on Islamist political is-
sues (Hafez 2008). Soon after, the Brotherhood began developing explicitly political
websites, including Egyptwindow.net in 2002 and their official website Ikhwanonline.
net in 2003, which became one of the most popular websites in Egypt (Awad 2017). After
the Egyptian regime closed down the official Brotherhood website in 2004, the move-
ment adapted by decentralizing its web presence. In this period, Brotherhood activists
also began using the Internet to organize demonstrations (Ajemian 2008). Brotherhood
blogs began to proliferate in 2007. As a result, the younger wing of the organization
gained visibility, using blogs to bring international attention to the plight of jailed
Brotherhood members (Ajemian 2008; Awad 2017).
As social media platforms gained popularity in Egypt, the Muslim Brotherhood
worked to grow its online audience. Hoping to supervise the discussions and activ-
ities of the organization’s youth, the older generation of Muslim Brotherhood lead-
ership developed alternative spaces for group members to interact. These included
Ikhwanbook, Ikhwan Twitter, and Ikhwantube, initially developed between 2009
and 2010. These platforms enabled members of the Brotherhood to share thousands
of documents, speeches, statements, and videos with constituents (Herrera 2014).
In the years before the Arab Spring, the Brotherhood also expanded its presence on
mainstream social media platforms, setting up official pages and accounts such as @
Ikhwanweb. By the time protests erupted in Tahrir Square in January 2011, the Muslim
Brotherhood already had established a presence on online platforms. As one media
strategist for the Freedom and Justice Party, the Brotherhood’s political party, put it,
“We used new media before the revolution to spread the message and people made fun
of us” (Burleigh 2012).

Tracking Islamist Online Mobilization

Islamists’ relatively centralized leadership, access to resources, and long history of

taking advantage of new media tools to organize suggests that they were especially well
566   Alexandra A. Siegel

positioned to mobilize on social media at the start of the January 25 revolution. But how
effective have Islamists been at gaining visibility and spreading narratives that advance
their goals on social media relative to other actors? To asses this, I draw on almost a
decade of public Egyptian Facebook posts and interactions from January 2011 to June
2020, collected using Facebook’s CrowdTangle API.3 Facebook is the most popular so-
cial media platform in Egypt, with 90 percent of Egyptian Internet users using the plat-
form as of 2019 (Dennis et al. 2019). Actors across Egypt’s political spectrum have public
Facebook pages and groups, including liberal activists, media outlets, military generals,
popular clerics, and prominent members of the Muslim Brotherhood. Facebook data
therefore allows me to compare how everyday Egyptian Facebook users engage with
these diverse actors over time.
To get a sense of the relative success of Islamist mobilization in the early days of the
Egyptian revolution, I compare engagement with the Rassd News Network, a popular
Facebook page used by the Muslim Brotherhood to organize protesters, to the Arabic-​
language “We are all Khaled Said” page, the Facebook page credited with helping spark
the Egyptian revolution and the epicenter of activity by liberal activists during the
protests (Alaimo 2015). This analysis is displayed in Figure 27.1.
Looking at data from the first year of the revolution, displayed in Figure 27.1, we see
that while posts on the “We are all Khaled Said” page generally received a higher level of

January 25 Revolution
1200000
Weekly Volume of Engagement with Public Posts

800000

400000

0
2010−07 2011−01 2011−07 2012−01
Date
page
Rasd News Network We’re All Khaled Said

Figure 27.1: Engagement with Rassd Network vs. We are all Khaled Said Facebook pages.
Note: Plot displays the weekly volume of engagement (interactions, including likes, shares,
comments, and reactions) with public posts on the Rassd News Network and We are all Khaled
Said Facebook pages from July 2010 to January 2012.
Source: Data was collected through the Crowdtangle API.
 New Media and Islamist Mobilization in Egypt    567

engagement (18,430,921 total interactions), posts on the Rassd Network page perform
surprisingly well by comparison (13,214,113 total interactions). Interestingly, the Rassd
Network received more engagement than the Khaled Said page in the first two weeks of
the Egyptian revolution. This is particularly noteworthy given that the Khaled Said page
is frequently credited as playing an important role in sparking the Egyptian revolution.
Moreover, the Rassd Network page was not created until January 25, the first day of the
revolution, whereas the Khaled Said page had been gaining followers since its creation
in June 10, 2010.
In the post-​revolution period, the Muslim Brotherhood continued to use social
media to disseminate their messages and broaden their appeal. Following the uprisings,
the use of social media across the Arab world grew dramatically (Liu, Kliman-​Silver,
and Mislove 2014), enabling the Brotherhood to reach a larger online audience than it
had previously. This bolstered the organization’s success as it seized the opportunity to
compete for power in Egypt’s first democratic elections.
The Muslim Brotherhood’s electoral success in 2012 is often—​rightly—​credited to its
impressive ground game, cohesion, and reputational advantage (Brooke 2019; Cammett
and Luong 2014). At the same time, the use of social media alongside this offline mobi-
lization helped to amplify the Brotherhood’s message and broaden its appeal in this pe-
riod. Beginning in April 2012, the Muslim Brotherhood began strategically promoting
its election campaign on Facebook and Twitter. While many of the Brotherhood’s
constituents remained offline and in rural areas, social media helped Islamists to mo-
bilize educated and urban Egyptians. As one spokesman for the Muslim Brotherhood
described in the lead-​up to the 2012 election, “It is important for me to know the opinion
of the educated classes of Egyptians and I can find those by mining the Internet . . .
I don’t want to lose them. . . . My state needs the educated people to drive the state—​but
under my rules. I have to cooperate with them, understand them and work with them”
(Srinivasan 2014). In the lead up to the elections, candidates from across the political
spectrum promoted their campaigns and attacked one another on Facebook, Twitter,
and YouTube.
Examining Facebook posts containing the Arabic term intikhabat or “elections”
over time offers some insight into Islamists’ successful use of social media during the
2012 elections. Figure 27.2 shows engagement with the most popular candidate and
party Facebook pages posting about the election in the lead-​up to and aftermath of
the 2012 presidential election. The solid line shows the weekly volume of interactions
with posts from Islamist party and candidate pages, while the dotted line shows
interactions with non-​Islamist candidates and pages. As the figure suggests, Islamists
dominated discussions of the election on Facebook both before and after the two elec-
tion rounds.
The Muslim Brotherhood’s online presence may also have contributed to its ability
to peacefully transition to power in the aftermath of Morsi’s electoral victory. Their
coordinated online and offline campaign in the period surrounding the election
made it very difficult for the election to be rigged or stolen and forced a new level
568   Alexandra A. Siegel

Round 1
Weekly Volume of Engagement with Public Posts

15000 Round 2

10000 page
Islamist
Non−Islamist

5000

0
Jan 2012 Apr 2012 Jul 2012 Oct 2012 Jan 2013
Date

Figure 27.2: Facebook engagement with posts by Islamist vs. non-​Islamist candidates and
parties referencing the election.
Note: Plot displays the weekly volume of engagement (interactions, including likes, shares,
comments, and reactions) with public posts by Islamist and non-​Islamist candidates from
January 2012 to January 2013.
Source: Data was collected through the Crowdtangle API.

of transparency. On the night of the election, members of the Muslim Brotherhood

systematically updated and publicized results in real time on Twitter. The Muslim
Brotherhood had people on the ground at polling stations who called electoral results
into Muslim Brotherhood headquarters as votes were counted. As Zeynep Tufekci
blogged in the aftermath of the election, if the old-​regime apparatus had rigged the
election or changed the results, a “Google spreadsheet would stare at them at every
turn. And the narrative laid by Ikhwan’s online and offline operation would be hard to
surmount.”4
The July 2013 military coup that ousted Mohamed Morsi sparked the largest wave
of Islamist mobilization in Egypt’s modern history. For months, in the face of vio-
lent repression, anti-​coup protesters turned out in the streets. Protests persisted as
Brotherhood leaders and members were jailed, tortured, and killed; financial assets
were frozen; associations and centers were closed; and the organization was desig-
nated a terrorist group. Despite massive setbacks, the Muslim Brotherhood continued
to organize. In this period, social media became a particularly effective resource for the
Brotherhood. It helped the organization to broadcast its messages in real time and or-
ganize collective action as its traditional lines of communication were shut down. On
the one hand, the Muslim Brotherhood’s use of social media in the post-​coup period
represents a continuation of its long history of exploiting new media channels in the
face of repression. On the other hand, the real-​time networked structure of social media
 New Media and Islamist Mobilization in Egypt    569

Death Morsi’s
Election Coup Sentence Coup Death
1000000
750000
500000
Weekly Volume of Engagement with Public Posts

250000
0
2012 2014 2016 2018 2020
Death Morsi’s
Election Coup Sentence Morsi Death
3000000
2000000
1000000
0
2012 2014 2016 2018 2020
Death Morsi’s
Election Coup Sentence Raba massacre Death
250000
200000
150000
100000
50000
0
2012 2014 2016 2018 2020
Date

Figure 27.3: Facebook engagement with posts mentioning Morsi, the coup, or the Rabaa
massacre.
Note: Plot displays the weekly volume of engagement (interactions, including likes, shares,
comments, and reactions) with public posts on Egyptian Facebook pages referencing Morsi, the
coup, or the Raba’a massacre from January 2012 to June 2020.
Source: Data was collected through the Crowdtangle API.

made these tools particularly well suited to mobilizing collective action on the ground
in a manner that previous technologies could not.
After Morsi was removed from power, his supporters used social media to condemn
the military coup, highlight military brutality, and call for democratization. Dozens of
popular pro-​Brotherhood Facebook pages were used to disseminate the Brotherhood’s
message after its satellite channels were blocked. For example, the Rassd News Network
page developed before the revolution had an audience of almost five million people by
the time of the coup. The site worked to “reveal the agenda” of the military through a
series of video leaks known as elSisileaks. Other pro-​Brotherhood Facebook networks,
such as Ahrar 25 TV and Egypt Online, also became very popular in this period
(Shehabat 2015). Pro-​Brotherhood Facebook pages and Twitter accounts worked to
document military brutality such as the violent crackdowns on peaceful protests held in
Rabaa and el-​Nahda Squares when many thousands of peaceful protesters were killed.
YouTube videos, photos, and voice memos were captured by protesters and posted daily
in an attempt to persuade followers to denounce military rule. The daily volume of posts
mentioning Morsi, the coup, and the Rabaa massacre on Egyptian Facebook pages are
displayed in Figure 27.3.
These social media platforms were similarly used to mobilize collective action on
the ground. The “Peace and Justice Party” Facebook page quickly gained almost two
million followers, and frequently called for people to respect the legitimacy of the
570   Alexandra A. Siegel

1500000

Weekly Volume of Engagement with Public Posts

1000000
Egyptian Coup

500000

2012 2014 2016 2018 2020

Date

Figure 27.4: Facebook engagement with Freedom and Justice Party page.
Note: Plot displays the weekly volume of engagement (interactions, including likes, shares,
comments, and reactions) with public posts on the Freedom and Justice Party Facebook page
from July 2010 to June 2020.
Source: Data was collected through the Crowdtangle API.

2012 elections and protest the coup. As Figure 27.4 demonstrates, engagement with the
page increased in the immediate aftermath of the coup and again spiked throughout
2014 and 2015 as the crackdown on the Brotherhood intensified. Similarly, the “Morsi’s
Ultras” page called for daily mobilization of Muslim Brotherhood supporters to bring
down military rule (Shehabat 2015). In addition to these public pages, anti-​coup
activists also used closed social media groups to coordinate national days of protest
and disseminate tactics. This was especially necessary as the regime began to monitor
the anti-​coup movement online and search protesters’ mobile phones or laptops. As a
safety measure, these private groups, some of which had many thousands of members,
were given misleading names such as “I love Christiano Ronaldo” or “Buy second-​hand
cars” (Ketchley 2017). Additionally, demonstrations were creatively organized online
to prevent surveillance, and false dates for protests were posted on public social media
platforms to throw off security forces monitoring the Brotherhood’s social media ac-
counts (El-​Sharif 2014).
While most of this mobilization initially focused on peaceful protests, calls for ac-
tion became increasingly violent over time. For example, postings about the third
 New Media and Islamist Mobilization in Egypt    571

anniversary of the Egyptian revolution celebration included warnings directed at the

military regime, claiming that protesters would respond violently to any military bru-
tality by burning police cars and armored vehicles, abducting military personnel, and
using Molotov cocktails and attacking police stations. Pro-​Brotherhood social media
users shared advice and tactics on these pages, including videos from activists as far
afield as the Ukraine, who, in the words of one Brotherhood supporter, knew “how to
use Molotovs to hold their square against the police” (Ketchley 2014).
Pro-​Brotherhood Facebook pages were also used to call for civil disobedience in this
period. Brotherhood supporters were mobilized to disrupt trains in the metro, boycott
military products, close businesses, and even set fire to public buildings (Pargeter 2018).
One civil disobedience campaign started on pro-​Brotherhood Facebook pages was
called “expose them” and worked to dox or identify police and military personnel who
had been involved in the killings of pro-​Morsi protesters. Participants posted photos,
phone numbers, and addresses of these individuals, urging supporters to seek revenge
(Shehabat 2015).
The transnational reach of social media further bolstered mobilization of the anti-​
coup movement both domestically and abroad. Following the violent dispersal of
anti-​coup protests in Rabaa al-​Adawiya Square in June 2013, the four-​fingered hand ges-
ture that came to be known as the “Rabaa Salute” went viral on social media (van de
Bildt 2015). Anti-​coup Facebook pages and Twitter accounts called for their followers
to use the Rabaa symbol in their profile pictures to show their solidarity with those
who were harmed during the government crackdown. The symbol quickly spread, not
only on Egyptian social media platforms, but also in the Gulf, Turkey, and across the
broader Muslim world as sympathizers expressed their solidarity. Turkish prime min-
ister Erdogan even played a role in the virality of the symbol, using the Rabaa salute
to greet Turkish crowds shortly after the massacre (El-​Shenawi 2013). In this same pe-
riod, popular pro-​Brotherhood clerics also expressed their support for the movement
online, bringing more transnational support to the movement. For example, the Saudi
cleric Mohamed al-​Arefe tweeted and posted YouTube videos calling for solidarity
with Morsi’s supporters, which quickly went viral among his millions of fans across the
Muslim world (Pan and Siegel 2020).
Social media also facilitated communication with exiled Brotherhood leader-
ship and access to official Brotherhood media abroad. Many members of the Muslim
Brotherhood sought refuge in Turkey after the coup. These exiled Islamists quickly
set up their own television channels in Turkey in order to continue to broadcast their
message. These pro-​Brotherhood channels, which air twenty-​four hours a day, in-
clude Rabaa, Misr Al’n, Mekammelyn, and Al Sharq (Magued 2018). Because each
of these channels maintains active social media profiles on YouTube, Facebook, and
Twitter, they were easily able to reach Muslim Brotherhood supporters in Egypt.
Social media also made these television channels interactive, so followers and
opponents could post content and interact with TV presenters in real time, directly
incorporating social media into broadcasts. For example, the Al Sharq channel began
572   Alexandra A. Siegel

Raba’a Inside Egypt

Weekly Volume of Engagement with Public Posts 250000 Raba’a massacre

200000

150000

100000

50000

Raba’a Outside Egypt

250000 Raba’a massacre

200000

150000

100000

50000

0
2012 2014 2016 2018 2020
Date
Figure 27.5: Facebook engagement with posts mentioning the Rabaa massacre inside and
outside Egypt.
Note: Plot displays the weekly volume of engagement (interactions, including likes, shares,
comments, and reactions) with public posts on Egyptian and Non-​Egyptian Facebook pages
referencing the Rabaa massacre from January 2011 to June 2020.
Source: Data was collected through the Crowdtangle API.

presenting programming where TV presenters discussed current issues in Egypt

with viewers by reading aloud their comments on the channel’s Facebook page and
Twitter account.
The Facebook pages associated with these TV programs were critical in drawing at-
tention to the anti-​coup protests. Videos showing large anti-​coup demonstrations
occurring on Friday nights in Egypt and around the world spread quickly online.
These videos were frequently live broadcasts of ongoing demonstrations and protests
taken by demonstrators on their smartphones. In addition to protest coverage, these
channels and their associated pages also documented the humanitarian conditions in
Egypt’s prisons as well as growing numbers of arrests, torture, and death sentences for
Brotherhood leaders and supporters.
Examining interactions with posts mentioning the Rabaa massacre, Figure 27.5
suggests that posts on pages outside of Egypt received more engagement than posts from
Egyptian Facebook pages, particularly after 2014 (11 million interactions vs. 9 million
interactions in total from 2011 to 2020). The most popular posts outside of Egypt came
from the Gulf, Turkey, and the UK, where many members of the Muslim Brotherhood
moved after the coup.
 New Media and Islamist Mobilization in Egypt    573

How New Media Undermined

Islamist Mobilization

As these examples suggest, social media played a key role in amplifying the Brotherhood’s
messages, spreading news and information, and facilitating collective action in the af-
termath of the Arab Spring, bolstering its ground mobilization strategies. But the same
features of social media that helped the Brotherhood to mobilize in the face of repres-
sion also undermined the organization by accelerating its fragmentation and polariza-
tion and amplifying extremist voices. Moreover, at the same time that the Brotherhood
was using social media to mobilize, the military regime worked to use the tool against the
movement.
While the proliferation of Brotherhood social media pages and accounts may have
facilitated the spread of information and mobilization of collective action after the
coup, by reducing barriers to entry and undermining hierarchical structures, social
media may have played a role in fragmenting the authority of the Egyptian Muslim
Brotherhood. The coup that removed Mohamed Morsi from office and the subsequent
crackdown on Brotherhood activities dealt a devastating blow to the organization’s
leadership. With the core Brotherhood leaders imprisoned, rival factions began to op-
erate separate websites with different spokespeople. Competing to secure the loyalty
of their base, senior leaders posted rival statements on websites, and their followers
responded in real time, frequently arguing on the leaders’ Facebook pages. Other
members set up independent Facebook pages to assert their demands or attack rival
factions (Awad 2017).
While social media was of course one of many factors that contributed to this frag-
mentation of the Muslim Brotherhood, it provided an avenue for new actors to fill the
void, amplifying extremist voices. For example, Brotherhood members courting Salafi
support frequently framed the state’s crackdown on the organization as a war against
Islam, encouraging angry youth to embrace violent Salafi jihadism. As Awad (2017)
argues, “[t]‌hrough these [social] media platforms, Brotherhood leaders and clerics
also attempt to acculturate a fragmented organization and win over a young rebellious
generation. When left unchecked, Brotherhood clerics and their Salafist allies with
Gulf funding quickly propagated the most hateful, violent, and sectarian expressions
of Islamism, arguably helping radicalize some youth by providing them with religious
justifications for violence against the state.”
Additionally, social media has several affordances that facilitated the amplifica-
tion of extremist voices within the movement. First, online platforms provide social
distance and anonymity, enabling people to produce and spread extreme or vio-
lent content with few consequences. Second, the amplifying nature of social media
makes extremist content—​once relegated to the dark corners of the Internet—​
highly visible (Siegel 2020). As a consequence, Egyptian extremist groups like the
Islamic State in the Sinai and Ansar-​B eit al-​Maqdis have been adept at using a wide
574   Alexandra A. Siegel

100 Egyptian Coup

75
Search hits

50

25

0
2012 2014 2016 2018 2020
Date

Figure 27.6: Searches for Ansar-​Beit al-​Maqdis YouTube videos.

Note: Plot displays the relative weekly volume of YouTube searches located in Egypt referencing
Ansar-​Beit al-​Maqdis from January 2011 to June 2020.
Source: Data was collected through the Google Trends API.

variety of social media tools to propagate their message and lure recruits. Groups
like the Islamic State have used a combination of “atrocity porn,” fear, and entertain-
ment to gain visibility on social media (Berger and Morgan 2015; Siegel and Tucker
2018). They have also effectively developed cross-​platform means of reaching
supporters and would-​be followers, radicalizing Brotherhood youth. For example,
as Figure 27.6 suggests, Egyptian YouTube searches for Ansar-​B eit al-​Maqdis spiked
after the coup, and then grew more dramatically in 2015, highlighting rising interest
in the extremist group.
The networked structure of social media also means that individuals who seek out
extremist content online often find themselves deeply embedded in echo chambers
where harmful content is normalized and encouraged (Siegel 2020). Research on the
Egyptian Twittersphere suggests that Egyptians became increasingly polarized into
Islamist and Secular or non-​Islamist networks after the coup (Borge-​Holthoefer et al.
2015; Lynch, Freelon, and Aday 2017). Indeed, the period after the coup was marked by
a “great unfriending” in which anti-​Brotherhood social media users stopped following
Islamists, and vice versa. This polarization appeared to push Egyptians into more ex-
treme camps. A recent study of political tolerance in Egyptian Twitter networks after
the coup demonstrates that individuals who spent an additional year in ideologi-
cally homogenous networks—​that is, networks where they were only exposed to con-
tent produced by either Islamists or non-​Islamist Egyptians—​expressed less tolerant
positions toward the out-​group over time (Siegel et al. 2020). In this way, as viral rumors
 New Media and Islamist Mobilization in Egypt    575

500000
Weekly Volume of Engagement with Public Posts

400000

300000

200000

100000

2012 2014 2016 2018 2020

Date

Figure 27.7: Facebook engagement with posts mentioning “Tahya Masr.”

Note: Plot displays the weekly volume of engagement (interactions including likes, shares,
comments and reactions) with public posts on Egyptian Facebook pages referencing the pro-​Sisi
and pro-​Military slogan “Tahya Masr” (Long Live Egypt) from January 2011 to June 2020.
Source: Data was collected through the Crowdtangle API.

and polarizing content flowed through segregated online networks, moderate voices
lost traction and extremist voices were amplified.
In addition to the acceleration of fragmentation, polarization, and extremism within
the organization, the Muslim Brotherhood also faced a direct challenge from the re-
gime in the online sphere. Authoritarian regimes in diverse contexts have not stood idly
by as citizens and movements have used new media technologies to mobilize against
them. Digital media provides regimes with an authoritarian toolkit, which they can use
to promote their messages while stifling opposition (Roberts 2018; Tucker et al. 2017). As
Figure 27.7 demonstrates, the volume of engagement with posts using the pro-​Sisi and
pro-​military slogan “Tahya Masr” (Long Live Egypt) have continued to rise since Sisi
came to power.
In the post-​coup period, Egypt’s military regime stepped up efforts to use social media
to surveil and target members of the Muslim Brotherhood. While many members were
arrested during protests on the ground, the government also began investing in new
technology to monitor their online communications on a large scale. Although offi-
cially the monitoring was intended to focus on terror attacks, as one Egyptian Interior
Ministry official described in 2014, “We are looking at any conversation, any interaction,
576   Alexandra A. Siegel

15000 Coup
Weekly Volume of Engagement with Public Posts

10000

5000

2012 2014 2016 2018 2020

Date

Figure 27.8: Facebook engagement with posts referring to Muslim Brotherhood members as
terrorists.
Note: Plot displays the weekly volume of engagement (interactions, including likes, shares,
comments, and reactions) with public posts on Egyptian Facebook pages referencing the Arabic
phrase “Brotherhood are Terrorists” from January 2011 to June 2020.
Source: Data was collected through the Crowdtangle API.

we might find worrying or would want to keep a closer eye on. . . . We are watching
conversations between Islamists, or those who discuss Islamism. We are watching
communities, which we consider at risk” (Frenkel and Atef 2014). Along these lines,
social media communications have increasingly been used as a pretext for arresting
members of the Muslim Brotherhood (Freedom House 2018), and recent legislation
considers any individual with over 5,000 followers on social media to be a media entity,
enabling the regime to further monitor and control online activity.
Beyond intimidation and censorship of online activists, the Egyptian regime
has also embraced the use of computational propaganda, with General Sisi even
stating publicly that he could sway social media agenda using “electronic brigades”
(Darwish et al. 2017). These messages pollute the online sphere, inflating perceptions
of grassroots support for the military regime and drowning out other voices. By
portraying members of the Muslim Brotherhood as extremists and terrorists in the
online sphere, the regime has effectively discredited the organization in the eyes of
many Egyptians. As Figure 27.8 demonstrates, the volume of engagement with posts
referring to members of the Muslim Brotherhood as terrorists increased after the
coup and spiked again in 2020.
 New Media and Islamist Mobilization in Egypt    577

Conclusions

Islamists have long capitalized on new media technologies to mobilize their followers
in the face of repression, and their embrace of social media is the latest manifestation of
this trend. As the examples detailed in this chapter suggest, the Muslim Brotherhood’s
early adoption of new media technologies, extensive offline networks and financial re-
sources, and hierarchical structure enabled the organization to be remarkably effec-
tive on social media relative to non-​Islamist actors. Additionally, the organization has
proven resilient, mobilizing online persistently in the face of severe repression in the
aftermath of the coup that ousted Mohamed Morsi.
However, the same tools that gave Islamists new opportunities to mobilize have
also accelerated the fragmentation of their leadership and amplified extremist voices.
Given how adept extremist actors are at exploiting online platforms to broadcast
their messages, this development is particularly troubling from a policy perspective.
Moreover, as Egypt’s military regime has increasingly harnessed social media to achieve
its goals of stifling opposition, the ability of Islamists to continue to effectively mobilize
on these platforms remains uncertain. If, as resource mobilization theory would suggest,
groups with more centralized leadership and access to resources will mobilize more ef-
fectively both on-​and offline, then we are likely to see the Egyptian regime, and repres-
sive regimes in diverse contexts, increasingly exert control over the online sphere. At the
same time, the flexible and fast-​evolving nature of social media platforms enables op-
position movements to constantly find new opportunities to mobilize. Given Islamists’
long history of creatively harnessing new media technologies in the face of repression,
we may instead continue to see the Muslim Brotherhood embrace alternative online
platforms for mobilization. Thus, the cat-​and-​mouse game between Islamists and au-
thoritarian regimes will likely continue to evolve across new platforms as Islamists con-
tinue their power struggle on-​and offline.

Notes
1. See Tucker et al. (2017) for an overview.
2. See Zhuravskaya, Petrova, and Enikolopov (2020) for a recent review of the literature.
3. CrowdTangle is a social media analytics platform owned by Facebook that tracks public
posts on Facebook made by public accounts or groups as well as public interactions (likes,
reactions, comments, shares, upvotes) to these posts. I received CrowdTangle data access
from Facebook in collaboration with Social Science One through NYU’s Center for Social
Media and Politics. Crowdtangle tracks 99 percent of public posts from pages with over
100,000 likes, as well as a large number of pages with smaller followings. This makes it ideal
for measuring engagement with posts over time as posts from more popular pages that re-
ceive high levels of engagement. It does not include paid ads unless those ads began as or-
ganic, nonpaid posts that were subsequently boosted using Facebook’s advertising tools.
CrowdTangle also does not track posts made visible only to specific groups of followers.
578   Alexandra A. Siegel

See https://​help.crowdtangle.com/​en/​articles/​1140930-​what-​is-​crowdtangle-​tracking for

an overview of what data is included through the API.
4. Available at: http://​technosociology.org/​?p=1064.

References
Ajemian, Pete. 2008. “The Islamist Opposition Online in Egypt and Jordan.” Arab Media &
Society 4. https://​www.arabmediasociety.com/​the-​islamist-​opposition-​online-​in-​egypt-​
and-​jordan/​.
Alaimo, Kara. 2015. “How the Facebook Arabic Page “We Are All Khaled Said” Helped Promote
the Egyptian revolution.” Social Media+ Society 1, no. 2. https://​journals.sagepub.com/​doi/​
full/​10.1177/​2056305115604854.
Awad, Mokhtar. 2017. “The Role of Traditional and New Media in the Egyptian Muslim
Brotherhood’s Internal Power Struggle.” POMEPS Briefing. Washintgon, DC: Project on
Middle East Political Science. https://​pomeps.org/​the-​role-​of-​traditional-​and-​new-​media-​
in-​the-​egyptian-​muslim-​brotherhoods-​internal-​power-​struggle.
Berger, J. M., and J. Morgan. 2015. “The ISIS Twitter Census: Defining and Describing the
Population of ISIS Supporters on Twitter.” Analysis Paper no. 20. Washington, DC: The
Brookings Project on US Relations with the Islamic World.
Borge-​Holthoefer, Javier, Walid Magdy, Kareem Darwish, and Ingmar Weber. 2015. “Content
and Network Dynamics behind Egyptian Political Polarization on Twitter.” In Proceedings of
the 18th ACM Conference on Computer Supported Cooperative Work and Social Computing,
700–​7 11. New York: Association for Computing Machinery.
Brooke, Steven. 2019. Winning Hearts and Votes: Social Services and the Islamist Political
Advantage. Ithaca, NY: Cornell University Press.
Burleigh, Nina. 2012. “Islam’s High-​ Tech Backlash.” Discover Magazine. https://​www.
discovermagazine.com/​technology/​islams-​high-​tech-​backlash
Cammett, Melani, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17: 187–​206.
Caren, Neal, Kenneth T. Andrews, and Todd Lu. 2020. “Contemporary Social Movements in a
Hybrid Media Environment.” Annual Review of Sociology 46: 443–​465.
Castells, Manuel. 2012. Networks of Outrage and Hope: Social Movements in the Internet Age.
Cambridge: Polity Press.
Darwish, Kareem, Dimitar Alexandrov, Preslav Nakov, and Yelena Mejova. 2017. “Seminar
Users in the Arabic Twitter Sphere.” In Proceedings of 9th International Conference on Social
Informatics, edited by Giovanni Luca Ciampaglia, Afra Mashhadi, Taha Yasseri, 91–​108.
Cham, Switzerland: Springer.
Dennis, E., J. Martin, Robb Wood, and M. Said. 2019. Media Use in the Middle East 2019: A
Seven-​Nation Survey. Research Report. Doha: Northwestern University in Qatar.
Diamond, Larry, and Marc F. Plattner. 2012. Liberation Technology: Social Media and the
Struggle for Democracy. Baltimore: Johns Hopkins University Press.
Earl, Jennifer, Heather McKee Hurwitz, Analicia Mejia Mesinas, Margaret Tolan, and Ashley
Arlotti. 2013. “This protest Will Be Tweeted: Twitter and Protest Policing during the
Pittsburgh G20.” Information, Communication & Society 16 no. 4: 459–​478.
Echchaibi, Nabil. 2011. “From Audio Tapes to Video Blogs: The Delocalisation of Authority in
Islam.” Nations and Nationalism 17, no. 1: 25–​44.
 New Media and Islamist Mobilization in Egypt    579

El-​Sharif, Ashraf. 2014. The Muslim Brotherhood and the Future of Political Islam in Egypt.
Washington, DC: Carnegie Endowment for International Peace.
El-​Shenawi, Eman. 2013. “Four-​Finger Salute: Egypt Rivals use ‘Rabaa Hand’ to Turn Facebook
Yellow.” Al-​Arabiya 6. http://​english.alarabiya.net/​en/​media/​2013/​08/​21/​Four-​finger-​salute-​
Egypt-​rivals-​use-​Rabaa-​symbol-​to-​turn-​Facebook-​yellow.html.
Fandy, Mamoun. 2000. “Information Technology, Trust, and Social Change in the Arab
World.” Middle East Journal 54, no. 3: 378–​394.
Freedom House. 2018. “Freedom on the Net 2018: The Rise of Digital Authoritarianism.”
Washington, DC: Freedom House.
Frenkel, Sheera, and Maged Atef. 2014. “Egypt Begins Surveillance of Facebook, Twitter, and
Skype on Unprecedented Scale.” BuzzFeed News, September 17, 2014.
Hafez, Ahmad. 2008. “Muslim Brotherhood Websites.” Ikhwanweb. https://​www.ikhwanweb.
com/​article.php?id=18865=search.php.
Herrera, Linda. 2014. Wired Citizenship: Youth Learning and Activism in the Middle East.
London: Routledge.
Hirschkind, Charles. 2006. The Ethical Soundscape: Cassette Sermons and Islamic Counterpublics.
New York: Columbia University Press.
Howard, Philip N., and Muzammil M. Hussain. 2013. Democracy’s Fourth Wave?: Digital Media
and the Arab Spring. Oxford: Oxford University Press.
Jenkins, J. Craig. 1983. “Resource Mobilization Theory and the Study of Social Movements.”
Annual Review of Sociology 9, no. 1: 527–​553.
Ketchley, Neil. 2014. “How Social Media Spreads Protest Tactics from Ukraine to Egypt.”
Washington Post, February 14, 2014.
Ketchley, Neil. 2017. Egypt in a Time of Revolution. Cambridge: Cambridge University Press.
Lerner, Melissa Y. 2010. “Connecting the Actual with the Virtual: The Internet and Social
Movement Theory in the Muslim World—​The Cases of Iran and Egypt.” Journal of Muslim
Minority Affairs 30, no. 4: 555–​574.
Liu, Yabing, Chloe Kliman-​ Silver, and Alan Mislove. 2014. “The Tweets They Are A-​
Changin’: Evolution of Twitter Users and Behavior.” In ICWSM 2014: Eighth International
AAAI Conference on Weblogs and Social Media. Menlo Park, CA: Association for the
Advancement of Artificial Intelligence.
Lynch, Marc. 2006. Voices of the New Arab Public: Iraq, Al-​Jazeera, and Middle East Politics
Today. New York: Columbia University Press.
Lynch, Marc, Deen Freelon, and Sean Aday. 2017. “Online Clustering, Fear and Uncertainty in
Egypt’s Transition.” Democratization 24, no. 6: 1159–​1177.
Magued, Shaimaa. 2018. “The Egyptian Muslim Brotherhood’s Transnational Advocacy in
Turkey: A New Means of Political Participation.” British Journal of Middle Eastern Studies 45,
no. 3: 480–​497.
McCarthy, John D., and Mayer N. Zald. 1977. “The Trend of Social Movements in America:
Professionalization and Resource Mobilization.” CRSO Working Paper 164. Ann Arbor, MI:
Center for Research on Social Organization.
Mecham, Quinn. 2017. Institutional Origins of Islamist Political Mobilization. Cambridge:
Cambridge University Press.
Pan, Jennifer, and Alexandra A. Siegel. 2020. “How Saudi Crackdowns Fail to Silence Online
Dissent.” American Political Science Review 114, no. 1: 109–​125.
Pargeter, Alison. 2018. “The Muslim Brotherhood in Egypt: Current Crisis and Future
Directions.” Terrorism Monitor 16, no. 16. https://​jamestown.org/​program/​the-​muslim-​
brotherhood-​in-​egypt-​current-​crisis-​and-​future-​directions/​.
580   Alexandra A. Siegel

Roberts, Margaret E. 2018. Censored: Distraction and Diversion inside China’s Great Firewall.
Princeton: Princeton University Press.
Rock, Aaron. 2010. “Amr Khaled: From Da‘wa to Political and Religious Leadership.” British
Journal of Middle Eastern Studies 37, no. 1: 15–​37.
Sakr, Naomi. 2001. Satellite Realms: Transnational Television, Globalization and the Middle
East. London: I.B. Tauris.
Sakr, Rehab. 2014. “The Power of Online Networks: Citizenship among Muslim Brotherhood
Cyber Youth.” In Wired Citizenship, edited by Linda Herrera, 179–​191. London: Routledge.
Schradie, Jen. 2019. The Revolution That Wasn’t: How Digital Activism Favors Conservatives.
Cambridge, MA: Harvard University Press.
Shehabat, Ahmad. 2015. “Arab 2.0 Revolutions: Investigating Social Media Networks during
Waves of the Egyptian Political Uprisings that Occur between 2011, 2012 and 2013.” PhD
diss., University of Western Sydney.
Siegel, Alexandra A. 2020. “Online Hate Speech.” In Social Media and Democracy: The State of
the Field. Cambridge: Cambridge University Press.
Siegel, Alexandra A., and Joshua A Tucker. 2018. “The Islamic State’s Information Warfare.”
Journal of Language and Politics 17, no. 2: 258–​280.
Siegel, Alexandra, Joshua Tucker, Jonathan Nagler, and Richard Bonneau. 2020. “Tweeting be-
yond Tahrir: Ideological Diversity and Political Tolerance in Egyptian Twitter Networks.”
Forthcoming, World Politics.
Sivan, Emanuel. 1990. “The Islamic Resurgence: Civil Society Strikes Back.” Journal of
Contemporary History 25, no. 2: 353–​364.
Skovgaard-​Petersen, Jakob. 2011. “Islamic Fundamentalism in Arab Television: Islamism and
Salafism in Competition.” In Fundamentalism in the Modern World. Vol. 2, Fundamentalism
and Communication: Culture, Media and the Public Sphere, edited by Ulrika Mårtensson,
Jennifer Bailey, Priscilla Ringrose & Asbjørn Dyrendal: 264–​291. London: Tauris Academic
Studies.
Sreberny-​Mohammadi, Annabelle. 1990. “Small Media for a Big Revolution: Iran.” International
Journal of Politics, Culture, and Society 3, no. 3: 341–​371.
Srinivasan, Ramesh. 2014. “What Tahrir Square Has Done for Social Media: A 2012 Snapshot in
the Struggle for Political Power in Egypt.” Information Society 30, no. 1: 71–​80.
Tarrow, Sidney G. 2011. Power in Movement: Social Movements and Contentious Politics.
Cambridge: Cambridge University Press.
Tilly, Charles. 1981. “Britain Creates the Social Movement.” CRSO Working Paper 232. Ann
Arbor, MI: Center for Research on Social Organization.
Tucker, Joshua A., Yannis Theocharis, Margaret E. Roberts, and Pablo Barber´a. 2017. “From
Liberation to Turmoil: Social Media and Democracy.” Journal of Democracy 28, no. 4: 46–​59.
Tufekci, Zeynep, and Christopher Wilson. 2012. “Social Media and the Decision to Participate
in Political Protest: Observations from Tahrir Square.” Journal of Communication 62, no.
2: 363–​379.
Ullah, Haroon K. 2017. Digital World War: Islamists, Extremists, and the Fight for Cyber
Supremacy. New Haven, CT: Yale University Press.
van de Bildt, Joyce. 2015. “The Quest for Legitimacy in Postrevolutionary Egypt: Propaganda
and Controlling Narratives.” Journal of the Middle East and Africa 6, no. 3–​4: 253–​274.
Zhuravskaya, Ekaterina, Maria Petrova, and Ruben Enikolopov. 2020. “Political Effects of the
Internet and Social Media.” Annual Review of Economics 12: 415–​438.
 Chapter 28

Isl am ically Fra me d

Mobiliz ation i n T u ni sia
Ansar al-​Sharia in the Aftermath of the Arab Uprisings

Frédéric Volpi

This chapter addresses two main aspects of Islamically framed social mobilization, with
a particular focus on the protest dynamics in the aftermath of the 2011 Arab uprisings
in Tunisia. By “Islamically framed mobilization,” I mean social activism visibly shaped
by discourses and practices claimed to be Islamic by their proponents and recognized
as such by those with whom they interact (regardless of whether or not they endorse
the Islamic referent put forward). The first issue under scrutiny is the evolution of such
forms of mobilization over time, as the state apparatus controls, more or less strictly, the
political and religious field. This feature is particularly prominent in Tunisia after 2011
with the relaxation of authoritarian surveillance during the democratic transition. The
second concern is how contextual factors in times of institutional change enable and/​
or constrain specific articulations and diffusions of Islamically framed movements. It
sheds light on the processes of mobilization within the religious sphere in a situation of
open competition between actors claiming the Islamic higher ground.
Empirically, the chapter uses the protest episodes of the 2011 Arab uprisings to shed
some light on how different types of Islamically framed social mobilization interacted
to produce both change and continuity. More specifically, I focus on Tunisia as the site
of the only successful democratic transition in the region resulting from this wave of
uprisings. The rise of two distinct religious forms of mobilization characterized the
Tunisian transition. The Islamists of Ennahda articulated a rather conventional expres-
sion of religious conservatism in the social and political sphere through their associative
network and their political party. The Salafists of Ansar al-​Sharia mobilized grass-​roots
activists mainly in the social sphere without participating in institutional politics.
582   Frédéric Volpi

Whereas significant attention has been and is still paid to Ennahda as a “Muslim
democrat” movement, less consideration has been given to the significance of Ansar
al-​Sharia, especially after its demise in 2013 when it was classified as a terrorist orga-
nization by the Tunisian state. This chapter delves therefore into the dynamics of this
Salafi organization, and the factors that contributed to its rapid rise to fame after the
Tunisian uprising, and its equally swift decline after 2013. The argument is grounded on
interviews and participant observations conducted in Tunisia in 2014–​2015, particularly
with former sympathizers of the Ansar al-​Sharia movement.
In the Tunisian case, I evaluate the effectiveness of Islamically framed mobilization
in relation to the practices that bring together grass-​roots actors in processes asserting
specific religious solidarities and differentiations. I evaluate the success of mobilization
in relation to the strategies of organized movements proposing a specific social and po-
litical agenda for state and society. From this perspective, the case of Ansar al-​Sharia
indicates that the religious agenda used by the organization to justify and explain mo-
bilization was not always the main driver behind the strategic choices and behaviors
of those underprivileged youths who most enthusiastically joined the movement.
For these new grass-​roots activists, a Salafi discourse helped them to articulate a new
public role for themselves and gave them an opportunity for upward social mobility in
a social system where they used to have a marginal role. The meaning of this mobiliza-
tion is viewed in relation to the dominant political order present at any one time in the
country—​i.e., the pre-​2011 authoritarian system and the post-​2011 democratic regime.
While the insights gained from studying mobilization during this particular episode
of institutional transformations cannot be directly generalized to other sequences of
institutional change or to longer periods of authoritarian or democratic stability in
the region, I argue that they underscore the malleability of Islamically framed mobi-
lization. Whereas established Islamist movements such as the Muslim Brotherhood
have often been shown to be constrained by quite rigid organizational and ideological
structures (al-​Anani 2015), the Arab uprisings illustrated that Islamically framed mo-
bilization more generally was more influenced by context (Zollner 2019). The evolu-
tion of the Tunisian mobilizations outlined here highlights some of the most relevant
contextual factors that intervene in these situations. In particular, it illustrates in the
case of the Salafi activism linked to Ansar al-​Sharia that the movement’s rise and fall
depended on creative appropriation of religious discourses and practices by would-​be
revolutionary actors.

Studying Social Mobilization

in Tunisia and the Middle East

This section outlines how the explanations of social mobilization in the Middle
East reflected the main debates and perspectives in social movement theory. These
 Islamically Framed Mobilization in Tunisia     583

approaches shed light on the evolution of social movements in the region—​particu-

larly Islamist mobilization—​and contribute to debunk the notion of an exceptional
social context in that part of the world.
The study of mobilization in the Middle East and North Africa only became an area of
serious interest for social movement theory in recent time. Until the 1990s, area studies
specialists duly investigated protest mobilization and religious movements, but their
approaches were not generally informed by the debates about social movements within
established Western democracies. Indirectly, this mutual ignorance contributed to the
perception that mobilization in the Muslim world, particularly by Islamically framed
movements, could be somewhat distinct from the rest. Whereas social movements in
Western democracies were generally seen as a positive feature of civil society, a new
wave of movements coming from post-​Communist countries and the developing world
appeared to be more ambiguously so. In the early 1990s, Tarrow worriedly noted the rise
of “ugly movements” rooted, among other things, in “religious fanaticism” (1994, 14).
By the early 2000s, Middle East specialists engaging directly with the mainstream so-
cial movement literature began to produce detailed investigations of Islamically framed
mobilization in Muslim-​majority countries. Their findings were generally supportive
of the notion that such mobilization was not that different from what social movement
theory had analyzed elsewhere. In-​depth studies of Muslim Brotherhood activism in
Egypt (Wickham 2002), in Egypt, Jordan, and Yemen (Clark 2004), and in Jordan and
Yemen (Schwedler 2006) illustrated the “normality” of these forms of social mobili-
zation, which were understandable in terms of “political opportunity structure,” “re-
sources mobilization,” and “framing.”
In Tunisia, the growing public appeal and political visibility of the Muslim
Brotherhood–​ inspired Movement of Islamic Tendency—​ subsequently renamed
Ennahda—​in the 1980s fit well within this general framework. In ideational terms, as the
nationalist discourse of the regime was losing its appeal, the Islamic frames proposed by
a newer generation of ideologues and activists such as Rached Ghannouchi were gaining
a wider audience. In addition, the neoliberal economic model adopted by the regime
indirectly contributed to strengthening socioeconomic networks of religious solidarity
and gave the Islamists more resources. Finally, an aging President Habib Bourguiba was
beginning to make political mistakes that not only emboldened the Islamist opposition
but also increased intra-​regime tensions. Ultimately, it was in the context of a wave of
unrest launched by the Movement of Islamic Tendency in 1987 that Ben Ali, then prime
minister, was able to stage a palace coup to replace Bourguiba (Anderson 1991).
In the late 1980s and early 1990s, several countries in the region experienced episodes
of social mobilization that facilitated a reconfiguration of authoritarian regimes and po-
litical openings that bolstered social movements (Alexander 2000). Tunisia appeared to
be transformed by the same dynamics that were visible across Eastern Europe and Latin
America, as well as in neighboring Algeria, where Islamist activism was also rapidly
growing (Volpi 2003). Yet the window of political opportunity for an Islamically framed
mobilization to reform the institutional system was short-​lived in Tunisia. It soon
generated an authoritarian backlash that produced a long period of state repression
584   Frédéric Volpi

of social movements, particularly but not exclusively of the Islamic kind. After briefly
having been authorized to participate (as independents) in parliamentary elections in
1989, the members and sympathizers of Ennahda began to be systematically repressed
in the 1990s as the Ben Ali regime closed all avenues of social and political contestation
(Camau and Geisser 2003).
By the 2000s, Islamically framed mobilization was well studied across the Muslim
world using the conventional tools of social movement theory (Wiktorowicz 2003).
Islamically framed movements, and by extension Muslim societies, were not seen to
be exceptionally prone to having “ugly” movements, or to be split between progressive
movements (right-​based, unions, environmental, etc.) and regressive religious ones.
Rather, the Islamic dimension of social mobilization was but one aspect of these polities
that produced movements which were behaving in relatively predictable ways. Islamic
social mobilization, like any other culturally or identity-​based mobilization was not
particularly challenging for the theory—​nor, it appears, for the authoritarian regimes
in place in most of these countries. If there was exceptionalism in the Middle East and
North Africa at the time, it was not about Islamically framed movements per se, but
more about the repeated inability of otherwise effective forms of social mobilization to
influence significantly the political system. Bayat’s (2007, 2010) post-​Islamist approach
and insistence on the role of nonmovements constituted one attempt at resolving the di-
lemma. In his analysis, the harsh authoritarian regimes prevailing in the region ensured
the structural failure of organized social movements and placed the onus for social
change on informal grass-​roots networks and practices without explicit structures and
ideologies.
For most of that decade, Tunisia was a vivid regional illustration of a well-​run police
state, where organized social movements hardly achieved any visible success against the
regime (Camau and Geisser 2003). The Ben Ali regime tightly controlled the social and
political sphere, and organized expressions of public dissent were rare. Protest mobili-
zation was limited due to the self-​limiting strategies adopted by structured movements,
such as the still influential Tunisian General Labour Union (UGTT) (Beinin 2016),
while informal networks had only a limited influence on the system due to pervasive
state control and repression (Chomiak 2011). Islamically framed forms of mobiliza-
tion were particularly scrutinized and targeted by the security apparatus. Unlike other
movements in the region that had opted for political violence as a strategy of resistance,
Ennahda sought to remain a nonviolent social and political movement. In this repres-
sive political context, Ennahda survived as an opposition network in exile but hardly
had any capability to mobilize its supporters in the country, as its infrastructure had
been dismantled by the regime (Wolf 2017). Its main advantage over other political
parties when it came back to Tunisia after the revolution was its ability to present itself as
a “clean” political actor that was not involved in the Ben Ali system.
By the time of the Arab uprisings, social mobilization in the region was increasingly
well studied and well understood, using the mainstream theoretical and analytical tools
of social movement theory. While Islamist mobilization was not viewed as a particularly
exceptional phenomenon, the harsh authoritarian context in the region was judged to
 Islamically Framed Mobilization in Tunisia     585

be particularly constraining for social movements, and the case of Tunisia illustrated
this situation quite well.

Mobilization and the Arab

Uprisings in Tunisia

This section outlines the social and political processes that were salient at the time of
the 2011 uprisings in Tunisia. It presents the general trends that unfolded in the country
and notes the particular challenges encountered by Islamist actors who had long been
prevented by the regime from developing grass-​roots activities among the population.
The 2011 Arab uprisings in Tunisia and more generally across the Arab world illus-
trated two main features of social mobilization in the region. First, they showed that
sizeable protest movements could organize themselves outside structured social
movements. Second, they quickly revealed that established organizations were able to
use the situation to reposition themselves at the heart of the process of social and po-
litical change. This dual process is clearly noticeable in the case of Islamically framed
mobilizations, as both old and new movements were able to take advantage, at least
temporarily, of the political transformations that occurred in the countries of the Arab
uprisings. Although in Tunisia, like elsewhere in the region, Islamist activists were not at
the forefront of the protests when mobilization started, they were able to mobilize effec-
tively once the process had gained momentum. In 2011 Islamically framed mobilization
was generally more significant in the aftermath of the uprisings, mainly due to the pre-
existing configuration of repressive measures by the state.
In Tunisia, during the uprising itself, the most significant mobilizations occurred pri-
marily outside organized social movements. Even when the latter mobilized, as in the
case of the UGTT, they were following the dynamics of the popular protests rather than
directing them (Hmed 2012). The importance of the leaderless/​leaderful movements
progressively decreased throughout 2011 as the institutionalization of a democratic po-
litical system facilitated the repositioning of structured organizations as intermediaries
between the people and the state (Volpi 2017). The most visible success of such an or-
ganization during this transitional period was the rebirth of the Islamist movement
Ennahda. Whereas at the start of the uprisings they were only present on the ground in
Tunisia via small underground networks of sympathizers, they became, a few months
later, the largest political party in terms of votes in the country.
Three processes were particularly important for the revival of the social movements
previously repressed by the regime. First, there was the return from exile or liberation
from jail of leaders and cadres of the movement. Second, there was the reactivation of
the grass-​roots networks that had fallen into abeyance due to the unfavorable political
environment. Third, there was an opening of the movement to attract new members
and sympathizers, particularly when the strategy was to participate in electoral contests
586   Frédéric Volpi

(McCarthy 2018). Along these three dimensions, Islamically framed mobilization was
not particular different from what took place with other types of opposition movements
in the country at the time.
While the 2011 Arab uprisings illustrated that conventional social movements could
still be very relevant and effective, they also indicated that these episodes of institutional
change were not simply an opportunity for them to implement a preexisting agenda.
Not only was an established Islamic movement like Ennahda not the main actor driving
institutional changes, but it also had to adapt itself to new demands and circumstances.
Mobilization processes under a predictably harsh authoritarian system are different
from those operating during an upheaval, an episode of regime change, or a sequence
of reinstitutionalization due to changing environmental dynamics (Volpi and Clark
2019). In particular, Ennahda, like other established Muslim Brotherhood–​inspired
movements across the region, was confronted with a reorganization of the religious field
and new competition from both religious and secularized social movements (Clarke
2014, Lacroix and Shalata 2016). While these processes produced some generic effects,
the local context and the specific strategic choices made by these actors in Tunisia gave
some effects more prominence than others (notably the pragmatism of Ennahda and the
lack of institutionalization of Ansar al-​Sharia).
In this context, Mische’s argument that to understand mobilization “we should shift
our attention away from cultural forms such as ‘identities’ or ‘frames,’ toward the study
of how these forms are shaped, deployed, and reformulated in conversation, as this
unfolds across social movement forums over the course of movement development”
(2003, 259–​260) is particularly pertinent. In the Tunisian context, the reconfiguration of
movements’ identity, practices, and strategies was partly a consequence of the loosening
of earlier practices of authoritarian control, but it was also, in part, the result of new
interactions between the individuals and groups in a revolutionary and then transi-
tional situation. The first dimension of mobilization around Islamic themes and actors
reflected the longer-​term structural dynamics that had shaped the conditions of public
and private discourse. The second dimension, by contrast, had more to do with the con-
textual, emotional, and interactive processes of reframing of identities and exploring
new possibilities in situations where normative and political authorities were weak and
changing (Jasper and Volpi 2018).
The reemergence, repositioning, and creation of Islamically framed social movements
in Tunisia in the aftermath of the 2011 uprising was characterized by increased compe-
tition between religious movements, as well as, in some case, between these movements
and the state (Donker 2019). At that time, a main religious competitor for Ennahda
emerged in the shape of a new Salafi organization, Ansar al-​Sharia. While the nucleus of
Ansar al-​Sharia preexisted the 2011 uprisings, it was the wave of protests and the process
of regime change that made this movement powerful during the democratic transition
(Merone 2017). As was the case with other established movements, the release from jail
of cadres and ideologues and the reactivation of those parts of the network that were in
abeyance was important to revive the organization. However, it was the recruitment and
mobilization of new members and sympathizers that constituted Ansar al-​Sharia as a
 Islamically Framed Mobilization in Tunisia     587

distinctly new and successful organization at that time. At that juncture, protest mobi-
lization shaped Ansar al-​Sharia, which in turn shaped (religious) protest mobilization.
This Salafi organization helped bring together an often-​younger crowd of people who
became religiously and politically active during the revolution and who endorsed the
ultra-​conservative religious discourses coming primarily from the Gulf. It facilitated the
coordination of their actions and the spread of Salafi discourses and practices in the
public space in a way that had not been experienced in the country before.
The empirical section of the argument that follows is based on fieldwork conducted in
Tunisia during the democratic transition in 2014 and 2015 among Islamic activists linked
to Ennahda and, for the largest part, linked to Ansar al-​Sharia. As the Tunisian govern-
ment classified Ansa al-​Sharia a terrorist organization in 2013, grass-​roots activists from
the movement subsequently took their distance from the movement to avoid police
repression. Hence, by 2014–​2015, most of the activists identified themselves as former
sympathizers of the movement, and most of them (9 out of 13) declared that they no
longer engaged in any religious activism. These thirteen semi-​structured interviews
were conducted over a two-​month period in 2014 and 2015 in the governorates of Tunis
and Medenine among young (aged 19–​33) males approached through snowball sam-
pling. Each lasted about an hour and took place in a cafe or at the interviewee’s home.
The argument highlights the frames and framing processes that were common among
these activists as they tried to make sense of the successes and failures of the Salafi
movement.
This section outlined the options and choices for mobilization that Islamically
minded actors faced when the Arab uprisings offered them new opportunities for ac-
tion. It introduced the different roles played by the Ennahda movement and by emerging
Salafi actors.

Salafi Mobilization and

Demobilization in Tunisia

This section, focusing on the “rise and fall” of Ansar al-​Sharia, illustrates that the shape
of religious mobilization was more a consequence of the process of political transi-
tion than the outcome of an opposition between preexisting religious approaches. It
highlights how the Salafi message became popular because it justified continuing mobi-
lization against the status quo by people who did not benefit much from the consolida-
tion of a new democratic order.
The Tunisian uprising reflected, among other things, a desire by the population to
take back control of those levers of state power that had been captured by a predatory
ruling elite for their own benefit. In the religious field in particular, the mosques were
controlled by state-​appointed imams who tended to follow strictly the religious and
political orientations given by the ministry of religious affairs of a secularized regime
588   Frédéric Volpi

(Feuer 2017). The 2011 uprising thus provided an opportunity for the faithful to evict
these religious officials and to nominate people who had, in their view, greater religious
credentials. This process was itself fractious, as the consensus that emerged over the re-
moval of state-​appointed imams did not usually extend to the choice of their successors
(K.D. interview, 2014).
In particular, there was a generational tension between an older generation of faithful
with a traditional, conservative interpretation of Islam and a younger generation more
attuned to the Salafi views and militant discourses coming from the Gulf region in par-
ticular (A.O. interview, 2014). These diverging opinions regarding religious leadership
also reflected some important differences in the approach to the Tunisian “revolution.”
For the older generation, the fall of the authoritarian system meant primarily getting
rid of the state surveillance and micromanagement of their religious practices. For the
youth, it was often much more than this: it was an opportunity to redefine more broadly
social and political relations in a way that empowered them (K.A. interview, 2015).
Although this divide was probably overemphasized by the activists, it helps explain
how a different diagnostic framing of the religious problem underpinned the rapid
emergence of a Salafi movement in Tunisia. It was overemphasized by the (former)
activists because they sought to downplay the religious commitments of an older gener-
ation of mosque-​goers who had accommodated themselves to the instrumentalization
of religious discourse by the state, thereby presenting themselves as more credible re-
ligious actors despite their youth and inexperience. The slightly older generation of
conservative mosque-​goers was already well established in life and had little inter-
ested in a dramatic transformation of the relationship between the state and religious
affairs beyond the removal of security personnel and imams who slavishly followed
the instructions of the ministry of religious affairs. The younger generation of Salafi-​
inspired activists saw the revolution and the change of religious personnel as a means
to mobilize human and normative resources to reorganize not only religious activities,
but also the (local) social order. They could place new (younger) people in positions of
authority in the mosques—​people who could then bring to the attention of a wider au-
dience the theological views and practices favored by the Salafis. In addition, controlling
or taking a leading role in the mosques gave the Salafi activists greater social recognition
and authority, which they could use to improve their own prospects for social mobility
(from marriage to employment).
In terms of political opportunity structure, religious mobilization at the time of the
Tunisian transition was thus more than a case of Islamic activism facilitated by the
collapse of an authoritarian system that previously excluded religious actors from the
public sphere. It was also, at least for a part of the population, a means of organizing
and legitimizing the local power that they had gained and exercised during the uprising.
From this perspective, the Salafi discourse proved useful to challenge the prevailing
social and religious hierarchies (Marks 2013). It provided ideational and normative re-
sources to a new generation of activists to challenge both religious and secularized ac-
tors (including the state). By invoking an “authentic” Islamic legitimacy based on their
own interpretation of the scriptures, they would set out to change the social practices
 Islamically Framed Mobilization in Tunisia     589

and codes that were commonplace under the previous regime—​e.g., sale of alcohol,
wearing of skirts, etc.—​and delegitimize the actors that continued to validate these
practices. Importantly, though, as will be made clear below, the Salafis’ use of the theo-
logical card to justify their action and obtain approval did not mean that they all agreed
on the same interpretation (or its practical implications). This process is not a case of a
preexisting ideology suddenly going “viral.” Recalling Mische’s (2003) argument about
the evolution of networks and frames, what took place among activists at that time was
a creative reinterpretation of existing ideological resources in order make them relevant
to their situation, rather than an attempt at implementing these prescriptions faithfully.
Eventually, most established Islamically framed movements in Tunisia ended up
supporting a return to the constitutional order, especially after the parliamentary
elections of October 2011. Despite some internal tensions, most movements did not seek
to challenge the established institutional system, but rather to advance their strategic
interests within the system (Donker 2013). In particular, Ennahda signaled very clearly
to its supporters and its opponents alike from the very beginning of the democratic
transition that it was not looking to introduce any dramatic “Islamic” transformation of
the political system. It presented itself as a legal political organization and a responsible
party of government that was able to build a dialogue and working coalitions with the
secularized political forces of the country. Ansar al-​Sharia, by contrast, discursively and
materially challenged the established order by calling for a continuation and deepening
of the Tunisian revolution—​a process that should be leading, in their view, to the estab-
lishment of a genuine Islamic order (Merone 2015). In this context, the discourse of the
organization was taken up by local activists and used to justify their strategies directed
against religious and nonreligious competitors (including representatives of the state) in
the public sphere.
The different types of Islamically framed mobilizations in Tunisia during the tran-
sitional period were the consequence of two related but distinct dynamics that reflect
both the generic political processes unfolding during the Arab uprisings and the local
specificities of the Tunisian context. The first one, which was most explicitly debated
in the literature about the Tunisian transition, is the process of redefinition of state-​
sanctioned secularism in the country. There was an overt political competition between
Islamist and secularist parties regarding the possibility of changing the existing insti-
tutional order to provide more formal recognition for religion and religious activities
(Boubekeur 2016). This confrontation was particularly intense before the adoption of
the new Tunisian Constitution in January 2014. At the time, Ansar al-​Sharia generated
a negative radical flank effect that facilitated the coming together of secularist forces
against Ennahda—​a situation that forced the Islamist party to call for early elections
in 2014. At that juncture, not only did Ennahda and Ansar al-​Sharia fail to build any
common strategy on religious and political reform, but the two movements also in-
creasingly blamed one another for the difficulties that they encountered. Unlike in
other countries of the region that had experienced political liberalization after the Arab
uprisings (notably Egypt), the main Salafi movement in Tunisia did not choose to pro-
duce a political party to advance its agenda in the formal political sphere. This lack of
590   Frédéric Volpi

inclusion in the institutional system made the interactions between Ansar al-​Sharia
and Ennahda (and the other political parties) less predictable and calibrated. This op-
position was eventually made official in the summer of 2013 when the Tunisian inte-
rior minister, a leading Ennahda cadre, formally classified Ansar al-​Sharia a terrorist
organization.
The second set of processes is less related to the religious underpinnings of mobili-
zation and more a consequence of the social opportunities for mobilization during the
transitional period. The discourse and practices promoted by Ansar al-​Sharia allowed
many young activists to reclaim the public sphere in the name of Islam (A.B. interview,
2014). At the local level in particular, they were able to carve out for themselves a space
of freedom and authority where none was available to them under the old regime. Yet
as structured social and political actors slowly began to reestablish themselves and en-
force the new rules and boundaries of a democratic state, the underprivileged youth in
particular, especially those with a strongly marked religious identity, saw their avenues
of social recognition increasingly disappear (T.I. interview, 2015). For the members of
Ennahda, on the other hand, formal inclusion in the political system enabled the party
(like other actors in government) to have some degree of leverage (and patronage) in
the state administration to help their supporters. The challenge to the system posed by
Ansar al-​Sharia may have been religious, but it was also one of the only visibly effective
models of grass-​roots activism available to these protest actors. Thus, religious mobili-
zation was principally a way to challenge the redistribution of authority and resources
produced by the new democratically elected elites by those who had gained the least
from the change of regime (Merone 2015).
These tensions over who had authority to regulate local (religious) practices were not
always directly the result of a strategy by Ansar al-​Sharia to systematically push back
the state (and other civil society actors). Instead, they were produced by local actors
presenting themselves as members or sympathizers of the movement. In practice, the
Ansar al-​Sharia leadership had limited oversight over many Salafi activists (Volpi 2018).
Not only was the movement’s leadership not able to produce a unified strategy for the
movement, but they were also not able to impose a strict discipline within the organi-
zation. The latter was a consequence of the rapid growth of the movement in 2011–​2012
as an umbrella organization for religious-​minded youth activists dissatisfied with the
gradualist approach adopted by other movements (especially Ennahda). It is noticeable
that a reformist movement like Ennahda also lost some of its support, as its performance
in government was hardly impressive. However, by retaining a sizeable number of rep-
resentatives in the parliament even as it went into opposition in 2014, the party was able
to maintain discipline within its rank and buttress its reformist agenda.
The influence that the movement acquired was at the cost of a relative autonomy of
action granted to the new groups and networks expressing their support for Ansar al-​
Sharia’s agenda. In this context, trying to impose a stricter discipline within the orga-
nization ran the risk of having these groups dissociate themselves from the movement
in order to pursue their local agenda and activities, thereby diminishing the reach of
the movement. Repeatedly, therefore, the leadership of the organization had to make
 Islamically Framed Mobilization in Tunisia     591

public statements to the effect that activists misunderstood or misapplied the religious
precepts advocated by Ansar al-​Sharia. However, grass-​roots sympathizers repeatedly
ignored these calls whenever they thought they knew best. As indicated by one of those
interviewed (K.A.), although the organization’s leaders told them not to use violence
against people selling alcohol, some of the local activists assaulted the salesmen anyway
to ensure an alcohol-​free neighborhood, because they judged it was their duty.
The most spectacular failure of Ansar al-​Sharia to control its supporters came during
the 2012 demonstration in front of the US Consulate that turned into a partial ransacking
and torching of the building complex. This episode forced one of the main leaders of the
movement, and a key organizer of the demonstration, to go into hiding, as the govern-
ment attributed the responsibility of the attack to him (Wolf 2017). The US Consulate
event marked the beginning of an acceleration of the process of fragmentation of mobi-
lization in the religious field, as Ansar al-​Sharia confronted the hubris it helped to create.
This point was dramatically illustrated in January 2013 by the assassination of a leftist
parliamentarian by members of a Salafi network linked to Ansar al-​Sharia who decided
that armed struggle was now the most appropriate strategy for them. Although some
of the leading members of Ansar al-​Sharia publicly denounced the action of the armed
network, this event showed the limitations of the movement’s strategy of retaining a
loose umbrella organization for all Salafi-​inspired activists in the country.
The last annual conference of Ansar al-​Sharia, in May 2013, witnessed intense last-​
minute debates among the leaders of the various tendencies of the movement to define
a clear strategy and agenda for the organization. This focused, in particular, on whether
the movement should accept the legal and institutional framework of the Tunisian state
and become a legally recognized political party or civil society organization (Merone
2017). Their failure to reach an agreement revealed the inability of the movement to
adapt to a situation of consolidation of the democratic institutional order, as well as the
exhaustion of revolutionary fervor. The model of grass-​roots activism that did make
sense in 2011, and that retained some relevance for many in 2012 (as illustrated by the
continuing growth of Ansar al-​Sharia), increasingly became less relevant with time. In
the early transitional period, the main role that Ansar al-​Sharia played in the process
of Islamically framed mobilization was not due to an effective recruitment and organ-
izational strategy, but rather the consequence of the loose control over ideology and
practices that the leadership exercised (Volpi 2018).
This mobilization around Salafi themes was noticeable but not so widespread that it
could create the kind of popular groundswell that could have overturned the newly es-
tablished democratic institutions of the country. As the revolutionary fervor diminished
over time, ordinary citizens looked for better state institutions and policies, rather than
for a complete overhaul of the system. In this situation, Ennahda too had lost some of
its appeal, but because of its institutionalization it offered its members and voters access
to the state and an opportunity to influence policymaking that the Salafi organization
could not provide. The inability of the movement to adapt to changing patterns of mo-
bilization (i.e., reduced activism), and the continuing nuisance created by the armed
jihadi (who assassinated another parliamentarian in June 2013), facilitated the choice of
592   Frédéric Volpi

the Ennahda-​led Tunisian government to classify the entire movement a terrorist orga-
nization. The security clampdown that ensued directly contributed to the implosion of
Ansar al-​Sharia. Some of the activists went underground and joined jihadi networks at
home or abroad in order to continue their struggle. Others cut their links with the move-
ment and reorganized themselves as legal Islamic associations approved by the Tunisian
state (Volpi 2018). Many simply stopped mobilizing in the face of increased police re-
pression and downgraded or stopped their religious activism (B.N. interview, 2015).
After a heightened period of confrontation between Salafi actors, state actors, and
secularized civil society actors in 2013, the patterns of Islamically framed mobilization
in Tunisia began to follow a more routine and predictable course. Politically, Ennahda
remained the main engine of electoral mobilization via its promotion of a pragmatic
Islamic vision that balanced elements of traditionalism and modernism in order to at-
tract a wider audience. In the social sphere, religious competitors to Ennahda increas-
ingly adopted a legalist, reformist approach to ensure that their activities and their
members were not targeted by police repression (Donker 2019). The occasional outburst
of jihadi violence in the country, and the ensuing repressive backlash, ensured that a
distinct Salafi-​oriented, protest-​focused, grass-​roots activism was increasingly less and
less able to attract followers in large numbers. After two to three years, when a distinc-
tive form of religious mobilization raised the possibility of the social and political in-
stitutionalization of an Islamic model of governance, societal dynamics eventually
buttressed routine reformist orientations in the religious field.
This section illustrated that the “direct action” appeal of the Salafi movement
corresponded to the aspirations of a particular constituency at a specific moment of
the country’s political and institutional transformation. However, it pointed out that
due to the different strategic choices and objectives of its members and sympathizers,
the movement was not able to institutionalize this mobilization and attract other
constituencies.

Conclusion

The current outlook and trajectory of social movements in Tunisia, be they framed
around Islamic themes or not, increasingly correspond to trends observed in
consolidating democracies. Structured organizations with clear hierarchies and formal
strategies and discourses mobilize specific constituencies and/​or specific themes in
order to influence national or local policymakers. Islamically framed movements dis-
play, like many other nonreligious movements, a tendency toward specialization and
bureaucratization. The process of democratic consolidation provides a legal and po-
litical structure that underpins predictable interactions between the state and social
movements, between the movements themselves, and between the movements and the
population. Even though new, commonly issue-​based movements still occasionally
 Islamically Framed Mobilization in Tunisia     593

erupt on the scene—​e.g., “winou el petrol” (where is the oil)—​the identity and repertoire
of established social movements now evolve quite predictably.
Religious mobilization continues to play a significant role in Tunisia. However, reli-
gious and secularized movements alike converge in developing a reformist agenda. Only
during a brief episode of institutional instability and transition was Islamically framed
mobilization significantly different in form and content from both prior religious mo-
bilization and from nonreligious mobilization. During this period, the “revolutionary”
character of a brand of Salafi activism that challenged established religious views and
practices, as well as the secularized institutions of the state, illustrated itself. This ac-
tivism proved quite successful at a time when a significant number of individuals, par-
ticularly among the less well-​off youth, who had successfully mobilized to overthrow the
authoritarian regime, sought out new avenues for social advancement. At that juncture,
Salafi mobilization competed effectively with more established religious and secularized
movements.
Promoting an Islamic agenda was unmistakably an underlying factor of success for
Salafi and other Islamically framed movements after years of authoritarian control of
the Tunisian religious sphere. Both Ansar al-​Sharia and Ennahda initially benefited
from this revival in religious activism, albeit through different channels of mobilization.
Initially, Salafi claims grounded on religious authority served to strengthen the legiti-
macy of various forms of “revolutionary” activism more effectively than the reformist
discourse of Ennahda and its strategy of political institutionalization. Over time, how-
ever, the lack of internal coherence of the Ansar al-​Sharia project and organization,
coupled to a lessening of revolutionary fervor and the return of routine state adminis-
trative functions, made sure that this particular framing of Islamic identity did not gain
ascendency in the religious or political sphere.

References
Alexander, Christopher. 2000. “Opportunities, Organizations, and Ideas: Islamists and
Workers in Tunisia and Algeria.” International Journal of Middle East Studies 32, no.
4: 465–​490.
al-​Anani, Khalil. 2015. “Upended Path: The Rise and Fall of Egypt’s Muslim Brotherhood.”
Middle East Journal 69, no. 4: 527–​543.
Anderson, Lisa. 1991. “Political Pacts, Liberalism, and Democracy: The Tunisian National Pact
of 1988.” Government and Opposition 26, no. 2: 244–​260.
Bayat, Asef. 2007. Making Islam Democratic: Social Movements and the Post-​Islamist Turn.
Stanford, CA: Stanford University Press.
Bayat, Asef. 2010. Life as Politics: How Ordinary People Change the Middle East. Stanford,
CA: Stanford University Press.
Beinin, Joel. 2016. Workers and Thieves: Labor Movements and Popular Uprisings in Tunisia and
Egypt. Stanford, CA: Stanford University Press.
Boubekeur, Amel. 2016. “Islamists, Secularists and Old Regime Elites in Tunisia: Bargained
Competition.” Mediterranean Politics 21, no. 1: 107–​127.
594   Frédéric Volpi

Camau, Michel, and Vincent Geisser. 2003. Le Syndrome Autoritaire: Politique en Tunisie de
Bourguiba à Ben Ali. Paris: Presse de Sciences Po, 2003.
Chomiak, Larissa. 2011. “The Making of a Revolution in Tunisia.” Middle East Law and
Governance 3, no. 1–​2: 68–​83.
Clark, Janine A. 2004. Islam, Social Welfare, and the Middle Class: Networks, Activism, and
Charity in Egypt, Jordan, and Yemen. Bloomington: Indiana University Press.
Clarke, Killian. 2014. “Unexpected Brokers of Mobilization: Contingency and Networks in the
2011 Egyptian Uprising.” Comparative Politics 46, no. 4: 379–​397.
Donker Teije H. 2013. “Re-​emerging Islamism in Tunisia: Repositioning Religion in Politics
and Society.” Mediterranean Politics 18, no. 2: 207–​224.
Donker, Teije H. 2019. “The Sacred as Secular: State Control and Mosques Neutrality in Post-​
revolutionary Tunisia.” Politics and Religion 12, no. 3: 501–​523.
Feuer, Sarah J. 2017. Regulating Islam: Religion and the State in Contemporary Morocco and
Tunisia. Cambridge: Cambridge University Press.
Hmed, Choukri. 2012. “Réseaux dormants, contingence et structures: Genèses de la révolution
tunisienne.” Revue française de science politique 62, no. 5: 797–​820.
Jasper, James M., and Frédéric Volpi. 2018. “Introduction: Rethinking Mobilization after
the Arab Uprisings.” In Microfoundations of the Arab Uprisings: Mapping Interactions
between Regimes and Protesters, edited by Frédéric Volpi and James M. Jasper, 11–​39.
Amsterdam: Amsterdam University Press.
Lacroix, Stéphane, and Ahmed Z. Shalata. 2016. “The Rise of Revolutionary Salafism in Post-​
Mubarak Egypt.” In Egypt’s Revolutions: Politics, Religion, and Social Movements, edited by
Bernard Rougier and Stéphane Lacroix, 163–​178. New York: Palgrave Macmillan
Marks, Monica. 2013. “Youth Politics and Tunisian Salafism: Understanding the Jihadi
Current.” Mediterranean Politics 18, no. 1: 104–​111.
McCarthy, Rory. 2018. Inside Tunisia’s al-​Nahda: Between Politics and Preaching. Cambridge:
Cambridge University Press.
Merone, Fabio. 2015. “Enduring Class Struggle in Tunisia: The Fight for Identity beyond
Political Islam.” British Journal of Middle Eastern Studies 42, no. 1: 74–​87.
Merone, Fabio. 2017. “Between Social Contention and Takfirism: The Evolution of the Salafi-​
Jihadi Movement in Tunisia.” Mediterranean Politics 22, no. 1: 71–​90.
Mische, Ann. 2003. “Cross-​Talk in Movements: Reconceiving the Culture-​Network Link.” In
Social Movements and Networks: Relational Approaches to Collective Action, edited by Mario
Diani and Doug McAdam, 258–​280. New York: Oxford University Press.
Schwedler, Jillian. 2006. Faith in Moderation: Islamist Parties in Jordan and Yemen. Cambridge:
Cambridge University Press.
Tarrow, Sidney G. 1994. Power in Movement: Social Movements and Contentious Politics.
Cambridge: Cambridge University Press.
Volpi, Frédéric. 2003. Islam and Democracy: The Failure of Dialogue in Algeria. London: Pluto
Press.
Volpi, Frédéric. 2017. Revolution and Authoritarianism in North Africa. New York: Oxford
University Press.
Volpi, Frédéric. 2018. “Shaping Contention as a Salafi Movement: The Rise and Fall of Ansar al-​
Sharia in Post-​Revolutionary Tunisia.” In Microfoundations of the Arab Uprisings: Mapping
Interactions between Regimes and Protesters, edited by Frédéric Volpi and James M. Jasper,
89–​111. Amsterdam: Amsterdam University Press.
 Islamically Framed Mobilization in Tunisia     595

Volpi, Frédéric, and Janine A. Clark. 2019. “Activism in the Middle East and North Africa in
Times of Upheaval: Social Networks’ Actions and Interactions.” Social Movement Studies 18,
no. 1: 1–​16.
Wickham, Carrie Rosefsky. 2002. Mobilizing Islam: Religion, Activism and Political Change in
Egypt. New York: Columbia University Press
Wiktorowicz, Quintan, ed. 2003. Islamic Activism: A Social Movement Theory Approach.
Bloomington: Indiana University Press.
Wolf, Anne. 2017. Political Islam in Tunisia: The History of Ennahda. New York: Oxford
University Press.
Zollner, Barbara H. E. 2019. “The Metamorphosis of Social Movements into Political
Parties: The Egyptian Muslim Brotherhood and the Tunisian al-​Nahda as Cases for a
Reflection on Party Institutionalisation Theory.” British Journal of Middle Eastern Studies
(Online First, May 22). doi: 10.1080/​13530194.2019.1609413.

Interviews
A.B, Medenine governorate, October 2014.
A.O., Medenine governorate, October 2014.
B.N., Tunis Governorate, October 2015.
K.A., Tunis Governorate, October 2015.
K.D., Medenine governorate, October 2014.
T.I., Tunis Governorate, October 2015.
 Chapter 29

Isl amist Mobi l i z at i on

during th e A ra b
U prising s

Chantal Berman

The study of political Islam and the 2011 revolutionary wave in the Middle East presents
a conundrum. Islamist parties were arguably the Arab Spring’s largest immediate
beneficiaries. In Tunisia and Egypt, where incumbent regimes collapsed in the face of
protests, Islamist parties went on to dominate their countries’ first democratic elections,
winning pluralities in 2011 in Tunisia and 2012 in Egypt. In Morocco, where the Alaouite
monarchy rode out the February 20th wave largely intact, and where members of the
Islamist Parti de la Justice et du Dévéloppement (PJD) were already a major force in par-
liament, early elections in October 2011 nevertheless resulted in the first PJD plurality
and the first Islamist prime minister. At least until the cataclysmic roll-​back of Egypt’s
democratic transition in 2013, it became common for scholars and pundits to remark
that Islamists had “won” the Arab Spring.
Yet principle Islamist parties in each case responded quite differently to the escalating
protest movements that swept across North Africa in early 2011. While the Egyptian
Muslim Brotherhood eventually threw its weight behind the anti-​Mubarak protest
movement, the Moroccan PJD urged its members not to protest, and leaders of Tunisia’s
clandestine Ennahda movement largely watched the revolution play out from the
sidelines. Outside of Egypt, scholarly accounting has credited labor unions, youth or-
ganizations, and social media networks for the 2011 uprisings to a far greater extent than
Islam or Islamists. Further, survey research has shown that religious opposition to sec-
ular hegemony ranked quite low in self-​reported reasons for joining the 2011 protests.
As shown in Figure 29.1, motivations of social justice and economic opportunity often
dominate these polls, even among self-​identified Islamists.
In a context where Islam and Islamist movements have been long understood to
constitute “the opposition” to secular regimes, their ambivalent role in the largest
598   Chantal Berman

Top reasons for 2011 protests in Egypt (Arab Barometer 2011)

300

200

100

0
demands combating demands demands replacing objecting objecting
for corruption for for mubarak to to
improving authority civil with pro-israel pro-western
the not and an egyptian egyptian
economic to political islamic policy policy
situation pass freedoms regime
to
gamal
mubarak
Top reasons for 2011 protests in Tunisia (Arab Barometer 2011)

600

400

200

0
demands combating demands replacing objecting
for corruption for mubarak to
improving civil with pro-western
the and an egyptian
economic political islamic policy
situation freedoms regime

Strongly support shari’a Others

Figure 29.1: Reasons for anti-​regime protest in 2011.

Source: Data from Arab Barometer (2011).

oppositional events of the 21st century bears investigation. Why was Islamist mobi-
lization more visible in some cases than in others? Further, who counts as “Islamist”
in the question of Islamist mobilization during the Arab Spring? Studies of political
Islam in North Africa have tended to focus either on public attitudes or on Islamist
organizations and parties.1 Aiming to bridge the gap between organizational and indi-
vidual approaches to the study of mobilization, this chapter does both.
First, in Egypt, Tunisia, and Morocco, I pose the question of why principle Islamist or-
ganizations responded as they did to their respective revolutionary movements. I argue
that these varied responses were primarily strategic rather than ideological, reflecting
each organization’s experience of repression or inclusion under the incumbent regime
 Islamist Mobilization during the Arab Uprisings    599

and its resulting expectations regarding the benefits of a democratic future. Islamist
leaders played the strongest mobilizing role where their parties prior to 2011 occupied
a “middle ground” between repressed outsider and regime insider—​where autocrats
permitted above-​ground Islamist organizations to flourish but constrained these groups
from achieving true political influence.
Second, I shift the focus from organizations to citizens, using survey data ask how
these elite strategies mapped on to patterns of mobilization of the rank-​and-​file. Cross-​
national gaps in protest participation suggest that many Islamist-​leaning individuals did
follow their leaders’ edicts to either join the protests or abstain. A great many others, no-
tably in Tunisia, protested in the absence of formal Islamist leadership, galvanized likely
by economic grievance. Collective religious behaviors, namely mosque attendance, lead
to higher protest in general, while private religious behaviors like prayer made an indi-
vidual more likely to stay home.
These analyses underscore the diversity of Islamist actors in the MENA region and
the extent to which prior experiences of authoritarianism shaped Islamists’ capacities,
strategies, and expectations during these revolutionary moments. They also re-
veal a complex relationship between organizational strategy and individual behavior.
Organizational guidance matters, but it is far from determinative in accounting for citi-
zens’ decisions to protest.

Organizations, Individuals, and

Regimes: What We Know about
Revolutionary Actors

Among other debates, scholars have disagreed—​both implicitly and explicitly—​

about which actors are the primary units of revolutionary mobilization. Many studies
focus on the role of revolutionary organizations2—​formal, named organizations that
recruit and mobilize like-​minded citizens to challenge the sovereign power of the
incumbent. Revolutionary organizations may be formed for explicitly pro-​democ-
racy purposes, such as Solidarity in Poland (Mason 1989) or the African National
Congress in South Africa (Wood 2001). Alternatively, they may be preexisting groups
that pivot toward a revolutionary role given the opportunity.3 Some studies ask ex-
plicitly how revolutionary leaders mobilize citizens and maintain public confidence
over the course of a rebellion (Clarke 2014; Weinstein 2005), while others dwell more
on the prerogatives and decisions of revolutionary leaders, setting aside questions
of mass participation. In general, organizational studies tend to focus on actors al-
ready coded as “revolutionary”; the question of why existing organizations may
support or oppose a revolutionary movement, given the opportunity, has received
less attention.
600   Chantal Berman

An alternate branch of protest scholarship examines the prerogatives of individuals.

What social, demographic, or ideational fact ors make some citizens more likely than
others to join protests? Wider availability of nationally representative surveys, many
of which ask explicitly about protest participation,4 coupled with new methodolog-
ical approaches to surveying protesters themselves (Berman 2018; Tufekci and Wilson
2012), have allowed scholars to test hypotheses about the individual-​level determinants
of protest participation. This approach has been used to great effect in the field of Arab
Spring studies. Tufekci and Wilson (2012), for example, used a protest survey fielded
during the Egyptian uprising to study the ways in which protesters consumed media.
Beissinger, Jamal, and Mazur (2015) used Arab Barometer surveys to characterize and
compare revolutionary coalitions in Egypt and Tunisia. Cammett and Salti (2018) used
Gallup polls to investigate the prevalence of different economic grievances in the run-​
up to the uprisings.
By and large, the organizational perspectives assume that revolutionary organiza-
tions, once activated, are representative of revolutionary interests in society. Often, these
organizations become shorthand for the populations they represent (i.e. “the Islamists,”
or “organized labor”). Survey-​based approaches, correspondingly, tend to side-​step
questions of revolutionary leadership and organization, leaving the reader with a fairly
atomized view of the process by which individuals come to stand collectively. Nearly all
revolutionary movements, of course, rely both on organizations and on the decisions
of individuals. Curiously, few studies have explicitly asked how individual and organi-
zational prerogatives during a revolutionary episode relate to one another. To what ex-
tent do organizational edicts determine their members’ decisions to protest or not? How
do organizational affiliations stack up against other individual-​level drivers of protest
participation? To what extent are organizations’ positions toward emergent uprisings
shaped “from below”?
These questions may be particularly important when it comes to assessing the
role of Islam and Islamism in contemporary anti-​regime movements. Though news
reports and policy documents are rife with assumptions about Islamists, broadly
construed, the question of who “speaks” for Islam, particularly during moments of
political upheaval, is complex and unsettled. The movements that challenged incum-
bent regimes in Tunisia, Egypt, and Morocco each represented broad coalitions of
civil society organizations, joined by even larger crowds of citizens mobilized in-
formally through social networks and new media platforms (Beissinger, Jamal, and
Mazur 2015; Brownlee, Masoud, and Reynolds 2015). Labor unions, human rights ad-
vocacy groups, student organizations, and social media groups all played a role. This
chapter examines the role of Islamist organizations and Islamist-​leaning individuals
within these diverse coalitions. In so doing, I also aim to shed light on the relation-
ship between party-​based mobilization (or the lack thereof) and individual-​level
protest behavior. I begin by investigating the role of Islamist organizations in the 2011
uprisings, before moving on to survey analyses and, finally, to discussion of the links
between these two.
 Islamist Mobilization during the Arab Uprisings    601

Islamist Organizations and the Arab

Uprisings: Comparative Analysis
of Egypt, Tunisia, and Morocco

What role did Islamist organizations and parties play in the popular uprisings that swept
across North Africa in 2011? Though Islamists did not initiate the revolutionary protests
in any case, principle Islamist parties in Morocco, Tunisia, and Egypt responded quite
differently as the protests escalated in early 2011. In Tunisia, where networks of Harakat
Ennahda were totally clandestine, Islamist leaders in prison or exile were unable to play
a large mobilizing role, particularly in comparison with other societal actors like labor
unions. In Morocco, where the PJD by 2010 played a well-​defined role of parliamentary
opposition to monarchical forces, the party came out against the February 20th move-
ment and urged its followers not to participate. In Egypt, where the Muslim Brothers
were permitted to operate large networks of social service provision but were also sub-
ject to military trials and electoral manipulation, Brotherhood leaders threw their
weight behind protests planned initially by secular groups. I argue in this section that
these varied responses were primarily strategic rather than ideological, reflecting each
organization’s experience under the incumbent regime and, thus, its expectations over
the marginal benefits of a democratic future.
On what grounds are the Muslim Brotherhood, the Ennahda movement, and the PJD
comparable? Although these organizations differ in important ways—​as the following
sections will attest—​these groups have key organizational similarities that facilitate
comparative analysis. All three are designed as nationwide mass-​membership organ-
izations, though their capacity to operate as such has been periodically diminished by
repression. Each organization advocates for a strong influence of shariʿa over law and
public policy, though time and circumstance have proven these ideologies flexible, with
matters of Islamic doctrine as well as views toward democracy subject to rigorous in-
ternal debate (Wickham 2013). Broadly speaking, all three organizations have navigated
a transition from “outsider” movement toward major political organization, and have,
when permitted, fielded candidates for public office. However, all three organizations
have also practiced various forms of what Vairel (2011) calls “auto-​limitation”—​for ex-
ample, purposefully limiting the number of parliamentary candidates fielded in order to
limit confrontation with the regime (Masoud 2011; Willis 2004). Finally, all three parties
were extremely successful as electoral forces during the post-​revolutionary contests of
2011–​2013.
The relationship between Arab regimes and Islamist parties, like the workings of op-
position groups under authoritarianism more generally, has inspired large amounts of
scholarly writing.5 Scholars have also begun drawing connections between pre-​and
post-​revolutionary periods, asking, for example, how prior experiences of repression
shaped Islamists’ approach to governance and constitution-​writing after regime change
602   Chantal Berman

(Nugent 2020). Here, I leverage these historical perspectives to draw connections be-
tween each Islamist organization’s experience of late-​stage authoritarianism and their
response to anti-​regime mobilization in early 2011.

Coming Late to the Party: How the Muslim Brotherhood

Joined the Egyptian Revolution
Out of the major Islamist organizations in North Africa, the Egyptian Muslim
Brotherhood played the largest role in the 2011 revolutionary uprisings. Following the
successful protests of January 25, the Brotherhood joined a coalition of left-​leaning op-
position groups and alternative labor syndicates in calling for countrywide protests on
Friday, January 28. Though Brotherhood leaders were at first reticent to protest, younger
activists within the organization were able to convince the Guidance Bureau to throw its
lot in with the protest movement. The Brothers were able to mobilize adherents to pro-
test and to provide organization and protection to the flagship sit-​in at Tahrir Square.
The Brotherhood-​affiliated Freedom and Justice Party would go on to win 47 percent of
seats in Egypt’s 2011 parliamentary election, far more than any other party.
Historically, the Brotherhood’s relationship to the authoritarian state has been a pen-
dulum between repression and toleration. Neither acceptance nor co-​optation have
ever been on the table. Founded in 1928 by the Islamic scholar Hassan al-​Banna, the
Brotherhood is the region’s oldest and most prominent mass-​membership Islamist or-
ganization. During the interwar period, the Brotherhood grew in popularity through
extensive mosque outreach before succumbing to severe repression under the Nasser
regime. During the measured opening offered by the Sadat regime, the Brothers rebuilt
their influence through membership in professional associations and greatly expanded
their networks of social service provision—​a strategy that would continue to win the
Brotherhood widespread public support through the late Mubarak era (Brooke 2019;
Clark 2004).
Hosni Mubarak’s inauguration as Egypt’s president in 1981 coincided with a shift in
the Brotherhood’s strategy and political orientation. Following intense internal debates
over the value of “partyism,” the Brotherhood opted to contest the 1984 and 1987
elections by entering into strategic alliance with a series of secular nationalist and leftist
parties (El-​Ghobashy 2005). Wickham (2013) refers to the 1980s as the “high point” of
Muslim Brotherhood participation within the authoritarian system. By the 1990s, re-
gime officials, alarmed by the Brotherhood’s popularity, employed increasingly se-
vere and blatant measures to manipulate parliamentary elections.6 At the same time,
mass arrests and military trials of Brotherhood members sought to force the organiza-
tion out of political life. Nugent (2020) recounts nearly 100,000 arrests of Brotherhood
members, often candidates for office, during the final decade of Mubarak’s rule.
As the Brotherhood was slowly squeezed out of formal politics, the repression of the
late Mubarak era also rendered Brotherhood leaders uneasy about demonstrations and
 Islamist Mobilization during the Arab Uprisings    603

other forms of street politics. Demonstrations involving Brotherhood members were

often swiftly repressed, and Brotherhood leaders grew to expect that participation in
protests would lead to greater targeting of Brotherhood members, even if other, non-​
Islamist organizations were also involved. Here, we should note an essential irony. In
the years leading up to the 2011 uprisings, the Muslim Brotherhood was Egypt’s largest
and best-​organized opposition group. It was also the most reticent about street poli-
tics. Important Egyptian protest movements of the 2000s—​such as the Kefaya (Enough)
movement and the labor strikes at Mahalla al-​Kubra (Duboc 2011) did not involve the
Muslim Brotherhood.
Thus, Muslim Brothers were not involved in organizing the first major protests of the
2011 revolutionary wave. The set of initial revolutionary organizations were predomi-
nantly secular. Left-​leaning urban opposition groups and independent labor syndicates
came together to organize multisited demonstrations against police brutality on January
25, 2011. As elsewhere, younger Islamists were early supporters of the protest move-
ment. Student leaders within the Muslim Brotherhood formed alliances with secular
anti-​regime activists and demanded the right to participate in demonstrations under
the banner of Muslim Brotherhood Youth. After initially declining, the Brotherhood’s
Guidance Bureau heeded the advice of younger members on January 27 and officially
endorsed the protests (Clarke 2014). Brotherhood branches were active during the
January 28 Day of Rage and also during street battles with regime-​sponsored thugs in
Cairo the following week.
The Brotherhood’s entry into street protests arguably lent the Egyptian revolutionary
coalition its critical mass. Why did the Brotherhood’s leadership pivot toward the rev-
olutionary movement? The protests of January 25 were larger and more widespread
than most Egyptians had expected. Coupled with the example of Tunisia, where the
dictator Ben Ali had fled only ten days prior, the January 25 protests convinced many
Egyptians—​Brotherhood members included—​of the possibility of deposing Mubarak
through street protests.
Crucially, Wickham (2011) notes that Brotherhood leaders expected to “be the first
target of the regime’s wrath if the uprising failed.” Shortly after the protests on January
25, the Ministry of Interior issued a statement blaming the Muslim Brothers for the un-
rest. The regime arrested more than a thousand demonstrators and moved to disrupt
Internet and cell phone service, demonstrating intent to follow through on its repressive
threats once again (Ketchley 2017). In an important contrast with the Moroccan PJD—​
as the following section will explore—​the Brotherhood wagered that the organization
would suffer most from a wounded-​yet-​still-​intact Mubarak regime. Mistrust of the se-
curity state coupled with years of inflexible barriers to entry in electoral politics led the
Guidance Bureau to conclude that revolution, rather than continued negotiation and
self-​limitation, was the best way forward.
For the sake of comparison, we must note that that the Brotherhood was capable as a
mobilizing force because it was able to draw on an extensive internal organization and
hundreds of thousands of active members. Had the Brotherhood’s public organization
been dismantled through repression as extensive as that visited on Tunisia’s Islamists
604   Chantal Berman

in the prior decade, its leaders’ change of heart would have meant little. Yet, as Albrecht
(2005) points out, the Mubarak regime never went so far as to destroy the Brotherhoods’
organizational structure. Its clinics and schools were permitted to operate without ex-
cessive interference, likely because they provided a crucial supplement to the state’s
own meager social services. Thus, the Brotherhood’s “mixed” experience of repression
and toleration left the organization with adequate capacity to act as a mobilizing force,
coupled with the expectation that regime change would remove impediments to the
organization’s future growth.
Finally, the Muslim Brotherhood during the Egyptian revolution represents a case
where less-​senior members successfully prevailed upon senior leadership, swaying their
stance toward the protest movement. Had Brotherhood youth not lobbied the Guidance
Bureau to support their participation in anti-​regime protests, the Muslim Brotherhood
may never have formally endorsed the revolutionary movement against Hosni
Mubarak. Though this case does not demonstrate influence of the “rank-​and-​file” per
se—​the youth in question were already participating in positions of leadership within
the hierarchical organization—​it does affirm the heterogeneity of Egyptian Islamists’
responses to the uprisings and the collective nature of organizational decision-​making, a
dynamic often lost when revolutionary organizations are treated as unitary actors.

Loyal Opposition: The PJD’s Insider Game during M20F

In contrast to the Egyptian Muslim Brotherhood, the Moroccan PJD sided publicly
against the February 20th movement (M20F). Though some younger PJD members
supported the protests, PJD leaders—​like other “opposition” parties in parliament—​
declined to lend their support and advised their members not to protest in the streets.
Instead, Moroccan Islamist leaders leveraged this rupture as an opportunity to extract
specific concessions and increase its prominence within the Moroccan electoral sphere.
Like the Muslim Brotherhood, by early 2011 the PJD was a large and capacious
formal organization with significant political influence and an extensive network of
supporters. Like Islamist leaders throughout the Arab world, many of the PJD’s leading
figures were initially inspired by the Egyptian Muslim Brotherhood. The PJD (formerly
the Movement for Unity and Reform, At-​Tawhid wa al-​Islah) was founded in 1967. As
in Egypt, Moroccan Islamists were barred from formal politics and faced devastating
repression during the postcolonial decades. During the “Years of Lead” under King
Hassan II, Islamists not jailed or disappeared were forced to work through clandestine
networks (Vairel 2011).
The advance of Moroccan multipartyism in the late 1990s and the ascension of
the more “liberal” monarch Mohammed VI allowed the PJD to slowly expand its in-
fluence and to branch, tepidly, into electoral politics. Known PJD members were
permitted to join an existing political party (the Mouvement Populaire Démocratique
et Constitutionnel, MPDC) and to file in this capacity as parliamentary candidates in
1997. After changing its name to PJD, a move that cemented Islamists’ dominion over
 Islamist Mobilization during the Arab Uprisings    605

the party, the party again fielded candidates in the 2002 and 2006 elections, while pur-
posefully contesting a limited number of seats in order to avoid confrontation with the
regime (Willis 2004). The PJD over this period attracted a wide support base with its
conservative Islamic principles, but also with its critiques of the regime in the arena of
economic governance and social welfare provision (Daadaoui 2017).
In Moroccan politics, the parliament maintains limited constitutional authority,
and political power remains concentrated around the monarch and his close circle of
advisors, known as the makhzen (Yom 2017). Nevertheless, the monarchy watched PJD’s
rising popularity and organizational infrastructure with trepidation. The Moroccan
regime’s main strategy for tempering the PJD’s rise in the decade preceding the uprisings
rested not on direct repression or banning the party from participation in formal poli-
tics, but rather the creation of a novel monarchist party intended to serve as a counter-
weight to PJD’s electoral popularity. The monarchist Parti de l’Authenticité et Modernité
(PAM) was founded in 2008 by a coalition of political figures with close ties to the
palace. PAM worked to undermine the PJD’s role in Moroccan politics, for example,
working to muscle the party out of local multiparty coalitions following the 2009 local
elections (Buehler 2013a). As Wegner (2011) notes, the regime also attempted to inter-
cede in the PJD’s internal hierarchy in to neutralize its more outspoken critics.
Notably, the regime’s approach to managing the PJD has involved little direct, physical
violence. While PJD-​related gatherings are sometimes policed heavily (Berman 2020),
the PJD in the 20th century has faced far less repression than the Egyptian Muslim
Brotherhood and, importantly, far less repression than Morocco’s secondary Islamist
movement, al-​ʿAdl wa al-​Ihsan, which eschews formal politics in favor of mosque-​based
organization (Cavatorta and Durac 2011). In return, the PJD has not been reticent about
using demonstrations of its mass base as a form of bargaining with the makhzen and the
loyalist parties. However, the PJD has also fallen short of directly critiquing the mon-
archy or calling for regime change in Morocco. As Daadaoui (2017) writes, the PJD has
“walked a tightrope of appearing loyal to the regime while also reproaching the palace
for tahakoum [political manipulation].”
This strategy was in full force as the PJD leadership confronted the possibility of
mass demonstrations in the week leading up to February 20, 2011, when a coalition
of Moroccan activists called for nationwide protests. The February 20th movement
(M20F), much like initial Egyptian protests, represented a coalition of organizations fo-
cused on social justice and human rights, notably including the Moroccan Association
of Human Rights (AMDH). February 20th leaders did not explicitly call for regime
change—​the movement’s stance toward monarchism and the precise nature of the
Moroccan state was internally debated throughout its existence—​but its anti-​status-​
quo messaging combined with the recent revolutions in Tunisia and Egypt prompted
parties in the “loyal opposition” to circle the wagons. As in Egypt, many younger PJD
members supported the protests, even publishing a supportive statement on February
17 before bowing to organizational pressure to retract it. That same day, PJD secre-
tary general Abdelilah Benkirane, who would become Morocco’s first Islamist prime
minister in November 2011, announced that the party would boycott M20F. Despite
606   Chantal Berman

record-​breaking protests on February 20—​demonstrations occurred in more than

fifty Moroccan cities and towns (Lawrence 2017)—​and demonstrations that continued
through May, the PJD did not reverse its position (Bennani-​Chraibi and Jeghllaly 2012).
The PJD’s firm anti-​revolutionary stance reflected more than simple loyalty to the
king or wholesale co-​optation. Buehler (2013b), for example, argues that the PJD viewed
the February 20th movement—​and the Arab uprisings more broadly—​as a political op-
portunity to advance its own interests. As the regime and the monarch moved to placate
Moroccans with proposed constitutional revisions and redistributive projects (Berman
2020), the PJD also gained important targeted concessions in exchange for its restraint.
PJD Secretariat General member Jama’a Moatassim, previously arrested on corruption
charges, was released from prison. The regime also dropped proposed changes to the
electoral code designed to limit PJD’s vote share. Finally, PJD took advantage of the po-
litical moment to roll back self-​imposed limits on its electoral bids. In early elections in
October 2011, the PJD increased its parliamentary presence by sixty-​one seats, leading
to an Islamist plurality.
Comparison of the PJD with Egypt’s Muslim Brotherhood is instructive. The PJD
was also a capacious, well-​liked organization with extensive popular networks. Had the
PJD thrown its weight behind the anti-​regime protests, as the Brotherhood did, it likely
would have counted among the most important revolutionary organizations. Yet, unlike
the Brotherhood, the PJD’s strategic calculations rested on the projection that the party
stood to benefit more from an intact (if politically weakened) Moroccan state than from
a popular revolution. Based on their prior experience working within the system, PJD
leaders were willing to gamble that the monarchy and makhzen would not use the M20F
as pretext to ban or repress the Islamist party. Rather, they expected their own influence
to expand—​an expectation that would be unlikely in a context like Egypt or Tunisia,
where Islamist leaders clearly viewed Mubarak and Ben Ali as inflexible barriers to the
Islamist political project.

Observing from Exile: Why Ennahda Lacked

Capacity to Revolt
Unlike their counterparts in Egypt and Morocco, Tunisian Islamists did not have formal,
above-​ground mass-​ membership organizations at the onset of the 2011 uprisings.
Prominent Islamists and leaders of the Ennahda (Islamist renaissance) movement
watched revolutionary events unfold from foreign exile or from prison. Tunisian
Islamists remained popular, as attested by their quick reassembly and political domina-
tion of the early democratic transition. They certainly opposed Ben Ali’s administration
and supported the prospect of revolutionary change. Yet Ben Ali’s systemic dismantling
of Ennahda’s domestic organization during the decade prior to the Jasmine revolution
rendered the party unable to act as a major mobilizing force. Left and labor organizations,
in particular the syndicates of the Union Générale des Travialleurs Tunisiens (UGTT),
picked up the slack, steering Tunisia’s revolutionary movement to victory.
 Islamist Mobilization during the Arab Uprisings    607

As in prewar Egypt, Islamist organizing in Tunisia operated mainly through informal

mosque networks until 1981, when Rachid al-​Ghannouchi founded the Mouvement
de la Tendance Islamique (MTI) with the objective of working through the country’s
electoral system to promote Islamic law. Like the Muslim Brothers, the MTI also forged
links with other associations, in particular the Tunisian student union UGET. The or-
ganization also applied for a license to form an official political party. Observing MTI’s
growing popular and organizational infrastructure, the regime of Habib Bourguiba
moved to repress the nascent Islamist movement, carrying out successive waves of
arrests and interrogations against MTI members in 1981, 1983, and 1987 (Perkins 2014).
Based on these experiences of repression, Tunisian Islamists initially welcomed former
prime minister Zine el-​Abidine Ben Ali’s coup again Bourguiba in November 1987. Gearing
up to participate in the 1989 parliamentary elections, MTI signed the National Pact and
changed its name to Harakat Ennahda in a bid to comply with a new law banning parties
based on religion, race, or language (McCarthy 2018). Unlike Islamist parties elsewhere,
who were periodically permitted to stand as political candidates as either party members
or independents, the 1989 elections would prove to be Ennahda’s sole opportunity for elec-
toral participation until the 2011 revolution. Ennahda contested a large number of seats
and emerged as the largest opposition force, with 14.5 percent of the vote. The Ennahda
deputies were prohibited from taking their seats in parliament and were instead subject to
a massive wave of repression. In the years that followed, up to 30,000 Ennahda members
were arrested, and many were subjected to torture. Thousands of Islamists who were able
fled into exile. Ghannouchi himself was jailed 1981 and again in 1988; shortly thereafter, he
fled to the UK, where he remained until the fall of the Ben Ali regime.
During the 1990s and 2000s, Ennahda leaders remained committed to their orga-
nization. They relaunched the party’s newspaper al-​Fajr and even participated in a se-
ries of rapprochement initiatives with leaders of Tunisia’s secular opposition parties,
most of whom remained territorially in Tunisia (Cavatorta and Haugbolle 2012). Yet as
Wolf (2017) points out, maintaining a mass-​membership organization with members
dispersed in more than seventy countries and in prison remained a massive challenge.
Members who remained at large in Tunisia lived in the shadow of fear. Thus, the orga-
nization itself atrophied. Further, Tunisian Islamists lacked affiliated social service or-
ganizations or other public-​facing institutions that may have recruited more members.
Some conservative Tunisian Islamists turned to Salafism and other ideologies that dis-
courage engagement with the state.
Thus, when the initial protests of the Tunisian revolution broke out in December
2010, there were few Islamist leaders or institutions on hand to provide leader-
ship. Ghannouchi himself would not return to Tunis until late January 2011, shortly
after Ennahda was re-​legalized. All indicators point to the fact that Ennahda leaders
would likely have supported the anti-​regime protests. Many of Ennahda’s intellectual
leaders had been endorsing democracy since the 1980s (Wolf 2017). Yet by virtue of
the dismantling of their domestic political organization, their views counted for little.
Ennahda was unable to play a significant role in mobilizing Tunisians to protest against
the Ben Ali regime.
608   Chantal Berman

What did revolution in Tunisia look like in the absence of a strong Islamist presence?
Previous research has examined difference between popular protest coalitions in Egypt
and Tunisia. Beissinger, Jamal, and Mazur (2015) argue that the Egyptian revolution was
more middle-​class, middle-​aged, and relied more on the intermediation of “civil society
organizations.” The Tunisian revolution, by contrast, relied more on unemployed youth
and on spontaneous forms of networked mobilization, including the use of Facebook
pages to organize protests. These differences likely reflect, in no small part, the influence of
the Muslim Brotherhood in the Egyptian revolution.7 The absence of a parallel organized
Islamist sector in Tunisia likely left these same middle-​aged professionals un-​mobilized,
throwing the balance of the revolutionary coalition toward students and youth.
Yet the Tunisian uprising did unfold entirely without the aid of revolutionary or-
ganizations. Protests spread spontaneously at first, as citizens of Tunisia’s marginalized
southern governorates echoed protests in the town of Sidi Bouzid. Tunisia’s leftist
and secular opposition groups also contributed. If the Muslim Brotherhood’s arrival
at Egyptian protests provided the single biggest injection of organizational strength
and mass membership, this role was played in the Tunisia by the UGTT. The UGTT
is Tunisia’s largest mass-​membership organization, counting more than half a mil-
lion members. Historically a militant force in Tunisian politics, during Ben Ali’s final
decade of rule the UGTT had shied away from its political role. As such, the UGTT
National Committee did not jump to endorse the nascent protest movement spilling
out of Tunisia’s rural periphery in mid-​December. As protests continued into January,
however, the union leadership pivoted. Yielding to pressure from historically militant
syndicates in the public sector, on January 11 the UGTT called for a general strike in
Sfax, Tunisia’s second city (Angrist 2013). A UGTT strike call in Tunis for January14
helped to catalyze Ben Ali’s departure from Tunis that same day.

Islam and Protest on the Popular

Level: Survey Analysis

Principle Islamist organizations in North Africa responded differently to the 2011 Arab
revolutions, reflecting each party’s unique experience of authoritarian rule. Yet elite
pronouncements are, in times of uprising, only part of the story. To what extent did the
directives of Islamist leaders translate into protest participation (or lack thereof) among
the rank-​and-​file? How did Islamist beliefs and religious commitments affect citizens’
individual choices to protest, and how did these responses vary across countries? This
section uses descriptive and regression analyses of the 2013 Arab Barometer Survey8 to
examine the role that Islamist beliefs and affiliations played in driving individuals to
protest in 2011. These analyses help to draw out the individual-​level implications of the
country-​level divergences discussed in the previous section.
For a snapshot of the revolutionary coalitions, Table 29.1 presents raw proportions
of protest participation among Islamists, secularists, and the population at large. For
 Islamist Mobilization during the Arab Uprisings    609

Table 29.1: Percent of Islamist and Secularist Citizens Mobilized in 2011

Egypt Morocco Tunisia
Islamist party role in uprisings Supportive Opposed No strong role

% Islamists protested 14.3% 8.2% 18.0%

% Secularists protested 12.1% 16.7% 22.0%
% Protested in the general population 13.5% 9.1% 18.6%

parsimony in this table, I use a rough dichotomous measure, coding “Islamists” as those
citizens who either “agree” or “strongly agree” that parliament should enact law and
policy in keeping with Islamist law, while “secularists” are those who either “disagree” or
“strongly disagree.” Protest participation overall was higher in the successful cases, top-
ping 13 percent in Egypt and 18 percent in Tunisia, compared with 9 percent in Morocco.
These breakdowns indicate that Islamist party guidance may have held some sway
over individual protest decisions. Egypt, where the Muslim Brotherhood became in-
volved in revolutionary protests, is the only case where Islamist participation appears
(slightly) higher than secularist participation, at 14.3 percent. By contrast, in Morocco,
where the PJD implored members not to protest, Islamists participated at roughly half
the rate of secularists. Given that Moroccan secularist participation (16 percent) was
well within the range of Tunisia and Egypt, it is possible that low Islamist participation
(8 percent) was a key factor in preventing the Moroccan February 20th movement from
gaining critical revolutionary mass.
Still, nearly one in ten Islamist-​leaning Moroccans protested against the instruction
of the main Islamist party. Some of these were likely adherents to al-​ʿAdl wa al-​Ihsan
(Justice and Charity movement), PJD’s more anti-​state, clandestine counterpart, whose
leadership endorsed the February 20th protests. Others were likely PJD adherents for
whom the M20F’s message of social and civic justice nonetheless resonated. Though the
Morocco Arab Barometer surveys unfortunately contain no questions as to why citizens
protested, the M20F’s public messaging and its alliance with social justice organizations
indicates a strong parallel role of socioeconomic grievance in mobilizing Moroccans of
all ideological bents.
Tunisia provides an opportunity to examine protest rates in the absence of definitive
guidance from an Islamist party. Islamist-​leaning citizens appear to have protested at a
slightly lower rate than secularists, but nevertheless made a strong showing at 18 per-
cent. The near-​parity of Islamist and secularist participation speaks, again, to the broad-
ness of the revolutionary coalition that deposed Ben Ali, and to the primacy of other
incentives—​in particular, socioeconomic grievances—​in mobilizing Tunisians in 2011.
(See Figure 29.1.)
Table 29.2 complicates this picture by performing a set of multivariate regressions. The
dependent variable is a binary measure of revolutionary protest participation. For each
country (as well as the pooled three-​country sample), I include one reduced regression, in-
cluding indicators for Islamist beliefs, trust in Islamist parties, and religious practice, and
Table 29.2: Political Islam and Revolutionary Protest (Logit Regressions)
Dependent variable: revolutionary protest participation (binary)

Pooled Egypt Morocco Tunisia

(1) (2) (3) (4) (5) (6) (7) (8)

Law should -​0.049 0.017 0.159 0.151 -​0.323** -​0.255* -​0.103 -​0.022
be based on (0.060) (0.062) (0.107) (0.111) (0.149) (0.154) (0.084) (0.089)
shariʿa
Trust Islamist -​0.067 -​0.054 0.013 0.040 -​0.212* -​0.181 -​0.038 -​0.038
party (0.051) (0.052) (0.093) (0.098) (0.111) (0.113) (0.071) (0.074)
Attend 0.268*** 0.325*** 0.472*** 0.479*** 0.129 0.133 0.236*** 0.351***
Friday (0.050) (0.054) (0.118) (0.127) (0.130) (0.134) (0.063) (0.070)
mosque
Read Qurʾan -​0.191*** -​0.139** -​0.321*** -​0.222** -​0.090 -​0.064 -​0.136 -​0.116
(0.060) (0.063) (0.101) (0.107) (0.140) (0.143) (0.088) (0.095)

Urban 0.583*** 0.663*** 0.719*** 0.437**

(0.119) (0.192) (0.274) (0.186)

Education 0.362*** 0.577*** 0.175** 0.327***

(0.041) (0.008) (0.080) (0.065)

Age -​0.026*** -​0.007 -​0.031*** -​0.034***

(0.004) (0.008) (0.011) (0.006)

Country: -​0.462*** -​0.268

Morocco (0.155) (0.165)

Country: 0.766*** 1.030***

Tunisia (0.131) (0.144)

Constant -​1.884*** -​3.240*** -​2.938*** -​5.732*** -​1.051* -​1.343* -​1.127*** -​1.733***

(0.281) (0.380) (0.561) (0.768) (0.600) (0.766) (0.365) (0.524)

Observations 3019 3019 1095 1095 897 897 1027 1027

Log -​1219.371 -1​ 121.670 -​430.022 -​384.938 -​264.508 -​253.079 -​516.130 -​467.436
likelihood
Akaike 2452.741 2263.340 870.044 785.876 539.016 522.158 1042.261 950.871
Information
Critreion
McFadded 0.13 0.20 0.09 0.18 0.22 0.26 0.10 0.19
Psuedo-​R
squared

Note: * p<0.1;
** p<0.05;
*** p<0.01
 Islamist Mobilization during the Arab Uprisings    611

one expanded regression, including common demographic covariates. These regressions

don’t aim to causally identify any singular effect, but rather to consider a range of beliefs and
behaviors that may prove positively or negatively associated with the decision to protest.
In the simple regressions, the effect of Islamist beliefs—​measured by support for
the statement, “The government and parliament should enact laws in accordance with
Islamic law”—​is negative and significant in Morocco, and positive and nearly significant
(p = 0.13) in Egypt. Also in Morocco, trust in the PJD is a significant negative predictor
of protest. Significance is attenuated by the addition of the demographic covariates in
both cases, although less so in Morocco. These results point, again, to the divergent
implications of the Muslim Brotherhood’s decision to endorse the protests and the PJD’s
decision to withhold its support. In Tunisia, we see little significance associated with ei-
ther Islamist beliefs or trust in Islamist parties, again indicating relative parity of partici-
pation between Islamist and secularist citizens.
These regressions also consider the effects of nonpolitical religious behaviors. Were
pious citizens more or less inclined to protests, regardless of their adherence to polit-
ical Islam? Here, I find some strong yet divergent effects. Collective religious behavior
(Friday mosque attendance) proves a significant predictor of revolutionary protest,
particularly in Egypt and Tunisia. By contrast, individual religious behavior (Qurʾan
readership) appears to exert a negative effect on protest throughout the sample.9 These
findings correspond to contextual evidence—​as Ketchley (2017) points out, Friday
services were often an important focal point for protests, and citizens would leave the
mosque to protest together. These results also support the general view that individuals
who participate in other collective endeavors—​political or otherwise—​are also more
likely to protest, perhaps because they are more social or civic-​minded.
In aggregate, these results suggest that the individual-​level relationship between
Islamist ideology and revolutionary protest is highly context-​specific. Holding Islamist
views doesn’t automatically incline citizens toward anti-​statist behavior, such as protest,
even where regimes are avowedly secular. Further, faith in Islamist leaders and personal
piety are shown to matter, but only up to a point. Here, I refer to calculated McFadden’s
pseudo-​R squared, a statistic capturing the amount of variation explained by each re-
gression. In Table 29.2, the pseudo-​R squared statistics associated with the reduced
regressions are, in each case, quite low. (The addition of the demographic covariates raises
pseudo-​R squared in each case, highlighting the importance of age, education, urban ge-
ography in predicting protest behavior.) These low values indicate that the factors under
consideration in this chapter, while meaningful, constitute only some among many
factors driving individuals to protest in each country. When it comes to explaining revo-
lution in the Middle East, political Islam is one piece of a large and complex puzzle.

Conclusion

In conclusion, I will draw out key implications from the analyses presented in this
chapter and address several alternative factors that may also have contributed to our
observed outcomes.
612   Chantal Berman

First, we must recognize important variations in the political role of Islamists during
the North African uprisings. Each principle Islamist party responded differently to the
revolutionary wave. I have argued that Islamist parties’ decisions to enter into revolu-
tionary protests (or not) were primarily strategic, rather than ideological, or even the-
ological. There is scant evidence to suggest that parties’ decisions were based on any
particular interpretation of religious text or tradition. In other words, there was nothing
particularly Islamic about the ways in which Islamist organizations responded to the
Arab Spring. That the Muslim Brotherhood’s entry into the uprisings mirrored the pro-
cess by which the UGTT came to participate in Tunisian protests reinforces the notion
that Islamist and secular organizations, confronted with similar circumstances, made
similar calculations.
Further, none of the Islamists in question were monolithic in their response. Decisions
to participate or not participate were contested among leadership and among the rank-​
and-​file. Generational splits were germane in many cases, with younger Islamists
advocating for a more active oppositional role. In Egypt, younger members were able
to sway senior leadership to join the protests. In Morocco, they failed to do so—​largely
because of the PJD’s strategic calculations outlined above. Nonetheless, these internal
debates speak to tremendous range in the ways that Islamists and Islamist organiza-
tions responded to the prospect of political revolution. Cross-​national studies that use
the prevalence of Islam as a “dummy” in regression analyses, particularly in addressing
questions like regime change and democracy, should take note of this essential diversity.
Second, the multivariate survey analyses serve to remind us that Islamist beliefs
and adherence to an Islamist organization represent only two among many factors
shaping an individual’s protest behavior. This chapter has not focused on the nonpo-
litical components of religious life in North Africa, but nonetheless we see evidence of
mosque attendance associated positively with protest behavior. These regressions like-
wise affirm the crucial role played by demographic variables: across the board, citizens
who are younger, more educated, and live in urban areas were far more likely to pro-
test. Organizational studies (much like the first half of this chapter) tend to focus on
one associational sector—​the role of religious parties, the role of labor unions, and so
on. Survey analyses by nature allow for multiple “explanations” for behavioral outcomes
to coexist. This multifactor framework resonates with recent research on revolutions,
which has stressed the diversity of revolutionary coalitions and the multiplicity of revo-
lutionary actors involved in successful movements (Beissinger 2013).
Third, the role of Islamist parties in revolutionary events did not appear to determine
their popularity during post-​revolutionary elections. Islamists across the three cases
swept the next round of elections in 2011–​2012.10 This dynamic is particularly interesting
in the case of Tunisia, where Islamists won the 2011 National Constituent Assembly
elections despite the much larger role of secular organizations in the 2011 revolution.
Ennahda’s political dominance of the early transitional period—​a coda to its impressive
performance in the 1989 elections—​helps to dispel a potential challenge to the argu-
ment that state repression is to blame for the nonexistence of organized political Islam in
 Islamist Mobilization during the Arab Uprisings    613

late-​twentieth-​century Tunisia. There exists a popular perception that Tunisian society

is more secular-​minded than other Arab countries, dating back to the Westernizing pro-
gram of the postcolonial president Habib Bourguiba. Tunisia’s purported secularity may
pose a challenge to the argument that the Ennahda movement was weakened prima-
rily by repression. Such a culturalist argument may hold instead that Ennahda did not
maintain or mobilize formal networks of supporters in 2011 simply because Tunisians
lacked interest in Islamist ideas. The 2011 elections demonstrated that the Islamist poli-
tics maintained significant popular appeal, and—​most likely—​that citizens understood
that repression was to blame for Ennahda’s low profile during the revolution.
Finally, the comparative analysis sheds light on which technologies of authoritarian
rule may lead opposition parties to join the revolutionary coalition. Many scholars of
political Islam have advanced and refuted the “inclusion-​moderation” hypothesis—​the
notion that by incorporating Islamist organizations and citizens within the political
system, authoritarian rulers could render these parties less likely to revolt against the
state (Schwedler 2011). Whatever its empirical merits, autocrats in many cases seem to
have embraced this wisdom, as regimes across the region began to accept Islamist elec-
toral candidates in the 1980s and 1990s. Many regimes also tempered inclusion with
violence, producing an ambivalent stasis of accommodation and repression. Islamists
were neither fully accepted nor entirely routed. Yet, per the case of Egypt, these precise
conditions appear to have led to the highest odds of Islamist participation in the revo-
lution. Though far from definitive, this chapter provides some novel indication that the
“mixed strategy” of carrots and sticks for opposition movements—​in some ways, the
gold standard for entrenched authoritarians in the early 21st century—​may eventually
backfire.

Notes
1. For the former, see Grewal et al 2019; Nugent, Masoud, and Jamal 2016; Hoffman and Jamal
2014. For the latter, see Schwedler 2006; Clark 2004; Wolf 2017; Storm 2007; Brooke 2019;
Nugent 2020.
2. Revolutionary organizations are conceptualized per McCarthy and Zald’s (1977) more ge-
neral criteria social movement organizations: “formal and complex organizations which
align their interests with the movement and attempts to implement these goals.”
3. Labor unions in particular have played such a role in turning organized interests toward a
revolutionary posture, as in the cases of Peru and Argentina (Collier and Mahoney 1997).
4. Most Arab Barometer surveys since 2013, for example, have asked citizens about their par-
ticipation in the 2011 uprisings.
5. See, for example, Blaydes 2011; Masoud 2014; Wickham 2013; Wolf 2017; Storm 2007;
Schwedler 2006; Clark 2004.
6. Masoud (2011) refers to Mubarak-​era elections as “stage-​managed to generate victories for
the ruling party.”
7. Brotherhood scholars have argued that the Brotherhood’s main constituency comprises
middle-​class professionals ( Clark 2004).
614   Chantal Berman

8. Wave III of the Arab Barometer includes surveys from Morocco, Tunisia, and Egypt.
Further information and data downloads for Wave III may be found at https://​www.
arabbarometer.org/​waves/​arab-​barometer-​wave-​iii/​.
9. These findings contradict prior findings based on 2011 Arab Barometer data. Hoffman and
Jamal (2014) find that Qurʾan reading positively predicts protest, and that mosque attend-
ance does not.
10. Though the pendulum has since swung back somewhat—​ as tends to happen in
democracy—​PJD and Ennahda remain potent political forces in Morocco and Tunisia.
The Brotherhood in Egypt has been heavily repressed since General Abdel Fattah al-​Sisi
came to power, in a situation reminiscent of 1990s Tunisia. Electoral developments since
2011 are beyond the scope of this chapter; see Nugent’s chapter.

References
Albrecht, Holger. 2005. “How Can Opposition Support Authoritarianism? Lessons from
Egypt.” Democratization 12, no. 3: 378–​397.
Angrist, Michelle Penner. 2013. “Understanding the Success of Mass Civic Protest in Tunisia.”
Middle East Journal 67, no. 4: 547–​564.
Beissinger, Mark R. 2013. “The Semblance of Democratic Revolution: Coalitions in Ukraine’s
Orange Revolution.” American Political Science Review 107, no. 3: 574–​592.
Beissinger, Mark R., Amaney A. Jamal, and Kevin Mazur. 2015. “Explaining Divergent
Revolutionary Coalitions: Regime Strategies and the Structuring of Participation in the
Tunisian and Egyptian Revolutions.” Comparative Political Studies 48, no. 1: 1–​21.
Bennani-​Chraibi, Mounia, and Mohamed Jeghllaly. 2012 “La dynamique protestataire du
movement du 20 février à Casablanca.” Revue francaise de science politique 62, no. 5: 867–​894.
Berman, Chantal E. 2018. “When Revolutionary Coalitions Break Down: Polarization, Protest,
and the Tunisian Political Crisis of August 2013.” Middle East Law and Governance 11,
no. 2: 136–​179.
Berman, Chantal E. 2020. “Policing the Organizational Threat in Morocco: Protest and Public
Violence in Liberal Autocracies.” Forthcoming, American Journal of Political Science.
Blaydes, Lisa. 2011. Elections and Distributive Politics in Mubarak’s Egypt. Cambridge:
Cambridge University Press.
Brooke, Steven. 2019. Winning Hearts and Votes: Social Services and the Islamist Political
Advantage. Ithaca, NY: Cornell University Press.
Buehler, Matt. 2013a. “Safety-​Valve Elections and the Arab Spring: The Weakening (and
Resurgence) of Morocco’s Islamist Opposition Party.” Terrorism and Political Violence 25, no.
1: 137–​156.
Buehler, Matt. 2013b. “The Threat to ‘Un-​moderate’: Moroccan Islamists and the Arab Spring.”
Middle East Law and Governance 5, no. 3: 231–​257.
Brownlee, Jason, Tarek Masoud, and Andrew Reynolds. 2015. The Arab Spring: Pathways of
Repression and Reform. Oxford: Oxford University Press.
Cammett, Melani, and Nisreen Salti. 2018. “Popular Grievances in the Arab Region: Evaluating
Explanation for Discontent in the lead-​up to the Uprisings.” Middle East Development
Journal 10, no. 1: 64–​96.
Cavatorta, Francesco, and Vincent Durac. 2011. Civil Society and Democratization in the Arab
World: The Dynamics of Activism. New York: Routledge.
 Islamist Mobilization during the Arab Uprisings    615

Cavatorta, Franceso, and Rikke Hostrup Haugbolle. 2012. “The End of Authoritarian Ruel and
the Mythology of Tunisia under Ben Ali.” Mediterranean Politics 17, no. 2: 179–​195.
Cavatorta, Francesco and Fabio Merone. 2013. “Moderation through Exclusion? The Journey
of the Tunisian Ennahda from Fundamentalist to Conservative Party.” Democratization 20,
no. 5: 857–​875.
Clark, Janine A. 2004. Islam, Charity, and Activism: Middle Class Networks and Social Welfare
in Egypt, Jordan, and Yemen. Bloomington: Indiana University Press.
Clarke, Killian. 2014. “Unexpected Brokers of Mobilization: Contingency and Networks in the
2011 Egyptian Uprising.” Comparative Politics 46, no. 4: 379–​397.
Collier, Ruth Berins, and James Mahoney. 1997. “Adding Collective Actors to Collective
Outcomes: Labor and Recent Democratization in South America and Southern Europe.”
Comparative Politics 29, no. 3: 285–​303.
Daadaoui, Mohamed. 2017. “Of Monarchs and Islamists: The ‘Refo-​lutionary’ Promise of the
PJD Islamists and Regime Control in Morocco.” Middle East Critique 26, no. 1: 355–​371.
Duboc, Marie. 2011. “Social Upheaval in Egypt Since 2002: Instability and Local Mobilization
of Textiel Factory Workers.” Revue Tiers Monde 5: 95–​115.
El-​
Ghobashy, Mona. 2005. “The Metamorphosis of the Egyptian Muslim Brothers.”
International Journal of Middle East Studies 37: 373–​395.
Grewal, Sharan, Amaney A. Jamal, Tarek Masoud, and Elizabeth R. Nugent. 2019. “Poverty and
Divine Rewards: The Electoral Advantage of Islamist Political Parties.” American Journal of
Political Science 63, no. 4: 859–​874.
Hoffmann, Anja, and Christoph Konig. 2013. “Scratching the Democratic Façade: Framing
Strategies of the February 20th Movement. Mediterranean Politics 18, no. 1: 1–​22.
Hoffman, Michael, and Amaney Jamal. 2014. “Religion in the Arab Spring: Between Two
Competing Narratives.” The Journal of Politics 76, no. 3: 593–​606.
Ketchley, Neil. 2017. Egypt in a Time of Revolution: Contentious Politics and the Arab Spring.
New York: Cambridge University Press.
Lawrence, Adria K. 2017. “Repression and Activism among the Arab Spring’s First Movers:
Evidence from Morocco’s February 20th Movement.” British Journal of Political Science 47,
no. 3: 699–​7 18.
Mason, David S. 1989. “Solidarity as a New Social Movement.” Political Science Quarterly. 104,
no. 1: 41–​58.
Masoud, Tarek. 2011. “The Upheavals in Egypt and Tunisia: The Road to (and from) Liberation
Square.” Journal of Democracy 22, no. 3: 20–​34.
Masoud, Tarek. 2014. Counting Islam: Religion, Class, and Elections in Egypt. New York:
Cambridge University Press.
McCarthy, John D., and Mayer N. Zald. 1977. “Resource Mobilization and Social Movements: A
Partial Theory.” The American Journal of Sociology 82, no. 6: 1212–​1241.
McCarthy, Rory. 2018. “When Islamists Lose: The Politicization of Tunisia’s Ennahda
Movement.” Middle East Journal 72, no. 3: 365–​384.
Nugent, Elizabeth R. 2020. After Repression: How Polarization Derails Democratic Transition.
Princeton: Princeton University Press.
Nugent, Elizabeth R. 2020. “The Political Psychology of Repression and Polarization.” World
Politics 72, no. 2: 291–​334.
Nugent, Elizabeth, Tarek Masoud, and Amaney A. Jamal. 2016. “Arab Responses to
Western Hegemony: Experimental Evidence from Egypt.” Journal of Conflict Resolution 62,
no. 2: 254–​288.
616   Chantal Berman

Perkins, Kenneth. 2014. A History of Modern Tunisia. 2nd ed. New York: Cambridge University
Press.
Schwedler, Jillian. 2006. Faith in Moderation: Islamist Parties in Jordan and Yemen. Cambridge:
Cambridge University Press.
Schwedler, Jillian. 2011. “Can Islamists Become Moderates? Rethinking the Inclusion-​
Moderation Hypothesis.” World Politics 63, no. 2: 347–​376.
Storm, Lise. 2007. Democratization in Morocco: The Political Elite and Struggles for Power in the
Post-​Independence State. New York: Routledge.
Tufekci, Zeynep, and Christopher Wilson. 2012. “Social Media and the Decision to Participate
in Political Protest: Observations from Tahrir Sqaure.” Journal of Communication 62:
363–​379.
Vairel, Frédéric. 2011. Politique et mouvements sociaux au Maroc. Paris: Sciences Po.
Wegner, Eva, and Miquel Pellicer. 2009. “Islamist Moderation without Democratization: The
Coming of Age of the Moroccan Party of Justice and Development?” Democratization 16,
no.1: 157–​175.
Wegner, Eva. 2011. Islamist Opposition in Authoritarian Regimes: The Party of Justice and
Development in Morocco. Syracuse: Syracuse University Press.
Weinstein, Jeremy M. 2005. “Resources and the Information Problem in Rebel Recruitment.”
Journal of Conflict Resolution 49, no. 4: 598–​624.
Wickham, Carrie Rosefsky. 2011. “The Muslim Brotherhood and Democratic Transition in
Egypt.” Middle East Law and Government 3: 204–​223.
Wickham, Carrie Rosefsky. 2013. The Muslim Brotherhood: Evolution of an Islamist Movement.
Princeton: Princeton University Press.
Willis, Michael J. 2004. “Morocco’s Islamists and the Legislative Elections of 2002: The Strange
Case of the Party That Did Not Want to Win.” Mediterranean Politics 9, no. 1: 53–​81.
Wolf, Anne. 2017. Political Islamist in Tunisia: The History of Ennahda. New York: Oxford
University Press.
Wood, Elisabeth Jean. 2001. “An Insurgent Path to Democracy: Popular Mobilization,
Economic Interests, and Regime Transition in South Africa and El Salvador.” Comparative
Political Studies 34, no. 8: 863–​888.
Yom, Sean. 2017. “Jordan and Morocco: The Palace Gambit.” Journal of Democracy 28, no. 2:
132–​146.
 Pa rt V

E C ON OM IC
P E R F OR M A N C E
A N D DE V E L OP M E N T
OU TC OM E S
 Chapter 30

Re ligious Leg i t i mac y

an d L ong-​R un E c onomi c
Grow th i n t h e
Middle E ast

Jared Rubin

For centuries following the spread of Islam, the Middle East was the economic, tech-
nological, and cultural leader of western Eurasia. Middle Eastern trade flowed much
more widely, its cities were larger, its urbanization rates were higher, and its scientific
production was greater than the “Christian West” (Bosker, Buringh, and van Zanden
2013; Chaney 2016; Kuru 2019). This can easily be seen in Table 30.1, which lists the top
eight most populous cities in western Eurasia from 800 to 1800. Urban population is
one of the best proxies we have for premodern economic development. Large urban
populations were indicative of a significant agricultural surplus, and it was mostly in
urban areas where luxury goods were produced, traded, and consumed. These data
tell a tale of a reversal of fortunes. Between 800 and 1100, the only city that could rival
the Abbasid capital of Baghdad was the Byzantine capital Constantinople. The rest of
the largest cities were Muslim-​ruled, either in the Abbasid homeland (in Iraq) or in
Muslim Spain.
This began to change in the late medieval and early modern periods, and at some
point there was a reversal of fortunes. This reversal occurred well before the Industrial
Revolution set parts of Western Europe on a fundamentally different economic trajec-
tory. Middle Eastern wages fell behind those of Europe’s leading economies, European
technology well surpassed that of the Middle East, and Middle Eastern states struggled
to collect the revenues to rival their increasingly wealthy European counterparts (Mokyr
1990; Özmucur and Pamuk 2002; Karaman and Pamuk 2013). Why did this reversal of
fortunes occur? That is the question this chapter aims to shed some light on. It overviews
some of the key theories of divergence in the literature and considers how they interact
with each other to paint a fuller picture of the causes of economic divergence.
Table 30.1: Top Eight Most Populous Cities in Western Eurasia, 800–​1800
Population Population Population
City (1000s) Religion City (1000s) Religion City (1000s) Religion
800 CE 900 CE 1000 CE
Baghdad 350 Muslim Baghdad 450 Muslim Baghdad 300 Muslim
Constantinople 250 Christian Constantinople 300 Christian Constantinople 300 Christian
Basra 100 Muslim Fustat (Cairo) 150 Muslim Fustat (Cairo) 135 Muslim
Fustat (Cairo) 100 Muslim Alexandria 100 Muslim Cordoba 100 Muslim
Kayrawan 100 Muslim Cordoba 95 Muslim Sevilla 90 Muslim
Kufa 100 Muslim Basra 80 Muslim Tinnis 83 Muslim
Alexandria 95 Muslim Kufa 80 Muslim Damietta 80 Muslim
Cordoba 75 Muslim Wasit 80 Muslim Alexandria 60 Muslim
1100 CE 1200 CE 1300 CE
Baghdad 250 Muslim Baghdad 200 Muslim Paris 250 Christian
Constantinople 200 Christian Fustat (Cairo) 200 Muslim Fustat (Cairo) 220 Muslim
Fustat (Cairo) 150 Muslim Paris 110 Christian Granada 150 Muslim
Tinnis 110 Muslim Constantinople 100 Christian Venice 110 Christian
Damietta 100 Muslim Damietta 100 Muslim Damietta 108 Muslim
Sevilla 85 Muslim Fez 80 Muslim Constantinople 100 Christian
Cordoba 80 Muslim Sevilla 80 Muslim Genoa 100 Christian
Marakesh 70 Muslim Marakesh 70 Muslim Milan 100 Christian
1400 CE 1500 CE 1600 CE
Fustat (Cairo) 250 Muslim Constantinople 280 Muslim Constantinople 700 Muslim
Paris 200 Christian Paris 200 Christian Paris 300 Christian
Genoa 100 Christian Fustat (Cairo) 180 Muslim Naples 275 Christian
Granada 100 Muslim Adrianople 127 Muslim Fustat (Cairo) 250 Muslim
Tunis 100 Muslim Naples 125 Christian London 200 Christian
Venice 100 Christian Milan 100 Christian Venice 151 Christian
Prague 95 Christian Tunis 100 Muslim Adrianople 150 Muslim
Baghdad 90 Muslim Venice 100 Christian Sevilla 135 Christian
1700 CE 1800 CE
Constantinople 700 Muslim London 948 Christian
London 575 Christian Paris 550 Christian
Paris 500 Christian Constantinople 500 Muslim
Fustat (Cairo) 330 Muslim Naples 430 Christian
Naples 300 Christian Fustat (Cairo) 263 Muslim
Amsterdam 200 Christian Vienna 247 Christian
Lisbon 180 Christian Amsterdam 217 Christian
Madrid 140 Christian Dublin 200 Christian

Source: Data from Bosker et al. (2013).

622   Jared Rubin

Before addressing possible causes of divergence, it is necessary to dismiss the old

trope of there being something “inherent” in the nature of Islam that made it more
conservative, less conducive to change, or more antithetical to economic growth. This
was the standard hypothesis of Western scholars in the early 20th century (Weber
1922; von Grunebaum 1966; Lewis 1982, 2002; Landes 1998). The biggest problem such
explanations face is that they have a difficult time explaining the reversal of fortunes: if
Islam is so detrimental to economic development, how can we explain the “Golden
Age,” when the Islamic world was far ahead of Western Europe? In any case, placing the
blame on religion fails to recognize that, for all their differences, Islam and Christianity
are much closer in doctrine and practice than they are apart.
Four recent theories of divergence attempt to account for the reversal of fortunes in
a more nuanced, satisfying manner than simple explanations based on solely cultural
or religious differences. I briefly discuss them here before returning to them in the
concluding section. Timur Kuran has argued in a series of articles and his 2011 book The
Long Divergence that aspects of Islamic law held the Middle East back. His argument
does not merely regurgitate the simplistic argument that Islam is inimical to economic
success. Instead, he argues that aspects of Islamic law arose in a premodern setting and
were at one point in time conducive to economic growth. These parts of Islamic law
became stifling over time for unforeseeable reasons. Examples, which we will detail
later in this chapter, include laws on inheritance, partnerships, and legal biases favoring
Muslims. Kuran argues that these laws—​and the institutions that emerged in response
to them, such as the waqf—​stifled demand for changes to the law, even as changing eco-
nomic conditions made such changes more beneficial. This in turn held the Middle East
back just as Western Europe was beginning to re-​emerge from its long economic dol-
drums following the fall of the Western Roman Empire.
Blaydes and Chaney (2013) offer a theory of political divergence that has implications
for the economic divergence. They argue that the fact that Middle Eastern rulers had
access to slave soldiers (mamluks) meant that they did not need to bargain with other
elites in society to stay in power. In Europe, following the fall of the Western Roman
Empire, rulers were relatively weak, forcing them to bargain with local (feudal) lords
for revenue and military power. One result of these differing institutional arrangements
was that power was more dispersed in Western Europe, meaning there were fewer pow-
erful groups that were dissatisfied with the prevailing political authority. As a result,
European rulers were more secure in their tenure following the European feudal revo-
lution. Perhaps even more importantly from an economic perspective, this institutional
arrangement entailed that European leaders were more constrained in their lawmaking
than were Middle Eastern rulers. A large literature suggests that greater executive con-
straint in Europe permitted its economies to flourish (e.g., North and Weingast 1989;
Tilly 1990).
A third theory is put forth by Rubin (2017), who argues that the reversal of economic
fortunes was a result of the different degrees to which rulers employed religious legit-
imacy in the Middle East and Western Europe. This theory, which will be elaborated
upon in the remainder of this chapter, argues that religious legitimacy played an
 Religious Legitimacy and Long-Run Economic Growth    623

economically beneficial role in the centuries following the spread of Islam, as the spread
of Islamic-​based legal and economic institutions over a previously politically fractured
land helped lower transaction costs and facilitated interpersonal trust, which is essential
for economic exchange. Over time, however, reliance on political legitimacy meant that
economic elites were kept away from the political bargaining table. The opposite was
true in late medieval Western Europe, where the economic elite gained a political voice
via their role in parliaments. Rubin argues that this helped place parts of Europe (specif-
ically England and the Dutch Republic, where parliaments gained the most autonomy)
on the path to economic success. This was not because the economic elite were more
altruistic or intelligent, but because their own self-​interest tended to coincide with the
types of outcomes that yield economic success. They desired things like strong property
rights (for themselves, at least) and expensive public goods such as transport networks.
Since the English and Dutch economic elite had a much stronger seat at the political bar-
gaining table, they tended to achieve their desired outcomes.
A fourth theory, highly related to the one proposed by Rubin (2017), is put forth by
Kuru (2019). Kuru largely has the same premise as Rubin. He argues that the “ulema-​state
alliance” which formed in the eleventh century, is responsible for the ultimate economic
decline of the Middle East. Like Rubin, Kuru argues that this institutional arrangement
is not a natural state of things, but instead is something that emerged as the Golden Age
of Islam concluded. He argues that this alliance helped place Muslim intellectuals and
merchants on the sidelines, while raising the status of religious authorities. Ultimately,
there is significant overlap with Rubin—​and the theory overviewed in this chapter.
This chapter will primarily focus on laying out theories of political legitimacy, how
political legitimacy can affect economic development, and how this can help us under-
stand the economic divergence between Western Europe and the Middle East. It will
briefly conclude by recounting the four theories highlighted above, noting how they
complement each other.

A Theory of Political Legitimacy

Rubin (2017) presents a framework for understanding how political legitimacy affects
economic outcomes. Its intuition is straightforward. It considers the vexing problem
faced by any ruler: How does one stay in power and have one’s subjects follow the
preferred laws and policies? Different societies have answered this question in dif-
ferent ways, and the framework proposed by Rubin attempts to flesh out the economic
implications of how this question is answered.1
The framework begins by noting that there are two primary reasons why people might
follow a ruler: there is either some extrinsic motivation (coercion) or some intrinsic
motivation. The latter motivation is what we shall denote as legitimacy: people follow
a ruler because they believe that she or he has the right to rule and thus be obeyed. This
conceptualization of legitimacy is similar to the dominant Weberian idea of legitimacy,
624   Jared Rubin

which states that legitimacy stems from either charismatic, traditional, or legal-​rational
grounds (Weber 1947). It also builds off of the framework proposed by Lipset (1959, 86),
who defines legitimacy as “the capacity of a political system to engender and maintain
the belief that existing political institutions are the most appropriate or proper ones for
the society.”2 In other words, how one derives legitimacy is dependent on a society’s his-
tory and its informal institutions that facilitate beliefs in a ruler’s right to rule.
The Weberian definition largely views legitimacy as exogenous. That is, legitimacy is
a result of the traits of the person or the traditions of the society. While Rubin does not
deny the importance of such exogenous characteristics, his theory focuses on how legit-
imacy is endogenously generated.3
How does a ruler obtain legitimacy (and coercion)? Importantly, there are people
in society who can provide both of these. Rubin calls these legitimating (and coercive)
agents.4 For some reason in a society’s past, certain individuals—​whom we shall call
elites—​have the power to influence the beliefs of others. Indeed, this is what makes them
elite. This power can be used in many ways, including augmenting beliefs about who is
the rightful ruler. Examples include religious authorities, local lords, tribal chiefs, mil-
itary heroes, and celebrities. The endorsement of these individuals can be a powerful
means of keeping a ruler in power.
Of course, elites do not provide legitimacy or coercion free of charge. They expect
something in return. Sometimes this is something small, like tax exemptions or mon-
etary payouts. However, if these agents have a powerful enough bargaining position,
they can ask for a bigger prize: a say in the society’s laws and policies. Hence, we can
view a society’s laws and policies as an outcome of a bargaining game between a ruler
and the various agents that provide the ruler with legitimacy and coercion. Bentzen and
Gokmen (2020) show, for instance, that a society is more likely to have religious laws
if it were historically hierarchical and held widespread beliefs in punishing gods. Such
societies were likely to have institutionalized religion and given political power to reli-
gious authorities.
An interesting implication of this political bargain can be seen in medieval Egypt,
where rulers relied heavily on religious authorities for legitimacy. Chaney (2013) shows
that religious authorities gained power and were more secure in their rule when the Nile
River was either dangerously high or low (meaning flood or drought conditions). This
is precisely when a ruler would need legitimacy, and thus when the bargaining power of
the religious establishment would be greatest. But why, more generally, do political bar-
gaining outcomes differ over time and space?
The outcome of any bargain relies heavily on two features: the bargaining power of
the players and what they want to get out of the bargain. The latter is dependent on the
identities of the parties at the bargaining table. The economic elite, religious authorities,
and military all want different laws and policies. In each case, we can assume that they
desire outcomes that benefit their self-​interest. But what determines who gets a seat at
the bargaining table? And how is their bargaining power determined?
A society’s institutions determine the costs and benefits of various forms of political
legitimacy and coercion. These institutions differ by society and evolve endogenously
 Religious Legitimacy and Long-Run Economic Growth    625

over time, meaning that the task of finding the “root causes” of differing political
arrangements is inherently a historical one. For the purposes of the current framework,
we will focus on institutions that make various sources of legitimacy or coercion par-
ticularly effective or inexpensive from the perspective of the ruler. For instance, reli-
gious doctrine and institutions that are conducive to legitimating rule make religious
legitimacy more effective. Likewise, institutional features that give rulers power inde-
pendent of the elite make them less dependent on other agents to keep them in power,
thus strengthening the bargaining power of the ruler and weakening that of all others.
This is one dimension on which Blaydes and Chaney’s (2013) argument sheds signifi-
cant light. Muslim rulers had substantial power independent of other elites because they
had access to slave soldiers. Rulers typically did not need to appeal to other elites for
resources or military power, and they could thus cut these elites out of the political bar-
gaining table. On the other hand, weaker medieval European rulers depended on their
feudal lords for military service and revenue, without which rule would have been im-
possible and the threat of conflict constant. This meant that medieval European feudal
lords had substantial influence at the political bargaining table.
Political outcomes in any society are therefore determined by the identities of the
various players at the political bargaining table and their bargaining power, both of
which are in turn determined by a society’s institutional and historical past. But who
were the people who were actually at the political bargaining table in the medieval and
early modern Middle East and Western Europe? The answer to this question matters
for the outcome of the political bargain, because this outcome reflects both the bar-
gaining power and the interests of those at the table. Religious authorities desire dif-
ferent outcomes than do economic elites, who desire different outcomes than military
elites, and so on.
In both Islamic and Christian histories, religion has played an important role in
legitimating rule. There is one important reason this has been such a persistent phe-
nomenon: religious legitimacy is inexpensive relative to other forms of legitimacy or co-
ercion (Gill 1998). Military expenditures require resources, and economic elites tend to
want the “big prize”—​property rights—​as well as expensive investments like transport
networks or security (to facilitate trade). Religious authorities can often be “bought”
more cheaply; for example, tax exemptions, some policy favoring their interests (which
are often orthogonal to economic development), and suppression of rival religions
are often enough to win favor with religious authorities. Religious authorities are also
effective at legitimating rule (although the next section makes the case that their effi-
cacy differed in Islam and Christianity). There are many reasons this is the case. First
and foremost, in both Islam and Christianity, religious authorities are the primary
interpreters of religious doctrine. To the extent that the population has internalized
religious rules as those that are the appropriate ones to live by,5 this gives religious
authorities power to affect the beliefs of the population. What is pertinent here is that
this allows religious authorities to affect beliefs about the ruler’s right to rule. Moreover,
religious authorities tend to have more personal contact with the population. In the
medieval and early modern Islamic and Christian worlds, people would regularly see a
626   Jared Rubin

local religious figure, often multiple times a week. This would be the closest that many
people would come to an “elite” in their lives (although not all religious figures were
“elite”). Friday or Sunday sermons gave religious authorities a pulpit to spread ideas that
were common knowledge to everyone in attendance. Of course, this pulpit could also
be used to support the ruler. The Ottomans were well known for doing just this: it was
common that the Friday sermon would include a prayer for the sultan, thus linking him
with religious authority.6
Since religious authorities were effective at providing legitimacy while also being rel-
atively inexpensive, rulers in both Islam and Christianity brought religious authorities
to the bargaining table. This strengthened their own capacity to rule while also
strengthening their own hand vis-​à-​vis alternative sources of legitimacy and coer-
cion. Yet although religious legitimacy is relatively inexpensive, the next section makes
the case that its efficacy differed between the two religions because of differing histor-
ical circumstances. This meant that Islam was more effective at legitimating rule, and
Islamic religious authorities thus tended to have a larger seat at the political bargaining
table. The remainder of the chapter lays out the consequences of this historical and insti-
tutional arrangement.

The Lure of Islamic

Religious Legitimacy

The efficacy of religious legitimacy in Islam was a result of the circumstances under
which it was founded. Islam formed conterminously with empire—​the early Islamic
empires were among the largest (in land mass) the world had seen up to that point
(see Table 30.2). In the span of less than a century, the “Islamic world” stretched from
South Asia in the east to the Iberian Peninsula in the west. At the same time, the corpus
of Islamic law was being formulated, synthesizing existing traditions with dictates
attributed to Muhammad. Importantly, the corpus of Islamic law arose at precisely the
time that various rulers were attempting to legitimate their rule in the context of an
Islamic empire.
The first four caliphs after Muhammad (the Rashidun Caliphate) derived legitimacy
from the fact that they were Muhammad’s companions. At this time, the institutions
of Islamic law were hardly established and there was no clerical class to serve as a
legitimating agent of their rule. For this reason, the first four caliphs faced major le-
gitimacy crises: three of the four were assassinated, and succession disputes caused the
Sunni-​Shiʿa split. Without a set of agents who could bolster their legitimacy, the first
caliphs claimed religious authority for themselves. Under this conception, the fact that
they had ties to Muhammad was enough to make them the rightful rulers of all Muslims.
None of their successors could claim such ties, although the Umayyad Empire (661–​750
CE), the Sunni successors to the first four caliphs, employed the legitimating concept
 Religious Legitimacy and Long-Run Economic Growth    627

Table 30.2: Largest Empires in World History, through 1750

Peak Land Mass

Empire Birth Year Death Year (million km2) Muslim

Mongol 1206 CE 1502 CE 33.2 No

Russian 1462 CE 1795 CE 16.5 No
Umayyad 661 CE 750 CE 13.2 Yes
Qing 1644 CE 1911 CE 12 No
Qin 247 bce 209 bce 12 No
Abbasid 750 CE 861 CE 11 Yes
Rashidun 632 CE 661 CE 9 Yes

Note: replicated from Rubin (2017, Table 3.1); data source is Iyigun (2010).

that they were the “deputies of God” (khalifat Allah) to justify their right to rule (Crone
and Hinds 1986).
Unlike the stereotype of Islamic religious authority exemplified by the first four
caliphs and the Umayyad Caliphate, in which all religious authority is vested in the ca-
liph (Lewis 2002), religious authorities from the eighth century onward had authority
that was independent of the ruler. It was precisely in this period that the sunna (ex-
emplary actions of Muhammad) and hadith (reports of the teachings of Muhammad)
were codified. Ultimately, these religious doctrines were codified as a source of law, al-
though this took centuries (Hallaq 2005, especially ­chapter 5). As a result, the Islamic
legitimating concept changed as the religious establishment began to be consolidated by
the turn of the eighth century (Masud, Messick, and Powers 1996; Berkey 2003; Hallaq
2005; Coşgel et al. 2009). Leading jurists would regularly hand out opinions in accord-
ance with Islam and would employ independent reasoning (ijtihād) to mold Islamic
law in response to new exigencies. Once the clerical class consolidated, they formed a
group capable of serving as legitimating agents of Muslim rulers. This process began
in earnest in the mid-​tenth century when the weakness of the Abbasid Empire became
apparent. Their political weakness increased the value of religious legitimacy (Coşgel
et al. 2009). The preeminence of the clerical class was not achieved, however, until the
eleventh century, when the four schools of Sunni Islam consolidated, and a ruler-​cleric
alliance emerged throughout much of the (Sunni) Islamic world (Kuru 2019). On the
one hand, this was good for Muslim rulers—​independent legitimating agents are more
effective because they are not under the ruler’s thumb (Greif and Rubin 2020). On the
other hand, it meant that Muslim rulers had to cede more to religious authorities, whom
they could not control, in return for legitimacy.
Importantly, the confluence of the consolidation of the religious establishment
along with expanding empires seeking to justify their rule meant that there was much
to be gained from a unifying ideology that associated Islam with political legitimacy.
628   Jared Rubin

Consequently, there are numerous passages of Islamic doctrine that demand obedi-
ence to a ruler who acts in accordance with Islam. This is perhaps most vividly seen in
the hadith of al-​Bukhari, which are among the most important and influential in Islam
(quoted in Rubin 2017, 52):

The Prophet said, “A Muslim has to listen to and obey (the order of his ruler) whether
he likes it or not, as long as his orders involve not one in disobedience (to Allah), but
if an act of disobedience (to Allah) is imposed one should not listen to it or obey it.
(Vol. 9, Book 89, No. 258)

This is not to say that doctrine is the reason that Muslim religious authorities have the
capacity to legitimate rule. Doctrine can always be interpreted to fit the needs of the day,
as indeed it has in both Islamic and Christian history (Noonan 2005). In fact, most reli-
gious doctrine should be viewed as an outcome of a bargaining process between religious
authorities and other interested parties (political authorities, economic elites, military
elites, etc., depending on the case; see Rubin 2011 for a theoretical exposition). Islamic
doctrine in support of associating political legitimacy with “being a good Muslim” is
no different. It was a manifestation of the institutional environment that pervaded the
earliest Islamic polities: the Rashidun Caliphate, the Umayyad Empire, and the Abbasid
Empire.
Even though the institutions of later Muslim empires were different from their
predecessors, their institutions did not arise in a vacuum. The elements that worked
in the past tended to be used in the future, albeit in different forms. Indeed, Turkic
empires like the Ottomans could clearly not claim a bloodline to the Prophet (as could,
say, the Rashidun and Umayyad caliphs), but were clearly aware of the potency of re-
ligious legitimacy. The Ottomans brought the religious establishment into the state. In
return for their patronage, the sultanate received fatwas from their Grand Muftis on all
sorts of issues. This gave the sultan religious cover for what would have otherwise been
controversial actions. The point here is simply that the Ottomans (and many other
medieval and early modern Muslim states) were acting in accordance with a long-​
established tradition of appealing to religious authority to legitimate various aspects of
their rule.
This stands in stark contrast with Christianity. While it is not worth going into early
Christian history in detail, the differences of the origins of the two religions are key to
why Islam is more effective at legitimating rule than Christianity.7 The latter was born in
the Roman Empire, which was hardly a newly expanding empire. The Roman political
elite had numerous mechanisms for legitimating their rule and did not need the support
of a fledgling religion. In any case, Christianity was a minority religion for centuries.
For this reason, early Christian doctrine is largely silent on the capacity of religion to
legitimate secular power, and where it does speak its overarching premise is that secular
authorities rule over their own sphere, separate from that of God. The most famous pas-
sage along these lines is attributable to Jesus: “Render unto Caesar the things which are
Caesar’s, and unto God the things that are God’s” (Matthew 22:21).8
 Religious Legitimacy and Long-Run Economic Growth    629

The fact that Islam is so effective at legitimating rule means that it has historically
been—​and continues to be today—​attractive for rulers to use to legitimate their rule.
Consider what this means in the context of the framework laid out in the previous sec-
tion. In that framework, rulers choose to negotiate with agents who can provide some
combination of legitimacy and coercion based on the costs and benefits that each can
provide. Since Islamic doctrine is so explicit in stating that one should follow a ruler who
“acts in accordance with Islam,” Muslim religious authorities are an extremely effective
source of legitimacy. They are, after all, the ones who can determine what it means to act
in accordance with Islam. This is all the more the case when compared with Christianity,
for which there is less doctrinal and institutional basis for legitimating rule. This is not
to say that Christianity cannot be used to legitimate rule—​it clearly has been used in
that way, especially in the medieval period. It is just that Islam and its institutions are
more effective. This means that the benefits to rulers of negotiating with Muslim clerics
exceeds that of negotiating with Christian clerics. In the context of the framework
proposed previously, this means that (all else being equal) Muslim rulers will employ re-
ligious legitimacy to a greater degree than their Christian counterparts. In turn, Muslim
religious authorities will have a greater seat at the political bargaining table, and their
desires will be reflected to a greater extent in the laws and policies of society.
To be clear, this does not mean that “church-​ state separation” is inherent in
Christianity or inherently opposed in Islam. Such separation, or lack thereof, is a choice
made by various elites at various points in time, subject to the institutional environ-
ment they find themselves in and the incentives provided by these institutions. It does
mean, however, that Muslim rulers historically had greater incentive to bring religious
authorities to the political bargaining table. This can explain the relatively large role of
the clerical class in historical Muslim polities. But there is nothing in the origins of Islam
(or Christianity) that predetermined this outcome. What this means for the long-​run
economic trajectory of Muslim societies is the issue we will turn to next.

Religious Legitimacy and the

“Long Divergence”

The theoretical framework proposed earlier, combined with the historical “lure” of
Islamic religious legitimacy overviewed in the previous section, sheds light on nu-
merous aspects of the “Long Divergence” between the economies of the Middle East
and Western Europe. After the Sunni religious establishment consolidated in the eighth
and ninth centuries and the major schools of Islamic thought solidified by the end of
the tenth century, reinterpretation of Islamic law slowed in spite of changing economic
circumstances. Western scholars denoted this phenomenon as the “closing of the gate
of ijtihād” (independent reasoning). While there is plenty of evidence that the gate
was hardly closed in theory or in practice, there is also evidence, overviewed by Rubin
630   Jared Rubin

(2011, 2017), that reinterpretation did slow after this period. Yet, as the framework above
emphasizes, this was not due to some inherent “conservatism” or backward-​looking na-
ture of Islam. Instead, it can be seen as an equilibrium outcome in which rulers derived
legitimacy from clerics, clerics received favorable laws and policies in return, and the
masses risked transgressing both secular and religious authorities should they violate
religious law. In this interpretation, the “gate of ijtihād” may have been closed (or ajar),
but it was not locked. The political equilibrium simply dictated that there was little de-
mand for change—​and thus little change was supplied.
This outcome was reinforced and institutionalized in the eleventh century with
the rise of madrasas as centers for Islamic learning (Kuru 2019). The rise of madrasas
signaled a changing landscape, one that favored religious learning at the expense of sci-
entific learning. Yet the increased importance of madrasas was not a cause of Islamic
scientific decline; it was rather a manifestation of the changing political equilibrium that
occurred over the previous century, in which religious legitimacy gained paramount
importance. The rise of madrasas had numerous important implications for Muslim ec-
onomic and scientific fortunes. Chaney (2016), in an extensive study of Islamic book
production, finds that Islamic production of scientific works well outpaced that of
Christendom during the “Islamic Golden Age,” roughly the period from the seventh
through tenth centuries. However, consistent with the theory espoused above, there
was significant decline in Islamic scientific output beginning in the eleventh and twelfth
centuries, as religious authorities gained greater political power throughout the Muslim
world. Chaney notes that the rise of the madrasa system and the increased emphasis on
religious works are the likely culprit. This, in turn, was a direct result of the increased
importance of religious legitimacy in Muslim polities in this period. This finding is con-
sistent with the one proposed by Kuru (2019), who argues that the strengthening of the
cleric-​state alliance in the eleventh century was responsible for the slowing down (al-
though not elimination) of Muslim scientific and philosophical thought. The study of
Blaydes, Grimmer, and McQueen (2018) reveals a related pattern. They find that texts
offering advice on governance in the Muslim world increasingly emphasized religion
over time, with an inflection point in the eleventh–​thirteenth centuries. Meanwhile, late
medieval European texts placed a decreasing emphasis on religious appeals over time.
All of these findings suggest a rise in the importance of religious legitimacy toward the
end of the Islamic Golden Age, which was reinforced and institutionalized by the spread
of the madrasa system.
But why did this matter for economic outcomes? The framework proposed above
indicates one possible reason. To the extent that religious authorities kept their place
at the political bargaining table—​at the expense of the economic elite—​by refraining
from accepting economically beneficial laws and policies, a dampening of economic ac-
tivity should follow. There is reason to believe this is what happened. Among the most
important mechanisms through which Muslim religious authorities kept their hold on
power was their monopoly over numerous aspects of law, including commercial law, via
Islamic courts. Kuran (2005, 2011) spells out numerous unforeseeable implications of
this arrangement. One consequence resulted from Islamic law regarding partnerships
 Religious Legitimacy and Long-Run Economic Growth    631

and inheritance. Under Islamic partnership law, partnerships were disbanded upon the
death of any member. This would only be a problem if the partnership were not imme-
diately reformed; if it were not, the fixed (sunk) costs associated with the venture would
be lost. However, Islamic inheritance laws placed a major impediment to partnerships
reforming after the death of a member. The law provided specific guidance for who the
rightful heirs of one’s estate were, and the split was often among many individuals. If any
of these heirs needed their inheritance for a pressing issue, they might not be willing to
reinvest their inheritance in the partnership. As Kuran notes, this dampened incentives
for Muslim merchants and others engaged in commerce to enter into long-​lived
partnerships with numerous individuals in the first place. Kuran contrasts this with
Europe, where such impediments were not in place. Over time, marginal changes made
to the partnerships and partnership law resulted in vastly more complex organizational
forms, eventually culminating in the corporation.
There are many more unintended and unforeseeable economic consequences of the
equilibrium in which Islamic law provided the dominant form of commercial law. For
instance, Balla and Johnson (2009) note that the absence of the corporate form in the
Islamic world meant that Ottoman tax farmers had no means of banding together to
constrain sultans if the sultan encroached on their property rights (by confiscating
their tax farms). Balla and Johnson compare this to the situation in France at the time,
in which tax farmers did precisely this in order to have some bargaining power vis-​
à-​vis the autocratic Bourbon crown. This was a consequence of Islamic law, which
disincentivized individuals from building large and long-​lasting enterprises. Kuran
and Lustig (2012) show another negative effect that Islamic law had on commerce: it
was biased towards individuals with some degree of power. In particular, they find that
those with elite titles were more likely to win court cases, holding constant a number
of factors that might affect the merit of the case. Kuran and Rubin (2018) show that
these biases directly affected the interest rates paid by those with privilege in Islamic
courts—​not just the elites, but also men and Muslims, both of whom are favored by
various proscriptions in Islamic law. They show that the privileged paid a price for
their privilege: the interest rates they paid on private loans were 2–​4 percentage points
higher than their less privileged counterparts. Since the privileged were the most likely
to have the means to invest in capital and business enterprises, the fact that they paid a
higher cost for credit must have had a negative effect on the long-​run trajectory of the
Ottoman economy. Rubin (2010) shows yet another consequence of the dominance of
Islamic law in commerce: the Islamic bill of exchange (suftaja) remained an instru-
ment for merchants who wished to avoid the costs of carrying specie. This is only an
obvious impediment to economic growth when seen in the light of the evolution of
the European bill of exchange, which transformed into the most important financial
instrument of late medieval Europe. The schools of Islamic law that considered the
suftaja legal in the first place forbade their use as an instrument of finance, since this
would have been too obviously in violation of the usury prohibition. The framework
described earlier in this chapter suggests why such laws persisted for so long despite
being economically harmful.
632   Jared Rubin

More generally, the economic elite had little seat at the political bargaining table for
much of Middle Eastern history, especially after the consolidation of the clerical class in
the tenth and eleventh centuries (Kuru 2019). This was particularly true in the Ottoman
Empire (1299–​1922), where the religious establishment was brought into the apparatus
of the state. This provided great flexibility for the Ottoman ruling elite—​the Grand Mufti
would generally find a way to make the sultan’s desired policies consistent with some
interpretation of Islam (Imber 1997)—​even if it meant that the religious establishment
was less independent and thus had a weaker capacity to legitimate rule. As a result, the
Ottoman economic elite—​merchants, money-​changers, manufacturers—​had little say
in governance (Pamuk 2004a, 2004b). The reason is simple, and it follows directly from
the framework proposed above: if Ottoman rulers did not have to bargain with the ec-
onomic elite in order to have legitimacy or resources, there was little incentive for them
to do so. As a result, the desires of the religious establishment were reflected in Ottoman
laws and policies. One example of the consequences of this arrangement will suffice.
Coşgel, Miceli, and Rubin (2012a) argue that the Ottomans failed to adopt the printing
press for nearly 250 years, despite knowing about it, precisely because it would have
undermined the monopoly of the clerical class on religious knowledge (a suspicion that
proved prescient, as suggested by the Protestant Reformation). The Ottoman sultanate
made printing in the Arabic script punishable by death. It is clear why the Ottomans
would forgo the “free lunch” offered by the new and important technology in light of the
framework proposed above. If the press threatened the power of one of the sultan’s key
legitimating agents, the sultan stood to lose an important source of his own power by
permitting the press.
The failure to give a seat to the economic elite at the political bargaining table was
the fundamental reason the Middle East fell behind Western Europe. In the leading
European economies of the early modern period (England and the Dutch Republic),
a series of events dramatically increased the bargaining power of economic elites in
parliaments. Not surprisingly, a series of laws and policies aligned with their interests
followed, including stronger protections of property rights; investments in public goods
such as canals, poor relief, and turnpikes; and the strengthening of the rule of law. To
be clear, the economic elite did not bargain for these laws out of some altruistic motive.
They desired these laws for self-​interested reasons. It just so happens that many (but not
all) of the types of laws and policies desired by the economic elite were also beneficial for
economic success.
Meanwhile, the dominant Middle Eastern power of the period—​ the Ottoman
Empire—​slowly fell behind its European rivals. This outcome was not obviously
preordained. For most of the fifteenth and sixteenth centuries, the Ottomans were just as
powerful as the leading European powers and not obviously behind economically. This
changed in the seventeenth and eighteenth centuries, and by the nineteenth century
the Ottoman Empire was known as the “sick man of Europe.” This chapter suggests that
there were deep historical roots to this reversal of fortunes. The failure of the Ottomans
(and other Middle Eastern polities) to bring the economic elite to the political bar-
gaining table was not a foolish mistake, nor was it a stroke of bad luck. The Ottoman
 Religious Legitimacy and Long-Run Economic Growth    633

ruling establishment was simply incentivized to not bring the economic elite to the bar-
gaining table. What began as small economic differences were therefore magnified over
time, and by the nineteenth century there was a clear economic and technological diver-
gence between the Middle East and Western Europe.

Bringing the Theories Together

In the introduction, I noted that there have been four recent theories of the economic
divergence between Western Europe and the Middle East: those proposed by Timur
Kuran, Lisa Blaydes and Eric Chaney, myself, and Ahmet Kuru. I conclude this chapter
by considering how these theories fit with each other.
It is my impression that the four theories complement each other. The most obvious
similarity between the four is that all focus on the role of politics in the economic diver-
gence. The closest connection is between Rubin (2017) and Kuru (2019), who both focus
on the role that religious legitimacy has played in the political and economic fortunes
of the Middle East. They differ a bit on why religious elites were able to achieve such
power—​Rubin focuses on the institutions that emerged in early Muslim states, while
Kuru focuses on the institutionalization of madrasas in the eleventh century—​but both
recognize the role that religious authorities can play in dampening economic devel-
opment when they are given a position of political power. The theory of Blaydes and
Chaney (2013) similarly focuses on how rulers are constrained and what that means
for economic development. Unlike Rubin and Kuru, whose primary (though not sole)
focus is on religious authorities, Blaydes and Chaney focus on the role that military elites
play in constraining rule. They emphasize the important distinction between European
feudal lords and Muslim slave soldiers and their capacities to constrain and negotiate
with their rulers. This theory is highly complementary to the ones proposed by Rubin
and Kuru. Religious authorities were not the only group to constrain medieval and early
modern rulers. Nor were military elites. Hence, while none of these three theories paint
a complete picture of Middle Eastern and Western European governance, combined
they come much closer to doing so. Finally, the arguments of Kuran (2011) reveal one
of the most important reasons why religious legitimation had such a dampening effect
on economic activity. In this light, Kuran is most closely related to Rubin and Kuru.
As discussed above, the fact that religious authorities (ulema [or ʿulamaʾ]) had pur-
view over most aspects of commercial law for much of the medieval and early modern
periods had numerous unforeseeable, yet damaging effects, on economic development.
Kuran highlights why this was the case. Yet it was the very fact that religious authorities
played such an important political role that permitted them to maintain their purview
over commercial law in the first place.
In short, each of the four theories has something unique to say about the long-​run
economic divergence between Western Europe and the Middle East. Such a long-​run,
wide-​spanning event is bound to have multiple causes, and these theories focus on three
634   Jared Rubin

of those causes. In fact, each theory helps shed some light on the others. Combined, they
provide a complex, nondeterministic story for why the Middle East fell behind Europe
despite being so far ahead in the centuries following the initial spread of Islam.

Notes
1. For a similar framework that also considers how the cultural/​religious evolution of a society
interacts with legitimating and institutional changes in that society over time, see Bisin et al.
(2020).
2. For similar definitions of political legitimacy, see Razi (1990); Levi, Sacks, and Tyler (2009);
Greif and Tadelis (2010); and Greif and Rubin (2020).
3. Greif and Rubin (2020) expand significantly on the notion of endogenously generated legit-
imacy. Their paper, however, primarily focuses on early modern England.
4. For more on the theory of legitimating agents, see Coşgel, Miceli, and Rubin (2012a, 2012b).
5. The degree to which society is religious is endogenous to the degree to which religious
authorities have political power. See Bisin et al. (2020) for a theoretical exposition of this
insight.
6. Religious authorities in both Islam and Christianity had many other sources of power vis-​
à-​vis local populations. For one, they were the primary providers of poor relief. They could
also be used to maintain social control, especially during periods of political turmoil.
7. For much more on the differences between the legitimating conceptions in Islam and
Christianity and their effects on economic outcomes, see Rubin (2011, 2017).
8. Platteau (2017) and Auriol and Platteau (2017) argue that a different, but related, institu-
tional difference between Islam and Christianity is salient with respect to economic and po-
litical outcomes: the (de)centralization of religious institutions. Their theoretical insight is
that in a decentralized system, like that of Sunni Islam, a ruler will have to co-​opt (for legiti-
macy) the marginal cleric, leaving some, more radical, clerics outside of the ruling coalition.
Meanwhile, under a centralized system, like Catholicism, rulers only have to co-​opt the av-
erage cleric, and the church will make sure all others fall in line. One insight that stems from
this logic is that Muslim rulers are more likely to have challenges from radical authorities.
This means that inter-​clerical competition could (and did) arise in Islam, but only those
clerics sufficiently in line with the state would be co-​opted by the state. Platteau and Auriol’s
arguments are hardly inconsistent with the ones proposed in this chapter. Both view reli-
gious legitimacy as an essential part of Middle Eastern and European political history, and
they provide complementary explanations regarding how legitimacy affects political and
economic outcomes.

References
Auriol, Emmanuelle, and Jean-​Philippe Platteau. 2017. “Religious Co-​option in Autocracy: A
Theory Inspired by History.” Journal of Development Economics 127: 395–​412.
Balla, Eliana, and Noel D. Johnson. 2009. “Fiscal Crisis and Institutional Change in the
Ottoman Empire and France.” Journal of Economic History 69, no. 3: 809–​845.
Bentzen, Jeanet, and Gunes Gokmen. 2020. “The Power of Religion.” CEPR discussion
paper 14706.
 Religious Legitimacy and Long-Run Economic Growth    635

Berkey, Jonathan P. 2003. The Formation of Islam: Religion and Society in the Near East, 600–​
1800. Cambridge: Cambridge University Press.
Bisin, Alberto, Jared Rubin, Avner Seror, and Thierry Verdier. 2020. “Culture, Institutions and
the Long Divergence between Western Europe and the Middle East.” Chapman University
Working Paper.
Blaydes, Lisa, and Eric Chaney. 2013. “The Feudal Revolution and Europe’s Rise: Political
Divergence of the Christian West and the Muslim World before 1500 CE.” American Political
Science Review 107, no. 1: 16–​34.
Blaydes, Lisa, Justin Grimmer, and Alison McQueen. 2018. “Mirrors for Princes and
Sultans: Advice on the Art of Governance in the Medieval Christian and Islamic Worlds.”
Journal of Politics 80, no. 4: 1150–​1167.
Bosker, Maarten, Eltjo Buringh, and Jan Luiten van Zanden. 2013. “From Baghdad to
London: Unraveling Urban Development in Europe, the Middle East, and North Africa,
800–​1800.” Review of Economics and Statistics 95, no. 4: 1418–​1437.
Chaney, Eric. 2013. “Revolt on the Nile: Economic Shocks, Religion, and Political Power.”
Econometrica 81, no. 5: 2033–​2053.
Chaney, Eric. 2016. “Religion and the Rise and Fall of Islamic Science.” Working Paper.
Coşgel, Metin M., Thomas J. Miceli, and Rasha Ahmed. 2009. “Law, State Power, and Taxation
in Islamic History.” Journal of Economic Behavior and Organization 71, no. 3: 704–​7 17.
Coşgel, Metin M., Thomas J. Miceli, and Jared Rubin. 2012a. “The Political Economy of
Mass Printing: Legitimacy and Technological Change in the Ottoman Empire.” Journal of
Comparative Economics 40, no. 3: 357–​371.
Coşgel, Metin M., Thomas J. Miceli, and Jared Rubin. 2012b. “Political Legitimacy and
Technology Adoption.” Journal of Institutional and Theoretical Economics 168, no. 3: 339–​361.
Crone, Patricia, and Martin Hinds. 1986. God’s Caliph: Religious Authority in the First Centuries
of Islam. Cambridge: Cambridge University Press.
Gill, Anthony. 1998. Rendering unto Caesar: The Catholic Church and the State in Latin America.
Chicago: University of Chicago Press.
Greif, Avner, and Jared Rubin. 2020. “Political Legitimacy and the Institutional Foundations
of Constitutional Government: The Case of England.” Stanford University Working Paper.
Greif, Avner, and Steven Tadelis. 2010. “A Theory of Moral Persistence: Crypto-​morality and
Political Legitimacy.” Journal of Comparative Economics 38, no. 3: 229–​244.
Hallaq, Wael B. 2005. The Origins and Evolution of Islamic Law. Cambridge: Cambridge
University Press.
Imber, Colin. 1997. Ebu’s-​su‘ud: The Islamic Legal Tradition. Stanford, CA: Stanford University Press.
Iyigun, Murat. 2010. “Monotheism (From a Sociopolitical & Economic Perspective).”
University of Colorado Working Paper.
Karaman, K. Kivanc, and Şevket Pamuk. 2013. “Different Paths to the Modern State in
Europe: The Interaction between Warfare, Economic Structure, and Political Regime.”
American Political Science Review 107, no. 3: 603–​626.
Kuran, Timur. 2005. “The Absence of the Corporation in Islamic Law: Origins and Persistence.”
American Journal of Comparative Law 53: 785–​834.
Kuran, Timur. 2011. The Long Divergence: How Islamic Law Held Back the Middle East.
Princeton, NJ: Princeton University Press.
Kuran, Timur, and Scott Lustig. 2012. “Judicial Biases in Ottoman Istanbul: Islamic Justice
and Its Compatibility with Modern Economic Life.” Journal of Law and Economics 55, no.
3: 631–​666.
636   Jared Rubin

Kuran, Timur, and Jared Rubin. 2018. “The Financial Power of the Powerless: Socio-​Economic
Status and Interest Rates under Partial Rule of Law.” Economic Journal 128, no. 609: 758–​796.
Kuru, Ahmet T. 2019. Islam, Authoritarianism, and Underdevelopment: A Global and Historical
Comparison. Cambridge: Cambridge University Press.
Landes, David S. 1998. The Wealth and Poverty of Nations: Why Some Are So Rich and Some So
Poor. New York: Norton.
Levi, Margaret, Audrey Sacks, and Tom Tyler. 2009. “Conceptualizing Legitimacy, Measuring
Legitimating Beliefs.” American Behavioral Scientist 53, no. 3: 354–​375.
Lewis, Bernard. 1982. The Muslim Discovery of Europe. New York: Norton.
Lewis, Bernard. 2002. What Went Wrong? The Clash between Islam and Modernity in the Middle
East. New York: HarperCollins.
Lipset, Seymour Martin. 1959. “Some Social Requisites of Democracy: Economic Development
and Political Legitimacy.” American Political Science Review 53, no. 1: 69–​105.
Masud, Muhammad K., Brinkley Messick, and David S. Powers. 1996. “Muftis, Fatwas, and
Islamic Legal Interpretation.” In Islamic Legal Interpretation: Muftis and their Fatwas, ed-
ited by M. K. Masud, B. Messick, and D. S. Powers, 3–​32. Cambridge, MA: Harvard
University Press.
Mokyr, Joel. 1990. The Lever of Riches. Oxford: Oxford University Press.
Noonan, John T. 2005. A Church That Can and Cannot Change: The Development of Catholic
Moral Teaching. Notre Dame, IN: University of Notre Dame Press.
North, Douglass C., and Barry R. Weingast. 1989. “Constitutions and Commitment: The
Evolution of Institutional Governing Public Choice in Seventeenth-​Century England.”
Journal of Economic History 49, no. 4: 803–​832.
Özmucur, Süleyman, and Şevket Pamuk. 2002. “Real Wages and Standards of Living in the
Ottoman Empire, 1489–​1914.” Journal of Economic History 62, no. 2: 293–​321.
Pamuk, Şevket. 2004a. “The Evolution of Financial Institutions in the Ottoman Empire, 1600–​
1914.” Financial History Review 11, no. 1: 7–​32.
Pamuk, Şevket. 2004b. “Institutional Change and the Longevity of the Ottoman Empire, 1500–​
1800.” Journal of Interdisciplinary History 35, no. 2: 225–​247.
Platteau, Jean-​ Phillipe. 2017. Islam Instrumentalized: Religion and Politics in Historical
Perspective. Cambridge: Cambridge University Press.
Razi, G. Hossein. 1990. “Legitimacy, Religion, and Nationalism in the Middle East.” American
Political Science Review 84, no. 1: 69–​91.
Rubin, Jared. 2010. “Bills of Exchange, Interest Bans, and Impersonal Exchange in Islam and
Christianity.” Explorations in Economic History 47, no. 2: 213–​227.
Rubin, Jared. 2011. “Institutions, the Rise of Commerce, and the Persistence of Laws: Interest
Restrictions in Islam and Christianity.” Economic Journal 557: 1310–​1339.
Rubin, Jared. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the Middle East Did
Not. New York: Cambridge University Press.
Tilly, Charles. 1990. Coercion, Capital, and European States, AD 990–​1990. Oxford: Blackwell.
von Grunebaum, Gustave E. 1966. Medieval Islam: A Study in Cultural Orientation.
Chicago: University of Chicago Press.
Weber, Max. 1922. Economy and Society: An Outline of Interpretive Sociology. Berkeley: University
of California Press.
Weber, Max. 1947. The Theory of Social and Economic Organization. New York: Simon & Schuster.
 Chapter 31

Isl am and Ec onomi c

Devel op me nt
The Case of Non-​Muslim Minorities in the Middle East
and North Africa

Mohamed Saleh

The conjecture that Islam has a detrimental impact on economic development in the
Middle East and North Africa (henceforth, MENA) is an old one, dating back to the
pioneering European scholarship on the region in the nineteenth century. During
the twentieth century, Bernard Lewis put this conjecture in a more concrete form in
his influential study of the region (Lewis 2002). There are typically two main historical
puzzles to which this conjecture offers an answer. First, MENA is much poorer than
Europe, which is known as the “Divergence Puzzle”: After the so-​called Golden Age of
Islam, which lasted roughly from the seventh to the eleventh century, why did MENA
lag behind Europe at some point in the early modern period, or during the late Middle
Ages? Second, within MENA, Muslims tend to be poorer, more rural, and less educated,
on average, than the native (non-​European) non-​Muslim minorities, such as Coptic
Christians, Armenians, Greek Orthodox, Levantines, Rabbinic Jews, and Karaite Jews.
The “Islam-​based” set of explanations argues that there is something about Islam, either
as a set of beliefs or as a set of institutions, that hindered the development of the MENA
region as a whole, or at least of its Muslim population, vis-​à-​vis Europe or non-​Muslim
minorities in MENA.
An alternative conjecture traces the two phenomena to European colonialism.
Instead of tracing the cause of MENA’s or Muslims’ underdevelopment to Islam, broadly
defined, this conjecture seeks the cause in colonialization. Due to Europe’s increased
influence in the Ottoman Empire starting from the eighteenth century, and the subse-
quent direct colonization of the region in the nineteenth and twentieth centuries, the
638   Mohamed Saleh

region was kept behind Europe. European powers also favored the region’s non-​Muslim
minorities in various domains, including the capitulations and access to European legal
structures (Issawi 1981; Kuran 2004), as well as access to European missionary or secular
schooling. The colonialism explanation tends to be the preferred explanation by the na-
tionalist postcolonial historiography in the MENA region.
While both conjectures have their own merits and are consistent with some of the em-
pirical facts, there is a surmounting need for empirical evidence that puts these theories
under close scrutiny. This chapter thus has three objectives. First, it attempts to pro-
vide the reader with a critical evaluation of these two theories by focusing on evidence
on the second puzzle: the socioeconomic (henceforth, SES) inequality between non-​
Muslim minorities and the Muslim majority in the MENA region. The chapter’s argu-
ment is based on the author’s published research on SES inequality between Copts and
Muslims in Egypt, the most populous MENA country, with its Coptic Christian pop-
ulation being the largest non-​Muslim population (in absolute number) in the region.
Using a wide range of novel primary data sources, I argue that the Coptic-​Muslim SES
gap cannot be explained by a causal negative impact on Muslims’ economic develop-
ment of “Islam,” defined as a set of beliefs (cultural explanation) or a set of institutions
(institutional explanation). Nor is it explained by the negative impact of colonization on
Muslims. Instead, I trace the phenomenon to selection on SES during Egypt’s conver-
sion from Coptic Christianity to Islam following the Arab Conquest of the then-​Coptic
Egypt in 641 CE. Specifically, Arabs introduced a tax system that imposed a poll tax on
every adult male Copt (initially, all Egyptians), which was removed only if he converted
to Islam, and this poll tax remained until 1856. Because of the (quasi) lump-​sum na-
ture of the poll tax, it arguably led to the shrinkage of non-​convert Copts into a better-​
off minority. I trace the subsequent persistence of the Coptic-​Muslim SES gap to group
restrictions on access to artisanal and white-​collar occupations, which led the initial
positive selection of non-​converts to perpetuate.
Two remarks are in order. First, while Islamic taxation is an Islamic institution, the
tax-​induced conversion explanation of the Coptic-​Muslim SES gap differs in an impor-
tant way from the institutional explanation that traces Muslims’ underdevelopment to
“Islamic institutions.” Specifically, the impact of taxation on the Coptic-​Muslim SES
gap operates only via screening/​sorting of converts on SES during the historical for-
mation phase of religious groups (Muslims and Copts). However, Islamic taxation
does not impact converts’ SES, once they become Muslims. This is in contrast to the
institutional hypothesis that we find in the previous literature (e.g., Kuran 2004), where
Islamic institutions affect Muslims’ SES, taking their religious affiliation itself as exoge-
nous. Second, taxation is not the sole cause of conversions or of the Coptic-​Muslim SES
gap. There are many other processes that impacted the conversion process, and that al-
tered the Coptic-​Muslim SES gap, such as modern schooling and state industrialization.
However, unlike the other explanations, the tax-​induced conversion explanation offers a
coherent explanation to both conversions and the Coptic-​Muslim SES inequality. More
importantly, it emphasizes the endogeneity of the formation of religious groups, and the
ever-​changing membership in these groups via conversions. This is often an overlooked
 Islam and Economic Development     639

fact in the recent literature on the economics of religion, which, in its attempt to identify
the causal impact of religion on economic outcomes, tends to understudy the (self-​)se-
lection of individuals into religions.
The second objective of the chapter is to highlight the importance of relying on pri-
mary archival sources and the need for large-​scale research projects that digitize
the unearthed and under-​studied data sources of the MENA region. There is a long-​
standing belief among researchers in Western academia that MENA is poor with re-
spect to data. This is an unfounded belief. By and large, MENA possesses rich sources
of data that are even on par with Western Europe and North America. Specifically, the
chapter employs a wide range of novel data sources for Egypt (digitized by the author),
including papyrological poll tax records in 641–​1100, a dataset on occupations and re-
ligious affiliation constructed from the papyri in 641–​969, a dataset on locations of
Christian churches and monasteries in 1200 and 1500, and two nationally representa-
tive individual-​level samples from Egypt’s population censuses of 1848 and 1868. It also
employs the subsequent village-​level population census reports in 1897–​1986 (CEDEJ
2003), and individual-​level samples of Egypt’s population censuses of 1986, 1996, and
2006 (available on IPUMS-​International).
The third and final objective of the chapter is to introduce the reader (in the
Conclusion section) to future areas of research and other archival data sources that
remain undigitized and can potentially shed light, not only on the question of inter-​
religion SES inequality in the MENA region, but also on within-​group inequality.

Copts’ Population Share and

the Coptic-​Muslim SES Gap

Egypt was almost 100 percent Copt on the eve of the Arab Conquest in 641 CE. Tracing
Copts’ population share, and the Coptic-​Muslim SES gap, is feasible for the period 1800–​
2000 using both archival and published population census data. Tracing both variables
between 641 and 1800 is more difficult, though, in the absence of population censuses,
which requires using nonstandard primary data sources. I discuss in this section the de-
scriptive statistics in 1800–​2000, before proceeding to the pre-​1800 period.

Descriptive Statistics in 1800–​2000

The proportion of non-​Muslim minorities in the MENA region is a topic of heated and
probably never-​ending debate. In Egypt, the widely accepted figure of 10 percent is not
based on any population census or household survey. In Lebanon, the last population
census was conducted in 1932, due to the inter-​religious tensions over the population
share of each religious group.
640   Mohamed Saleh

In a similar vein, the socioeconomic superiority of the native non-​Muslim minorities

in the region in comparison to the Muslim majority has been long noted, albeit qualita-
tively, by various scholars, including Tagher ([1951] 1998), Issawi (1981), and Courbage
and Fargues (1997). Copts, Maronites, Greek Orthodox and Catholic Levantines,
Armenians, and Jews all tend to outperform Muslims in educational and occupational
attainment during the nineteenth and twentieth centuries. They tend to have better
health outcomes and to have witnessed the mortality and fertility decline earlier than
Muslims (Courbage and Fargues 1997).
Understanding the facts about this phenomenon—​its existence, magnitude, causes,
and how it was impacted by colonization and state-​led modernization over the last two
centuries—​is of major interest. This is not driven by mere historical curiosity, but also
because of its implications for today’s economic and political problems. Religious affilia-
tion is still among the major sources of social segmentation in the region. In many ways,
it is similar to race in the Americas. It is an inherited social marker that individuals do
not typically choose. Inter-​religion tensions, often caused by conflict over economic re-
sources, have recently resurfaced and have made the democratic transition following
the uprisings in 2010–​2011 challenging (along with the Sunni-​Shiʿite division among
Muslims).1
Yet we lack comprehensive individual-​level evidence on this phenomenon. What
we have are historical narratives that tend to over-​represent the urban elite of each re-
ligious group. We also have scattered statistics based on the published aggregate sta-
tistics in population census reports or statistical yearbooks that mostly come from the
colonial/​mandate period. Even the latter sources remain for the most part undigitized
and under-​studied. Saleh (2013) attempted to fill this gap in the literature by digitizing
two nationally representative individual-​level samples of the Egyptian population
censuses of 1848 and 1868, from the original Arabic manuscripts that are preserved at
the National Archives of Egypt. These are the earliest precolonial modern population
censuses in the MENA region. They include information on every household member,
including females, children, and slaves, and not only adult free males, as was the case
with the Ottoman tax registers from the period.2 They are in fact two of the earliest pre-
colonial population censuses from any non-​Western country.
Using this source, Saleh (2015) documents the educational and occupational
differences across religious groups in mid-​nineteenth-​century Egypt. Egypt was an
autonomous Ottoman province at the time, under the rule of Muhammad Ali Pasha
(reigned from 1805 to 1848) and his successors. The two censuses come from the pre-
colonial period before the British colonization of Egypt in 1882. About 7 percent of the
population was non-​Muslim; 94 percent of these were Coptic Christians, 4 percent were
non-​Coptic Christians who included Levantines, Greeks, and Armenians, and 2 percent
were Jews, both Rabbinic and Karaite.
The population census samples reveal that school enrollment in 1848 was the highest
among Jews, followed by non-​Coptic Christians, Coptic Christians, and Muslims. The
data also reveal that non-​Muslims were over-​represented among white-​collar workers
and artisans and under-​ represented among farmers and unskilled (non-​ farmer)
 Islam and Economic Development     641

workers. But there were important differences across non-​Muslim minorities. Whereas
non-​Coptic Christians and Jews were entirely urban minorities, Copts were mostly
rural. And while the occupational advantage of non-​Coptic Christians and Jews over
Muslims stemmed from their over-​representation among merchants, moneychangers,
and jewelers, Copts’ occupational advantage stemmed from their over-​representation
in the fiscal bureaucracy, in both urban and rural Egypt, and in artisanship, such as
jewelers, carpenters, tailors, and weavers. If anything, Muslims were more likely than
Copts to be merchants.
Saleh (2016) examines the urban quarter/​ village-​level data in the subsequent
published population census reports for Egypt from 1897 to 1986. These data enable
us to document the long-​term evolution of the proportion of non-​Muslims and of the
inter-​religious educational and occupational differences during the twentieth cen-
tury.3 First, the proportion of non-​Muslims increased slightly between 1897 and 1927,
due to European immigration into Egypt, and then declined to reach 6 percent of the
population in 2006, which is not very different from the proportion of non-​Muslims
in 1848. Second, examining the correlation at the quarter/​village level between the pro-
portion of non-​Muslims and each of the male literacy rate (the proportion of the male
population that is able to read and write), and the proportion of males who work in the
non-​agricultural sector, enables us to estimate the difference between non-​Muslims and
Muslims with respect to these outcomes.4 The data reveal that both the educational and
occupational gaps between non-​Muslims and Muslims declined during the second half
of the twentieth century.
A further piece of evidence comes from the 10 percent individual-​level population
census samples for 1986, 1996, and 2006. We are able to examine the differences in lit-
eracy, years of schooling, and being in a white-​collar occupation by cohort of birth and
religious affiliation.5 Hence, we can trace the long-​term evolution of the Christian-​
Muslim gap with respect to these outcomes among survivors in 1986, 1996, and 2006.
This exercise reveals similar trends to those that we find in the urban quarter/​village-​
level exercise. The literacy and years of schooling differentials went down among
cohorts born in the second half of the twentieth century. The white-​collar gap also
declined. However, there is an interesting pattern here. While the proportion of white-​
collar workers was going up among both Christians and Muslims, it started to decline
among the 1950s and 1960s cohorts of birth. This is likely because of the abolition in the
1980s of the employment guarantees in the government and public sectors for university
and secondary schools graduates.

Descriptive Statistics before 1800

A more daunting task is to trace the population share of non-​Muslims, let alone the
inter-​religious educational and occupational gap, before 1800. We know that non-​
Muslims constituted 100 percent of MENA’s population before the Arab conquests of
the region, which started in 632. But we do not know the evolution of this share between
632 and 1848 for Egypt (or 1881–​1892 for the Levant and Iraq). The reason is that there
are no population censuses before 1848 in Egypt, and before 1881–​1892 in the Ottoman
642   Mohamed Saleh

Empire. The sixteenth-​century Ottoman tax registers provide a further data point on the
religious composition of households at the village level in the Levant ad Iraq. The aggre-
gate statistics from this source reveal that non-​Muslims were around 5–​9 percent in Iraq
and the Levant by 1600. The sixteenth-​century Ottoman tax registers for Egypt are pre-
served at the National Archives of Egypt but have not yet been digitized.
To produce plausible estimates of MENA’s religious composition during the mil-
lennium that elapsed between the Arab and the Ottoman conquests, we need to rely
on other (non-​census) sources. For Egypt, Saleh (2018) employs two medieval Coptic
chronicles in 1200 and 1500 that list Christian churches and monasteries by location.
They include both Coptic and non-​Coptic ones. Using these lists, it is possible to es-
timate the proportion of villages that had at least one Christian church or monastery
in 1200 and 1500, which can be used as an estimate of the proportion of non-​Muslims
under certain assumptions. The findings suggest that villages that had at least one
Christian church or monastery were 16 percent of the total number of villages in 1200
and about 3 percent in 1500. This suggests that the majority of Copts’ conversions to
Islam happened between 641 and 1200.
Studying the Christian-​Muslim occupational gap is more difficult because it requires
having data on both religious affiliation and occupational title. The Arabic Papyrology
Database (APD) offers some solution. Saleh (2018) collected all mentions of occupa-
tional titles and names of workers in the papyri between 641 and 969 (date of the Fatimid
conquest). I then inferred religious affiliation from names. This data set reveals two
findings. First, the Coptic-​Muslim occupational gap first emerged between 641 and 969
and persisted through 1848 and 1868, as Copts are over-​represented among white-​collar
workers and artisans and are under-​represented among farmers and unskilled workers.
Second, Copts were not a top political elite but rather an apolitical (technocratic) middle
class in both rural and urban areas. The top white-​collar jobs, such as high bureaucracy,
judiciary, military, police, and (Muslim) clergy, were actually monopolized by Muslims
by law. The source of Copts’ occupational advantage over Muslims stemmed from their
over-​representation in the mid-​low (fiscal) bureaucracy, such as scribes, tax collectors,
accountants, and land surveyors, and in certain skilled artisanal jobs, such as carpenters,
jewelers, tailors, and weavers.

Does “Islam” Explain Why Muslims Are

Worse Off?

Why were non-​Muslim minorities better off than the Muslim majority in the nine-
teenth and twentieth centuries, and even before? The first set of explanations traces
the inter-​religious SES inequality to Islam, either as a set of religious beliefs or as a set
of institutions. If Muslims are poorer than non-​Muslims, it is plausible that this is be-
cause Muslims endorse a specific set of beliefs about life that hinders economic success.
 Islam and Economic Development     643

Three beliefs generally stand out in the literature as being pro-​economic success: (1)
the work ethic, and related beliefs such as thrift and the view of wealth as God’s reward
to individuals for exerting effort; (2) individualism, which facilitates the emergence of
impersonal exchange; and (3) the salience of human capital formation. The work ethic
explanation is due to Max Weber ([1905] 1930), who traced Protestants’ superior eco-
nomic outcomes in Europe to their stronger work ethic in comparison to Catholics.
Extrapolating his thesis to other regions, he argued that the latter were not conducive
to capitalism. The individualist/​collectivist cultural belief can be found in Greif (1994),
who postulated that collectivist cultures, including Islam, were less conducive to the
emergence of impersonal exchange and third-​party contract enforcement. The human
capital formation belief has been suggested by recent scholarship in the economics of
religion literature. Botticini and Eckstein (2005) argued that the literacy requirement
introduced by Rabbinic Judaism in the second century CE led to the conversion of Jews
with a weaker taste for education out of Judaism, and thus to the shrinkage of Rabbinic
Jews into a better-​off and educated minority. Becker and Woesmann (2009) showed that
Protestants were better off than Catholics in nineteenth-​century Prussia, not because of
their stronger work ethic, but rather because of Protestants’ emphasis on literacy and the
formation of human capital.
The second explanation of the underdevelopment of Muslims traces the phenom-
enon to Islam, not as a set of beliefs, but rather as a set of institutions. The argument is
that Islamic institutions subjected Muslims to certain constraints or rules that prevented
them from economic growth, at least after a certain point in time. Lewis’s (2002) in-
fluential thesis on Islam and economic development traces MENA’s underdevelopment
to the incompatibility between Islam and modernization, due to the inseparability of
religion from the state in Islam, which (according to Lewis) dates from the Prophet
Muhammad’s lifetime. The pioneering work of Timur Kuran offers a more nuanced
institution-​based explanation, where Islamic institutions may have been pro-​economic
growth in the premodern period but hindered the economic growth of the region in the
modern period. Kuran (2004) traced the rise of non-​Muslim minorities to the fact that
they had access to European legal systems via the capitulations system, unlike Muslims
who did not have this access and thus had to resort to Islamic courts.
I argue that both the cultural and institutional explanations do not account for the supe-
rior socioeconomic outcomes of Copts in Egypt. Starting with the cultural explanation, it
is widely accepted that Coptic Christianity, as a set of beliefs, does not put a particular em-
phasis on work ethic, individualism, or the formation of human capital, in comparison to
Islam. First, Copts actually share with Egyptian Muslims (who were mostly Sufi Sunnis up
to the mid-​twentieth century) a mystic outlook on life that attributes materialistic success
to metaphysical factors rather than to effort. The salience of saints and martyrs in popular
Coptic Christianity resembles the salience of Sufi saints among Egyptian Muslims, and
in fact many of the festivals of Coptic and Muslim saints are attended by both Muslims
and Copts, especially in rural Egypt. Second, family values and collectivism are arguably
equally salient among Copts. Third, the Coptic Church does not place any particular em-
phasis on literacy or human capital formation. Unlike non-​Coptic Christians, such as
644   Mohamed Saleh

Armenians, Greeks, Jews, and Levantines, who are urban and educated minorities and are
thus over-​studied in the literature, Copts are mostly rural and less educated. In particular,
despite Copts’ superior educational level compared to Muslims, illiteracy remains sizable
among Copts, reaching 21 percent of Coptic males in 2006.
The institutional interpretation is not capable either of explaining why Copts are
better off than Muslims for two reasons. First, Copts’ superior occupational status long
predates the rise of Europe. As shown in the previous section, papyrological evidence
suggests that the occupational gap between Copts and Muslims emerged between
641 and 969. Second, I attempt to test Kuran’s (2004) explanation that traces non-​
Muslims’ advantage to their access to European legal systems. The 1848 and 1868 pop-
ulation census samples indicate that among non-​Muslims, only Jews and non-​Coptic
Christians had a relatively sizable proportion of “protégés,” those who purchased ac-
cess to European consulates’ courts. Copts and Muslims had essentially no protégés.
Moreover, the proportion of protégés among non-​Coptic Christians and Jews is low,
between 10 and 20 percent, and cannot on its own explain their economic advantage
over Muslims. The reason for this difference in protégés across non-​Muslim minorities
is likely driven by their occupational differences. While non-​Coptic Christians and
Jews were over-​represented among merchants, where European law mattered to resolve
conflicts with European traders, Copts’ advantage stemmed from artisanship and mid-​
low fiscal bureaucracy, where European law played less of a role.
This is not to say, though, that the cultural and institutional explanations of why
Muslims fell behind non-​Muslims are not consistent with any empirical fact about non-​
Muslim minorities in the MENA region. It is only to say that both explanations do not
appear to explain the Coptic-​Muslim SES difference in Egypt. However, Kuran’s (2004)
hypothesis may explain why certain Jews and non-​Coptic Christians in Egypt were
better off than both Copts and Muslims. But even for non-​Coptic Christians and Jews,
we need to collect data on the pre-​1800 period in order to understand how their SES ad-
vantage evolved both before and after the emergence of the capitulations system.

Does Colonization Explain Why

Muslims Are Worse Off?

The second set of explanations trace the SES superiority of non-​Muslim minorities to
colonization. European powers favored non-​Muslims, whether during the “indirect
colonization” episode under the Ottoman Empire or under the direct colonization/​
mandate episodes in the late nineteenth century and the first half of the twentieth cen-
tury (Issawi 1981). One version of this theory overlaps with Kuran’s (2004) institution-​
based explanation outlined above. It postulates that the capitulations system favored
non-​Muslims because they were able to purchase access to European consulates’ court
system, but Muslims were not.
 Islam and Economic Development     645

Given that the Coptic-​Muslim SES gap emerged long before colonization, this expla-
nation does not fully account for the gap that we observe in the nineteenth and twen-
tieth centuries. However, this is not to say that colonization did not have any effect, but
that we rather need to be specific on what is meant by colonization, which is in fact a
large bundle of institutions and policies. For example, one aspect of colonization is the
rise of European schools, whether missionary or secular. The 1848 and 1868 popula-
tion census data reveal that non-​Muslim children in Egypt were more likely to enroll in
school than Muslim children in 1848. Almost all students (Muslims and non-​Muslims)
were enrolled in religious traditional schools, with a tiny proportion of Muslim students
enrolled in public modern schools, which were established by Muhammad Ali Pasha
starting from 1818 (non-​Muslim students were not allowed to enter public modern
schools until 1873). However, non-​Muslim students significantly shifted from religious
schools to modern private schools in the subsequent decades: In 1868, 40 percent of
non-​Muslim students were enrolled in modern private schools, in comparison to only
4 percent of Muslim students who were enrolled in public modern schools. But contrary
to the colonialism explanation, these modern private schools were not only European,
let alone missionary, schools. In fact, Egyptian non-​Muslim minorities started to re-
spond to European missionary and secular schools by establishing their own communal
(private) modern schools, and this response was faster than that of the Muslim com-
munity. The school censuses in 1906/​07–​1951/​52 reveal that most Muslim students con-
tinued to rely on religious schools until 1951, whereas most non-​Muslim students in 1951
were enrolled in modern schools, both public and private, European and native. Hence,
while European schools may have indeed contributed to increasing the educational gap
between non-​Muslims and Muslims between 1848 and 1951, the full effect of modern
schools was not only due to European schools but also included the (under-​studied) pri-
vate native schools.

Colonization and the Role of the Local Institutions

The colonization explanation implicitly presumes that colonization emerged within
a vacuum, and typically underestimates the effect of the precolonial local institutions
or the local institutions that emerged (or evolved) in response to colonization. First,
non-​Muslim religious schools were of better quality than those of Muslims long be-
fore colonization. Coptic and Jewish schools in Egypt taught arithmetic and geometry
to students, whereas Muslim schools lacked this training. Second, as indicated above,
modern private schools that enrolled non-​Muslim students in Egypt were not only
European schools. Copts, Jews, and non-​Coptic Christians responded to European
schools by establishing their own modern schools.
The role of the local state between 1800 and independence in the post-​WWII period
is typically absent in the colonization explanation. In Egypt, Muhammad Ali and his
successors embarked on an ambitious, yet coercive, program of state-​led moderniza-
tion that spanned agriculture, education, manufacturing, and the military, long before
646   Mohamed Saleh

the British colonization in 1882. Saleh (2015) examines the impact of state industrializa-
tion in 1816–​1868 on the Christian-​Muslim occupational difference. The paper finds that
while the first wave of state industrialization in 1816–​1848, which focused on textiles,
was upskilling among Christians, it was deskilling among Muslims. The second wave, in
1848–​1882, which focused on transportation, was upskilling among both groups. Both
waves of state industrialization did not reduce the Christian-​Muslim occupational gap
that we observe in the traditional sector. The reason is that state manufacturing and
transportation firms in both waves hired people based on their preexisting skills that
they acquired from the traditional sector. State firms did not train (Muslim or Christian)
workers to improve their occupational status, though.
In a similar vein, the role of the postcolonial state is important to examine. In the
schooling domain, Egypt’s postcolonial expansion of public mass modern education in
1951–​1953 contributed to the decline of the Christian-​Muslim educational and occupa-
tional differences in the second half of the twentieth century (Saleh 2016).

Taxation and the Self-​Selection

of Converts on SES

The two explanations discussed above implicitly assume that “Muslims” and “non-​
Muslims” are exogenously formed groups that are fixed over time. This is not valid
historically, though. While in today’s world, religious affiliation is mostly inherited
(although non-​Muslims have the choice of converting to Islam), this was not the case
historically. Any religious group is, by definition, formed via the conversion of an ini-
tial population to that religion. MENA’s local populations were entirely non-​Muslims
in the early seventh century: mostly Christian, with a small Jewish minority. Over the
centuries that followed the Arab conquests, MENA was Islamized, meaning that a ma-
jority of its local population became Muslim. This process can be due to either conver-
sion of the local population to Islam or to Arab immigration from the Arab peninsula
into other parts of MENA.6 Historical evidence suggests, though, that Islamization was
driven by voluntary conversion of the local population to Islam. The reason is that Arab
immigration was small in comparison to the local populations of MENA. The full popu-
lation of the Arab peninsula was much smaller than that of the Levant, Iraq, and Egypt.
Within these territories, Arab immigrants were not a sizable proportion.
Saleh (2018) argues that the Coptic-​Muslim socioeconomic gap can be partially
attributed to the tax system that was imposed upon the Arab conquest of the then-​Coptic
Christian Egypt in 641 CE, and that was enforced until 1856. Arabs imposed a discrim-
inatory poll tax (jizya) on adult non-​Muslim males, a tax that was removed upon an
individual’s conversion to Islam. Due to the (quasi) lump-​sum feature of the poll tax,7
the incentive to convert to Islam was likely stronger among poorer Copts, holding other
factors (notably, Coptic religiosity) constant. The paper thus documents that the poll tax
 Islam and Economic Development     647

led to the conversion of poorer Copts to Islam (who since then became Egypt’s “Muslims”)
and to the consequent shrinkage of non-​convert Copts into a better-​off minority.
The empirical evidence for this hypothesis is based on exploiting the local variation in
the poll tax rate. A simple model in which a Coptic Christian chooses whether to con-
vert or not by comparing the (psychological) cost of conversion to the lump-​sum poll
tax predicts that poorer Copts are more likely to convert at a given level of religiosity,
and that less religious Copts are more likely to convert at a given level of income.8 It thus
follows (unsurprisingly) that the proportion of non-​convert Copts is decreasing in the
poll tax rate. It also follows that the magnitude of the Coptic-​Muslim average (before-​
tax) income gap is increasing in the poll tax rate.9 Put differently, high-​poll-​tax areas
are expected to have relatively fewer non-​convert Copts but a greater Coptic-​Muslim
income gap.
Constructing a novel data set from papyrological poll tax registers and receipts in
Egypt in 641–​1100, Saleh (2018) first documents that Egypt’s high-​poll-​tax districts in
641–​1100 were less likely to have any Coptic churches or monasteries in 1200. Given that
all districts were 100 percent Copt in 641, this finding suggests that these areas witnessed
relatively more conversions to Islam among Copts between 641 and 1100. Second, these
districts had a larger Coptic-​Muslim occupational gap in the 1848 and 1868 popula-
tion census samples, which are the earliest data sources with localized individual-​level
data on both religious affiliation and occupational titles. Specifically, while Copts in all
districts were more likely than Muslims to be white-​collar workers and artisans, and less
likely to be farmers and unskilled non-​agricultural workers, the Coptic-​Muslim occu-
pational gap was greater in high-​poll-​tax districts.
This theory emphasizes the endogenous process of the historical formation of re-
ligious groups that may be characterized by (self-​)selection on SES. A deeper under-
standing of the inter-​religious SES differences in the MENA region thus arguably
requires studying how non-​Muslim groups were formed (and evolved) historically.
This stands in contrast to the existing literature on non-​Muslim minorities in MENA,
which treats Copts, Armenians, Greeks, Jews, and Levantines as exogenously formed
groups that do not change over time (except via migration). In this regard, self-​selection
of converts is also implied by Boticinni and Eckstein’s (2005) hypothesis about Rabbinic
Jews’ higher SES status: Rabbinic Jews with a weaker taste for education convert out of
Judaism (and thus lose their Jewish identity), whereas those with a stronger educational
taste continue to hold the group’s identity.

Persistence of the Coptic-​Muslim

Socioeconomic Gap

The empirical evidence in Saleh (2018) suggests that non-​convert Copts shrank into
a minority by 1200 and that the initial positive selection of non-​converts emerged in
648   Mohamed Saleh

641–​969 and persisted through 1868. (Note that this section is taken from Saleh 2018,
419–​424.) The persistence of the Coptic-​Muslim SES gap for over a millennium poses a
theoretical dilemma, though. Economic models of intergenerational mobility trace the
correlation between parents’ and children’s outcomes to the parental investment in a
child’s human capital (nurture) and the inheritance of biological and cultural traits (na-
ture). These models typically predict that dynasties eventually regress to the mean after
a few generations, even at high rates (<1) of intergenerational persistence (Becker and
Tomes 1979). To be sure, since the poll tax was enforced from 641 to 1856, persistence
can be possibly explained by repeated (negatively) selected conversion waves, even in
the absence of poll tax rises, due to idiosyncratic shocks to SES and religiosity. Indeed,
Copts’ population share declined, albeit slowly, between 1200 and 1868, and so it is plau-
sible that selected conversion on SES continued throughout the whole period, leading
the SES gap to perpetuate. Yet this explanation fails to account for why non-​convert
Copts did not vanish. An additional mechanism, besides intergenerational transmis-
sion of SES within dynasties, is thus needed to explain not only the persistence of the gap
but also the survival of non-​convert Copts.
Saleh (2018) explains both the persistence of the Coptic-​Muslim SES gap and the sur-
vival of Copts by group effects on children’s SES. Group effects operated by allowing
the in-​group children’s human capital accumulation (and blocking that of the out-​
group children), via group control over apprenticeships and schooling. In medieval
Egypt, white-​collar and artisanal occupations required learning occupation-​specific
skills from a young age. As human capital was job-​specific, it was acquired primarily via
apprenticeships and, to a lesser extent, schooling. The supply of both apprenticeships
and schooling was mostly restricted, though, to the social networks of white-​collar
workers and artisans. For one, obtaining an apprenticeship, the gateway to most arti-
sanal and white-​collar occupations, required the approval of a master in a specific oc-
cupation. Masters were more likely to admit their family members and acquaintances.
For another, school enrollment that provided training for white-​collar jobs (which had
to be later augmented by apprenticeships) was limited to social networks of workers
in these jobs. Conversions of Copts to Islam between 641 and 1200 arguably redefined
social networks along religious lines. As conversions sorted Copts and Muslims on
occupations, each group then attempted to exclude the other from apprenticeships and
schooling that would qualify a child to the white-​collar and artisanal jobs in which the
group was over-​represented. Put differently, a child’s occupational attainment depended
on two factors: intergenerational transmission of SES via both nature and nurture, and
the family’s religious group.
This mechanism is supported by historical evidence. Copts restricted access to skills
that were required for jobs in the mid-​low bureaucracy. While Coptic elementary
schools taught arithmetic and geometry in order to train Coptic children for jobs in the
mid-​low bureaucracy, Muslim schools failed to provide this training (Heyworth-​Dunne
1938, 2–​7, 84–​92). However, it was primarily apprenticeships, not schools, that trained
Coptic children for bureaucratic jobs. (Samir (1996, 190) notes that in Fatimid Egypt
(969–​1171),
 Islam and Economic Development     649

the persistence of Coptic administrative personnel [was because] the agrarian ad-
ministration was very complex and not easily mastered. In it the Copts played an
important role at the local level as well as at the central offices in the capital. . . . The
administrative knowledge was passed on by the officials in their families when fa-
thers employed their sons, thus maintaining the hold of the family over posts.

In the words of Lord Cromer, the British consul of Egypt in 1883–​1908, the Coptic
accounting system was “archaic” and “incomprehensible to anyone but themselves”
(Tagher 1951, 213). Copts used fractions and “ambiguous abbreviations” in accounting
based on units of measurement in use in rural Egypt. Group effects on acquiring human
capital were not limited to Copts in the mid-​low bureaucracy, though. Copts were le-
gally banned from the judiciary, military, police, and Muslim clergy, and these jobs
were thus monopolized by Muslims. Muslims were banned from brewing, which be-
came a Coptic specialization. The 1848 and 1868 censuses reveal that Copts were over-​
represented among jewelers, dyers, carpenters, weavers, and tailors, whereas Muslims
were over-​represented among blacksmiths, sawyers, bakers, and butchers. Raymond
(1973, 544–​51) suggests that the reason for the persistence of this occupational special-
ization was restricting apprenticeships to group members. The Coptic-​Muslim occu-
pational gap may have increased in the late nineteenth century with the expansion of
modern (European and private non-​Muslim) schools, but it later declined in the late
twentieth century with the expansion of public mass modern education in 1951–​1953,
which relaxed each group’s restrictions on access to skills (Saleh 2016).

Conclusion

This chapter has examined a well-​known phenomenon in the MENA region: the supe-
rior SES of its native non-​Muslim minorities, in comparison to the Muslim majority,
both historically and today. Employing a wide range of novel data sources, I argue
that the Coptic-​Muslim SES gap in Egypt is not driven by a negative causal impact on
Muslims’ SES of Islamic beliefs, Islamic institutions, or colonization. Instead, I trace the
phenomenon to the Islamic tax system, which arguably provided stronger pecuniary
incentive among poorer Copts to convert to Islam, leading to the shrinkage of non-​con-
vert Copts into a better-​off minority. The Coptic-​Muslim SES gap then persisted be-
cause of group restrictions on skill formation.
The chapter opens new and exciting areas of research. First, there is a question of the
external validity of the tax-​induced conversion argument beyond the case of Coptic
Christians in Egypt to other non-​Muslim minorities of the region and beyond. There are
two remarks to make here: (1) The poll tax was not limited to Egypt. It was introduced
in all the conquered territories of the Arab Caliphate including Iran and India, and sub-
sequently in all Muslim-​ruled territories up to the nineteenth century. It is thus impor-
tant to investigate its effect on the historical formation of other non-​Muslim minorities
650   Mohamed Saleh

(e.g., when and why did Christians shrink into a minority in Lebanon, why did certain
Hindus and Zoroastrians convert to Islam, and why was Christianity but not Judaism
wiped out from North Africa, west of Egypt?). (2) When evaluating the Islam-​based
and colonization-​based theories, it is important to examine each historical context in
depth, allowing for potentially context-​specific conclusions. For example, MENA coun-
tries went through different forms of colonization. French colonization of Algeria and
Tunisia was probably different from the mandate colonization of Syria and Lebanon
during the interwar period. Egypt had been an autonomous Ottoman province since
the early 1800s, and its British colonization was primarily focused on controlling the
Suez Canal. Capitulations may have also had different, potentially larger, effects on non-​
Coptic Christians and Jews in Egypt and elsewhere.
Second, and relatedly, studying this topic requires measuring the long-​term trends
of the religious composition, and of the socioeconomic advantage, of other native
non-​Muslim minorities in MENA, both before and after 1800. A promising future
area of research is to expand the descriptive analysis to non-​Muslim minorities in
other MENA countries with sizable non-​Muslim populations during the nineteenth
and twentieth centuries. This includes Syria, Palestine, Israel, Jordan, Lebanon, Iraq,
Algeria, and Tunisia. We know that Christians and Jews in these countries were all
better off than Muslims (Courbage and Fargues 1997). The individual-​level (or at least
sub-​country level) data sources that can be used to quantify this argument for the
Levant and Iraq are the Ottoman population censuses of 1881–​1892 and 1914. Later
statistics for the Levant and Iraq include the French mandate statistical yearbooks and
population censuses for Syria and Lebanon, and the British mandate statistics for Iraq,
Jordan, and Palestine. Sources for Tunisia and Algeria include the French coloniza-
tion statistical yearbooks and population censuses.10 Finally, the national population
censuses and statistical yearbooks are available for the postcolonial period for most
MENA countries.
Third, while this chapter focuses on SES inequality across religious groups, within-​
group inequality is an important area of future research. While intergroup inequality
typically receives more attention in the public debate, probably because of the salience
of inter-​religious tensions, it is not obvious if its contribution to overall inequality actu-
ally exceeds that of within-​group inequality. It is thus important to evaluate the relative
weight of intergroup and within-​group inequality in overall inequality.
Fourth, this chapter has examined educational and occupational inequality across
Muslims and Copts. Yet there are other important outcomes that are worth studying.
For example, income and wealth inequality are of utmost significance, both across and
within groups, although they are more challenging to study. In Egypt, data on earnings
are only available for the most recent period (1988–​2012), and even then they do not
record religious affiliation. Data sources on wealth are primarily the agricultural and
real estate tax registers (cadasters) that extend from 1813 to the present day, but these
are challenging to digitize and require a collaborative data collection effort. Equivalent
cadastral data for the Levant and Iraq are available from the sixteenth-​and nineteenth-​
century Ottoman tax registers in the precolonial period, the British and French
 Islam and Economic Development     651

cadasters during the mandate period, and the national cadasters in the postcolonial
period. French cadasters for Algeria and Tunisia are important to investigate. Another
example is health outcomes, which can be observed in the Demographic and Health
Surveys, such as the female genital-​cutting practice, which went down faster among
Copts than among Muslims during the second half of the twentieth century (Blaydes
and Platas, forthcoming).
Fifth, the chapter has examined two colonization institutions or policies: capitulations
and schooling. A more complete examination of the impact of colonization is an exciting
area for future research, though. This research agenda must study one colonial institu-
tion or policy at a time, and examine in depth how it interacted with local institutions,
both in the precolonial period and during colonization.
Finally, the analysis has focused on outcomes of men. This is justified because enu-
meration of women’s schooling and employment was extremely low in the pre-​1900
population censuses. However, women’s outcomes improved drastically during the
twentieth century, and it is an exciting area of research to examine how the evolution of
women’s outcomes may have varied across religious groups in the MENA region.

Notes
1. This is not to say, though, that socioeconomic inequality across religious groups is more sa-
lient that inequality within groups. In fact, intragroup inequality may well be much higher.
However, intergroup inequality has been often used by the elites of each group to main-
tain their socioeconomic privileges vis-​à-​vis the masses. I come back to this point in the
Conclusion.
2. The earliest Ottoman population census was conducted in 1881–​1892. It covered the
Levant and Iraq within the MENA region, but not North Africa, which had already been
under direct colonial rule (Egypt, Algeria, and Tunisia) by 1881. I do not know if Libya was
enumerated in this census. The pre-​nineteenth-​century Ottoman tax registers are from the
sixteenth century and are at the village level.
3. The 1882, 1996, 2006, and 2017 population censuses do not report the religious composition
at the urban quarter/​village level, although the religious affiliation is available at the indi-
vidual level in the IPUMS 10-​percent individual-​level samples for 1996 and 2006 (urban
quarter/​village of residence is not recorded though).
4. The urban quarter/​village-​level tabulations in the population census reports from 1897 to
1986 do not report literacy and occupational composition by religious group. They only
report the religious composition and the proportion of the population that is literate (or
works in a specific sector). Regressing these outcomes on the proportion of non-​Muslims
gives us an estimate of the inter-​religious gap with respect to these outcomes.
5. The 1986, 1996, and 2006 population census samples do not distinguish among Coptic and
non-​Coptic Christians. The proportion of the population who have “other” religious affilia-
tion is negligible. This is because of Jewish exodus and deportation from Egypt in the after-
math of the 1956 Suez crisis and the 1967 Egyptian-​Israeli war.
6. Other demographic processes that can drive Islamization are mortality and fertility
differences between Muslims and non-​Muslims and intermarriage between Muslim males
and non-​Muslim females (opposite scenario is prohibited under Islamic law). Demographic
652   Mohamed Saleh

evidence from the 1848 and 1868 censuses suggests that there were no statistically sig-
nificant mortality and fertility differences between Copts and Muslims. Intermarriage
(without prior conversion of one of the partners) was extremely rare.
7. Following the emergence of Islamic jurisprudence circa 750 CE, the poll tax was imposed
in three brackets: 1 dinar on the poor, 2 dinars on the middle income, and 4 dinars on the
rich. Despite this variation, the poll tax was regressive in income.
8. The model predicts that the poorest non-​convert will be richer than the richest convert,
at a given level of religiosity. This prediction does not hold, though, if religiosity varies
in the population. In particular, the model allows for poor Copts who do not convert be-
cause of their high religiosity, and for rich Copts who convert because of their low religi-
osity. The model generates the prediction that non-​converts are richer, on average, than
converts, which is the main empirical fact that the paper is interested in. Importantly, rich
Copts may have converted to Islam for non-​tax reasons (to access high bureaucracy, for
example). However, their population share was presumably not large enough to offset
non-​converts’ SES privilege over converts.
9. The latter result is not trivial. As the poll tax rate rises, the proportion of non-​convert
Copts will go down. Non-​converts will be richer on average as they lose their poorest
members. The same holds for converts, though, who will be richer on average as they gain
new converts who are richer than any previous convert. Hence, the Coptic-​Muslim in-
come gap may go up or down depending on the shape of the income distribution. It is
increasing in the poll tax rate if the density of income is everywhere decreasing.
10. I exclude Morocco because I was not able to locate its pre-​independence statistics.
Morocco was not under Ottoman rule. It fell under French/​Spanish joint rule in 1912, but
I am not aware of French/​Spanish population statistics for Morocco between 1912 and its
independence in 1956. To the best of my knowledge, Italian statistics for Libya did not enu-
merate the local population. The Arab peninsula did not have a sizable non-​Muslim popu-
lation, perhaps with the exception of Yemen.

References
Becker, Gary S., and Nigel Tomes. 1979. “An Equilibrium Theory of the Distribution of Income
and Intergenerational Mobility.” Journal of Political Economy 87, no. 6: 1153–​1189.
Becker, Sasha O., and Ludger Woessmann. 2009. “Was Weber Wrong? A Human Capital
Theory of Protestant Economic History.” Quarterly Journal of Economics 124, no. 2: 531–​596.
Blaydes, Lisa, and Melina R. Platas. Forthcoming. “Religion, Patriarchy, and the Perpetuation
of Harmful Social Conventions: Explaining Women’s Attitudes toward Female Genital
Cutting in Egypt.” Journal of Demographic Economics.
Botticini, Maristella, and Zvi Eckstein. “Jewish Occupational Selection: Education,
Restrictions, or Minorities.” Journal of Economic History 65.4 (2005): 922–​948.
CEDEJ (Centre d’Études et de Documentation Économiques, Juridiques, et Sociales). Century
Census CD-​ROM: Egypt 1882–​1996, Cairo: CEDEJ, 2003.
Courbage, Youssef, and Philippe Fargues. 1997. Christians and Jews under Islam. Translated by
J. Mabro. London: I.B. Tauris.
Greif, Avner. 1994. “Cultural Beliefs and the Organization of Society: A Historical and
Theoretical Reflection on Collectivist and Individualist Societies.” Journal of Political
Economy 102, no. 5: 912–​950.
 Islam and Economic Development     653

Heyworth-​Dunne, James. 1938. An Introduction to the History of Education in Modern Egypt.

London: Luzac & Co.
Issawi, Charles. 1981. “The Transformation of the Position of the Millets in the Nineteenth
Century.” In The Arab World’s Legacy¸ 199–​230. Princeton, NJ: Darwin Press.
Kuran, Timur. 2004. “The Economic Ascent of the Middle East’s Religious Minorities: The
Role of Islamic Legal Pluralism.” Journal of Legal Studies 33: 475–​515.
Lewis, Bernard. 2002. What Went Wrong?: Western Impact and Middle Eastern Response.
Oxford: Oxford University Press.
Raymond, André. 1973. Artisans et Commerçants au Caire au XVIIIe Siècle. Damascus: Institut
Français de Damas.
Saleh, Mohamed. 2013. “A Pre-​colonial Population Brought to Light: Digitization of the
Nineteenth-​Century Egyptian Censuses.” Historical Methods: A Journal of Quantitative and
Interdisciplinary History 46, no. 1: 5–​18.
Saleh, Mohamed. 2015. “The Reluctant Transformation: State Industrialization, Religion, and
Human Capital in Nineteenth-​Century Egypt.” Journal of Economic History 75, no. 1: 65–​94.
Saleh, Mohamed. 2016. “Public Mass Modern Education, Religion, and Human Capital in
Twentieth-​Century Egypt.” Journal of Economic History 76, no. 3: 697–​735.
Saleh, Mohamed. 2018. “On the Road to Heaven: Taxation, Conversions, and the Coptic-​
Muslim Socioeconomic Gap in Medieval Egypt.” Journal of Economic History 78, no.
2: 394–​434.
Samir, Samir K. 1996. “The Role of Christians in the Fatimid Government Services of Egypt to
the Reign of Al-​Hafiz.” Medieval Encounters, 2, no. 3: 177–​199.
Tagher, Jacques. (1951) 1998. Christians in Muslim Egypt: A Historical Study of the Relations be-
tween Copts and Muslims from 640 to 1922. Altenberge, Germany: Oros Verlage.
Weber, Max. (1905) 1930. The Protestant Ethic and the Spirit of Capitalism. Translated by Talcott
Parsons. New York: Charles Scribner’s Sons.
 Chapter 32

State Institu t i ons a nd

Ec onom ic Performa nc e
in Nineteenth -​C e nt u ry
Egy p t

Lisa Blaydes and Safinaz El Tarouty

Leading theories in political economy suggest that state institutions are essential
prerequisites for generating high levels of economic growth and prosperity. North and
Thomas (1973) argue that secure property rights regimes provide incentives for inno-
vation and investment, with positive effects on development. Formal markets expand
when the costs of economic transactions decline and property rights are made secure.
North (1981, 20) argues that “the existence of the state is essential for economic growth,”
since states provide the institutional basis for property rights protection. Epstein (2000)
suggests that political centralization put Europe on the path to high levels of economic
growth by lowering transaction costs and the costs of organizing new institutions. For
Epstein (2000, 169), the “centralization of government” was the main contributor to pre-
modern economic growth in Europe.
At the same time, however, scholars have argued that when states are too strong, rulers
can predate on the citizenry or impose growth-​stifling taxes. Because both weak and
strong states generate distortions that can damage development, the best outcome from
an economic growth perspective is a “consensually strong state equilibrium,” whereby
the state is strong enough to impose taxes and provide public goods, but politically
weak enough that society exercises constraint over the government (Acemoglu 2005).1
Dincecco (2011) argues that to generate a prosperous economy, states need to first over-
come the challenge of fiscal decentralization; after that, national parliaments are vital for
holding the monarch accountable and avoiding out of control spending.2 Dincecco and
Katz (2014) find that fiscal centralization and limited government, together, improved
economic performance, but that centralization was even more impactful than limited
government in generating development.3 These existing arguments suggest that states
656    Lisa Blaydes and Safinaz El Tarouty

not only need to achieve “goldilocks” levels of state strength and executive constraint
(e.g., Acemoglu 2005), but also that the ordering of state fiscal centralization relative to
the emergence of limited government institutions also matters in generating economic
growth and prosperity (e.g., Dincecco 2011; Dincecco and Katz 2014).
If state centralization was a boon for economic development in the European con-
text, what lessons might be drawn from the Middle Eastern experience, and particularly
the history of institutional development in Egypt? Middle Eastern states long enjoyed
forms of centralization, even if the precise definition of a “centralized” state has varied
over time.4 Blaydes (2017) argues that the relatively early transition to agriculture in the
Middle East meant that states in the region tended to have relatively high levels of fiscal
and bureaucratic capacity relative to other world regions, particularly Europe. For ex-
ample, Blaydes (2019) shows that during the Mamluk Sultanate (1250–​1517 CE), a large
percentage of agricultural land in Egypt was held by the state, with a relatively sophis-
ticated bureaucracy that administered land grants offered as payment for military or
other service to state servants. Over the course of the sixteenth to eighteenth centuries,
however, Egypt became a vassal of the Ottoman Empire and witnessed a decentraliza-
tion of political control. During this time, military and religious elites enjoyed consider-
able influence over relatively weak Ottoman envoys, who exercised only limited control
over Egypt’s fiscal apparatus.
In this essay, we examine Egypt in the nineteenth century through the lens of the de-
velopment of state centralization and executive constraint. While national governments
in Europe were undertaking centralizing reforms over the course of the “long” nine-
teenth century (Dincecco and Katz 2014), similar efforts were occurring contempora-
neously in Egypt. Indeed, the reign of Muhammad Ali (1805–​1848) is often cited as a
turning point in the emergence of modern Egypt, ushering in an era of fiscal central-
ization and nation-​state building. During this period, Egypt developed institutions
in the image of European models of state development, such as a standing military, a
centralized tax system, and a professionalized bureaucracy.5
Yet centralization of fiscal authority in Egypt was also accompanied by an increasing
concentration of power in the hands of the executive. Levi (1988) argues that rulers
are predatory, constantly requiring revenue to provide collective goods associated
with the states they control. Military, political, and economic constraints, however,
limit the amount of revenue that they can collect. Our findings suggest that the order
of modernizing reforms matter for economic outcomes, particularly insofar as they
affect the political strength of societal elites. Fiscal centralization without an institu-
tional legacy of executive constraint can sideline those elites who may have been best
positioned to reign in excessive ruler spending. Because no power center emerged that
was able to effectively constrain Muhammad Ali and his successors, rent-​seeking and
international borrowing eventually made the Egyptian state politically vulnerable to
European colonial intervention.
The rest of the essay proceeds as follows: First, we discuss Egypt’s transition from an
Ottoman province with a fragmented political elite to a more centralized state following
 State Institutions and Economic Performance    657

the reforms of Muhammad Ali. Next, we consider the effects of these centralizing
reforms on Egypt’s economic development in the first half of the nineteenth century. We
argue that state centralization eliminated those elite actors—​particularly slave soldiers
and Islamic clerics—​who were best positioned to constrain the executive in the future.
Third, we consider the development of parliamentary institutions in Egypt and find
that they were stacked with elites beholden to the Egyptian khedives, who were poorly
positioned to effectively constrain successive rulers. Indeed, the lack of constraints on
the executive may have contributed to costly and unsustainable international borrowing
that damaged Egypt’s longer-​term development prospects.

Egyptian Political Institutions

before Muhammad Ali

How can we characterize the institutional status quo in Egypt before the rise of
Mohammad Ali, Egypt’s celebrated state-​building leader? The Mamluk Sultanate ended
in 1517, following the conclusion of the Ottoman-​Mamluk War (1516–​1517). After that
point, Egypt was administered as a relatively unruly Ottoman province, with a brief
French occupation between 1798 and 1801. This section provides a broad overview of the
political institutions in Egypt that predominated between the sixteenth and eighteenth
centuries.

Mamluks and Pashas

During Egypt’s period as an Ottoman province, the Ottoman sultan would send
a viceroy, or governor, to Cairo, typically with the title of pasha.6 Several of these
governors came directly from the sultan’s imperial court. For example, six viceroys were
eunuchs who began their careers in the household of the sultan. These viceroys often
emphasized their personal piety in a bid to enhance their public image.7
Although the Mamluk Sultanate was overthrown, successive Ottoman governors
incorporated members of the Mamluk forces into the state administration of Egypt.8
For example, a Mamluk was appointed by the sultan and funded by the state to lead
the yearly pilgrimage caravan to Mecca (amir al-​hajj) and protect it against bedouins
(Winter 1998, 13).9 A Mamluk was also named the leader of the caravan that carried
the yearly tribute to Istanbul, a highly prestigious position. The head of the financial
administration (defterdar) was also appointed from among the Mamluks (Marsot
1985, 42).
Under early Ottoman control, imperial possessions in Egypt were divided into parcels
(muqata’at) and distributed to civilian officers, chosen by the sultan, who would be
658    Lisa Blaydes and Safinaz El Tarouty

responsible for tax collection. These civilian officers were paid annual salaries and were
responsible for delivering the taxes collected (Shaw 1958, 31). By the seventeenth cen-
tury, the Ottoman authorities could not easily direct affairs autonomously in Egypt and
prominent positions in the administration increasingly went to Mamluks (Crecelius
1998, 60). For example, Mamluks came to play an important role in tax collection. Tax
farming became the norm for revenue collection whereby each tax farm (iltizam) was
auctioned to the highest bidder (multazim), who paid for the right to collect taxes; many
of these tax farmers were Mamluks. By the eighteenth century, tax farming came to in-
creasingly resemble a form of land ownership whereby the heirs of tax farmers were able
to inherit the right to a tax farm (Marsot 1985, 41; Shaw 1958, 37–​38). Many of these tax-​
farming landlords were absentee, relying on local village leaders to administer on their
behalf (Marsot 1984, 139).10
Beginning in the mid-​eighteenth century, Ottoman Egypt fell under the influence of
a group of Mamluks of Georgian origin. For example, Ibrahim Kahya al-​Qazdaghl es-
tablished the Qazdaghl household that exercised de facto control over Egypt from 1748
to 1754. Most of the Georgian Mamluk rulers of the late eighteenth century belonged to
this household (Hathaway 1995, 44).11
During the period of Ottoman control of Egypt, there was considerable turnover of
Ottoman governors (see Figure 32.1).12 These Ottoman governors left office in a variety
of ways. Rulers were withdrawn due to corruption; executed for declaring themselves
sultan; dismissed after failing to collect enough tax revenue; appointed to a different po-
sition, including Grand Vizier; murdered during a mutiny; or dismissed for misman-
agement of Nile River flooding. Ruler duration in Egypt was relatively unstable during
this period, with high levels of Ottoman elite turnover.

10

0
1500 1550 1600 1650 1700 1750 1800

Figure 32.1: Smoothed duration of rule (in years) for Ottoman governors to Egypt.
Source: Data from Öztuna (1994) and Sureyya ([1890] 1996).
 State Institutions and Economic Performance    659

Religious Elites
During the final decades of Ottoman rule in Egypt, religious elites (ulema) came to ex-
ercise unprecedented power and influence. These religious elites included teachers,
scholars, and judges, as well as those at the top of the religious and political hierarchy
like the Rector of al-​Azhar, the muftis of the major schools of Islamic thought (sing.
madhhab, pl. madhahib), the chief representative of the descendants of the Prophet
(naqib al-​ashraf), and the heads of two major Sufi orders (Marsot 1968, 267). Religious
elites played the role of mediator between rulers and ruled, with ordinary Egyptians
submitting complaints and petitions through religious channels that might then be
conveyed to the political elite (Marsot 1973, 133).13
The political power of religious elites was magnified by their economic influence.
Religious elites enjoyed economic rents as a result of their role assisting in the admin-
istration of religious endowments (sing. waqf, pl. awqaf). While many tax farms were
controlled by Mamluks, religious elites began to purchase the rights to collect revenue
(Marsot 1973, 136).14 In some cases, the economic influence of religious elites served
to balance the economic strength of Mamluks (Marsot 1973, 137). For example, in one
case a Mamluk dignitary sought to increase taxes on a particular agricultural region;
the local tax farmers in that area appealed to the rector of al-​Azhar, who was also a tax
farmer in that area, and the rector was able to appeal to the Mamluk dignitary to limit
the increase in taxes (al-​Jabarti 2013a, 309–​310).

Muhammad Ali and

State Centralization

In July 1798, Napoleon Bonaparte and his forces landed in Alexandria. The army of
Ibrahim Bey and Murad Bey was defeated on July 21 during the Battle of the Pyramids.
While Ibrahim fled to Gaza and Murad to the South, the French established rule across
Egypt for a period of three years (Bjorneboe 2007, 68). During the occupation, the
French relied on religious elites from across Egypt in a bid to exert political control; on
the other hand, Mamluk forces were largely excluded from power (Fahmy 1998, 140;
Marsot 1968, 272; al-​Rafei 1981, 22–​25).
Despite the brevity of the French occupation, this period witnessed a weakening
of Mamluk power. War deaths and disease led to a reduction in the overall number of
Mamluks from between 10,000 and 12,000 to just 1,200 by 1802 (Bjorneboe 2007, 72;
Marsot 1984, 38). After the French departure, two Mamluk dignitaries from Murad’s
household, Uthman Bey al-​Bardisi and Mohamed Bey al-​Alfi, took on leadership for the
group, though internecine struggles for power were common (Fahmy 1998, 141).
During this period, Egypt experienced a breakdown of law and order. In urban areas,
soldiers attacked merchants, forcing them to lend money and to purchase goods with
660    Lisa Blaydes and Safinaz El Tarouty

forged currency. In rural areas, unpaid soldiers attacked peasants (Fahmy 1998, 141). The
economic deterioration was such that citizens accused Uthman Bey al-​Bardisi of living
off their economic ruin (Marsot 1968, 272).

Muhammad Ali and the Mamluks

With the French departure and a weakening of the Mamluks, the Ottoman Sultan
attempted to fill the power vacuum by sending a relatively small force of well-​trained
Albanian troops to re-​establish control over Egypt (Fahmy 1998, 141–​143). Mohammed
Ali was second-​in-​command of this unit and recognized the opportunity to exercise in-
fluence. Leveraging his control over the Albanian troops, as well as the Mamluks under
al-​Bardisi, he worked to deport the Ottoman governor to Istanbul. When various so-
cial forces—​like the religious elites, merchants, and local notables—​complained about
the high taxes needed to sustain the Mamluks and other military forces, Muhammad
Ali suggested that were he to rule, such taxes would not be required in the future (143).
Muhammad Ali was eventually named the Ottoman viceroy in Egypt with the rank of
pasha. He next focused his gaze upon the Mamluks.
In 1806 and 1807, two leading Mamluk dignitaries—​ al-​
Bardisi and al-​Alfi,
respectively—​died, leaving the various Mamluk factions relatively leaderless in the
face of Muhammad Ali’s strategic political maneuvers (Fahmy 1998, 145). Shortly
thereafter, Muhammad Ali expelled the Mamluks from the Nile Delta, giving them
authority in Upper Egypt. After failing to pay overdue taxes, however, Muhammad
Ali removed the Mamluks from Upper Egypt in 1810, returned them to Cairo, and
consolidated political control over all of Egypt.
The infamous event to follow led to a nearly complete destruction of the Mamluks.
In March of 1811, Muhammad Ali held a celebratory event at his residence in the
Citadel, to which all Mamluk chieftains were invited. While trapped in a passageway,
Muhammad Ali’s troops opened fire on the Mamluks, resulting in the deaths of
450 Mamluks; a campaign was then waged against those who managed to escape.15
Muhammad Ali went on to expropriate the tax farms held by Mamluks (Fahmy 1998,
148–​149). One result of Muhammad Ali’s consolidation of tax collection capacity
through expropriation of the Mamluks was that he was able to increase his annual rev-
enue considerably (149).

Fiscal Centralization
Muhammad Ali undertook a series of centralizing reforms that significantly increased
the capacity of the state. Most significantly, he abolished the system of tax farming. How
was this achieved? Expropriation of fiscal rights and resources rarely occurs without re-
sistance. Muhammad Ali first raised taxes on tax farmers, leading many of them to de-
fault; this action then providing him with an excuse to confiscate their properties. In
 State Institutions and Economic Performance    661

1809 Muhammad Ali imposed taxes on some land types that had previously been desig-
nated tax-​free properties (Rivlin 1961, 47–​48).
By 1815 most of Egypt’s agricultural land had been converted into state land. Profits
from agriculture became available to the ruler, and Muhammad Ali reorganized the ad-
ministrative structure of the government to ensure strict control of the economy.16 In
addition, a new tax on the waqf was implemented, though it had been previously tax-​
free.17 In sum, Muhammad Ali changed the land tenure system, abolished tax farming,
created a more centralized bureaucracy, and forced peasants to pay their taxes directly to
the government, all of which created a more centralized political and fiscal order (Issawi
1961, 4–​5; Panza and Williamson 2015).
Muhammad Ali’s fiscal centralization policies were accompanied by other
centralizing reforms. For example, by 1816 the government had come to control farming
to a much greater extent, providing peasants with capital and buying their crops at fixed
prices (Hunter 1984, 15). Beginning in 1821, long-​staple cotton was planted and exported
to Europe under the auspices of a state-​controlled monopoly; this pattern was extended
to other industries as well (Issawi 1961, 5). The increasing centralization of fiscal au-
thority allowed Muhammad Ali to invest in a professional, standing military. Panza and
Williamson (2015, 81) summarize these reforms:

[Muhammad Ali] intervened in an effort to augment state revenues to finance

his expansionary military agenda. . . . Ali adopted a series of state led fiscal and
trade policies aimed at moving resources into industry and increasing government
revenues. One of the most important reforms was the replacement of tax farming
with a system whereby land revenues accrued directly to the state. The reform also
extended taxable land through the introduction of a tax on waqfs, landholdings
managed by religious authorities and devoted to charitable purposes.

Panza and Williamson (2015) argue that although Muhammad Ali picked a difficult
time to pursue his economic agenda, his policies were largely successful at growing
the economy.18 Egypt was transformed from a primarily agrarian society to one with
growing industries in textiles, food processing and metallurgy.19

Muhammad Ali and the Religious Elite

Muhammad Ali developed a strong, early relationship with Egypt’s religious elites,
only to see that relationship deteriorate over time. The religious elite initially backed
Muhammad Ali over Ottoman alternatives, saying that “we will have none but you; you
will become wali over us, on our conditions and what we have seen from you of jus-
tice” (al-​Jabarti 2013b, 403). After initially demurring, Muhammad Ali accepted calls for
his formal accession to leadership. After accepting both the rector of al-​Azhar—​Shaykh
al-​Sharqawi—​and the naqib al-​ashraf—​Umar Makram—​these religious leaders placed
Muhammad Ali in a ceremonial kaftan as an official act of investiture (al-​Jabarti 2013b,
662    Lisa Blaydes and Safinaz El Tarouty

403). Muhammad Ali pledged an oath to rule justly and in consultation with the reli-
gious elite (bi mashurat al-​ulema) (Marsot 1968, 275).
In the years to follow, Muhammad Ali regularly consulted with the clerics of al-​
Azhar before imposing new taxes to pay for state expenses (al-​Rafei 1982, 32–​33). For
example, in September 1805, Muhammad Ali sought revenue to pay for military costs
and approached the leadership of al-​Azhar for an endorsement of higher tax rates. With
the consent of al-​Azhar, the state seized a third of the revenue collected by tax farmers
during that year. After the tax farmers objected, Muhammad Ali said that the confisca-
tion of revenue was an exceptional case related to rising military expenses. Despite the
fact that Muhammad Ali—​in consultation with the religious elites at al-​Azhar—​issued
an order prohibiting higher taxes, shortly thereafter the government imposed new taxes
on merchants (al-​Rafei 1982, 32–​33). The merchants complained to the religious elites
and asked them to negotiate a tax reduction (33). In 1806 and 1807, the taxes collected
by Muhammad Ali were insufficient to cover government expenditures and the clerics
of al-​Azhar agreed that he should be able to collect reduced taxes but also that religious
elites were to be exempted (Rivlin 1961, 47).
During the initial years of his rule, Muhammad Ali completely exempted the reli-
gious elites from taxes (Marsot 1984, 140). Beginning in October 1807, however, religious
elites were no longer exempt (al-​Rafei 1982, 84).20 Muhammad Ali was also able to effec-
tively cultivate divisions between members of the religious elites, particularly through
the exile of Umar Makram, an act that decreased the collective influence of the cler-
ical establishment (Marsot 1968, 276–​277).21 Marsot (1968, 277) writes that with Umar
Makram gone, the religious elites were unable to mount an effective opposition to the
power of Muhammad Ali; and once he confiscated the religious endowments, the ulema
“became dependent on the pasha’s bounty for survival, and in a greater position of sub-
servience than they had ever been.”

Creating (and Recreating) Elites

After eliminating the Mamluks and reducing the power of the clerics, there was a void
in the class of political elites that came to be associated with a new landed aristocracy.
In this spirit, Muhammad Ali granted land (ib’adiya) to high state officials and re-
gime insiders, who were exempted from paying taxes on the condition they cultivate
the land.22 Those who received the land only enjoyed usufruct rights, though in 1842
Mohammad Ali granted complete rights of ownership, including sale and transfer rights
(Baer 1962, 17). He also granted estates to royal family members (jiflik), including to
himself and members of his family (18).23 One of the main groups to benefit from land
grants were an emerging class of landed elites who did not simultaneously serve as vil-
lage mayors.24 Many of these individuals came from foreign, particularly Circassian and
Turkic, backgrounds.25 Often, those who benefited from the land grants were former
Mamluks who had reconciled with Muhammad Ali and opted for a permanent settle-
ment with the state (Abu-​Lughod 1967, 332).
 State Institutions and Economic Performance    663

Ethnicity—​particularly Turkish or Circassian background—​was no longer a decisive

factor in the assignment of high-​level positions and elite status, however (EzzelArab
2004). Although these ethnic groups continued to be well represented at the elite level,
indigenous Egyptians increasingly came to enjoy elite status as large landowners and
members of the military or bureaucracy elite. EzzelArab (2004) argues that both indig-
enous Egyptians and elites from a Turkish or Circassian background came to enjoy a
stake in the existing regime, as it served their economic interests.
This section has argued that Muhammad Ali oversaw the centralization of polit-
ical authority and a decrease in the power of the Mamluks. Early during his time in
power, religious elites—​like the rector of al-​Azhar—​were able to exert influence over
policy by channeling complaints from various societal classes and bringing those issues
to Muhammad Ali (al-​Rafei 1982, 85); in exchange, religious elites enjoyed special tax
status and other privileges. Over time, however, religious elites saw their influence de-
cline. Muhammad Ali eliminated the former ruling elite, expropriated the tax-​farming
class, and subordinated religious elites to the state. In addition, Muhammad Ali oversaw
the fiscal centralization of the Egyptian state through a channeling of agricultural re-
sources into the hands of an increasingly centralized state authority. The net result was
that “no group within Egyptian society was capable of forcing fundamental changes
upon him [Muhammad Ali] . . . elements that might have served as the instruments of
change had been crushed at the outset of his regime.”26

Weak Parliamentary Institutions and

the Limits of Executive Constraint

In 1829 Muhammad Ali established the Council of Consultation (Majlis al-​Mashura),

which was composed of senior government employees, religious elites, and local
notables, and headed by his eldest son, Ibrahim Pasha. The council was created with
the nominal purpose of advising the government regarding how to best administer the
country (Marsot 1984, 108; Sami 2009, 349–​352). Despite this mandate, the council met
just once a year and had only an advisory role (al-​Rafei 1982, 517). Under Muhammad
Ali’s successors—​Khedive Abbas and Khedive Sa’id—​the Majlis al-​Mashura did not
meet at all (al-​Rafei 2016, 328).
Beginning in the late 1850s, Egypt’s external debt began to balloon, and existing
institutions were poorly positioned to halt this development. For example, construc-
tion of the Suez Canal was financed, in part, with foreign loans that grew considerably
between 1858 and 1860 (Issawi 1961). In 1862, under Khedive Sa’id the Egyptian govern-
ment took out the first state loan (rather than a loan to the khedive), with repayment
guaranteed by revenue generated by the provinces (Issawi 1961). Egyptian economic
fortunes looked very positive at this period as the US Civil War meant that demand for
Egyptian cotton was especially high.
664    Lisa Blaydes and Safinaz El Tarouty

Changes in the international economic environment also contributed to the growth

of Egyptian foreign debt. The globalization of international finance accelerated over
the course of the nineteenth century, a trend that coincided with the development of
nation-​states in world regions outside of Europe (Queralt 2019). The ready availability
of external finance may have dampened the need for governments to increase domestic
taxation, with associated implications for the development of state institutions, like ro-
bust parliaments (Queralt 2019).
In 1866, Khedive Ismail (1863–​1879), grandson of Muhammad Ali, established the
Advisory Council of Delegates (Majlis Shura al-​Nuwwab). Al-​Rafei (2001, 88) writes
that this council was similarly focused on a consultative role.27 Despite the state’s con-
tinued borrowing and major efforts at state-​building, there was no opportunity for
the council to exercise control over the decisions of the executive. Scholch (1981, 15),
for example, argues that the Advisory Council of Delegates was not an instrument for
constraining the power of the khedive. Indeed, the by-​laws of the council gave the khe-
dive the power to dissolve the body, to modify the duration of parliamentary sessions,
to call for elections, and to appoint both the chair and deputy chair (al-​Rafei 2001,
85–​88).28
During the first two sessions of the Advisory Council of Delegates, discussion fo-
cused on issues that reflected the interests of the delegates rather than the country as a
whole or the heavily burdened Egyptian peasantry.29 Parliamentary actions dealt with
the distribution of unused agricultural parcels to the landed elites, rescheduling of tax
payments for landholders, amendments to the inheritance law that sought to prevent
fragmentation of land, and repairing of canals and bridges to provide irrigation water
for landed elite (Barakat 1977, 380–​381).
In 1868, the minister of finance, Ismail Pasha Seddiq, went to the Advisory Council of
Delegates seeking support for an increase in taxes as well as the adoption of a new loan
in the value of 5 million Egyptian pounds. Despite the large value of previous unpaid
loans, the members of the council did not oppose the new loan request. In fact, Ismail
Pasha Seddiq went as far as to borrow almost 12 million Egyptian pounds, more than
what he had pledged to the deputies (al-​Rafei 2001, 110–​111). Salem (2009, 138) finds
that Khedive Ismail sent an envoy who would inform the members of his will; deliber-
ation took place, but delegates did not typically oppose the khedive’s directive. In other
cases, the council was not consulted on high-​salience policy decisions. The net result
was that the Egyptian deficit grew further, reaching increasingly unsustainable levels
(Tuner 2015).
In response to the debt crisis, the Egyptian government began exploring options to in-
crease government revenue. In addition to the Advisory Council of Delegates, the khe-
dive also convened a Privy Council (Majlis Khususi) made up of ministers at the head of
nine government departments, in addition to a small number of select individuals.30 In
1871 the Privy Council submitted the Muqabala Law (i.e., the Law for Exchange) to the
khedive for confirmation (Abd al-​Moaty and Hosny 2006, 8). The law was designed to
trade off a 50 percent tax reduction on landed property in return for the landholders’ ad-
vance payment of land taxes.31 The khedive’s advisors believed that Egypt’s growing state
 State Institutions and Economic Performance    665

debt could be eliminated in a relatively short period of time through the adoption of this
law, even if it meant borrowing against the future (Hunter 1984, 180–​181).
In practice, the Muqabala Law failed to raise the amount of funding needed and
only seemed to benefit large landowners (Hunter 1984, 180–​181). Elites expressed little
concern for how these payments would be used or the long-​term implications of the
Muqabala Law for Egypt’s financial solvency (Weipert-​Fenner 2020, 50). During a
meeting called by Khedive Ismail to discuss the Muqabala Law, delegates speculated
about the future of the law while also raising questions about the 1875 budget (Hosny
and Abd al-​Hamid Mohammed 2011, 10–​12). In 1876 the Advisory Council of Delegates
affirmed the Muqabala Law, an outcome deemed in the interest of the delegates
(Ezzelarab 2009, 317).
Egypt’s economic situation became increasingly grim over the course of Khedive
Ismail’s reign. It is worth pointing out that the Egyptian experience was consistent with
financial crises taking place in states around the world. Abetted by European financiers,
countries like Egypt struggled to service their debt, a pattern that eventually led to waves
of default in several postcolonial contexts (Taylor 2006). Many of the loans secured were
used to enhance the khedive’s personal power and interests. For example, a portion of
one loan was used bribe the Ottoman Sultan to change a succession rule to allow po-
litical control to pass to Khedive Ismail’s eldest son instead of his brother or his uncle
(al-​Rafei 2001, 35).
By 1876, Egypt’s external debt (including the Khedive’s private debt) reached 91 mil-
lion Egyptian pounds, an amount equivalent to 80 percent of total revenues for that
year (Amin 2014, 27–​28). This was the point at which European economic and polit-
ical actors began to become more involved in the administration of Egyptian finances.
In December 1875 a committee headed by a senior British official was invited by
Khedive Ismail to report on the state of finances in Egypt. In May 1876 the Caisse de
la Dette was created for direct receipt of public revenues allocated for debt repayment
(EzzelArab 2002, 11). Figure 32.2 suggests that expenditure far outstripped revenue for
more than a decade.32
In January 1878, Khedive Ismail established a commission of inquiry (Commission
Superieure d’Enquete) to investigate the sources of Egypt’s insolvency, as well as to
recommend financial reforms. By August 1878 the commission had submitted its pre-
liminary report, which included recommendations regarding corruption, the need to
increase taxes, as well as the curbing the economic and political powers of the khedive
(EzzelArab 2002, 12–​13). The report recommended that the khedive delegate all respon-
sibility to an independent cabinet—​a council of ministers. In this context, the khedive
would have the role of endorsing the cabinet’s decisions.
Khedive Ismail appointed Nubar Pasha as Egypt’s first prime minister.33 Two of the
most important portfolios in Nubar’s cabinet—​minister of finance and minister of
public works—​were given to Europeans who were members of the Commission of
Inquiry (EzzelArab 2002, 44–​45). While members of the Advisory Council of Delegates
tried to constrain the khedive’s power, they were largely unsuccessful. In fact, European
influence may have led the delegates to be less likely to oppose the khedive. Scholch
666    Lisa Blaydes and Safinaz El Tarouty

20

15

10

0
1850 1855 1860 1865 1870 1875 1880 1885 1890 1895
revenue expenditure

Figure 32.2: Revenue and expenditure for Egypt (in millions of Egyptian pounds).
Source: Data from Henin (2008).

(1981, 80) writes that “the vehement opposition exercised by the delegates was . . .
against the Council of Ministers . . . (and) the delegates aligned themselves with the
Khedive against the ‘European government.’”
Early 1879 witnessed a growing confrontation between traditional elites and European
cabinet members over policy measures related to taxes and resource access (EzzelArab
2004). In March 1879 additional executive powers were delegated to Europeans. These
ministers had the right to veto any decision, provided the veto was issued by both.34
When the two European ministers decided to end the Advisory Council of Delegates
session held on the 27th of March (and Prime Minister Riyad Pasha supported their
decision), members organized opposition. In particular, the delegates refused to leave
the assembly without inspecting the government’s financial policy. This act represented
the first time that delegates directly opposed the government (Abduh 1961, 60; al-​Rafei
2001, 191–​194). Days later, the delegates signed a petition of grievances to present to the
khedive, protesting disregard of the assembly in addition to the cabinet’s proposal to
declare Egypt bankrupt (al-​Rafei 2001, 194–​195). According to EzzelArab (2004), these
efforts represented a meaningful attempt to derail European-​proposed policy measures.
In April 1879 a national convention was held that included a wide range of
participants, including traditional elites, religious dignitaries, merchants, and army
officers. The group promulgated a “national program” (al-​laiha al-​wataniyya) that was
sent to the khedive with over three hundred signatories (al-​Rafei 2001, 199). Members of
the Advisory Council of Delegates contributed to less than 20 percent of the signatories,
however, suggesting that the national program was supported by social forces distinct
from council (EzzelArab 2009, 309). Debt settlement featured prominently in the na-
tional program as well as a proposed requirement that the Council of Ministers submit
to the authority of the Advisory Council of Delegates (al-​Rafei 2001, 197–​199). EzzelArab
 State Institutions and Economic Performance    667

(2009, 320) argues that, at its core, the national program was “an attempt to redistri-
bute power among and within groups of contenders to political control who claimed
the space created by the decline in khedivial powers, namely Europe and the traditional
(Egyptian) elites.”
Khedive Ismail was sympathetic to the demands of the national program, but in June
1879, European and Ottoman administrators forced Khedive Ismail to abdicate.35 After
Ismail was replaced by Khedive Tawfiq, the Advisory Council of Delegates met on July
6, 1879, and was disbanded after it was suggested that its deliberations were proving too
long (Landau 1954, 25).36 In 1881, 1,600 people presented a petition to Khedive Tawfiq
asking the government to convoke a new meeting of the council.37
The British occupation of Egypt began in 1882, and Lord Dufferin was sent to Egypt
to reorganize its administrative machine. Dufferin’s final report suggested the creation
of two semi-​parliamentary institutions—​the Consultative Legislative Council (Majlis
Shura al-​Qawanin) and the General Assembly (al-​Jam’iyya al-​Umumiyya). Under the
British occupation, the principle of subjecting the cabinet to the parliament was not
mentioned in the Egyptian constitution until 1923 (al-​Rafei 2001, 203). Despite the exist-
ence of multiple Egyptian consultative and parliamentary institutions, none were able to
impose forms of executive constraint on the khedives of the nineteenth century and the
early 20th century.38

Implications and Conclusions

The idea that the wealthy countries of the world today were both fiscally centralized
and politically limited is intuitive and powerful. Prominent theories focused on eco-
nomic development in Europe suggest that state centralization was a key precondition
for achieving higher levels of growth and prosperity. Epstein (2000, 8), writing of the
European experience, suggests that “limitations to, rather than excesses of, state sover-
eignty are what restrained the rise of competitive markets.” In this setting, only states are
in a position to provide the institutional framework required for high levels of economic
growth. Dincecco (2011) develops this theoretical line by arguing that, in historical
Europe, the interaction of state fiscal centralization and limited government combined
to encourage economic development.
This essay examines these arguments in the context of state and economic devel-
opment in the Middle East, particularly nineteenth-​century Egypt. Prior to the nine-
teenth century, local military and religious actors exerted meaningful forms of political
influence during Egypt’s period of relatively weak and decentralized Ottoman control.
Over the course of the nineteenth century, however, Muhammad Ali introduced a set
of modernizing reform that included centralization of fiscal authority. Scholars have
suggested that these reforms did encourage forms of economic development for Egypt
(Panza and Williamson 2015), though not at levels comparable to what was happening
contemporaneously within Europe.
668    Lisa Blaydes and Safinaz El Tarouty

Over the longer term, what were the implications of state centralization for economic
growth? In particular, we consider the extent to which elites were able to exert forms of
executive constraint within Egypt. We find that the centralizing policies introduced by
Muhammad Ali represented an important break when compared to Egypt’s traditional
pattern of governance. His reforms decreased the influence of the Mamluks and reli-
gious elites—​the two most important power brokers in the previous period. As these
influential actors were sidelined, nascent representative institutions were created. These
institutions lacked influence, however, as they tended to be consultative, and delegates
used their institutional platform to pursue their own rent-​seeking agendas rather than
to provide a check on the executive.
During this time, spending initiated by the Egyptian khedives increased consider-
ably. A growing Egyptian debt took place in a global context characterized by external
loans proffered by European financiers. While access to credit allowed for investment
in state-​building efforts, these loans also financed unproductive investments, including
khedival vanity projects. Egyptian elites, however, were poorly positioned to rein in for-
eign borrowing, even though they came to strongly oppose losses of political power to
European actors due to debt. Queralt (2019) stresses the dangers of external financing,
arguing that domestically collected taxes have different implications for state-​building
compared to politically “cheap” foreign loans.
European actors began to play a larger role in governance of Egypt, and by 1882,
Egypt’s inability to repay its debt had opened the door to foreign domination by Britain.
The evidence we have presented suggests that, at least for the Egyptian case, the order of
reforms matters for achieving economic development. If centralizing reforms eliminate
the strength of elite power brokers, the process of centralization itself may limit the pos-
sibility for future executive constraint. Traditional elites in Egypt were concerned about
losing political decision-​making and administrative power to Europeans (EzzelArab
2004), but they were unable to stop the state from taking on crippling levels of foreign
debt that came to damage their interests.
This essay has pointed to limits associated with the external validity of existing re-
search on development in Europe as applied to the premodern Middle East. There are a
number of reasons why these ideas do not extend easily to the Egyptian context. First,
the nature of centralizing reforms may have been different in Egypt as compared to the
European context. Second, perhaps Egypt’s economic development trajectory was de-
termined by factors that long preceded debates regarding the importance of nineteenth-​
century centralization and executive constraint (e.g., Blaydes and Chaney 2013; Kuran
2010; Rubin 2017). Finally, even if Acemoglu (2005) was fundamentally right about the
importance of balancing state strength and state weakness, perhaps such an outcome is
simply difficult to calibrate. When states are too strong, rulers impose taxes that stifle ec-
onomic activity, but when states are excessively weak, the ruler does not believe he will
be able to extract rents in the future and fails to provide public goods. Egypt was not able
to achieve this balance over the course of nineteenth-​century reforms, though the pre-
cise reasons why remain opaque and debated.
 State Institutions and Economic Performance    669

Notes
1. For Acemoglu (2005), weak states underinvest in public goods because rulers are self-​
interested and do not expect to derive any future private rewards from that action.
2. For Dincecco (2011), fiscal centralization is operationalized as the year that a country
secured revenue through a tax system with uniform rates across the country; executive
constraint (i.e., limited government) is generally coded as the year that a country estab-
lished a constitutional monarchy.
3. England is coded as having limited government in 1688, the majority of European country
cases see the arrival of limited government between 1831 and 1876 with the establishment of
constitutional monarchies and stable parliamentary regimes. For example, Dincecco (2011)
sees fiscal centralization taking place in France starting in 1790 and in Denmark in 1903.
4. Even the most prosperous of premodern states lacked highly centralized institutions by
modern standards.
5. Besley and Persson (2009) find that common interest public goods, along with political
stability and inclusive political institutions, are conducive to building state capacity.
6. There are a variety of terms used to describe the Ottoman viceroys to Egypt, some of which
overlap with the terminology used to describe Egypt’s slave soldiers, the Mamluks (sing.
mamluk, pl. mamalik). For example, Khayr Bey ruled as the Ottoman viceroy (wali) until
he passed away in 1522. With the exception of Khayr Bey, all viceroys who followed him
were given the title of pasha. The Mamluks were collectively referred to as the beys (Marsot
1985, 40).
7. For example, Khadim Hafiz Muhammad (1591–​1595) knew the Qurʾan and read it every
Friday. Most of the pashas adhered to Islamic law (shariʿa). But Dugakin Oglu Mohammed
(1554–​1556) was deemed to be wanton and immoral, and he was executed by the sultan
(Winter 1998, 7–​9).
8. A number of defeated Mamluks were pardoned and permitted to remain in Egypt
(Hathaway 1998, 37).
9. According to Abdel Qadir al-​Jaziri, from 1517 to 1550, there were fifteen umara al-​hajj, six
of whom were Mamluk amirs (Winter 1998, 13).
10. Marsot (1984, 139) writes that “society at that time was built on intermediaries acting as
buffers between the population and the administration . . . the authority of the central
government was gradually usurped by the intermediaries, especially in time of crisis.”
11. After the death of Ibrahim Kahya al-​Qazdaghl in 1754, one of his Georgian Mamluks, Ali
Bey al-​Kabir, continued his predecessor’s policies in controlling political and economic
life. Ali Bey al-​Kabir’s most trusted retainer, Mohammed Abu al-​Dahab, followed him.
Upon the death of Mohammed Abu al-​Dahab in 1774, leadership was split between two
of his retainers, Ibrahim Bey and Murad Bey, who together dominated political life in
Ottoman Egypt until the arrival of the French in 1798 (Bjorneboe 2007, 30).
12. The duration of Ottoman governors in Egypt has been well documented in Ottoman
sources, including Öztuna (1994) and Sureyya (1996). The data presented in Figure 32.1 are
smoothed with a ten-​year moving average.
13. Rubin (2017, this volume) argues that religious elites long enjoyed forms of political legiti-
macy in Muslim societies.
14. Moreover, some tax farmers were transforming their landholdings into non-​taxed reli-
gious properties that could be passed on to heirs of the tax farmer. For example, according
670    Lisa Blaydes and Safinaz El Tarouty

to the 1812 cadastral survey of agriculture lands, waqf properties were estimated to com-
prise 20 percent of all cultivated land (Marsot 1984, 143).
15. In 1812 Ibrahim Pasha was sent by Muhammad Ali to Upper Egypt to search for the
Mamluks who escaped the Citadel massacre (Fahmy 1998, 146–​148). While some scholars
argue that the Citadel massacre constituted the complete end of Mamluks in Egypt, Abu-​
Lughod (1967, 331–​332) argues that “many mamluks who had fled to Upper Egypt opted
for a more benign alternative.” Indeed, Hunter (1984, 25) argues that Muhammad Ali em-
ployed a large number of Mamluks, including 500 who worked for him directly in the civil
and military administration or were tutors within his household.
16. Helen Rivlin, “Muhammad Ali: Pasha and Viceroy of Egypt,” Encyclopedia Britannica,
https://​www.britannica.com/​biography/​Muhammad-​Ali-​pasha-​and-​viceroy-​of-​Egypt.
17. The status of waqf properties in light of a centralizing state was increasingly precar-
ious. In 1846 Muhammad Ali sought to issue a decree preventing the creation of new
waqf properties from agricultural land. He justified this policy change by arguing waqf
properties worked against Islamic law regarding inheritance and also prevented the gov-
ernment from collecting taxes. The head mufti refused to issue a religious edict (fatwa)
endorsing Muhammad Ali’s desire to place limits on the waqf. Ibrahim Pasha, the son
of Muhammad Ali, advocated for the removal of the mufti; instead, Muhammad Ali
approached the mufti of Alexandria to issue the fatwa instead. The fatwa was revoked a
short time later (Hillal 2016, 1835–​1838).
18. See Saleh (2015, this volume) for more on the impact of this economic agenda on socioeco-
nomic inequality between non-​Muslim minorities and the Muslim majority in Egypt. For
example, Saleh (2015) shows the impact of state industrialization in the Christian-​Muslim
occupational difference between 1816 and 1868.
19. In contrast to other analysts, Panza and Williamson (2015) consider the counterfactual of
what would have happened to the Egyptian economy were Muhammad Ali to have not
intervened, in broader, comparative context.
20. Muhammad Ali sparred with religious elites over questions related to economic jus-
tice. For example, in September 1808, the rector of al-​Azhar (Shaykh al-​Sharqawi) asked
Muhammad Ali to show more kindness to the Egyptian people through an easing of the
sources of economic and societal injustice. He pushed back against the characterization
that he alone was responsible for injustice in Egypt and accused the religious elites of
abusing the peasantry. In particular, he pointed to the fact that when he eased taxes on the
religious elites, they then turned around and increased their own taxes on the peasants (al-​
Jabarti 2013c, 93–​94).
21. Following the exile of Umar Makram, Muhammad Ali was no longer on good terms with
the rector of al-​Azhar but dared not remove him; instead he was put under house arrest.
See Hillal (2015, 1286) for more information on this subject.
22. The first decree by Muhammad Ali for ib’adiya was in 1826 (Baer 1962, 16; Barakat 1977).
23. For example, the first jiflik was granted to Mohammed Ali’s daughter in 1838 (Barakat
1977, 36).
24. See Baer (1962, 47) for a discussion of this class of elites, known as zawat.
25. For example, Khorsheid Pasha—​the governor of Dakhaliyya—​was originally Georgian
and owned more than one-​fifth of the villages in that province. Mohammed Sharif Pasha,
son of a Turkish judge, was appointed prime minister under Khedives Ismail and Tawfik
and was granted 1,000 feddans from Ismail in addition to an ib’adiya he held. Mahu Pasha
 State Institutions and Economic Performance    671

was appointed the governor of Sudan in the 1820s; by the end of the 19th century, his heirs
owned over 1,000 feddans of property (Baer 1962, 47).
26. Helen Rivlin, “Muhammad Ali: Pasha and Viceroy of Egypt.”
27. For example, in his opening speech to the council on November 25, 1866, Khedive Ismail
said, “I often thought of establishing Majlis Shura al-​Nuwwab . . . no one can deny the use-
fulness and advantages that the matter should be a consultation between the rulers and the
ruled.” In the speech, Ismail also referenced Qurʾanic verses about the importance of con-
sultation. See al-​Rafei (2001, 92–​93) for more details.
28. Even at the time that the council was established, elites were aware that their influence over
policy would be limited. For example, Lady Duff Gordon relates a relevant anecdote in
her “Letters from Egypt” (Ross 1902). She meets a group of Nubian deputies who are on a
boat down the Nile on their way to the inauguration of the council. She says to them, “Now
you will help to govern the country, what a fine thing for you,” to which the deputies re-
spond, “What words are these? Who is there on the banks of the Nile who can say anything
but hadir (i.e., to assent)?” (Ross 1902, 315). According to another anecdote, during the
opening of the council in 1866, deputies refused to sit on the left of the chamber because
this would suggest that they were in opposition to the government, an impossible position.
One of the deputies reportedly said, “We are all your servants, so how can we resist your
government?” This anecdote is disputed, however, and there are no surviving minutes of
this meeting in the parliamentary archive. See Rizk (1995, 129) and al-​Rafei (2001, 105–​
106) for more on this subject.
29. The peasants were burdened by the heavy taxes imposed upon them. For more informa-
tion on this topic, see Salem (2009, 28–​30) and Barakat (2018, 208).
30. The Privy Council was established in 1847, before the creation of the Advisory Council of
Delegates. For more information on the Privy Council, see al-​Rafei (1982, 523–​524).
31. In some cases, the khedive used the Advisory Council of Delegates to endorse policies to
increase the wealth of his family. This was evident when the council ratified Ismail’s deci-
sion to confiscate land that was subsequently allocated to members of his family (Hunter
1984, 53).
32. These data come from Henin (2008).
33. The Council of Ministers (Majlis al-​Nuzzar), was created by Khedive Ismail in 1878 to re-
place the Privy Council.
34. Following this declaration, Prince Mohamed Tawfiq was appointed prime minister in
March 1879 (al-​Rafei 2001, 189–​190). By late March, the confrontation between the Council
and the European cabinet reached its highest level. Forty-​nine members of Council wrote
a memorandum against the minister of finance for failing to report fiscal plans.
35. See al-​Rafei (2001, 242–​250) for a complete discussion.
36. The name of the council was changed from Majlis Shura al-​Nuwwab to Majlis al-​Nuwwab
by the end of Khedive Ismail’s rule.
37. The khedive convened the council, electing members based on the 1866 Electoral Law de-
spite controversy regarding the appropriate rules to use.
38. This is not to say that representative institutions represented the only locales from which
executive constraint might emanate. For example, by the late 1870s Gamal al-​Din al-​
Afghani had formed a national freemasons lodge that had over 300 members, including
journalists, intellectuals, notables, and religious elites, as well as some members of the
Assembly and the military (Kudsi-​Zadeh 1972, 30).
672    Lisa Blaydes and Safinaz El Tarouty

References
Abd al-​Moaty, Hossam, and Saeida Hosny. 2006. Mahadar Majlis Shura al-​Nuwwab, 1870–​1873
[Minutes of the meetings of the Advisory Council of Delegates, 1870–​1873]. Cairo: Dar
al-​Kutub wal Watha’iq al-​Qawmiya.
Abduh, Mohamed. 1961. Muzakart Mohamed Abduh [Memoirs of Mohamed Abduh]. Cairo:
Dar al-​Hilal.
Abu-​Lughod, Ibrahim. 1967. “The Transformation of the Egyptian Elite: Prelude to the Urabi
Revolt.” Middle East Journal 21, no. 3: 325–​344.
Acemoglu, Daron. 2005. “Politics and Economics in Weak and Strong States.” Journal of
Monetary Economics 52, no. 7: 1199–​1226.
Amin, Galal. 2014. Qisat a-​Iqtisad al-​Masri min Ahd Mohammed Ali ila Ahd Mubarak [The
story of the Egyptian economy from Mohammed Ali to the Mubarak Era). Cairo: Dar
al-​Shorouk.
Baer, Gabriel. 1962. A History of Landownership in Modern Egypt, 1800–​1950. Oxford: Oxford
University Press.
Barakat, Ali. 1977. Tatawur al-​Malkiya al-​Zara’ayia fi Misr, 1813–​1914 [The development of agri-
culture ownership in Egypt, 1813–​1914). Cairo: Dar al-​Thakafa al-​Jadida.
Barakat, Ali. 2018. Al-​Qariya wa al-​Solta fi Misr fi al-​Qarn al-​Tasa’ Ashr [Village and power in
19th-​century Egypt). Cairo: Al-​Haiy’a al-​Amma li al-​Kitab.
Besley, Timothy, and Torsten Persson. 2009. “The Origins of State Capacity: Property Rights,
Taxation and Politics.” American Economic Review. 99, no. 4: 1218–​1244.
Bjorneboe, Lars. 2007. In Search of the True Political Position of the ‘Ulama. Aarhus, Denmark:
Aarhus University Press.
Blaydes, Lisa. 2017. “State Building in the Middle East.” Annual Review of Political Science
20: 487–​504.
Blaydes, Lisa. 2019. “Mamluks, Property Rights and Institutional Durability in Medieval
Egypt.” Politics & Society 47, no. 3: 395–​424.
Blaydes, Lisa, and Eric Chaney. 2013. “The Feudal Revolution and Europe’s Rise: Political
Divergence of the Christian West and the Muslim World before 1500 CE.” American Political
Science Review 107, no. 1: 16–​34.
Crecelius, Daniel. 1998. “Egypt in the Eighteenth Century.” In The Cambridge History of Egypt.
Vol. 2, Modern Egypt from 1517 to the End of the Twentieth Century, edited by M. W. Daly,
59–​86. New York: Cambridge University Press.
Dincecco, Mark. 2011. Political Transformations and Public Finances: Europe 1650–​ 1913.
New York: Cambridge University Press.
Dincecco, Mark, and Gabriel Katz. 2014. “State Capacity and Long-​Run Economic Performance.”
Economic Journal 126: 189–​218.
Epstein, S. R. 2000. Freedom and Growth: The Rise of States and Markets in Europe, 1300–​1750.
London: Routledge.
EzzelArab, AbdelAziz. 2002. European Control and Egypt’s Traditional Elites: A Case Study in
Elite Economic Nationalism. Lewiston, NY: Edwin Mellen Press.
EzzelArab, AbdelAziz. 2004. “The Experiment of Sharif Pasha’s Cabinet: An Inquiry into the
Historiography of Egypt’s Elite Movement.” International Journal of Middle East Studies 36,
no. 4: 561–​589.
EzzelArab, AbdelAziz. 2009. “The Fiscal and Constitutional Program of Egypt’s Traditional Elites
in 1879: A Documentary and Contextual Analysis of ‘al-​La’iha al-​Wataniyaa’ (‘The National
Program’).” Journal of the Economic and Social History of the Orient 52, no. 2: 301–​324.
 State Institutions and Economic Performance    673

Fahmy, Khaled. 1998. “The Era of Muhammad Ali Pasha.” In The Cambridge History of Egypt.
Vol. 2, Modern Egypt from 1517 to the end of the Twentieth Century, edited by M. W. Daly,
139–​179. New York: Cambridge University Press.
Hathaway, Jane. 1995. “The Military Household in Ottoman Egypt.” International Journal of
Middle East Studies 27, no. 1: 39–​52.
Hathaway, Jane. 1998. “Egypt in the Seventeenth Century.” In The Cambridge History of Egypt.
Vol. 2, Modern Egypt from 1517 to the end of the Twentieth Century, edited by M. W. Daly,
34–​58. New York: Cambridge University Press.
Henin, Gerges. 2008. Al-​Atyan wal Dara’ib fil Qutur al-​Masri [Land and taxes in Egypt].
Cairo: al-​Hayat al-​Amma lil Kitab.
Hillal, Emad. 2015. Al-​Ifta’a al-​Masry min al-​Sahabi Okba ibn Amer ila al-​Doctor Ali Gomaa
[The Egyptian Ifta’a from the Companion Okba ibn Amer to Doctor Ali Gomaa). Vol.
3. Cairo: Dar al-​Ifta’a al-​Masriyya and Dar al-​Kutub wal Watha’iq al-​Qawmiya.
Hillal, Emad. 2016. Al-​Ifta’a al-​Masry min al-​Sahabi Okba ibn Amer ila al-​Doctor Ali Gomaa
[The Egyptian Ifta’a from the Companion Okba ibn Amer to Doctor Ali Gomaa]. Vol.
4. Cairo: Dar al-​Ifta’a al-​Masriyya and Dar al-​Kutub wal Watha’iq al-​Qawmiya.
Hosny, Saeida, and Dina Abd al-​Hamid Mohammed. 2011. Mahadar Majlis Shura al-​Nuwwab,
1876–​1879 [Minutes of the meetings of the Delegates Consultative Assembly, 1876–​1879].
Vol. 3. Cairo: Dar al-​Kutub wal Watha’iq al-​Qawmiya.
Hunter, Robert. 1984. Egypt under the Khedives: 1805–​1879: From Household Government to
Modern Bureaucracy. Pittsburgh: University of Pittsburgh Press.
Issawi, Charles. 1961. “Egypt since 1800: A Study in Lop-​sided Development.” Journal of
Economic History 21, no. 1: 1–​25.
al-​Jabarti, Abd al-​Rahman b. Hasan, 2013a. Aja’ib al-​Athar fil Tarajim wal Akhbar [The mar-
velous chronicles, biographies and events]. Vol. 2. Jerusalem: Printiv Press.
al-​Jabarti, Abd al-​Rahman b. Hasan, 2013b. Aja’ib al-​Athar fil Tarajim wal Akhbar [The mar-
velous chronicles, biographies and events). Vol. 3. Jerusalem: Printiv Press.
al-​Jabarti, Abd al-​Rahman b. Hasan, 2013c. Aja’ib al-​Athar fil Tarajim wal Akhbar [The mar-
velous chronicles, biographies and events]. Vol. 4. Jerusalem: Printiv Press.
Kudsi-​Zadeh, Albert. 1972. “Afghani and Freemasonry in Egypt.” Journal of the American
Oriental Society 92, no. 1: 25–​35.
Kuran, Timur. 2010. The Long Divergence: How Islamic Law Held Back the Middle East.
Princeton, NJ: Princeton University Press.
Landau, Jacob M. 1954. Parliaments and Parties in Egypt. New York: Frederick A. Praeger.
Levi, Margaret. 1988. Of Rule and Revenue. Berkeley: University of California Press.
Marsot, Afaf Lutfi al-​Sayyid. 1968. “The Role of Ulama in Egypt during the Early Nineteenth
Century.” In Political and Social Change, edited by P. M. Holt, 264–​280. Oxford: Oxford
University Press.
Marsot, Afaf Lutfi al-​Sayyid. 1973. “The Political and Economic Functions of the Ulama in 18th
Century Egypt.” Journal of Economic History of the Orient 16, no. 2/​3: 130–​154.
Marsot, Afaf Lutfi al-​Sayyid. 1984. Egypt in the Reign of Mohammed Ali. New York: Cambridge
University Press
Marsot, Afaf Lutfi al-​Sayyid. 1985. A Short History of Modern Egypt. New York: Cambridge
University Press.
North, Douglass. 1981. Structure and Change in Economic History. New York: Cambridge
University Press.
North, Douglass, and Robert Paul Thomas. 1973. The Rise of the Western World: A New
Economic History. New York: Cambridge University Press.
674    Lisa Blaydes and Safinaz El Tarouty

Öztuna, Yılmaz. 1994. Büyük Osmanlı Tarihi. Istanbul: Ötüken Neşriyat A.S.
Panza, Laura, and Jeffrey Williamson. 2015. “Did Muhammad Ali Foster Industrialization in
Early Nineteenth-​Century Egypt?” Economic History Review 68, no. 1: 79–​100.
Queralt, Didac. 2019. “War, International Finance, and Fiscal Capacity in the Long Run.”
International Organization 73, no. 4: 1–​41.
al-​Rafei, Abdel Rahman. 1981. Tarikh al-​Haraka al-​Qawmiyya wa Tatawur Nizam al-​Hukm fi
Misr [The history of the National Movement and the development of governance in Egypt).
Vol. 2. Cairo: Dar El Ma’araf.
al-​Rafei, Abdel Rahman. 1982. Asr Mohammed Ali [The era of Mohammed Ali]. Cairo: Dar
al-​Ma’araf.
al-​Rafei, Abdel Rahman. 2001. Asr Ismail [Era of Ismail]. Cairo: Maktabat al-​Osra.
al-​Rafei, Abdel Rahman. 2016. Misr al-​Mojahada [Hardworking Egypt]. Cairo: Dar al-​Hilal.
Rivlin, Helen. 1961. The Agriculture Policy of Muhammad Ali in Egypt. Cambridge, MA:
Harvard University Press.
Rizk, Younan Labib. 1995. Al-​Ahram Diwan al-​Hawiya al-​Mu’asira, 1876–​1882 [Diwan al-​
Ahram in contemporary life]. Vol. 1. Cairo: Al-​Ahram.
Ross, Janet. 1902. Lady Duff Gordon’s Letters from Egypt. London: R. Brimley Johnson.
Rubin, Jared. 2017. Rulers, Religion and Riches: Why the West Got Rich and the Middle East Did
Not. New York: Cambridge University Press.
Saleh, Mohamed. 2015. “The Reluctant Transformation: State Industrialization, Religion, and
Human Capital in Nineteenth-​Century Egypt.” Journal of Economic History 75, no. 1: 65–​94.
Salem, Latifa Mohamed. 2009. Al-​Quwa al-​Ijtimayya fi al-​Thawra al-​Urabiyya [Social forces in
the Urabi Revolution]. Cairo: Al-​Sherouk.
Sami, Amin 2009. Taqwim al-​Nil [Almanac of the Nile]. Cairo: Ministry of Culture.
Scholch, Alexander. 1981. Egypt for the Egyptians! The Socio-​political Crisis in Egypt, 1878–​1882.
London: Ithaca Press.
Shaw, Stanford. 1958. The Financial and Administrative Organization and Development of
Ottoman Egypt, 1517–​1798. Princeton, NJ: Princeton University Press.
Sureyya (Bey), Mehmet. (1890) 1996. Sicill-​i Osmanı̂. Istanbul: Kültür Bakanlığı ile Türkiye
Ekonomik ve Toplumsal Tarih Vakfı’nın.
Taylor, Alan. 2006. “Foreign Capital Flows.” In The Cambridge Economic History of Latin
America, edited by Victor Bulmer-​Thomas, John Coatsworth, and Roberto Cortes-​Conde,
57–​100. New York: Cambridge University Press.
Tuner, Ali Coskun. 2015. Sovereign Debt and International Financial Control: The Middle East
and the Balkans, 1870–​1914. London: Palgrave.
Weipert-​Fenner, Irene. 2020. The Autocratic Parliament: Power and Legitimacy in Egypt,
1866–​2011. Syracuse, NY: Syracuse University Press.
Winter, Michael. 1998. “Ottoman Egypt, 1525–​1609.” In The Cambridge History of Egypt. Vol. 2,
Modern Egypt from 1517 to the end of the Twentieth Century, edited by M. W. Daly, 1–​33.
New York: Cambridge University Press.
 Chapter 33

Isl am and th e P ol i t i c s
of Devel opme nt
Shrines and Literacy in Pakistan

Adeel Malik AND Rinchan Mirza

Sufism has been involved in all that we think of as politics: conception of au-
thority and power, legitimacy and contestation of rule, formation of socio-​
moral order of a community or nation, competition for patronage, prestige
and control of a society’s wealth, the mobilization of people and resources in
support or against the status-​quo, and so on.
—​Paul L. Heck, Sufism and Politics (2007)

A growingly influential field of research has probed the impact of religious beliefs and
institutions on economic development (Carvalho, Iyer, and Rubin 2019). Considerable
scholarly effort has been expended on studying the impact of religion on human capital
through a focus on the role of Protestantism in Europe (Becker and Woessmann 2009),
Catholicism in France (Squicciarini 2020), and Christian missionaries in former colo-
nies (Chaudhry and Rubin 2011; Woodberry 2004,). However, the primary focus of this
research has remained on non-​Muslim contexts. Furthermore, as Iyer (2016) argues in
her wide-​ranging review of the literature on economics of religion, the relationship be-
tween religion, politics, and economic processes remains significantly under-​studied.
In an ongoing project (Malik and Mirza 2018), we seek to address both of these gaps by
examining the role of religious elites on literacy in the second-​most populous Muslim
country, Pakistan.
While past scholarship has mainly focused on the impact of religious beliefs,
practices, and norms on long-​run development, we offer a slightly different empir-
ical take by emphasizing the role of religious elites. Elite power plays a crucial role in
676    Adeel Malik AND Rinchan Mirza

dominant political economy accounts and helps to explain the strong patterns of insti-
tutional persistence witnessed globally (Acemoglu and Robinson 2008). The interests
and incentives of elites can matter for development, since they can impede or facilitate
prosperity-​inducing reforms. Elites can oppose technological change if they perceive it
as a threat to their future political power (Acemoglu and Robinson 2006; Cosgel, Miceli,
and Rubin 2012). Similarly, elite opposition to schooling in the context of high land in-
equality has received considerable empirical attention. For example, landed elites have
obstructed the spread of mass education in the United States (Galor, Moav, and Vollrath
2009), Latin America (Sokoloff and Engerman 2000), and India (Banerjee, Iyer, and
Somnathan 2005). Favorable elite incentives can, on the other hand, improve public
goods provision. Drawing on evidence from rural China, Tsai (2007) has shown that
in regions where local solidarity groups are stronger, elites are incentivized to improve
public goods provision to maintain and enhance their “moral standing.”
Despite this emerging emphasis on elites, the literature on religion and develop-
ment lacks a systematic inquiry on the role of religious elites in shaping development.
In the context of Muslim societies, the role of religious clergy (ulama) figures promi-
nently in broader historical accounts of divergence in development between Western
Europe and the Middle East (Kuru 2019; Rubin 2017). However, there are few system-
atic empirical attempts to capture the impact of Muslim religious elites on development.
Using historic flood data on the River Nile, Chaney (2013) demonstrated that Egypt’s
central religious authority assumed greater political power in the face of economic
shocks. Similarly, Bazzi, Koehler-​Derrick, and Marx (2020) shed light on the origins
of Islamic politics in Indonesia and demonstrate that the risk of land expropriation in
the wake of Indonesia’s land reforms in 1960 increased the transfer of lands into waqf
endowments, and these transfers left a deep imprint on the size and electoral support of
Islamic parties. However, these studies are principally concerned with understanding
the deep institutional foundations of Islam in the public domain and their focus remains
on religious clergy (ulama) and Islamic political parties. We study the direct impact of a
third dimension, Islamic mystical orders (Sufism), on long-​run development.
Sufism has long been understood as a central force in the transmission of Islam across
vast swathes of Africa, Asia, and the Middle East. The missionary activities of mystical
Islamic brotherhoods not only extended the borders of Islam, but also left a deep im-
print on local economic and political structures. Yet it is surprising that the emerging
political economy scholarship on religion has completely neglected this central aspect
of Islamic civilization. Our research provides a first stab at addressing this important
gap by studying the developmental impact of religious elites whose authority and power
is derived from distinguished Sufi shrines that were recognized and rewarded by British
colonial administration. Our empirical material is drawn from Pakistan’s largest prov-
ince, Punjab. In this chapter we synthesize evidence from two related papers, Malik and
Mirza (2018) and Malik and Malik (2017). Interested readers can find the full technical
scope of our argument in these papers.
Interspersed between Central Asia and the heartlands of India, Punjab is home to
some of the oldest Islamic mystical orders of South Asia. Traveling from Central Asia
 Islam and the Politics of Development     677

and the Middle East, early Muslim saints settled down in different parts of Punjab and
were credited with bringing entire tribes and communities into the fold of Islam. The
tombs constructed to celebrate the sacred tradition of these saints became the “sym-
bolic cultural outposts” of Islam (Gilmartin 1979). Guardians of these shrines acquired
immense legitimacy and prominence in the eyes of the local populace and became im-
portant controllers of economic and political power over time. Shrine caretakers have
historically acted as brokers between the centralized state and its peripheral subjects.
Whether Mughal rule or Sikh interregnum, British colonial administration or post-​
independence Pakistan, no ruler has passed without extending favor to powerful shrine
families to seek their support.
To define a measure of the power of shrine elites, we compile a novel data set on the
presence of historically significant shrines from colonial-​era district gazetteers. To
investigate their impact on literacy, we utilize a coup-​induced administrative shock
to public goods provision that gave elected politicians greater power over spending
and allocation of public goods in their constituencies. We show that the impact
of shrine elites on literacy only becomes salient after General Zia’s coup when the
system of public goods provision underwent a radical change. We argue that, with
their greater aversion to mass education, shrine elites leveraged their political power
to suppress literacy expansion in shrine-​dominated regions after the coup-​induced
reforms.
While our analysis substantiates prior evidence on the opposition of landed elites to
schooling, we emphasize religious, as opposed to economic, motives. We argue that,
compared with landed elites, shrine elites have both a greater incentive and capacity to
oppose schooling. The standard explanation rests on elites suppressing education by di-
rectly restricting educational spending or blocking the construction of new schools. We
offer a more nuanced mechanism whereby shrine elites make existing school provision
defective by influencing the location of schools, the quality of school infrastructure, and
school size.
In relating shrine elites with literacy, we are inspired by a well-​established literature
that presents education as an existential threat to the power of shrine elites since it can
interfere with the structure of voluntary obedience and compliance that they seek to
promote. Our focus on literacy as a development outcome connects with a vast liter-
ature on the role of human capital is a determinant of the wealth of nations (Glaeser
et al. 2004; Klenow and Rodriguez-​Clare 2005). In thinking about constraints to human
capital development, our work is directly related with studies that argue that schooling
differences across countries are grounded in history and political economy (Acemoglu
et al. 2014; Gallego 2010).
The remainder of this chapter is organized as follows: We start by offering a brief
background on the historically persistent power of shrine elites and synthesize the
core empirical insights of Malik and Mirza (2018). Subsequently, we discuss the main
mechanisms through which shrine elites influence literacy and highlight the salience of
the political dimension. The chapter concludes by drawing out the broader significance
of our analysis for the study of Islam and development.
678    Adeel Malik AND Rinchan Mirza

The Power of Shrine Elites

With regard to the influence of shrine families, three aspects are pertinent to our anal-
ysis. First, the spiritual and material influence of shrine caretakers is mainly derived
from their sacred lineage and their association with the shrine as a seat of religious
power. Second, the power of shrine families is historically embedded. While courted
by all rulers, the British colonial rule radically transformed their power through the
introduction of formal property rights. By barring the sale of land from “agrarian”
to “non-​agrarian castes,” the Land Alienation Act of 1900 effectively preserved and
expanded elite control over land (Ali 1987). Exercising their discretion, the British
colonial administration treated prominent shrine families as “agrarian castes,” which
made them fit to receive landed gentry grants. Large tracts of land were awarded
to prominent families in return for preserving social order and lending support to
British rule.
While past rulers similarly awarded land grants to key shrine families, these usually
reverted back to the state upon the death of a shrine guardian. However, with the estab-
lishment of property rights under the British, hereditary succession acquired a powerful
economic motive and converted these families into both spiritual and feudal masters, or
pīr-​zamindārs. Third, given their religious and economic power and their ability to act
as intermediaries of rulers, shrine guardians were natural contenders for electoral poli-
tics when the British opened the gates for it in the early twentieth century. Entering the
electoral race from an early period, big shrine families were able to expand their pres-
ence. Their political influence has gained salience post-​partition in Pakistani Punjab
and persists to the present day.
This leaves us with a final introductory hook around the possible significance of these
shrine elites for development. Our interest lies in exploring whether shrine-​dominated
regions witnessed a substantially retarded trajectory of literacy rates over time rela-
tive to non-​shrine regions. Our focus on literacy is motivated by both historical and
contemporary literature that alludes to the aversion of shrine elites to education. In an
important monograph, The Punjab Peasant, British civil servant Sir Malcolm Darling
presciently observed that shrine elites are “instinctively opposed to the two movements
from the which the ordinary cultivator has most to hope. Neither education nor coop-
eration has their sympathy, for both strike at the regime which it is their one object to
maintain” (Darling 1928, 100).
Subsequent studies have affirmed this insight. In Pakistan: A Hard Country, Anatol
Lieven echoes the same concern: “in practice the pirs and their families cannot genu-
inely advance either local education or local democracy, as this would strike directly at
the cultural and social bases of their own power.” (Lieven 2012, 138) In Religion, Land,
and Politics, the noted historian K. K. Aziz argues that it is difficult for shrine elites to
“countenance any prospect for the education of the masses when his supremacy, status
 Islam and the Politics of Development     679

and income depended on their ignorance and superstition” (Aziz 2001, 27). Access to
education can unravel the hierarchical structure of power that is based on unquestioned
loyalty, obedience, and superstition.

Evidence on Shrines and Literacy

Our empirical analysis is based on two key innovations: the compilation of three com-
plementary data sets on the presence and influence of shrines across different regions of
Punjab, and a novel research design that uses a coup-​induced shock to the administra-
tion of public goods provision to identify the impact of shrine elites on literacy growth.
We spell out these two contributions and offer a quick summary of the main findings in
Malik and Mirza (2018).

Data
Our primary contribution is to compile three unique databases on shrines. The first
database covers the universe of shrines across tehsils, the lowest administrative unit of
Punjab. The primary source for this data is the Punjab Auqaf Department, which is re-
sponsible for the administration, construction, decoration, and management of shrines.
To fulfill these functions, it maintains a detailed list of shrines across different regions
of Punjab. Shrines with missing or incomplete location details were separately treated
through specialized interviews with informed respondents in each district circle.
A second source of shrine data is the colonial-​era district gazetteers of Punjab that
systematically provide information on major shrines in each district. Such information
was typically documented in a separate subsection titled, “Religious Fairs and Festivals.”
This is a crucial data source for our project, since British administrators were more likely
to recognize shrines that were influential in the precolonial era. It was often the same
shrines that later received colonial land grants, official titles, administrative posts, and
other instruments of colonial patronage. All shrines mentioned in district gazetteers
were separately recorded and consolidated with our original database. This required
mapping each historical shrine to the contemporary tehsil where it is located. A total of
146 shrines across Punjab were mentioned in district gazetteers, with forty-​seven tehsils
containing at least one shrine mention.
In a third data set we compiled supplementary information on the direct participa-
tion of shrine families in electoral politics. Using electoral records from fifteen waves of
election results since 1937 and a variety of primary and secondary sources, we identified
all shrine families that directly participated in electoral politics and entered in national
or provincial assemblies. We were thus able to identify sixty-​four shrines with a direct
680    Adeel Malik AND Rinchan Mirza

Figure 33.1: Spatial distribution of politically influential shrines in Punjab.

Source: Malik and Malik (2017), https://​doi.org/​10.1017/​S0026749X16000949,
© Cambridge University Press 2018. Reprinted with permission.

political connection. Figure 33.1 maps the spatial distribution of politically influential
shrines. A quick glance at Punjab’s sacred geography demonstrates substantial variation
in shrine presence across different regions. A greater number of shrines in southwest
Punjab were mentioned in district gazetteers and subsequently selected into politics.
However, a significant proportion, 39 percent of all politically influential shrines, are
situated in central Punjab.
Our main proxy for the historically embedded power of religious elites is the number
of shrines mentioned in colonial-​era district gazetteers per 10,000 persons in a tehsil,
the lowest administrative unit for which data is available. Shrines tend to vary in terms
of their number, size, and significance. Our argument is not based on the generalized
impact of shrines but on a subset of historically significant shrines, which bestowed on
their guardians tremendous power in both economic and political domains. For our
outcome variable, we compiled a long historical panel of literacy rates that spans over
a century (1901–​2011). The data has been compiled at the tehsil-​level for which archival
and contemporary data on literacy is available. Since some administrative boundaries
have changed over time, we conducted our empirical analysis while keeping historical
tehsil boundaries fixed at 1931.
 Islam and the Politics of Development     681

Coup-​Induced Shock
To identify the effect of shrine elites on literacy, we relied on a coup-​induced universal
shock to the administrative regime for allocation of public goods. After coming to power
in the wake of the military coup in 1977, General Mohammad Zia-​ul-​Haq radically al-
tered the administrative regime that governed the administration of public goods. Prior
to the coup, major decisions regarding public goods provision were in the hands of
bureaucrats who were relatively insulated from local political influence.1 After General
Zia’s coup, elected politicians were given direct control over allocation of public goods.
They now had a direct influence over the number and type of development schemes to
be pursued and the ensuing allocations and disbursements for these.
Besides this administrative change, three other dimensions of the Zia-​period (1977–​
1988) provide a crucial background to our empirical analysis. First, the military coup
was not driven by shrine elites and was largely exogenous to the domestic political mi-
lieu. While a national political agitation preceded the imposition of the coup, it was
spearheaded by an urban right-​wing alliance. Furthermore, while coup leaders typically
blame prevailing political crises as the main reason for the coup, such crises are often
manufactured to pave the way for such coups (Shah 2014). Furthermore, all of Pakistan’s
military coups coincided with major geopolitical projects and were probably influenced
by an interplay of domestic and international factors. Second, the Zia era saw an unprec-
edented expansion of schooling infrastructure. Education became a major area of devel-
opment priority and attracted more public spending.
Third, soon after ensconcing himself in power, and in the face of growing external
pressure for holding elections, Zia selectively opened the electoral arena. But he did so in
a manner that radically changed the political landscape for decades to come. Using the
Martial Law Order No. 65 and through a series of amendments in the Political Parties
Act of 1962, Zia disqualified members of a popular political party, the Pakistan People’s
Party, from participating in elections. The elections, held in 1985, disallowed candidates
from declaring any party affiliation, which cleared the electoral field for the entry of
new political actors who drew their strength purely from their wealth, status, and social
networks. This meant that factors constituting the society’s basic social formations, such
as tribe, clan, kinship, and religious authority, assumed greater salience in determining
electoral success. Shrine elites were important beneficiaries of this, as many shrine
families were incentivized to consolidate their political influence by increase their direct
electoral participation.
The Zia era thus set in motion a new politics of development both through changes
in the administration of public goods provision and electoral laws. This has profound
implications for our analysis. We have already argued that shrine elites are more averse
to education because it poses an existential threat to the regime of voluntary obedience
around shrine-​based religious authority. With these sweeping changes brought about by
General Zia, shrine elites also now possessed a greater ability to undermine educational
provision. We thus hypothesize that the interplay between shrine-​based religious power
682    Adeel Malik AND Rinchan Mirza

and the de jure change in the administrative regime for public goods provision should be
adversely associated with literacy.

Main Findings
Our empirical analysis, based on a difference-​in-​differences regression framework,
compares the impact of this policy treatment on the evolution of literacy rates across
regions that had greater historically embedded religious power versus those who lacked
such power—​see Malik and Mirza (2018) for technical details. Concretely, we expect
that tehsils with a higher number of historical shrines per capita would witness a dif-
ferential trajectory of literacy relative to low-​shrine tehsils. Our results lend support to
this. Areas with a greater concentration of historic shrines witnessed a substantially re-
tarded growth of literacy in the post-​coup period relative to low-​shrine areas. There is
no systematic difference in literacy trends between high-​and low-​shrine tehsils prior to
the 1977 military coup. Furthermore, shrine regions do not differ systematically from
low-​shrine regions in terms of levels of public infrastructure provision. The impact of
shrine elites remains latent until 1977. The divergence in literacy rates emerges only after
the military coup.

Alternative Explanations
Our main finding is robust to a variety of perturbations in the empirical set-​up, which
includes, among others, controlling for various literacy determinants, geographic
controls, and competing explanations. Two important dimensions of this analysis de-
serve some attention. First, we recognize the possibility that shrine location could itself
be determined by tehsil-​level characteristics that may have a direct bearing on literacy.
Second, to the extent that shrine elites act as controllers of both land and politics, their
influence on local political economy is similar to those of landed elites who are simi-
larly incentivized to suppress the expansion of schooling. Is the shrine effect then simply
proxying for the influence of landed elites?
It turns out that the negative effect of historic shrine presence on literacy holds up
when we account for the geographic selection of shrines and land inequality. Given
that some of the most prominent shrines in Punjab are situated in close vicinity to the
river, distance from river appears as a significant correlate of historic shrines. Analyzing
the diffusion of Islam in western Punjab, Eaton (2009) highlighted the role of agrarian
change, ecology, and larger socioeconomic changes. According to Eaton, the transmis-
sion of Islam during the period 1300–​1800 was gradual and mostly occurred in regions
considered peripheral to the Hindu cultural system and treated as frontier regions by
the Mughal sultanate.
 Islam and the Politics of Development     683

The riverine regions of Punjab were mostly populated by tribes that were untouched
by Hindu culture and religion. The riverine tracts also afforded superior agricultural
possibilities in an era when perennial irrigation through canals had not yet existed.
Access to the river also supported greater mobility of people at a time when rivers were
important modes of transportation for trade and the movement of people. For all these
reasons, riverine regions were excellent candidates for shrine location. Reassuringly,
controlling for both distance from river and distance from Delhi in our empirical anal-
ysis does not alter the result.
To ensure that our basic finding is not driven by land inequality, we rely on a variety
of empirical strategies. First, we use information on land tenure contracts from colonial
district gazetteers and construct a measure of initial land inequality, which is defined as
the proportion of villages in a colonial tehsil that are subjected to zamindari contracts
(landlord contracts). Given the absence of genuine land reforms in Pakistan and the per-
sistence of land inequality, initial land inequality could serve as a useful proxy. Second,
in one of the earliest investigations on land inequality in Punjab, Calvert (1925) estab-
lished rainfall as a leading correlate of the size and distribution of land holdings. While
such proxies of land inequality turn up as statistically significant correlates of literacy,
they do not eliminate the shrine effect. We also rule out other pertinent explanations
that could potentially interfere with our results, including the role of partition-​induced
displacement, the incidence of canal irrigation, urbanization, and migration.

The Primacy of Politics

We explore possible mechanisms that underpin the negative effect of historic shrines on
literacy in the post-​coup period. In doing so, politics emerges as a central intervening
variable. As argued earlier, since mass literacy poses a direct and existential threat to the
religious authority of shrine elites, they are more averse to education and have therefore
a greater incentive to suppress schooling than non-​shrine elites. We have also argued
that the administrative and electoral reforms under Zia placed greater financial and po-
litical power in the hands of elected politicians. Enjoying a stable and growing electoral
presence, shrine elites now had greater ability to undermine educational expansion after
the coup.
To put some meat behind this argument, we first present evidence on how shrine
elites have maintained—​and, in fact, increased—​their foothold on electoral politics in
their respective regions. We also show that regions with historically significant shrines
are more likely to witness the entry and presence of shrine elites into politics. Second,
we highlight some factors that are crucial for understanding the political resilience of
shrine elites. Finally, we identify the strategies that shrine elites seem to have deployed in
suppressing education after the 1977 military coup.
684    Adeel Malik AND Rinchan Mirza

Shrines and Politics

Historically serving as intermediaries between centralized rulers and their peripheral
subjects, shrine elites have enjoyed influence with most ruling powers. Their religious
authority confers on them a complementary power of intermediation: an intercession
between God and man. In his book, Religion, Land, and Politics in Pakistan, the noted
historian K. K. Aziz highlighted the political advantages associated with a sacred shrine
lineage:

Medium-​size shrine makes him a small landowner and a local squire. The big shrine
gives him an entrée into the zamindar club and makes a magnate of him. A leading
shrine is a gold mine, which catapults him into the aristocratic category and brings
him riches large enough to . . . enter politics directly at the highest level.
(Aziz 2001, 109)

Persistence
The influence of shrine elites on the local political economy, their capacity to act as
intermediaries, and their control of landed property serve as crucial political assets.
It is therefore not surprising that when the British opened the electoral arena in 1920,
shrine guardians eagerly participated in them. The 1937 and 1946 elections in British
India saw many prominent religious families taking part (see Table 34.1 for details).
When the demand for Pakistan gained strength, religious families readily joined
the ranks of the Muslim League and “played a decisive role in mobilizing support for
Pakistan” (Gilmartin 1979). Since then, shrine elites have remained a permanent fixture
of Pakistani politics. Their high degree of political persistence is evident from the fact
that while roughly 19 percent of total rural Muslim constituencies were held by shrine
families in the pre-​Partition period (1920 and 1946), the corresponding ratio only fell to
16 percent in 2013 (Malik and Mirza 2015).
Shrine elites have served in every political dispensation, whether a political party is
ideologically on the left or right or whether a military ruler supported “Islamization”
(General Zia-​ ul-​
Haq) or “enlightened moderation” (General Pervez Musharraf).
General Zia-​ul-​Haq reached out to key shrine families and inducted them into his
Majlis-​e-​Shura (consultative assembly). Some of the same families joined General
Pervez Musharraf ’s cabinet two decades later, this time for undoing Zia’s legacy. And,
even as dictatorial rule gave way to civilian elected governments, the sun has never set
on the political power of shrine notables.

Growing Electoral Presence after Zia

The new electoral restrictions introduced by General Zia’s military regime paved the
way for the entry of new families in electoral politics, including shrine aristocrats who
Table 33.1: Participation of Shrine Families in Elections
in Pre-​Partition Punjab
Candidates from shrine families

Constituency 1937 1946

Attock Syed Ameer Ahmed Shah Mohy-​ud-​din Lal Badshah

(Ittehad-​e-​Millat) (Unionist, Won)
Sargodha Syed Nawazish Ali Shah Sir Qalandar Shah
(Unionist, Won)
Pind Dadan Khan Raja Ghazanfar Ali Raja Ghazanfar Ali
(Unionist, Won) (Muslim League, Won)
Jhang Syed Mubarak Ali Shah Syed Mubarak Ali Shah
Jhang East (Unionist, Won) (Muslim League, Won)
Sardar Hasan Hussein Shah Syed Ghulam Mohammad Shah
(Muslim League, Won)
Toba Tek Singh Pir Nasir-​ud-​din Shah Pir Nasir-​ud-​din Shah
(Unionist, Won) (Unionist)
Shujabad Syed Razi Shah Gilani Mohammad Raza Shah Gilani
(Unionist, Won) (Muslim League, Won)
Mureed Hussain Qureshi
(Unionist)
Multan Major Ashiq Hussain Qureshi Major Ashiq Hussain Qureshi
(Unionist, Won) (Unionist, Won)
Lodhran Makhdum Wilayat Gilani Syed Mustafa Shah Gilani
(Unionist, Won) (Muslim League, Won)
Dera Ghazi Khan Khwaja Ghulam Murtaza
(Unionist, Won)
Ferozpur Syed Amjad Ali*
(Unionist, Won)
Gujar Khan Ali Raza Shah
Depalpur Syed Ashiq Hussain Shah
(Muslim League, Won)
Kabirwala Pir Syed Naubahar Shah
(Muslim League, Won)
Khanewal Pir Budhan Shah Khagga
(Muslim League, Won)
Pakpattan/​Arifwala Mohammad Masood Chishti

Source: Malik and Malik (2017). Data from Anjum 1995; Punjab Assembly Website.
* related to the caretakers of Daud Bandagi in Shergarh.
686    Adeel Malik AND Rinchan Mirza

100 Zia Rule Musharraf Rule

Number of shrine candidates in national assembly

80

60

40

20

1970 1980 1990 2000 2010

Figure 33.2: Shrine-​affiliated candidates for National Assembly elections, 1970–​2008.

used this opportunity to increase their direct presence in electoral politics. As Anatol
Lieven (2012, 137) correctly observes, in the Pakistani political milieu, “it is not wealth
alone, but wealth plus either kinship or spiritual prestige, or both, that gives political
power.” With their control of both religious and landed power, Punjab’s shrine notables
were well placed to increase their political fortunes under Zia’s military regime.
Indeed, even a cursory analysis of electoral data reveals the growing participation of
shrine elites in electoral politics since the Zia era. Figure 33.2 provides a preview of this
by plotting the number of shrine-​affiliated candidates during the period 1970–​2008, and
shows a noticeable spike in the number of shrine-​affiliated candidates since the 1980s.
Shrine families also seem to have received a major boost in their electoral fortunes in
the post-​Zia period (see Figure 33.3). Representation of major shrine families in the na-
tional legislature increased by roughly three times since 1970. Adjusting these shrine
victories by the number of constituencies (or seats) preserves these basic patterns (see
Figure 33.4). Analysis of provincial election results reveal the same pattern of rising po-
litical presence of shrine families. Temporally, we can thus notice a clear shift in the par-
ticipation and prospects of shrine-​affiliated families in electoral politics in the post-​Zia
period.
Beyond their direct participation in elections, shrine elites have always remained an
important indirect influence. In any region that hosts the tomb of a major Muslim saint,
the blessings of the shrine guardian are essential for candidates competing in that con-
stituency. Typically, election candidates try to win the confidence of the shrine guardian,
 20 Zia Rule Musharraf Rule
Number of shrine elites in national assembly

15

10

5
1970 1980 1990 2000 2010

Figure 33.3: Shrine-​affiliated winners for National Assembly elections, 1970–​2008.

% of seats that returned a member affliliated with a shrine

.16 Zia Rule Musharraf Rule

.14

.12

.1

.08

1970 1980 1990 2000 2010

Figure 33.4: Victory of shrine-​affiliated candidates per National Assembly constituency,

1970–​2008.
688    Adeel Malik AND Rinchan Mirza

who usually tends to control vast networks of devotees who can be swayed to support any
candidate of his choice. While such indirect political influence of shrine magnates tends
to vary across different regions, there are numerous instances where a shrine guardian’s
influence can play a decisive role in electoral contests. Sometimes, this influence also
spans several neighboring constituencies. Like the British, who were the masters of in-
direct rule, all military regimes from General Ayub Khan to General Pervez Musharraf
have cultivated ties with major shrine families and depended on their political brokerage.
Regardless of whether a shrine family exerts a direct or an indirect political influence, it
was able to partake in new patronage opportunities afforded by the political control of
development funds that General Zia’s administrative restructuring facilitated.

Factors Shaping Shrine Presence in Politics

To tie this discussion more firmly to our empirical analysis, we probe the determinants of
shrine presence in contemporary politics by asking whether regions with a greater con-
centration of historically significant shrines recognized by colonial-​era district gazetteers
saw a disproportionately higher presence of shrine electables in the post-​independence
period (1947–​2013). To do so, we ran a simple probit model where our binary-​dependent
variable was coded as one for tehsils where shrines have a direct electoral linkage, and as
zero otherwise. Our main explanatory variable was a dummy variable that picked out
tehsils with historically recognized shrines (i.e., shrines mentioned in colonial-​era dis-
trict gazetteers). Our analysis showed that as the historic shrine dummy changes from
zero to one, the probability that a tehsil has a political shrine increases by 42 percentage
points (Malik and Mirza 2015). In short, it is mostly regions with historical shrines that
have witnessed a greater presence of shrine families in politics.
Let us now turn to a larger question: What gives shrine elites political salience, and
what might explain their political persistence over time? In traditional rural settings,
guardians of notable shrines act as both religious and political figures and have sub-
stantial land holdings. The control of religious, economic, and political power is a lethal
combination in the electoral arena. It not only allows shrine elites to use the traditional
instruments of patronage and control typically deployed by landed gentry, but also
affords greater social control over local populations by shaping their cognitive beliefs
and mental constructs. In sociological terms, power relations demand compliance not
just through coercion and reciprocity, but also through voluntary obedience. It is with
regard to the latter that shrine elites, thanks to their established religious authority, enjoy
an additional political leverage. As Hans Gerth and C. W. Mills (1953, 193–​195) argue:

[T]‌he crux of the problem of power rests in understanding the origin, constitution,
and maintenance of voluntary obedience . . . . An adequate understanding of power
relations thus involves a knowledge of the grounds on which a power holder claims
obedience, and the terms in which the obedient feels an obligation to obey.

Shrine elites are able to elicit such voluntary obedience through their control of what the
American sociologist Robert Bierstedt (1950, 153) describes as “supernatural resources”
 Islam and the Politics of Development     689

of religious communities, which as “as agencies of a celestial government, apply super-

natural sanctions as instruments of control.” In rural settings, local populations fre-
quently believe in the ability of shrine guardians to offer divine intercession. Enjoying
an unquestionable obedience and loyalty of their devotees, shrine leaders usually ad-
vise believers on consequential decisions, such as the purchase of property, job selec-
tion, marriage, dispute resolution, and health decisions. Voting is just another item
on this list. It is common for shrine devotees to cast their vote in favor of a candidate
endorsed by the shrine guardian. Typically, before any electoral contest sets in, poten-
tial candidates start queuing up to seek the blessings of shrine families. Thus, even in
constituencies where shrine families are not directly contesting elections, their indirect
political role can be a decisive factor in electoral success (Malik and Malik 2017).
In short, through their control of both material and symbolic power, shrine elites are
a political force to reckon with. They possess all three constitutive elements of power
described by Bierstedt: numbers, organization, and resources. Influential shrines
have a vast network of devotees that often spans multiple constituencies. For example,
one major shrine of central Punjab alone (Sial Sharif) has a vote bank in over thirty
constituencies, and controls about 10 percent of the total registered votes in each con-
stituency, on average.2 Besides their numerical strength, shrines afford their guardians
superior organizational capacity by regularly holding congregations and festivals.
Turning to potential explanations for the political persistence of shrine families,
our work has highlighted at least five factors (see Malik and Malik 2017). First, shrine
families have been extremely adept at changing their political loyalties. In any elec-
toral season, shrine-​affiliated candidates are the first ones to change their party affili-
ation. Enjoying a stable and captive vote bank, shrines enjoy a portable network that
can be transferred with relative ease across political parties. Second, shrine elites tend
to penetrate party structures in ways that keeps political parties dependent on them for
accessing the electorate. In many constituencies, the electoral space remains impene-
trable without the political intercession offered by shrine guardians.
Shrine notables can act as political gatekeepers in these areas by influencing political
party decisions on who to award a party ticket for general elections. With their social and
religious capital cutting across party lines, they are also valued by political parties for their
support in coalition building inside the parliament. Frequent military interventions pro-
vide a third key reason for the staying power of shrine elites. After usurping power, every
military ruler faces a legitimacy gap that is filled by cultivating ties with traditional elites.
All three of Pakistan’s dictators tried to kill party politics and devolve power to the local
level in ways that allowed them to centralize political power, entrench clientelist politics,
and build a pliant and dependent class of political brokers. Extending the colonial legacy
of indirect rule, successive military regimes have, in turn, strengthened shrine elites
through their strategies for patronage and control. As part of this, the Awqaf departments
were systematically reorganized under Ayub Khan and Zia-​ul-​Haq to streamline their fi-
nancial and administrative resources and selectively channel them to shrine elites.
Fourth, strategically formed marital connections and extended family ties allow
shrine elites to play a much larger role in national politics beyond their constituencies
690    Adeel Malik AND Rinchan Mirza

and afford them greater relevance and durability in all political seasons. Some of the
country’s most prominent shrine families are related by marriage with other noted
shrine magnates and members of the landed gentry, bureaucracy, and business com-
munity. Figure 33.5 provides a stylized description of the extended marital ties of
Syed Yusuf Raza Gilani, a former prime minister and long-​time parliamentarian, and
a member of one of Pakistan’s most prominent shrine families, the Gilani Syeds of
Multan. As Figure 33.5 shows, the Gilani Syeds are related to at least eight other shrine
families controlling key seats of religious power in Punjab and Sind. Such dense fa-
milial networks define the larger shrine brotherhood.3 Thus, the Aristotelian logic that
“the whole is greater than the sum of its parts” befittingly applies to shrine elites.

Brother-in-law of Yusuf Raza Gilani is

Makhdum Syed Ibrar Hussain Shah

Sajjada Nasheen of
Sandhlianwali Sarkar

Syed Ghulam Tobatek Singh Pir Safiuddin

Mohiyuddin Gilani Shah Gilani
Sajjada Nasheen Sajjada Nasheen of
of Golra Sharif Makhad Sharif

Islamabad Attock
.. s
of. rrie
ily ma
am er
o f ist
int er’ s

Pir Syed Ijaz

Syed Wajahat Hussain
tb

Shah Gilani
Fa

Sajjada Nasheen of Hujra Sajjada Nasheen

Co of Pir-e-Piraan Musa
Shah Mugeem intousin ma .
fam rries rries.. Pak Shaheed
ily o r ma
f... siste
Okara Elder Multan

Syed Yusuf Raza Gilani

is
nd h .
ani a y of..
R a z a Gilto famil Cous
f in in is
Yusur marry ...
siste
Syed Wajahat Hussain
..
er i,

Pir Syed Israr

of.
gh an
au il

Hussain Shah Bukhari The Gardezi syeds

Co to f
-d ir G

Father married daughter of...

in
us am

of Bahawalpur
nd ad

Father, brother marries

Sajjada Nasheen of
in ily
gra l Q

ma o

Sandhlianwali Sarkar
es du

into family of...

rri f...
rri , Ab

es
ma Son

Tobatek Singh

Makhdum Syed
Shams-ud-din-Gilani
Pir Pagara
Sajjada Nasheen
Sind of Uch sharif
Makhdum-ul-Mulk
Ghulam Miran Shah Bahawalpur
Sajjada Nasheen of the
shrine of jamal al-Din Wali

Rahimyar Khan

Another daughter of makhdum-ul-mulk

Ghulam Miran Shah was married in
the family of Pir Pagara of Sind

Figure 33.5: Marital alliances of Syed Yusuf Raza Gilani.

Source: Malik and Malik (2017), https://​doi.org/​10.1017/​S0026749X16000949,
© Cambridge University Press 2018. Reprinted with permission.
 Islam and the Politics of Development     691

Finally, the instrument of hereditary succession, whereby guardianship of shrines

passes to members of the same family, allows shrine elites to spread religious power
across several generations and consolidates family interests, including property holdings.
Arguably, the gaddi (seat of shrine religious lineage) provides a more stabilizing source of
dynastic power:

In comparison with landed elites, the staying power of shrine guardians may also be
tied with the stabilizing association that shrine, as a fixed institution, offers to pīrs.
Conceptualizing shrine-​based gaddis as semi-​permanent “family seats” might ex-
plain as to why shrine families are usually spared from some of the most destructive
forms of factionalism. While internal strife, especially emanating from succession
battles for the gaddi, is not entirely uncommon, it is subdued by the role of a shrine
as the classic centripetal force that could generate incentives for cooperation among
family members.
(Malik and Malik 2017)

While this discussion offers a generalized description of the political influence of

shrines, it is pertinent to highlight that there is considerable temporal and spatial heter-
ogeneity in the sacred geography of Punjab. It is precisely such variation that allowed us
to investigate the differential effects of shrine penetration on development.

Strategy to Undermine Educational Expansion

We have so far argued that shrine elites have greater incentives to oppose mass education
and that shrine dense regions have witnessed a more retarded expansion of literacy since
the 1977 military coup. Our explanation for this rests on the greater capacity of shrine
elites to suppress educational expansion post-​Zia, due to both the changes in the admin-
istration of public goods and the greater political consolidation of shrine elites during
and after General Zia-​ul-​Haq’s military regime. This naturally leads us to pose another
important question: What specific strategies could these shrine elites have deployed in
retarding education? For example, they could restrain public spending, block the con-
struction of schools, make educational provision defective, and/​or influence the de-
mand for schooling in their regions of influence. Pinning down these mechanisms is
essential for completing the analysis—​and for situating the role of shrine elites in the
everyday politics of development.
Given that General Zia’s military regime expanded the educational effort and gave
control over development spending directly to elected politicians, some of the strategies
discussed here became more politically optimal for shrine elites than for others. With
more state resources now available for development projects (e.g., the construction
of schools, hiring of teachers, etc.) and indirectly administered through politicians,
Zia’s policies increased the opportunity cost for shrine elites to disengage from this
new competition for state patronage. We therefore expect that, rather than restraining
692    Adeel Malik AND Rinchan Mirza

development spending or outright opposing the construction of schools, shrine elites

would find it politically more optimal to reduce the quality of educational provision.
Evidence in Malik and Mirza (2018) tends to favor this prior. We find that shrine-​
affiliated members of the national assembly did not receive systematically lower
allocations and or have lower utilization rates of development funds compared to non-​
shrine parliamentarians. Similarly, shrine-​dense regions did not have a lower number
of schools or teachers per child in either primary or secondary schools relative to non-​
shrine regions. In fact, there is no statistically significant difference in school provision
between shrine and non-​shrine regions. It is also possible that shrine regions witnessed
a differential demand for schooling after the military coup. Conceivably, the coup could
have brought changes to the local economy, exposure to nonagricultural opportunities,
or increased the power of shrine guardians in ways that could influence the demand for
schooling. Exploring this demand-​side explanation using tehsil-​level data on urbaniza-
tion and economic migration, we find that neither of the two dimensions is fully able to
account for the negative shrine effect.
Instead, we find stronger evidence in favor of another supply-​side dimension: more
defective educational provision in shrine-​dominated regions. Our empirical anal-
ysis suggests that shrine-​dense regions have schools that are more distant from target
populations and lack essential facilities, such as electricity and boundary walls.
Furthermore, schooling resources are spread more thinly in high shrine regions. This
was reflected in the suboptimally lower size of primary schools and is a direct corollary
of political elites using school construction as a source of rents and a means to expand
personalized clientelist networks.
This evidence on defective educational provision chimes well with broader evidence
on the politicization of schooling that was amplified during the Zia-​period and con-
tinued thereafter. As Gazdar (2000, 31) argued, influential politicians vied to

get a disproportionate number of public works in the areas of their influence or in

areas where they wish to dispense patronage are, of course, quite damaging to the
performance of public services . . . . The use of government jobs as patronage implies
that employees are not selected for their abilities, and are also more difficult to hold
accountable. The take-​over of finished civil works (i.e. school buildings) for private
use implies that public services cannot be delivered. The issue appears, therefore, not
to be one of active opposition to schooling, but one of incidental damage due to the
loss of accountability and the private appropriation of public resources.

Our evidence suggests that such politicization left a deeper imprint on shrine-
​dense regions. This is also consistent with anthropological evidence from the field
(Martin 2016). Zia’s legacy of channeling development spending through elected
politicians continues to this day. Before being elected as prime minister, Imran Khan
was vehemently opposed to giving development funds to members of provincial
assemblies and the national assembly. However, once in power, he has been unable
to fulfill his promise and has faced stiff opposition from his own party members,
 Islam and the Politics of Development     693

including mostly prominently from Makhdoom Khusro Bakhtiar, an important

shrine guardian from southern Punjab.

Conclusion

We conclude this article by drawing out the broader implications of our analysis. Our
research on the political economy of shrines highlights the need to pay attention to
the interplay between Islam and the underlying distribution of economic and political
power. As Avner Grief argues, the impact of a society’s informal institutions is condi-
tioned by a historical process where political economy structures routinely interact with
cultural factors to shape outcomes. Religion is thus part of a complex institutional ma-
trix, a “system of social factors that conjointly generate a regularity of behaviour” (Grief
2006, 30). Grief ’s approach to institutional analysis thus emphasizes the role of both
timing and context. While modern political economy has highlighted the role of his-
tory for explaining contemporary development outcomes (Nunn 2009), the challenge
of “persistence studies is to figure out why and how history matters” (Fouka 2020). The
impact of a historical variable can remain latent for a long time until it interacts with
other factors to shape contemporary outcomes. Persistence, in other words, could be
“time-​varying.” Summarizing recent literature on the subject, Fouka (2020) notes that,
in many recent studies, “the time-​varying persistence of past events is engineered by
political elites.” This is a hugely pertinent observation for the study of historic Islamic
institutions on long-​run development. Like Grief, Fouka’s observation that “the way
history’s shadow manifests depends on timing and context” carries particular relevance
for examining the relationship between Islam, politics, and development.
In developing the external validity of our argument, it is therefore important to stress
that the impact of shrine elites in different societies will depend on how they are struc-
turally positioned within the prevailing power structure. In Morocco, authoritarian le-
gitimacy is partly derived through cultivating ties with powerful Sufi orders that have
recently pushed their influence on the political stage (Werenfels 2014). In northern
India, Sufi establishments were similarly patronized by past rulers through revenue-​
free land grants. But they subsequently witnessed a decline in their material fortunes,
first under British rule, due to succession battles in civil courts, and then after the intro-
duction of land reforms in the 1950s. Jafri (2006) shows how the financial fortunes of a
prominent shrine in southern Awadh suffered after the enactment of the Uttar Pradesh
Zamindari Abolition and Land Reforms Act 1952. While the connection between reli-
gion, land, and politics was decoupled after independence in India, it was structurally
consolidated in Pakistan.
The larger implication of this is clear. The impact of religion on development should
be studied in its relevant material context. Religion is but one variable in a complex
configuration of causes, and its effect can be highly nonlinear and mediated by other
694    Adeel Malik AND Rinchan Mirza

variables in the surrounding institutional environment. One illustration of this is the

work on education, politics, and religious transformation in Egypt by Gregory Starrett,
who shows that Islam is “put to work” through the “changing intersections of power, in-
terest, and circumstance” (Starrett 1998, 235).

Notes
1. The decline of civil services had already started during the 1970s but gathered greater mo-
mentum after General Zia’s coup.
2. Interview of authors with the caretaker of Bhera Sharif in Sargodha.
3. The Chishtis of Shaikh Fazil in Vehari, for example, are related by marriage to the Chishti
family of Pakpattan Sharif in Arifwala and the prominent shrine families of Chishtian in
Bahawalnagar. Daughters of Makhdum Sajjad Hussain Qureshi, the former governor of
Punjab, are married to Pīr Shujaat Hasnain Qureshi of Khanewal and Senator Iftikhar Ali
Bukhari of Rajoa Sadaat (related to caretakers of Shah Jewana, Jhang). See Malik and Malik
(2017) for more details.

References
Acemoglu, Daron, Francisco A. Gallego, and James A. Robinson. 2014. “Institutions, Human
Capital, and Development.” Annual Review of Economics 6, no. 1: 875–​912.
Acemoglu, Daron, and James Robinson. 2006. “Economic Backwardness in Political
Perspective.” American Political Science Review 100, no. 1: 115–​131.
Acemoglu, Daron, and James A. Robinson. 2008. “Persistence of Power, Elites, and
Institutions.” American Economic Review 98, no. 1267–​1293.
Ali, Imran. 1987. “Malign Growth? Agricultural Colonization and the Roots of Backwardness
in the Punjab.” Past & Present 114: 110–​132.
Anjum, Wakeel. 1995. Siasat Danaun Kee Kalabaziyan. 3 vols. Lahore, Pakistan: Feroz Sons
Publishers.
Aziz, Khursheed Kamal. 2001. Religion, Land, and Politics in Pakistan: a Study of Piri-​Muridi.
Lahore, Pakistan: Vanguard Books.
Banerjee, Abhijit, Lakshmi Iyer, and Rohini Somanathan. 2005. “History, Social Divisions,
and Public Goods in Rural India.” Journal of the European Economic Association 3, no.
2–​3: 639–​647.
Bazzi, Samuel, Gabriel Koehler-​ Derrick, and Benjamin Marx. 2020. “The Institutional
Foundations of Religious Politics: Evidence from Indonesia.” Quarterly Journal of Economics
135, no. 2: 845–​911.
Becker, Sascha O., and Ludger Woessmann. 2009. “Was Weber Wrong? A Human Capital
Theory of Protestant Economic History.” Quarterly Journal of Economics 124, no. 2: 531–​596.
Bierstedt, Robert. 1950. “An analysis of social power.” American Sociological Review 15, no.
6: 730–​738.
Calvert, Hubert. 1925. The Size and Distribution of Agricultural Holdings in the Punjab. Board of
Economic Inquiry, C&M, No 14. Lahore: Gazette Press.
Carvalho, Jean-​Paul, Sriya Iyer, and Jared Rubin, eds. 2019. Advances in the Economics of
Religion. Cham, Switzerland: Springer International.
 Islam and the Politics of Development     695

Chaney, Eric. 2013. “Revolt on the Nile: Economic Shocks, Religion, and Political Power.”
Econometrica 81, no. 5: 2033–​2053.
Chaudhary, Latika, and Jared Rubin. 2011. “Reading, Writing, and Religion: Institutions and
Human Capital Formation.” Journal of Comparative Economics 39, no. 1: 17–​33.
Coşgel, Metin M., Thomas J. Miceli, and Jared Rubin. 2012. “The Political Economy of Mass
Printing: Legitimacy and Technological Change in the Ottoman Empire.” Journal of
Comparative Economics 40, no. 3: 357–​371.
Darling, Malcolm Lyall. The Punjab Peasant in Prosperity and Debt. 2nd ed. London: Oxford
University Press, 1928.
Eaton, Richard M. 2009. “Shrines, Cultivators, and Muslim Conversion in Punjab and Bengal,
1300–​1700.” Medieval History Journal 12, no. 2: 191–​220.
Fouka, V. 2020. “Time-​Varying Persistence.” Broadstreet (blog), December 30, 2020. https://​
broadstreet.blog/​2020/​12/​30/​time-​varying-​persistence/​.
Gallego, Francisco A. 2010. “Historical Origins of Schooling: The Role of Democracy and
Political Decentralization.” Review of Economics and Statistics 92, no. 2: 228–​243.
Galor, Oded, Omer Moav, and Dietrich Vollrath. 2009. “Inequality in Landownership, the
Emergence of Human-​Capital Promoting Institutions, and the Great Divergence.” Review of
Economic Studies 76, no. 1: 143–​179.
Gazdar, H. 2000. “State, Community, and Universal Education: A Political Economy of Public
Schooling in Rural Pakistan (Mimeo).” London: London School of Economics and Political
Science, Asia Research Centre.
Gerth, Hans and C. Wright Mills. 1953. Character and Social Structure: The Psychology of Social
Institutions. New York: Harcourt, Brace.
Gilmartin, David. 1979. “Religious Leadership and the Pakistan Movement in the Punjab.”
Modern Asian Studies 13, no. 3: 485–​517.
Glaeser, Edward L., Rafael La Porta, Florencio Lopez-​de-​Silanes, and Andrei Shleifer. 2004.
“Do Institutions Cause Growth?” Journal of Economic Growth 9, no. 3: 271–​303.
Grief, Avner. 2006. Institutions and the Path to the Modern Economy: Lessons from Medieval
Trade. Cambridge: Cambridge University Press.
Heck, Paul L. 2007. Sufism and Politics: The Power of Spirituality. Princeton, NJ: Markus Wiener.
Iyer, Sriya. 2016. “The New Economics of Religion.” Journal of Economic Literature 54, no. 2: 395–​441.
Jafri, Saiyid Zaheer Husain. 2006. “Madrasa and Khânqâh or Madrasa in Khânqâh?
Education and Sufi Establishments in Northern India.” In Islamic Education, Diversity and
National Identity: Dīnī Madāris in India Post 9/​11, edited by Jan-​Peter Hartung, Helmut
Reifeld, 73–​104. Delhi. SAGE.
Klenow, Peter J., and Andres Rodriguez-​Clare. 2005. “Externalities and Growth.” Handbook
of Economic Growth. Vol 1A, edited by Philippe Aghion and Steven Durlauf, 817–​861.
Amsterdam: Elsevier.
Lieven, Anatol. 2012. Pakistan: A Hard Country. London: PublicAffairs.
Malik, Adeel, and Tahir Malik. 2017. “Pīrs and Politics in Punjab, 1937–​2013.” Modern Asian
Studies 51, no. 6, 1818–​1861.
Malik, Adeel, and Rinchan Ali Mirza. 2015. “Religion, Land and Politics: Shrines and Literacy
in Punjab, Pakistan.” PSPP Working Paper No. 030. Washington, DC: International Food
Policy Research Institute (IFPRI).
Malik, Adeel, and Rinchan Ali Mirza. 2018. “Pre-​colonial Religious Institutions and Development:
Evidence through a Military Coup.” CSEA Working Paper 2018-​04. Oxford: Centre for the
Study of African Economies, University of Oxford.
696    Adeel Malik AND Rinchan Mirza

Martin, Nicolas. 2016. Politics, Landlords and Islam in Pakistan. Abingdon, UK: Routledge.
Nunn, Nathan. 2009. “The Importance of History for Economic Development.” Annual Review
of Economics 1, no. 1: 65–​92.
Kuru, Ahmet T. 2019. Islam, Authoritarianism, and Underdevelopment: A Global and Historical
Comparison. Cambridge: Cambridge University Press.
Rubin, Jared. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the Middle East Did
Not. Cambridge: Cambridge University Press.
Shah, Aqil. 2014. The Army and Democracy: Military Politics in Pakistan. Cambridge, MA:
Harvard University Press.
Sokoloff, Kenneth L., and Stanley L. Engerman. 2000. “Institutions, Factor Endowments, and
Paths of Development in the New World.” Journal of Economic Perspectives 14, no. 3: 217–​232.
Squicciarini, Mara P. 2020. “Devotion and Development: Religiosity, Education, and Economic
Progress in Nineteenth-​Century France.” American Economic Review 110, no. 11: 3454–​3491.
Starrett, Gregory. 1998. Putting Islam to Work: Education, Politics, and Religious Transformation
in Egypt. Comparative Studies on Muslim Societies, vol. 25. Berkeley: University of California
Press.
Tsai, Lily L. 2007. “Solidary Groups, Informal Accountability, and Local Public Goods
Provision in Rural China.” American Political Science Review 101, no. 2: 355–​372.
Werenfels, Isabelle. 2014. “Beyond Authoritarian Upgrading: the Re-​emergence of Sufi Orders
in Maghrebi Politics.” Journal of North African Studies 19, no. 3: 275–​295.
Woodberry, Robert Dudley. 2004. “The Shadow of Empire: Christian Missions, Colonial
Policy, and Democracy in Postcolonial Societies.” PhD diss., University of North Carolina
at Chapel Hill.
 Chapter 34

Isl am and Ec onomi c

Devel opm e nt i n
Sub-​S ahara n A fri c a

Melina R. Platas

The world’s fastest-​growing Muslim population resides in sub-​Saharan Africa,1 a region

with the lowest levels of economic and human development globally. Within African
countries, Muslims often fare worse in terms of economic and human development
than their Christian counterparts. Across African countries, Muslim-​majority coun-
tries experience lower levels of socioeconomic development on average than countries
where Muslims are a minority. Islam in Africa has tended to be understudied with re-
spect to scholars of Islam and development as well as with respect to scholars of develop-
ment in Africa. In this chapter I bridge the gap between these rich bodies of scholarship,
demonstrating the importance of considering the role of religion in the study of eco-
nomic development in Africa, but not limiting the role of religion to religious beliefs
and institutions.
I use administrative and survey data to demonstrate the broad patterns that exist
across and within countries with respect to religious affiliation and economic develop-
ment. Focusing on several common development indicators, I show that within African
countries, Muslims are often worse off compared to Christians, particularly with respect
to educational outcomes. I also show that Muslim-​majority countries are poorer, expe-
rience higher rates of child mortality, and have lower rates of literacy and educational
attainment than countries where Muslims are a minority. What explains these patterns?
Does religion, and Islam in particular, affect development outcomes? If so, how, and
what is meant by “religion” in these explanations?
Drawing on existing literature regarding long-​term predictors of economic devel-
opment as well as the relationship between Islam and development, I discuss several
channels through which religion might affect development outcomes and, in particular,
generate a gap in development across religious groups. First, I discuss the role of insti-
tutional factors in shaping development trajectories across religious groups in Africa.
698   Melina R. Platas

There is little evidence that Muslim-​majority countries are less democratic than non-​
Muslim-​majority countries, and thus it is unlikely that variation in regime type explains
the patterns we observe. Further, while the features of Islamic economic institutions
have been posited as an explanation for the divergence in economic development be-
tween the Middle East and Europe, the political and economic institutions in Africa
today have been heavily influenced by European colonialism, such that Islamic eco-
nomic institutions may play less of a role in divergent development trajectories.
Instead, existing work suggests that areas inhabited by Muslim populations and areas
under Islamic political rule in the precolonial period affected the type and extent of colo-
nial investments in public goods and the type of political institutions developed during
the colonial period. Extensive reliance on Christian missionaries to provide formal ed-
ucation meant that Muslim-​majority areas and areas under Islamic rule may have re-
ceived fewer educational investments during the colonial period. Historical disparities
in educational investments have persisted to the present, and in turn likely affected ec-
onomic development more broadly. Evidence from case studies suggests that colonial
authorities may have also employed a more severe form of indirect rule in predomi-
nantly Muslim areas, which may have undermined the accountability of local leaders.
This argument builds on the idea that institutions matter for the different development
trajectories that Muslim and non-​Muslim societies have taken historically; it does not
focus on the role of Islamic institutions per se, but rather on the role of the colonial re-
sponse to the presence of Muslim populations and Islamic institutions in explaining
long-​term trajectories.
Second, I examine the potential role of norms and beliefs in shaping development
outcomes. I find that Muslims in Africa tend to have lower levels of support for gender
equality than Christians with respect to political participation and land inheritance,
which could in theory translate into gaps in development outcomes. I do not find that
Muslims exhibit lower levels of trust in medicine or higher rates of use of traditional
medicine, which provides suggestive evidence that there is not a gap in health-​seeking
behavior. However, compared to Christian women, Muslim women consistently report
higher values for their ideal number of children, and are also more likely to report that
their husbands want more children than they do. Muslim women exhibit higher fertility
rates than Christian women in Africa, and this disparity, perhaps supported in part by
differences in preferences for the number of children desired, could also affect devel-
opment outcomes and lead to a gap in health or education across religious groups. For
example, having a higher average number of children could result in having fewer re-
sources to spend per child on education or health resources.
Some evidence suggests differing preferences across Muslims and Christians
for formal education, positing that Muslims may view state-​led secular education
as less legitimate than do Christians (Manglos-​Weber 2017), or that Muslims have
a preference for Qurʾanic education (Dev et al. 2016), or that Muslims and Christian
communities have developed distinct norms about school attendance due to the his-
torical relationship between Christianity and formal education (Platas 2019). Together,
this work suggests norms and beliefs do play an important role in explaining divergent
 Islam and Economic Development in Sub-Saharan Africa    699

development trajectories; however, the beliefs and norms at play are not necessarily re-
ligious in nature.
Finally, the geographic distribution of Muslim populations could affect develop-
ment outcomes differentially if, for example, Muslims tended to live in more arid or
rural places on average, as compared to Christians. In examining education and health
outcomes, I include fixed effects for subnational regions within countries, to account
for the possible effects of spatial clustering. For the most part, the gap in outcomes is
still observed between Muslims and Christians, even taking into account their spatial
distribution, suggesting geography is unlikely to explain much of the development gap.
Further, areas exposed to Islamic rule historically, many of which have large Muslim
populations today, are not systematically worse off in terms of factors related to eco-
nomic development such as historical population density, suitability of land for agricul-
ture, rough terrain, or malaria prevalence.
These findings underscore the imperative in taking religion and religious affiliation seri-
ously in studies of comparative development and politics (Grzymala-​Busse 2012), particu-
larly in Africa, where many existing accounts have tended to focus much more on the role
of ethnicity and ethnic diversity than religion (Easterly and Levine 1997; Franck and Rainer
2012; Habyarimana et al. 2007; Miguel and Gugerty 2005). However, they also point to the
importance of thinking broadly about how “religion” matters for development. It is not
only the explicitly religious content of institutions or beliefs, such as religious teachings or
ideology, that matter for development, but also how religion shaped the design of colonial
institutions and distribution of public goods investments, particularly in a region where
Christian missionaries played such a large role in public goods provision historically.

Islam in Africa

I begin with an overview of present and historical religious demographics in Africa, a

region that is home to a Muslim population of around 300 million. By 2050 it is expected
that the African Muslim population will exceed that of the Middle East and North
Africa, Europe, North America, and Latin America combined, and will comprise about
one-​quarter of the world’s Muslims (Pew Research Center 2015).
There is substantial variation in the size of the Muslim population across African
countries, as shown in Figure 34.1. Seventeen have substantial Muslim-​ minority
populations (between 5 and 49 percent), while fourteen are Muslim-​majority countries.
Another sixteen have very small Muslim minorities (less than 5 percent). In all of these
countries, Christians make up the majority of the rest of the population, while African
traditional religions are practiced by very small minorities.
This degree of religious diversity reflects the process through which Islam spread in
Africa from the eighth century onward, culminating in the colonial period, which nearly
all African countries experienced simultaneously in the late nineteenth and twentieth
centuries (Levtzion and Pouwels 2000). Islam entered the region through two main
700   Melina R. Platas

Somalia
Mauritania
Comoros
Djibouti
Niger
Sudan
Senegal
Gambia
Mali
Guinea
Sierra Leone
Chad
Burkina Faso
Eritrea
Nigeria
Guinea-Bissau
Cote d'Ivoire
Ethiopia
Tanzania
Benin
Cameroon
Ghana
Togo
Mozambique
Mauritius
Liberia
Central African Republic
Malawi
Uganda
Kenya
South Sudan
Rwanda
Equatorial Guinea
Cape Verde
Madagascar
Burundi
South Africa
Congo DR
Congo
Angola
Zambia
Seychelles
Zimbabwe
Swaziland
Namibia
Lesotho
Sao Tome & Principe

0 .2 .4 .6 .8 1
Percent Muslim, 2010 (World Religions Database)

Figure 34.1: Percent Muslim by country, 2010.

Source: Data from World Religions Database.

routes, across the Sahara in West Africa and via the East African coast. Scholars have
noted that conversion to Islam in Africa followed a three-​stage process (Austen 2010;
Levtzion and Pouwels 2000). The first to convert were those directly engaged in trade,
and were concentrated in trading towns and cities. Local rulers then began to see Islam
as useful both politically and economically. Islam not only connected them to global
trading networks, but Muslim traders and scholars also carried useful skills, including
literacy, numeracy, and medicine (Loimeier 2013). A number of local political leaders
converted to Islam throughout the ninth–​thirteenth centuries, and both Sahelian West
Africa and coastal East Africa were home to kingdoms for which political authority was
Islamic in nature. The third stage of the spread of Islam in Africa was the conversion of
the masses. This last stage did not happen until relatively close to the colonial period in
Africa, such that by the time European colonizers and Christian missionaries arrived,
Islam remained a religion practiced by a minority of the population in the region.
Religious demographics changed most quickly during the colonial period. Estimates
from the World Religion Database, relying on historical government records, atlases,
and reports from religious organizations, indicate that while three-​quarters of Africans
practiced traditional religions in 1900, by 2010 this figure had fallen to less than one
in seven (Pew Research Center 2010). Christianity grew at a faster rate than Islam
during the same time period, and especially from the 1950s onward, but both Islam and
Christianity made large gains in the region during this hundred-​year period. Today, ap-
proximately two-​thirds of the population practices Christianity, while approximately
one-​third practices Islam.
 Islam and Economic Development in Sub-Saharan Africa    701

In most countries for which data are available, the plurality of Muslims report being
Sunni, followed by “just a Muslim” (Pew Research Center 2010). Those who identify as
Shiʿa are a minority in all countries where data are available, with the countries with the
largest Shiʿa populations being Chad (21 percent of all Muslims) and Tanzania (20 per-
cent). There are large minorities of Ahmadiyya in Ghana (16 percent of Muslims),
Tanzania (15 percent), and Cameroon (12 percent). Religious switching is relatively rare,
with the vast majority claiming adherence to the religion in which they were raised.
African countries today report some of the highest levels of religiosity around the
world. In a survey of nineteen countries (Pew Research Center 2010), the majority of
African respondents reported attending religious services at least once a week and
praying at least once a day.
More than 85 percent of Muslims in all but one country surveyed report fasting during
Ramadan, and between a quarter and two-​thirds of the Muslim population across coun-
tries surveyed reported high levels of support for religious jurisprudence.
Despite considerable religious diversity and high levels of religiosity, religious con-
flict has been relatively rare, though it has been increasing in recent years. A minority
of Muslims in all countries believe most, many, or all Christians are hostile toward
Muslims, and only a slightly larger minority of Christians report the same regarding
Muslims. In only three countries did a majority of respondents cite religious conflict
as a “very big problem” affecting their country: Rwanda, Nigeria, and Djibouti. Crime,
corrupt political leaders, and unemployment were far more frequently cited as major
problems.
The Muslim population in Africa is projected to grow faster than the Muslim popula-
tion in any other region over the next several decades. At 5.6 births per woman, Muslims
in Africa have substantially higher fertility rates than Muslims living in any other re-
gion, where the global average is 3.1 births among Muslim women. African Muslim
populations also have higher fertility rates than African Christian populations, where
the rate is 4.5 births per woman.

Islam and Development across and

within African Countries

In this section, I use administrative and household survey data to document the rela-
tionship between religious affiliation and development indicators across and within
African countries. Across African countries, Muslim-​minority countries tend to be
richer, and they experience lower rates of child mortality and higher rates of education.
Figure 34.2 shows this set of bivariate relationships, where a set of common develop-
ment indicators are plotted against percent Muslim at the country level. Specifically,
I examine the relationship between percent Muslim of a country and (a) income per
capita (2017), (b) child mortality (2018), (c) expected years of education (2017), and
(d) the Human Development Index (2017).2
 (a) 25000
Equatorial Guinea

20000
GDP per capita, 2017

Gabon
Botswana
15000

10000 Namibia

Swaziland

Angola
Nigeria
5000 Congo
Zambia Ghana Sudan
SSao Tome Cote D'Ivoire Mauritania
Lesotho Kenya Cameroon
Tanzania Senegal
Zimbabwe Uganda Guinea
Rwanda Togo Benin Mali
Mozambique Ethiopia Chad
CBurkina Faso Gambia
Madagascar Sierra Leone Comoros
DRC Malawi Niger
CAR
Liberia
0 Burundi

0 .2 .4 .6 .8 1
Country Percent Muslim, 1990 (Pew)
bandwidth = .8

(b) 120 Nigeria Chad

Somalia

CAR

Sierra Leone

100 Guinea
Mali

Benin

DRC
Child Mortality (2018)

Equatorial Guinea
Niger
Lesotho Cote D'Ivoire
80 Angola
Cameroon Burkina Faso Mauritania
Mozambique
Liberia
Togo Comoros

60 Burundi Sudan
Djibouti
Gambia
Zambia
Swaziland Ethiopia
Madagascar Tanzania
Congo Malawi
Zimbabwe Ghana
Gabon Uganda
Senegal
Kenya Eritrea
40 Namibia
Botswana
Rwanda
Sao Tome

20
0 .2 .4 .6 .8 1
Country Percent Muslim, 1990 (Pew)
bandwidth = .8

Figure 34.2: Islam and development outcomes cross-​nationally.

Source: Data on religious affiliation comes from Pew Research Center; data on economic
and human development indicators comes from various sources compiled by
the United Nations Development Programme.
 (c) 10

Zimbabwe

Swaziland

Namibia
8 Kenya
Congo
Gabon
Average years of education

Lesotho
Zambia
Ghana
Nigeria
Cameroon
DRC Comoros

6 Uganda
Tanzania

Sao Tome
Malawi
Angola
Rwanda
Madagascar Gambia
Liberia
Togo

4 Burundi
Mozambique
Sierra Leone Sudan
Cote D'Ivoire
Ethiopia
CAR

Benin Senegal
Guinea
Chad
2 Burkina Faso
Mali Niger

0 .2 .4 .6 .8 1
Country Percent Muslim, 1990 (Pew)
bandwidth = .8

(d)
Botswana

.7 Gabon

Namibia

Congo
Human Development Index 2017

.6 Equatorial Guinea
m nd Kenya
Zambia
Swaziland Ghana
SSao Tome
Angola
Cameroon
Zimbabwe Tanzania
Rwanda Nigeria
Madagascar
M Mauritania
Lesotho Uganda Benin
Senegal
al Comoros
.5 Togo
Cote D'Ivoire
Sudan

Malawi Djibouti
Ethiopia Guinea Gambia
DRC
Mozambique
Eritrea
Liberia
Burkina Faso Mali
Burundi Sierra LLeone
Chad
.4
CAR
Niger

.3
0 .2 .4 .6 .8 1
Country Percent Muslim, 1990 (Pew)
bandwidth = .8

Figure 34.2: Continued

704   Melina R. Platas

These figures suggest that Muslim-​majority countries within Africa have, on average,
worse development outcomes than those countries where Muslims are a minority. As
shown in the top left panel, there is a negative relationship between percent Muslim and
income per capita. Moving from 0 to 100 percent Muslim is associated with a decline of
about $3,700 in GDP per capita. Average income per capita among non-​Muslim ma-
jority countries ($4,550) is more than twice that of Muslim majority countries ($2,100).
There is a positive relationship between child mortality, measured as the number of
deaths of children under 5 per 1,000 live births, and percent Muslim of the country (top
right panel). The average child mortality rate among Muslim-​majority countries is 82
deaths per 1,000 births, compared to 62 deaths per 1,000 births among non-​Muslim-​
majority countries. With respect to education, there is a strong relationship between
percent Muslim and expected years of education (bottom left panel). Average expected
years of education is 10.8 for Muslim-​minority countries, and 8.4 for Muslim-​majority
countries, a gap of nearly two and a half years. The relationship between percent Muslim
and years of education is stronger among non-​Muslim majority countries. Among
Muslim majority countries, the percent Muslim of the country provides little to no
additional predictive power, which is similar to the pattern observed for the Human
Development Index (bottom right panel). These associations are not meant to indicate
causal relationships, but rather to document patterns worthy of further exploration.
Within countries, Muslim populations also tend to experience worse educational
and health outcomes than Christians. To document relationships between religious
affiliation and development outcomes at the subnational level, I use data from the
Demographic and Health Surveys (DHS). The DHS are conducted in over eighty low-​
and middle-​income countries around the world. These surveys typically target women
of childbearing age, but are also conducted with a subsample of men, and include dem-
ographic questions about all household members and all of the female respondents’
births. With a few exceptions, surveys include a question about the respondents’ reli-
gious affiliation. For the results presented below, I use the sample of respondents who
self-​identify as Muslim or Christian in the surveys.3
The primary variables of interest with respect to education are whether a respondent
has ever attended school, total years of education, and literacy. Any schooling is a dichot-
omous variable indicating whether or not the respondent reports having ever attended
school. Years of education indicates the number of years an individual has attended
school. In DHS questionnaires, questions about education frequently reference levels
of formal education (e.g., primary, secondary, etc.) and thus it is assumed that years of
school and any schooling refers to formal schooling. Literacy is a dichotomous variable
where a respondent is coded as literate if they can read part of a sentence or a full sen-
tence, and otherwise not. Because any schooling and years of education likely do not cap-
ture forms of schooling outside what is traditionally thought of as formal schooling, such
as religious schooling, literacy provides a way of assessing whether or not respondents
acquire skills associated with formal education regardless of whether or not they report
attending school, and also provides a basic measure of the quality and content of educa-
tion. For these analyses, I include adults who were born between 1965 and 1995.4
 Islam and Economic Development in Sub-Saharan Africa    705

Table 34.1: Religion and Educational Attainment

(1) (2) (3) (4) (5) (6)

Any School Any School School Years School Years Literate Literate

Muslim −0.182*** −0.191*** −1.921*** −1.997*** −0.176*** −0.182***

(0.003) (0.003) (0.036) (0.032) (0.003) (0.003)
*** ***
Female −0.128 −1.534 −0.179***
(0.001) (0.013) (0.001)
Rural −0.175*** −2.669*** −0.218***
(0.002) (0.025) (0.002)
Birth year 0.007*** 0.067*** 0.007***
(0.000) (0.001) (0.000)
*** *** *** *** ***
Constant 0.747 −13.838 5.928 –​122.865 0.651 −12.138***
(0.001) (0.010) (0.017) (1.334) (0.002) (0.141)
Observations 1,090,037 1,076,958 1,073,886 1,073,886 1,005,774 1,005,774
R2 0.360 0.406 0.339 0.422 0.298 0.362

Notes: All models include country and region fixed effects. Standard errors are clustered by survey cluster.
*** p«/​i>0.01;
** p«/​i>0.05;
* p«/​i>0.10

Table 34.1 shows that across all three measures, Muslims experience lower educa-
tional attainment than Christians. They are more likely to report never attending school,
and to attend for fewer years, and are less likely to be literate. All models include fixed
effects for country and subnational region, suggesting that the relationship between re-
ligious affiliation and education is unlikely to be driven by a few countries, nor by the
spatial clustering of Muslims in particular areas within countries. Women and those
who live in rural areas experience lower educational attainment. Birth year is positively
associated with educational attainment, suggesting that attainment is increasing with
time. These results also suggest that the relationship between religious affiliation and ed-
ucation is not driven by women, and, if anything, there may be a larger gap in attainment
between Christian and Muslim men than women.
To explore the role of gender further, Figure 34.3 shows the relationship between reli-
gious affiliation and years of education by country, plotting the coefficient for “Muslim”
for men and women, respectively. As in the models in Table 34.1, these country-​level
results include covariates for birth year and urban residence as well as fixed effects for
subnational regions. In many cases, the Muslim coefficient is quite similar for both men
and women. Where it differs significantly, it is usually the case that there is a larger gap
for men than for women. It is notable that while there are a handful of cases where there
706   Melina R. Platas

Chad
Gabon
Cote d’Iovire
Nigeria
Guinea
Cameroon
Congo
Gambia
Kenya
Ghana
Togo
Sierra Leone
Senegal
Liberia
Burkina
Mali
CAR
Ethiopia
Benin
Malawi
Burundi
Niger
Uganda
Tanzania
Mozambique
Rwanda Women
Madagascar
DRC Men
Zambia
–6 –4 –2 0 2
Difference in years of education

Covariates: Birth year, urban. Region FE. SE clustered by survey cluster.

Figure 34.3: Difference in years of education by religion and gender.

Source: Data from the Demographic and Health Surveys (DHS) Program.

is no significant association between religious affiliation and educational attainment,

there is only a single case (women in the Democratic Republic of Congo) in which there
is a positive and significant relationship between being Muslim and years of education.
Does the increase in educational attainment over time mean that the education gap
is closing between Christians and Muslims over time? Figure 34.4, which shows average
years of education by cohort and country, suggests this is not the case. Most countries
exhibit relatively parallel trends in average years of education over time across religious
groups. A few countries, such as Ghana, Liberia, and Sierra Leone, have an educational
gap that seems to be narrowing with time but remains substantively large, at one to two
years for the youngest cohort.
Next I examine the relationship between religious affiliation and child mortality, one
of the most common health indicators. The DHS ask women for a complete birth his-
tory, including whether each child born to her is still alive. Child mortality is an indi-
cator that is equal to one if a child is reported to have died before reaching the age of
five years, and zero otherwise. I include the sample of births that have taken place at
least five years before the year of the survey, again focusing on the sample of children
whose mothers report being Christian or Muslim. The sample includes nearly 2.4 mil-
lion observations, and I include controls for urban residence, mother’s education, total
births, mother’s age at birth and its square, child gender, whether the child is a multiple
 Benin Burkina Faso Cameroon Cote d’Ivoire Ethiopia
10 10 10 10 10
8 8 8 8 8
6 6 6 6 6
4 4 4 4 4
2 2 2 2 2
0 0 0 0 0
1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990

Gabon Ghana Guinea Kenya Liberia

10 10 10 10 10
8 8 8 8 8
6 6 6 6 6
4 4 4 4 4
2 2 2 2 2
0 0 0 0 0
1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990

Malawi Mali Mozambique Nigeria Rwanda

10 10 10 10 10
8 8 8 8 8
6 6 6 6 6
4 4 4 4 4
2 2 2 2 2
0 0 0 0 0
1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990

Senegal Sierra Leone Tanzania Togo Uganda

10 10 10 10 10
8 8 8 8 8
6 6 6 6 6
4 4 4 4 4
2 2 2 2 2
0 0 0 0 0
1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990 1965 1970 1975 1980 1985 1990

Figure 34.4: Years of education among Christians (open circle) and Muslims (closed circle) by country-​cohort.
Source: Data from the Demographic and Health Surveys (DHS) Program.
708   Melina R. Platas

Mali
Ethiopia
Burkina Faso
Ghana
Sierra Leone
Cote d’Ivoire
Nigeria
Kenya
Liberia
TOTAL
Senegal
Cameroon
Tanzania
Guinea
Malawi
Uganda
Togo
Benin
Chad
Rwanda
Mozambique
Madagascar
Burundi
Gabon
–.04 –.02 0 .02 .04 .06
Difference in child mortality

Figure 34.5: Child mortality difference by country.

Source: Data from the Demographic and Health Surveys (DHS) Program.

birth (e.g., twin), short birth spacing (less than 24 months), and child’s birth order and
its square, with fixed effects for subnational region and child birth year.5
Figure 34.5 shows the difference in child mortality rate for Christians and Muslims
by country, specifically, plotting the coefficient for “Muslim.” There are six countries
where Muslims experience significantly higher mortality rates than Christians, and
only two where Muslims experience significantly lower infant mortality. Small sample
sizes of Muslims in countries like Burundi and Rwanda combined with a relatively low-​
incidence event may make it difficult to observe significant differences in mortality rates
even if they exist. While the patterns with respect to child mortality and religion are not
as consistent as they are with respect to education, Muslim children are at higher risk of
death in many countries, and rarely at lower risk.
Further, in the pooled sample, Muslims experience approximately 8 additional
deaths per 1,000 births than Christians. While as a percentage this figure is relatively
small, it translates into a large number of children and families being affected in prac-
tice. There are more than thirty million children born in the region each year, and
if we assume at least one-​third of these are born to Muslim families, an additional
8 deaths per 1,000 births translates into around 80,000 deaths if past trends are ap-
plied to the future. This is almost certainly an upper bound, and there are a number
of caveats, including the fact that the DHS is not conducted in every country and that
child mortality rates have been falling steadily, but it is nevertheless indicative of the
magnitude of this effect.
 Islam and Economic Development in Sub-Saharan Africa    709

(How) Does Religion Matter

for Development?

As the results above demonstrate, there are systematic differences in common develop-
ment indicators across both Muslim-​majority and Muslim-​minority countries, as well
as differences across religious groups within countries. Does “religion” matter for these
development outcomes? If so, how, and what aspect of religion is responsible? I discuss
three channels through which religion may shape development outcomes: institutions,
norms and beliefs, and geography.
First, religion could affect development outcomes through political and eco-
nomic institutions. A prominent line of work in examining the relationship between
institutions and development evaluates the relationship between regime type and eco-
nomic development. For example, Acemoglu and Robinson (2012) argue that inclusive
political institutions foster development, while some, though not all, studies show that
democracy, or the stock of democracy, is associated with higher levels of human de-
velopment and with economic development (Acemoglu et al. 2019; Gerring et al. 2012;
Papaioannou and Siourounis 2008). One hypothesis, therefore, in explaining the cross-​
national patterns documented in this chapter is that Muslim-​majority countries are less
likely to be democratic than non-​Muslim majority countries, and that this explains their
relative underperformance in development. However, while it is well established that
Muslim-​majority countries are less likely to be democratic than non-​Muslim-​majority
countries globally, a phenomenon that has been explained by political history and soci-
oeconomic factors (see chapter by Nugent in this volume), this pattern does not seem to
hold in the African context. Since independence, Muslim-​majority countries in Africa
have experienced an average of 9.5 years (16 percent) under democracy, compared to
9.45 (17 percent) years in non-​Muslim majority countries. The average polity score in the
post-​independence period in Muslim majority countries is −1.9, and for non-​Muslim
majority countries is −2. Thus, there is no apparent relationship between regime type
and Muslim-​majority status in Africa, such that differences in regime type are unlikely
explain the disparity in development we observe between Muslim majority and non-​
Muslim majority countries.
Another vein of research on institutions and development, particularly in the African
context, examines precolonial and colonial legacies. Some of this work demonstrates
the long-​term effect of centralized precolonial institutions, typically finding higher rates
of contemporary economic development in areas that were part of precolonial states
(Michalopoulos and Papaioannou 2013). Bauer et al. (2019) qualify this general finding,
showing that within Africa, areas exposed to Islamic rule—​that is, areas that were part
of Islamic states or kingdoms in the precolonial period—​today experience lower levels
of education, higher rates of child mortality, and lower levels of economic development
than those that were not.
710   Melina R. Platas

This work points to an important source of institutional variation within Africa: the
extent to which precolonial societies were governed by Islamic political institutions.
Prior to colonization, Muslim societies in Africa varied in the presence and extent of
Islamic institutions, and centralized polities varied with respect to the religious basis
of political authority. Some precolonial Muslim societies were governed by institutions
that were more highly centralized and in which political leaders derived legitimacy
through Islam, while other Muslim societies were governed to a greater extent by tra-
ditional systems of authority, which were often less centralized. Meanwhile, there were
also a number of precolonial states that existed among non-​Muslim societies.
Both the extent of Islamic institutions and the distribution of the Muslim population
in the precolonial period could affect long-​term development across and within coun-
tries through their effect on investments by Christian missionaries during the colonial
period, as well as affect the design and implementation of colonial institutions.
There is substantial evidence that investments by colonial authorities and Christian
missionaries during the colonial period had long-​ term effects on development,
and particularly on educational outcomes (Huillery 2009; Nunn 2014). Colonial
investments in social services, especially those provided by Christian missionaries,
were likely to be lower in areas where Muslims were a majority or areas governed by
Islamic rule. Missionaries may have avoided areas already inhabited by large Muslim
populations, or they were explicitly banned from operating in these areas, as was the
case in parts of northern Nigeria and Sudan (Sharkey 2012). Lower levels of investment
in social services, especially schools, could have a lasting impact on contemporary de-
velopment through reduced access to education and lower levels of education among
earlier generations (Wantchekon et al. 2015). Alesina et al. (2018) find that intergen-
erational mobility in education also correlates strongly with proximity to Christian
missions. Bauer et al. (2019) find that missionaries were less likely to operate in areas
exposed to precolonial Islamic rule, and Platas (2019) finds that within African coun-
tries, areas where Muslims form a majority have lower levels of education than areas
where they are a minority.
The relatively low educational attainment among Muslims within countries as well as
in Muslim-​majority as compared to Muslim-​minority countries is thus likely explained
in part by historical differences in access to social services in areas that were or were not
predominantly Muslim during the colonial period, which persisted due to continued
differences in access to education as well as intergenerational factors. Disparities in his-
torical enrollment in formal education may explain education gaps in North Africa as
well. For example, Saleh (this volume) notes that in Egypt, non-​Muslim students were
enrolled in “modern” schools at a far higher rate than Muslim students, while Muslim
students were primarily enrolled in religious schools, which was also a result of the na-
ture and implementation of colonialism. This pattern persisted until the mid-​twentieth
century and appears to explain much of the education gap between Muslims and non-​
Muslims in Egypt.
In addition to differential investments in social services such as education, where
colonial authorities encountered Islamic political authority at the time of colonialism,
 Islam and Economic Development in Sub-Saharan Africa    711

they may have implemented a more extreme form of indirect rule. This would have
meant granting greater authority to local leaders as well as limiting colonial interference
and investment in these areas. Evidence from northern Nigeria, for example, suggests
that areas under centralized Islamic rule were granted greater autonomy both because
they tended to be hierarchical polities ready-​made for indirect rule, and also because
colonial authorities viewed Islam as a more legitimate religion than African tradi-
tional religions, and were thus reluctant to dismantle Islamic authority (Hailey 1951).
A policy of noninterference in the Muslim north of the country led to the increased
concentration of political powers among emirs, who became the sole native authority
within their jurisdictions (Archibong 2018). Even in the absence of centralized states,
colonial authorities could have relied more heavily on indirect rule where there were
large Muslim populations. For example, the Northern Territories of Ghana, with a large
Muslim population, experienced a higher degree of indirect rule than the centralized
(non-​Muslim) Ashanti kingdom in the Gold Coast (Michalopoulos and Papaioannou
2020). Greater indirect rule, in turn, could serve to undermine the accountability of
local leaders. In a cross-​country analysis, Lange (2009) finds that a greater extent of in-
direct rule is associated with lower bureaucratic quality and state capacity.
Finally, areas under Islamic rule may have been politically disadvantaged in the long
run, particularly if they ended up in countries that were non-​Muslim majority and were
excluded from majority Christian political coalitions. The relatively low rates of attend-
ance at colonial (Christian) schools among Muslim populations may have also placed
them at a disadvantage in terms of political participation. Archibong (2018) argues that
leaders in the highly centralized Muslim north of Nigeria chose not to comply with the
federal autocratic regime in the postcolonial period, which led them to be punished by
the regime, incurring costs with respect to public services. More broadly, Bauer et al.
(2019) find that ethnic groups that were historically part of Islamic kingdoms have less
political power in the postcolonial period, a finding that is driven by countries in which
Muslims today comprise a minority.
This type of institutional argument contrasts with other prominent arguments about
the relationship between religion, institutions, and development, in that the proposed
mechanisms are not the content of religious institutions, but rather how the presence of
Islam and Islamic institutions affected colonial governance. While existing accounts in
this literature focus on the role of religion, and specifically Islam, in shaping the content
of economic institutions, there are reasons to believe that this explanation may play less
of a role in the context of sub-​Saharan Africa. For example, Kuran (2004, 2011) argues
that legal and economic institutions that developed in Islamic societies in the Middle
East and North Africa were less globally efficient than those that developed in Europe,
even if they were efficient locally, ultimately disadvantaging these societies economi-
cally in the long run. However, the development of economic institutions in African
countries was somewhat different from those in the Middle East, due to the timing and
spread of Islam within Africa, as well as the effect of colonialism on both political and
economic institutions. Nevertheless, as much of the work on colonial institutions has
paid little attention to Islamic institutions (Michalopoulos and Papaioannou 2020), and
712   Melina R. Platas

as much of the work on Islamic economic institutions has focused little on sub-​Saharan
Africa, this is an area that is ripe for future research.
A second channel through which religion may matter for development, and which
may explain inequalities in development outcomes across religious groups, is that of
norms and beliefs. This includes beliefs that could be classified as “religious,” such as
beliefs based on religious texts, or beliefs that guide behavior among members of a reli-
gious group but are not necessarily religious in content. These beliefs include expecta-
tions about what is considered appropriate behavior (on religious or other grounds), as
well as empirical expectations about how others are likely to behave under a variety of
circumstances (Bicchieri 2016). Could norms and beliefs explain differences in behavior
across religious groups, such as attending formal education or health-​seeking behavior?
If so, what types of beliefs would affect these behaviors?
First, Muslims and Christians could have different beliefs about gender equality
with respect to political, social, and economic issues. For example, Fish (2011) finds
that societies that have a proportionally larger Muslim population have higher levels of
gender-​based inequality. Using data from recent Afrobarometer surveys I examine two
indicators relating to attitudes toward gender equality. Specifically, I examine the ex-
tent to which respondents agree with each of the statements: “Women should have the
same chance of being elected to political office as men,” and “Women should have the same
rights as men to own and inherit land.” I find that there are large differences in attitudes
both across religious groups and across gender. For example, 84 percent of Christian
women compared to 69 percent of Muslim women support women’s land inheritance,
while among men support is lower than their co-​religion female counterparts, but
there still exists a gap across religion—​75 percent of Christian men compared to 58 per-
cent of Muslim men support women’s land inheritance. The same pattern is observed
for support of women as political leaders, though both the gender and religion gaps are
slightly smaller than for land inheritance. These findings do suggest there are impor-
tant differences in beliefs about women’s rights, which could in theory translate into
differences in health and education outcomes and are worth exploring further in future
research. However, the fact that the education gap and the child mortality gap are found
across gender, and are often larger for men than for women, suggests that the overall gap
in development outcomes is not being driven solely by outcomes among Muslim women.
Another potential difference in norms and beliefs could exist with respect to
preferences for “modern” or “Western” medicine or education, especially to the ex-
tent to which these have been associated historically with Christianity or Christian
missionaries in Africa. The DHS includes questions about the use of traditional med-
icine as well as whether respondents would consent to a blood test, which has been
used as a measure of trust in medicine (Lowes and Montero 2018). Further, the DHS
also includes questions about respondents’ ideal number of children as well as their
husband’s preferences, which could contribute to the fertility gap that exists across reli-
gious groups, and could in theory affect health and education outcomes—​for example,
by creating greater resource constraints among Muslim families.
 Islam and Economic Development in Sub-Saharan Africa    713

Table 34.2: Religious Affiliation and Health-​Related Beliefs and

Behavior
Dependent Variable:

Blood Test Traditional Ideal Number Husband

Consent Medicine of Children Preference

Muslim −0.003 −0.015*** 0.520*** 0.073***

(0.002) (0.005) (0.018) (0.003)
*** *** ***
Constant 0.977 0.030 4.515 0.227***
(0.014) (0.010) (0.096) (0.033)
Observations 148,211 18,811 301,582 318,278
R2 0.315 0.040 0.393 0.137

Notes: All models include covariates for age, years of education, and wealth index,
and country and survey-​region fixed effects. The samples for each include the
most recent round of DHS surveys for which the dependent variables of interest
are available. Blood Test takes a value of one if the respondent provided consent to
a blood test for anemia and 0 in the case of a refusal. Traditional Medicine takes a
value of 1 if the respondent reported using traditional medicine to treat diarrhea
in a child in the past two weeks, and 0 otherwise, for the subset of respondents
who report diarrhea in a child in the past two weeks. Ideal Number of Children is
a variable indicating the number of children desired by the respondent. Husband
Preference takes a value of 1 if the respondent indicates that her husband desires
more children than herself, and 0 if the respondent desires more or the same
number as her husband. Standard errors are clustered by survey cluster.
*** p >0.01;
** p >0.05;
* p >0.10

Table 34.2 suggests that trust in medicine and the use of traditional medicine are un-
likely to explain differences in health outcomes between Christians and Muslims. There
is no difference in the rate of consenting to a blood test (column 1), and, if anything,
Muslims are less likely to report using traditional medicine to treat common illnesses
(column 2). However, there is a large gap between Christians and Muslims in both the
ideal number of children (column 3) and the likelihood that husbands want more chil-
dren than their wives (column 4).
With respect to education, a preference for Qurʾanic education over formal education
could reduce investments in the latter among Muslim populations, and there is some
evidence that this explains part but not all of the Muslim-​Christian education gap in
Nigeria (Dev et al. 2016). Manglos-​Weber (2017) suggests that Muslims may not priori-
tize state-​sponsored schooling as highly as Christians, given that in many African coun-
tries, national education systems are built upon the colonial-​era missionary education
714   Melina R. Platas

system and Muslims may view these schools as less legitimate than Qurʾanic schools and
other types of Islamic education.
Schooling choices may also have implications for political participation. For ex-
ample, within Francophone Africa, Bleck (2013) finds that parents who send their chil-
dren to Islamic schools are less politically active than those who send their children to
public schools. Possible mechanisms linking these two phenomena include that parents
who send their children to Islamic schools perceive state institutions as less legitimate
that religious institutions, that because they use nonstate services they are less inter-
ested in policy decisions related to state services, or that the use of religious institutions
integrates them into an alternative local community in which to participate.
Attendance at Islamic religious schools, especially if this takes place at the expense
of formal education, could also lead to the acquisition of Arabic to a greater extent
than the colonial languages that are often the official language of government in both
Muslim-​majority and Muslim-​minority countries in Africa. This could then mean that
those who attend religious schools are less likely to be employed in the formal sector,
have fewer job opportunities, and are less likely to hold government positions. In
turn, this could lead to lower incomes and less influence on political processes and
policymaking.
The relationship between religious education, language acquisition, employment,
and participation in politics is an area that should be investigated further, with the ca-
veat that cross-​national data on both the attendance of religious school and language ac-
quisition at the individual level is scarce. A survey conducted in southern Malawi found
a positive correlation between formal and religious education among Muslims—​those
that attended formal education were also more likely to attend religious school—​and
also that attendance of Islamic schools in this context was not strongly associated with
the acquisition of fluency in Arabic (Platas 2019). While Muslims in Malawi are not rep-
resentative of African Muslims generally, these findings point to the need to examine
the relationship between religious practice, development outcomes, and political partic-
ipation in a variety of contexts.
Platas (2019) also proposes another mechanism explaining persistent educational
inequalities: norms about the importance of formal school attendance. While there was
widespread distrust of missionary schools among Muslim populations in the colonial
period, this distrust does not extend into the contemporary period in some countries
where the Muslim-​Christian education gap is still observed. In addition to the institu-
tional legacies discussed above, this suggests that predominantly Muslim and Christian
communities have developed distinct social norms with respect to school attendance—​
due to the historical relationship between Christianity and formal education in many
African countries—​and that weaker norms enforcing school attendance in predomi-
nantly Muslim communities explain some of the continued gap in educational attain-
ment across religious groups.
All of these findings suggest an important role for beliefs and norms in explaining
differences in development outcomes between Christians and Muslims, but more work
is needed to examine the conditions under which institutional factors or norms play a
 Islam and Economic Development in Sub-Saharan Africa    715

relatively greater role in explaining development gaps, as well as how institutions and
norms may reinforce one another.
Finally, geography is another channel through which religion may matter for develop-
ment, and which may contribute to development gaps between Christians and Muslims.
Michalopoulos et al. (2018) find that the spread of Islam globally was shaped by Islamic
trade routes, but also that predominantly Muslim areas tend to be arid or semi-​arid with
relatively little fertile land. Residing in arid lands could affect economic development di-
rectly, for example, resulting in poor agricultural potential or greater food insecurity, and
could also affect the design of local institutions. While variation across Muslim and non-​
Muslim areas in terms of geography is a potential avenue to pursue for research, geography
is unlikely to be the primary explanation for the gaps observed. In their examination of
precolonial Islamic rule, Bauer et al. (2019) do not find that areas under Islamic rule within
Africa were necessarily worse off than areas not under Islamic rule. In fact, compared to
non-​Islamic kingdoms, areas under Islamic rule had higher population densities, greater
land suitability, and less rough terrain, with no differences in the degree of land inequality
or malaria prevalence. If anything, many of the factors on which areas under Islamic rule
differ might be expected to be positively associated with contemporary development.
Further, in the within-​country analyses for education and health, substantial disparities
persisted even within subnational regions—​in other words, holding geography constant.

Conclusion

Sub-​Saharan Africa is home to the fastest-​growing Muslim population in the world,

and for this reason the region will hold an increasingly important place in the study
of Muslim societies in the years to come. Globally, Africa experiences the highest
levels of poverty and low levels of human development. This chapter has shown that
within Africa, Muslim-​majority countries tend to perform worse on measures of eco-
nomic and human development than countries where Muslims are a minority. Further,
within countries, Muslims tend to experience lower levels of human development than
Christians, particularly in the area of formal education.
Together, these empirical patterns and the existing literature suggest the need to take
religion and religious affiliation seriously in studying development in Africa. However,
in examining how and why inequalities emerge and persist across religious groups, it
is important to consider not only factors that are explicitly religious in nature, such as
religious beliefs or institutions shaped by Islamic or other religious law, but also factors
such as the historical legacies of colonialism, along with norms and beliefs that differ
across religious groups but are not themselves religious in content. The fact that African
countries are highly diverse with respect to religious demographics, geography, and po-
litical institutions also makes the region an ideal setting for closely examining the rela-
tionship between Islam and development, and will complement the much larger body of
work that exists on this subject in the Middle East and North Africa.
716   Melina R. Platas

Notes
1. Henceforth sub-​Saharan Africa will be referred to as “Africa” for ease of exposition.
2. Data on religious affiliation come from the Pew Research Center; data on economic and
human development indicators come from various sources compiled by the United Nations
Development Programme.
3. Although there may be interesting differences across denominations within Christianity
or Islam, these are not coded systematically across countries and would also likely result
in very small sample sizes for some of the denominations. For this reason, I combine all
denominations within each religion.
4. Only survey waves where there are at least 100 respondents from the minority religious
group are included in the analysis.
5. Covariates modeled after Franck and Rainer (2012).

References
Acemoglu, Daron, Suresh Naidu, Pascual Restrepo, and James A. Robinson. 2019. “Democracy
Does Cause Growth.” Journal of Political Economy 127, no. 1: 47–​100.
Acemoglu, Daron, and James A. Robinson. 2012. Why Nations Fail: The Origins of Power,
Prosperity, and Poverty. London: Profile Books.
Alesina, Alberto, Sebastian Hohmann, Stelios Michalopoulos, and Elias Papaioannou. 2018.
“Educational Inequality and Intergenerational Mobility in Africa.” NBER working paper.
Cambridge, MA: National Bureau of Economic Research.
Archibong, Belinda. 2018. “Historical Origins of Persistent Inequality in Nigeria.” Oxford
Development Studies 46, no. 3: 325–​347.
Austen, Ralph A. 2010. Trans-​Saharan Africa in World History. Oxford: Oxford University
Press.
Bauer, Vincent, Melina R. Platas, and Jeremy M. Weinstein. 2019. “The Historical Legacies of
Islamic Rule in Africa.” Working paper.
Bicchieri, Cristina. 2016. Norms in the Wild: How to Diagnose, Measure, and Change Social
Norms. Oxford: Oxford University Press.
Bleck, Jaimie. 2013. “Do Francophone and Islamic Schooling Communities Participate
Differently? Disaggregating Parents’ Political Behaviour in Mali.” Journal of Modern African
Studies 51, no. 3: 377–​408.
Dev, Pritha, Blessing U. Mberu, and Roland Pongou. 2016. “Ethnic Inequality: Theory and
Evidence from Formal Education in Nigeria.” Economic Development and Cultural Change
64, no. 4: 603–​660.
Easterly, William, and Ross Levine. 1997. “Africa’s Growth Tragedy: Policies and Ethnic
Divisions.” Quarterly Journal of Economics 112, no. 4: 1203–​1250.
Fish, M. Steven. 2011. Are Muslims Distinctive? A Look at the Evidence. Oxford: Oxford University
Press.
Franck, Raphael, and Ilia Rainer. 2012. “Does the Leader’s Ethnicity Matter? Ethnic Favoritism,
Education, and Health in Sub-​Saharan Africa.” American Political Science Review 106, no.
2: 294–​325.
 Islam and Economic Development in Sub-Saharan Africa    717

Gerring, John, Strom C. Thacker, and Rodrigo Alfaro. 2012. “Democracy and Human
Development.” Journal of Politics 74, no. 1: 1–​17.
Grzymala-​ Busse, Anna. 2012. Why Comparative Politics Should Take Religion (More)
Seriously. Annual Review of Political Science 15: 421–​442.
Habyarimana, James, Macartan Humphreys, Daniel N. Posner, and Jeremy M. Weinstein.
2007. “Why Does Ethnic Diversity Undermine Public Goods Provision?” American Political
Science Review 101, no. 4: 709–​725.
Hailey, William Malcolm. 1951. Native Administration in the British African Territories. Part III:
West Africa: Nigeria, Gold Coast, Sierra Leone, Gambia. London: His Majesty’s Stationery Office.
Huillery, Elise. 2009. “History Matters: The Long-​Term Impact of Colonial Public Investments
in French West Africa.” American Economic Journal: Applied Economics 1, no. 2: 176–​215.
Kuran, Timur. 2004. “Why the Middle East Is Economically Underdeveloped: Historical
Mechanisms of Institutional Stagnation.” Journal of Economic Perspectives 18, no. 3: 71–​90.
Kuran, Timur. 2011. The Long Divergence: How Islamic Law Held Back the Middle East.
Princeton: Princeton University Press.
Lange, Matthew. 2009. Lineages of Despotism and Development: British Colonialism and State
Power. Chicago: University of Chicago Press.
Levtzion, Nehemia, and Randall L. Pouwels. 2000. The History of Islam in Africa. Athens,
OH: Ohio University Press.
Loimeier, Roman. 2013. Muslim Societies in Africa: A Historical Anthropology. Bloomington:
Indiana University Press.
Lowes, Sara Rachel, and Eduardo Montero. 2018. “The Legacy of Colonial Medicine in Central
Africa.” https://​scholar.harvard.edu/​files/​emontero/​files/​lowes_​montero_​colonialmedicine.pdf.
Manglos-​Weber, Nicolette D. 2017. “Identity, Inequality, and Legitimacy: Religious Differences
in Primary School Completion in Sub-​Saharan Africa.” Journal for the Scientific Study of
Religion 56, no. 2: 302–​322.
Michalopoulos, Stelios, and Elias Papaioannou. 2013. “Pre-​colonial Ethnic Institutions and
Contemporary African Development.” Econometrica 81, no. 1: 113–​152.
Michalopoulos, Stelios, and Elias Papaioannou. 2020. “Historical Legacies and African
Development.” Journal of Economic Literature 58, no. 1: 53–​128.
Michalopoulos, Stelios, Alireza Naghavi, and Giovanni Prarolo. 2018. “Trade and Geography
in the Spread of Islam.” Economic Journal 128, no. 616: 3210–​3241.
Miguel, Edward, and Mary Kay Gugerty. 2005. “Ethnic Diversity, Social Sanctions, and Public
Goods in Kenya.” Journal of Public Economics 89, no. 11: 2325–​2368.
Nunn, Nathan. 2014. “Gender and Missionary Influence in Colonial Africa.” In Africa’s
Development in Historical Perspective, edited by Emmanuel Akyeampong, Robert H. Bates,
Nathan Nunn, James Robinson, 489–​512. Cambridge: Cambridge University Press.
Papaioannou, Elias, and Gregorios Siourounis. 2008. “Democratisation and Growth.”
Economic Journal 118, no. 532: 1520–​1551.
Pew Research Center. 2010. Tension and Tolerance: Islam and Christianity in Sub-​Saharan
Africa. Technical report. Washington, DC: Pew Research Center. https://​www.pewforum.
org/​2010/​04/​15/​executive-​summary-​islam-​and-​christianity-​in-​sub-​saharan-​africa/​.
Pew Research Center. 2015. The Future of World Religions: Population Growth Projections, 2010–​
2050. Technical report. Washington, DC: Pew Research Center. https://​www.pewforum.org/​
2015/​04/​02/​religious-​projections-​2010-​2050/​.
718   Melina R. Platas

Platas, Melina R. 2019. Culture and the Persistence of Inequality: The Muslim-​Christian Education
Gap in Africa. Unpublished book manuscript.
Sharkey, Heather. 2012. “Jihads and Crusades in Sudan from 1881 to the Present.” In Just Wars,
Holy Wars, and Jihads: Christian, Jewish, and Muslim Encounters and Exchanges, edited by
Sohail H. Hashmi, 263–​282. Oxford: Oxford University Press.
Wantchekon, Leonard, Marko Klašnja, and Natalija Novta. 2015. “Education and Human
Capital Externalities: Evidence from Colonial Benin.” Quarterly Journal of Economics 130,
no. 2: 703–​757.
 Chapter 35

Isl amic Financ e a nd

Deve l opment i n Ma l aysia

Fulya Apaydin

In many ways, the banking industry is the lifeline of modern economic systems. This is
because banks are not only financial intermediaries, but also serve as private creators
of money when they extend credit, either for consumption or investment (Ingham
2004; Minsky 1986; Pettifor 2017; Ryan-​Collins, Greenham, and Werner 2011; Wray
2015).1 Banks can create money when they extend loans and simultaneously transfer this
amount into a deposit account, as the accounting rules followed by banks enable them
to increase the overall volume of transactions on their balance sheet without keeping
equivalent amount of hard cash in vaults.2 According to some scholars, the diffusion
of this model has been the necessary engine that enabled capitalist development in the
Global North (Arrighi 1994; Desan 2017).
Unlike conventional banks, Islamic scripts ban individuals or organizations from
extending credit without an underlying physical asset. In that sense, any bank that
operates on the principles of shariʿa would be limited in its ability to follow the con-
ventional route in money creation. According to these orthodox interpretations, only
God can create value out of nothing. Within shariʿa-​compliant finance, money only
exists in the form of gold, silver, and other mediums that could also be exchanged as
money, such as wheat, rice, barley, salt, and dates. These commodities derive their in-
trinsic value from God. The value of any other commodity can be measured using any of
these mediums. Together with the standard ban on interest, this “money as commodity”
interpretation seriously limits Islamic economic actors to be up to speed with capitalist
money creation, given the natural limitations on the rate of growth of these acceptable
mediums. This has led some scholars to argue that Islamic banks are not well-​suited to
follow the conventional blueprint in credit creation (Choudhury 1997).
Yet there are recent and surprising cases—​such as Malaysia—​where these theoret-
ical expectations face new challenges. Since the inception of the first Islamic bank in
1983, Malaysia has witnessed a rapid growth of this industry, where one-​third of all fi-
nancial transactions are now classified as shariʿa-​compliant. Islamic banks self-​style
720   Fulya Apaydin

themselves as moral harbingers of a risk-​sharing economy, one that avoids speculation,

asymmetric transfer of risk and, contractual ambiguity. Proponents of Islamic banks
often cite Malaysia to argue that these banks can indeed stimulate growth (Aziz 2011;
Iqbal and Mirakhor 2013). Others are more critical, suggesting that such techniques are
simply deceptive ruses to attract pious customers without any genuine potential to yield
positive returns, let alone risk-​sharing (Kuran 2004). This debate motivates the guiding
questions of the chapter: What is the role of Islamic finance in development? Are Islamic
banks well-​equipped to jump-​start investment in the Global South?
This chapter argues that the relationship between Islam and economic development
is contingent on the conditions under which they engage in private money creation.
Recent evidence suggests that a banking system that is solely based on an orthodox
interpretation of shariʿa principles is unlikely to foster capitalist development, since
their role in money creation is limited.3 Rather, arrangements where Islamic banking
and financial services exist side-​by-​side with secular counterparts are more common
in emerging market economies with predominantly Muslim populations. However, not
all Islamic banks have been commercially successful under competitive pressures. In
settings where the state supports the private creation of money by building new market
institutions, modern Islamic banks complement conventional finance in creating new
lines of credit. In the absence of such support, Islamic banks have remained marginal
and their role in economic development is inconsequential.4
At the same time, the emergence of a complex institutional infrastructure is insuf-
ficient to stimulate equitable development, especially when private money creation
is driven by consumer debt. The case of Islamic finance in Malaysia is particularly
telling: while an extensive network of Islamic banks attracts domestic and interna-
tional investors, much of the loans extended by these banks finance real-​estate and
consumer-​durables purchases, contributing to fast-​rising household debt levels. By
contrast, Islamic banks are less willing to finance small and medium sized businesses
on the grounds of perceived high risk. The key beneficiaries of the Islamic financial in-
dustry are large financial corporations that raise capital via issuing Islamic equity, and
governments that diversify their debt composition using shariʿa-​compliant bills and
bonds. There, Islamic banks have turned into a major conduit for transferring market
risks to individual borrowers as debt-​driven consumption has accelerated.
To elaborate these points, the next section begins with an overview of existing
theories of Islam and economic development. The chapter then takes a closer look
into the Malaysian experience and suggests that financial institution-​building along
liberal market principles does not guarantee the channeling of credit into high-​value-​
added industries. This discussion is based on primary and secondary resources col-
lected during fieldwork in Kuala Lumpur between November and December 2016. The
respondents include political and economic elites who have held key positions with
decision-​making powers, either in the Islamic financial industry or the government.
The participants were identified based on the key positions they held, and therefore
the sampling is non-​random. To address problems associated with potential sampling
bias, descriptive data obtained from the Central Bank of Malaysia is also included in the
 Islamic Finance and Development in Malaysia    721

analysis. The chapter concludes by discussing the implications of Islamic finance, debt-​
dependent growth, and development.

Islamic Finance and Development

in Historical Perspective

Today, the most vocal critics of Islamic finance argue that shariʿa-​compliant exchanges
are not capable of stimulating economic growth and development because the industry
practitioners deliberately transform simple procedures into complex, costly, and re-
source-​draining tasks (El-​Gamal 2008; Kuran 2018). At the heart of these allegations
lies the presumption that Islamic institutions are fundamentally incompatible with
capitalism, which arguably requires liberal institutions for development (Acemoglu
and Robinson 2012; North 2005). Among these, two perspectives stand out. The first
is the long divergence hypothesis (LDH), which posits that Islamic law is responsible for
holding back capitalist development in the Muslim world, especially in the Middle East
(Kuran 2011). According to this framework, Muslim societies are inhibited by Islamic
scriptures that prevent the formation of large corporations, which are necessary for cap-
ital accumulation and reinvestment in the long run. Underlying impediments include
the Islamic inheritance law, which requires a complete dissolution of a person’s estate
and distribution of possessions among the legal heirs. Moreover, because Islam does not
ascribe legal personhood to any entity other than individuals, the LDH argues that it
was not possible to set up European-​style corporations where Islam remained domi-
nant (Kuran 2011, 97–​117). Kuran (2018) goes further to argue that Islamic foundations
(waqfs) further inhibited the development of markets due to their rigidity and prioriti-
zation of their founders’ directives regarding the use of funds. Consequently, such in-
stitutional roadblocks prevented the early accumulation of capital, restricting Islamic
traders to simple and short-​term partnership arrangements that did not last longer than
a lifetime. This had further implications, leading to poor interpersonal trust due to the
short-​term nature of Islamic business contracts.
However, this emphasis on religious institutions is based on a somewhat orthodox
imagination of Muslim societies—​one where shariʿa remains uncontested. In practice,
this could not be further from the everyday experiences of economic actors. While most
equate Islamic institutions with Qurʾan and hadith, multiple reinterpretations of the
Holy Book and the Prophet’s sayings, as well as local customs have shaped the develop-
ment of Islamic economic institutions in myriad ways over centuries (Ercanbrack 2015).
Very importantly, the evolution of these institutions is not independent of the political
dynamics that surround them.
A second group of explanations brings politics back in and examines how the resulting
nonsecular political arrangements have shaped future development trajectories of
Muslim-​dominant societies. These works reveal that across much of the Muslim world
722   Fulya Apaydin

in the pre-​capitalist era, the ruler’s choice of institutions was largely driven by political
motives. For example, Rubin, (this volume) highlights political legitimacy concerns
as one mechanism through which religious authorities became influential actors in
the Ottoman Empire, enabling successive Sheikh-​ul-​Islams to issue fatwas on market
exchanges as well (Rubin 2017, 2020). Initially, the rise of religious figures into powerful
positions was driven by the sultans’ concerns to centralize religious training and thereby
control diverse medreses under their rule (Bulliet 1972).5 Most notably, the appointment
of Ebussuud Efendi by Suleiman the Magnificent in the sixteenth century was a game
changer in many ways—​also on the financial market building front—​following his de-
cree that legitimized cash waqfs as shariʿa-​compliant organizations for providing access
to credit.6 Prior to Suleiman’s rule, cash waqfs were deemed shariʿa-​incompliant by the
ulama; however, the growing need for financing economic activity under Suleiman’s
reign motivated this critical modification of the Islamic religious code.
The nonsecular political arrangements also had notable implications on the everyday
lending and borrowing practices of individuals. A related hypothesis highlights the fi-
nancial power of the powerless, with a focus on the role of the state and minority pol-
itics in shaping financial development. Looking at the creditor-​borrower relations in
Istanbul during the sixteenth century, Kuran and Rubin (2018) find that if the borrower
comes from the majority Sunni group with a privileged status, non-​Muslim creditors
impose higher interest rates to ensure the repayment of the loan. Because the courts sys-
tematically favored Muslims over non-​Muslims, and since Muslims were first among the
Ottoman hierarchy of subjects, they were charged higher interest rates when borrowing
from the private moneylenders (who were often non-​Muslim). Kuran and Rubin con-
clude that this unequal treatment prevented capital accumulation among the Muslim
majority, and partially explains why the Ottomans fell behind the West (Rubin 2017).
However, an exclusive focus on nonsecular dynamics comes at the expense of neglecting
the long-​term impact of institutional arrangements that were not primarily driven by re-
ligious concerns. Other economic choices crucial to the survival of the state also emerged
out of political necessities and had unintended consequences for financial market devel-
opment. For example, most Islamic states sought to expand their fiscal revenues by relying
on tax-​farming (iltizam) and life-​time farming contract (malikane) based systems (Broms
and Rothstein 2020; Fleet 2003). While these features were widely common in most
Islamic states, the arrangement is not exclusively informed by shariʿa principles. One con-
sequence of this was the absence of long-​term borrowing by the Ottoman state.7 Thus,
until the nineteenth century, private lenders were limited in their role as money creators in
the Ottoman Middle East, debt relations were mostly organized around short-​term bor-
rowing and lending in the sixteenth and seventeenth centuries (typically on high interest
rates), and long-​term borrowing from private moneylenders began to be the norm in the
eighteenth and nineteenth centuries as Galata bankers (sarrafs) became prominent fi-
nancial actors (Pamuk 2000, 200; Rubin 2017, 81–​82).8 Rather than the strictly prohibitive
Islamic institutions, then, this experience suggests that political preferences of the rulers
are responsible for the shortage of affordable financing options.
 Islamic Finance and Development in Malaysia    723

Because of these political dynamics, the social pacts that underlie lending and
borrowing relations remained very decentralized in the absence of any organized
banking system.9 Unlike in late seventeenth-​century Britain, moneylenders in Islamic
states lacked the capacity to organize into a corporation-​like entity (e.g., as in Bank of
England), since the Ottoman sultans did not seek long-​term loans from private lenders
well into the eighteenth century.10 Thus, financial contracts were mostly short-​term
arrangements between creditors and borrowers that ended as soon as debts were re-
paid. This practice substantially limited the money creation capacity of the lenders.11
Other attempts to address liquidity problems through new financial instruments (such
as trading of gedik—​tenancy or occupational rights—​in secondary markets in the
Ottoman Empire) remained limited to stimulate the development of capital markets
(Agir 2017; Agir and Yildirim 2015). In the long run, these dynamics prevented the stabi-
lization of lending rates as well.
The focus on unstable and irregular interest rates is an important staring point to un-
derstand the roots of economic underdevelopment in Muslim-​dominant settings. The
financial power of the powerless was further enhanced in an environment where a per-
sistent pattern of decentralized lending has prevented the stabilization of interest rates
as multiple creditors priced loans at greatly varying rates.12 However, this process was
not strictly driven by religious concerns or institutions. In the absence of regular public
borrowing by the state and a fiscal architecture that gave way too much power to the
military officers as tax collectors, Ottoman rulers had limited power to coerce private
moneylenders into standardizing rates. Thus, sarrafs in Ottoman Empire remained un-
coordinated in their day-​to-​day business operations until the formation of a financier’s
guild in the nineteenth century.13
These practices had further implications on the subsequent slowdown of economic
activity. For example, following the loss of control over major trade routes, long-​
distance transactions by Ottoman traders became especially risky, and the lack of af-
fordable funding options prevented entrepreneurs from pursuing profitable business
opportunities. In the case of the Ottoman Empire, the arrival of modern banking in
the late nineteenth and early twentieth centuries did little to lower lending interest
rates and stimulate private money creation, as many of these banks did not have an
extensive branch network beyond Istanbul (Clay 1994). This meant the majority of
the borrowers in rural areas relied exclusively on moneylenders for loans, who con-
tinued to charge high interest rates (Clay 1994). Overall, rather than strictly religious
impediments to capital accumulation, a series of political decisions locked the empire
onto a path where the systematic lack of access to loans in volatile markets without an
organized lending infrastructure eventually accelerated the economic decline. In the
late nineteenth century, the rulers tried to reverse this process by creating a centralized
commercial code informed by religious principles known as majalla; however, it was
far from effective in stimulating the much-​needed domestic capital accumulation.14 As
such, majalla was replaced by a new commercial code following the founding of the
new Turkish Republic in 1926.
724   Fulya Apaydin

While it may be difficult to generalize across a much more complex geography of

Muslim inhabitants based on the exclusive experience of the Ottomans, the imperial
outreach of the sultans and the reforms implemented in the nineteenth century had a
wider projection. In that sense, the economic development trajectories of Muslim-​
populated areas are not only influenced by religious dynamics, but also by subtle po-
litical networks that facilitate the transfer of ideas and institutions across diverse
geographies. For example, majalla was in use not only in former Ottoman territories like
Lebanon, Syria, and Jordan long after the disintegration of the empire, but also reached
Malaya via a transfer of political and economic institutions. Specifically, the sultan of
Johor’s travels to the Ottoman capital in the late nineteenth century prompted him to
the nearly identical adoption of majalla (Hussin 2016). Despite having a much different
historical experience under the British colonial rule, Malaya had a similar problem in
the financial marketplace: a decentralized money creation characterized by multiple
currencies in circulation (Singh 1986), and high interest rates on loans extended by
private moneylenders. By adopting a centralized commercial code in 1893, the sultan
of Johor sought to modernize a system where “Chinese mercantile practices, British
contracts and Malay maritime rules commingled” (Hussin 2016, 169).
Curiously, while the use of majalla was abandoned in the Middle East by the late
twentieth century, it became the basis for developing the legal framework of Islamic
banking and finance in Malaysia nearly a century after its adoption. Unlike most former
Ottoman territories that adopted new commercial codes based on civil law, Malaysia
combined common law with religious law following independence from the British.
This allowed practitioners sufficient flexibility to reinterpret existing codes in accord-
ance with the changing needs and demands of the marketplace. Following the global
diffusion of Islamic banking and finance, Malaysian policymakers benefited from this
malleable framework to develop a systematized set of rules to attract investment and
support the industry in line with a broader developmental agenda of the state.
The Malaysian experience with Islamic banking and finance gradually turned into a
model that many countries look up to in developing their unique standards.15 Today,
many Islamic banks can get around formal restrictions on money creation by relying on
the verdicts of shariʿa scholars (the legal experts on Islamic economic law). These range
from replacing “interest rate” with a “mark-​up” charge to devising halal products that
mimic conventional ones through incorporating the buying and selling of commodities
into shariʿa-​compliant financial contracts. The fatwas, or legal ordinances of these
experts, offer the legitimacy to become relevant actors in the global marketplace (Calder
2016; Rethel 2011). In that sense, these verdicts are more than tricky stratagems: shariʿa-​
compliant Islamic financial contracts enable these institutions to turn into active
participants in a capitalist market economy. Thus, Islamic banks also serve as conduits
of aggregation into a capitalist market, rather than challenging its fundamental prem-
ises (Calder 2016). However, as the Malaysian case reveals, the political conditions
under which the articulation of Islamic financial institutions to the capitalist economy
occurs have notable implications on their lending behavior and subsequent develop-
ment outcomes.
 Islamic Finance and Development in Malaysia    725

Islamic Finance and Market-​Building

in Malaysia

The fact that Malaysia has achieved impressive growth rates in the 20th century—​
despite the dual presence of common law and shariʿa—​raises additional questions about
the suggested limitations imposed by Islamic institutions on economic growth. An im-
portant factor that contributed to this was the stabilization of interest rates following
Malaya’s independence from British colonialism. Following independence, the new
government prioritized building financial institutions as part of its new development
policy. This was partially driven by ongoing tensions between Malaysia and Singapore
regarding the terms of their currency union and the political struggle over seigniorage
rights (Schenk 2013). Though the Central Bank (Bank Negara Malaysia, or BNM) was
founded in 1959, it operated alongside a currency board that issued notes known as
Malaya dollars. Though the BNM had no seigniorage powers during its first few years, it
became the sole authority for issuing ringgit following the dissolution of the monetary
union between Malaysia and Singapore in 1975. Thereafter, BNM played a key role in
reducing Malaysia’s financial dependence on Singapore and focused on building a na-
tional banking system to ensure access to credit on affordable terms. Along with other
institutional changes, this further contributed to year-​on-​year Malaysian growth on eq-
uitable terms (Kuhonta 2011).
Despite these changes, there remained pockets of pious individuals who refused to
engage in interest-​based transactions. The size of these groups was even larger in areas
where the state remained weak to consolidate its authority over indigenous communities
(Scott 2009). Often seeing them as a potential security threat, the Malaysian authorities
sought to control these groups (also known as ajaran sesat (deviant)–​including many
Sufi orders, Ahmadiyah and Darul Arkam) by issuing official fatwas (Shiozaki 2015).
Others who were part of the bumiputera (indigenous Malays following the Shafiʿi school
of Islam) were also left outside the formal market-​building efforts and effectively had no
power to influence conventional interest rates, even though they were indirectly affected
by market dynamics. Modern Islamic finance in Malaysia emerged in response to these
interweaving dynamics.
The predecessors of Islamic banks appeared in the form of credit and savings
unions—​one of the first being Lembaga Tabung Haji (LTH)—​during the second half
of the twentieth century, to finance the pilgrimage expenses of very poor Malays to
Mecca by way of extending interest-​free loans.16 However, because this organization
had no power to shield itself from the changing conditions of the marketplace, it faced
an existential threat. Within the first decade following its launch, LTH managers had
seen a rapid increase in the number of members, but they had no power to create
halal money in isolation from the rest of the economy. Most importantly, because
the organization had no financial instruments to hedge against market instabilities,
their assets faced the risk of losing value over time. This compelled the founders
726   Fulya Apaydin

to seek a different solution and lobby for political support in favor of a shariʿa-​
compliant banking system.17 The opportunity was seized by United Malays National
Organization (UMNO) leaders who were also concerned about losing political sup-
port to the rival Parti Islam Se-​Malaysia (PAS)–​a political party that found some sup-
port among the more conservative Sheiks and the ulama–​and Bank Islam Malaysia
Berhad was born (Apaydin 2018; Lai 2015).
As the experience of LTH shows, banks are more than savings associations. To be fi-
nancially viable, these organizations need political support to sustain their lending and
borrowing activities. Among other things, this support enables the creation of neces-
sary banking reserves, access to a lender of last resort, and interbank money markets to
manage their short-​term liquidity needs. The role of the state is crucial in setting up this
financial architecture, so that the Islamic banks can go beyond simple intermediation
and become part of the capitalist growth by extending credit.
Since the inception of the first Islamic bank in 1983, the Malaysian government has
nurtured a large Islamic financial industry with a well-​developed regulatory infrastruc-
ture (Rethel 2015), where shariʿa scholars rely on majalla as the blueprint to devise new
regulations. From the very first days, the state was heavily involved in developing the
industry: Bank Islam Malaysia Berhad founders were given an office within the Central
Bank (Bank Negara Malaysia, BNM) and had access to technical assistance when setting
up their operations.18 The government also played an active role in building the financial
architecture necessary to conduct day-​to-​day business operations of the Islamic banks.
Following the decision that enabled conventional banks to open up Islamic windows
in 1993, the Islamic Interbank Money Market was set up in 1994, and the BNM took
the lead in creating a shariʿa-​compliant platform to trade reserves. Thereafter, the state
began to issue shariʿa-​compliant short-​term instruments—​a necessary means to ad-
dress the everyday liquidity needs of these institutions.
Proponents showcase the Malaysian experience as a success story where a dedicated
state commitment has created a favorable institutional setting that investors desire and
dream of. This process was strongly driven by the political will of the dominant party
(UMNO) focused on building a shariʿa-​compliant financial market.19 Gradually, Islamic
banking and finance has become strongly embedded in the national development and
competitiveness agenda of the government, constituting a major priority in future plans
(Lai 2015; Rethel 2018). However, the creation of these institutions based on liberal eco-
nomic principles by no means guarantees the channeling of funds into productive and
high-​value-​added industries. Today, the impressive performance in shariʿa-​compliant
financial market-​building leaves much to be desired for equitable growth. While a well-​
regulated Islamic banking and finance industry enabled corporations to raise capital on
shariʿa-​compliant terms, and helped investors diversify their financial portfolios, retail
customers were encouraged to take on greater debt for their financial needs and face a
greater risk of penalty for any failure to repay the loan. Thus, rather than promoting risk-​
sharing, the Islamic banking industry played an important role in raising household
debt largely for motor vehicle, consumer durables, and real-​estate purchases. Overall,
 Islamic Finance and Development in Malaysia    727

the majority of the credit extended by Islamic banking institutions finance consumption
rather than production.

Rethinking Financial Market-​Building and Development:

Credit for What and Whom?
Until the Asian financial crisis in 1998, Islamic banking services in Malaysia were either
provided by Bank Islam Malaysia Berhad or through the Islamic windows of conven-
tional banks. However, as the country experienced significant growth rates during the
1990s, Islamic banks found themselves in an increasingly competitive environment to
attract customers. During this period, conventional banks became influential players
in enabling access to credit, as the ratio of bank lending to GDP rose from 69 percent
in 1994 to 99 percent in 1997 (Takayasu and Yokoe 1999). To reduce this overexposure
to risk via bank lending, the government rolled out a new policy to develop capital
markets and encourage nonfinancial firms to secure their long-​term needs on alterna-
tive platforms via equity financing. As part of this initiative, the securities commission
of Malaysia was tasked to oversee the issuing of first shariʿa-​compliant securities in the
stock market in 1996. Thereafter, the government began to promote the development of
an Islamic equities market, creating attractive incentives for corporate sukuk (Islamic
bonds) issuance and trading (COMCEC 2018, 77).
During the 1998 crisis, the average cost of bank loans increased (reaching to 12.1 per-
cent), making it quite difficult for small and medium-​sized firms to access affordable
funding.20 Consequently, the plan to move away from bank-​based financing to market-​
based financing accelerated in the aftermath of the crisis. Larger firms gradually moved
to capital markets for their funding needs, while banks increasingly focused on retail
consumers. Meanwhile, corporate sukuk issuances increased exponentially beginning
in 2000s. At the same time, the volume of shariʿa-​compliant shares traded in the stock
market has also risen. However, much of this happened at the expense of small and
medium-​sized enterprises, which found themselves increasingly excluded from afford-
able funding options by conventional and Islamic banks.
Indeed, a closer look at the distribution of loans given by Islamic banks reveals that
most of the credit is channeled into consumption-​oriented sectors, such as mortgage
financing, vehicle purchase, and consumer credit (Apaydin 2018, 475). Between 2007
and 2016, most of the shariʿa-​compliant banks were far from financing trade or produc-
tive industries and did not significantly contribute to the growth of economic activity
in Malaysia.21 Most of these shariʿa-​compliant loans—​including real-​estate and vehicle
purchases—​carry a lower risk in line with Basel Committee on Banking Supervision
guidelines. Therefore, Islamic banks have generally been less willing to change the
composition of their assets from consumption-​ oriented to production-​ oriented
loans.22 These dynamics also played an important role in the rising household debt in
Malaysia, where individuals carry the major burden of market risks in an increasingly
728   Fulya Apaydin

financialized environment, thanks to fast-​rising private debt levels (Rahman, Ertürk,

and Froud 2019; Rethel 2011). The share of household debt as a percentage of GDP is
consistently on the rise, and doubled from 43 percent in 1997 to 89 percent in 2016 (Soh,
Chong, and Chuah 2017, 39).
To encourage a change in the lending behavior of Islamic banks, the Malaysian gov-
ernment took several measures. In a first step, the Badawi government designated
the entire halal industry as a priority as part of its economic development policy and
launched the first Malaysia International Halal Showcase (MIHAS) in 2004.23 As the
government sought to transform Malaysia into a global hub of halal economy, the
policymakers also hoped to link small and medium-​sized firms operating in halal food,
cosmetics, pharmaceuticals, and the travel industry to Islamic banks for their financing
needs. In a second step, the BNM provided new incentives to encourage banks to
move away from markup-​based contracts (murabaha) to profit-​sharing arrangements
(mudaraba) and thereby finance productive activities informed by the principle of risk-​
sharing via the launching of the Investment Account Platform.24 The long-​term impact
of these measures on the composition of Islamic banking loans remains to be seen.
Thus far, the most notable beneficiaries of Islamic financial market growth include
the state and large private corporations that rely on equity financing. Since the inception
of Islamic bonds (sukuk) in 1990, sovereigns as well as corporations have increasingly
used this instrument to attract investors and borrow to meet their short-​and long-​term
needs.25 At a global level, Malaysia had the largest share of domestic sukuk issued be-
tween 2001 and 2017, where private and public borrowing on shariʿa-​compliant terms
in capital markets constituted around 73 percent of the global sukuk market (IIFM 2018,
45). While Malaysia’s sovereign sukuk issuances have become increasingly popular to
attract petro-​dollars from the Gulf countries, a substantial share of this debt is held by
public institutions—​most notably, the Employees Provident Fund (COMCEC 2018,
89) and Public Sector Retirement Fund (KWAP).26 In terms of sectoral distribution of
corporate sukuk in 2017, about 70 percent of issuances are by financial service providers
and infrastructure and utilities companies, while the manufacturing sector issues less
than 10 percent of all corporate sukuk combined (COMCEC 2018, 83).27
In sum, the Islamic banking sector in Malaysia complements the conventional
banking industry. It creates money by way of extending shariʿa-​compliant credit and
serves as an important conduit for the financial inclusion of previously unbanked
individuals or customers who feel uncomfortable with conventional finance. A well-​
developed network of financial market institutions and regulatory architecture provides
a sound base for the industry players to carry their day-​to-​day operations. However, in
terms of mobilizing funds for investment in development, the Islamic banking sector
has a negligible impact, as much of the loans extended by these banks are for domestic
consumption. Rather than practicing and promoting risk-​sharing, these entities shift
risks to borrowers by encouraging consumption-​driven debt growth.28 While the rapid
growth of Islamic capital markets enabled investors to diversify their risk portfolios, the
majority of corporate sukuk is issued by financial companies who seek to strengthen
their budgets via shariʿa-​compliant equity financing. Overall, the exponential growth
 Islamic Finance and Development in Malaysia    729

in the corporate sukuk market is not driven by demand from large manufacturing firms
but by a fast-​growing financial services industry.

Conclusion

Across much of the Global South, economic growth and development are contin-
gent on affordable financing and reinvestment options. In settings where money crea-
tion is highly decentralized—​as has been the case in the Middle East and the Ottoman
Empire—​countries were unable to build effective financial system to channel funds
into production and trade. These historical disadvantages were somewhat attenuated in
areas where the state took a leading role in building modern market institutions. Thus,
Islamic banking and finance grew faster in political settings where the government
exhibited a clear commitment to aggregate this industry into capitalist market building.
In Malaysia, this process was strongly influenced by local demands of financially
marginalized groups that sought to shield themselves against the vagaries of liberalizing
markets.29
The development of Islamic banks and halal money creation through the hand of the
state had three important implications. First, the growth of Islamic banks went hand in
hand with the creation of Islamic capital markets, including a shariʿa-​compliant secu-
rities exchange and insurance industry (takaful). A second development involves the
financing of rapidly growing shariʿa-​compliant markets (also known as the halal in-
dustry), including pharmaceuticals, cosmetics, and food. Most of the halal certification
agencies are based in Malaysia and collect global rents by providing shariʿa compli-
ance certificates to producers of these commodities, not only in Malaysia, but also in
Central Asian republics and the MENA economies (Rethel 2019). Third, in areas with
low financial inclusion rates, Islamic banking has turned into a vehicle for financial in-
clusion and the mobilization of savings. As the founders of Islamic banks encountered
an opportunity to push forward their agenda in the midst of a political competition be-
tween two Islamist political parties, previously unbanked populations were integrated
into the financial system through the increasing availability of shariʿa-​compliant
instruments.
Some suggest that Islamic banking and finance—​especially the national variant in
Malaysia—​is emerging as a credible alternative to conventional finance, with an am-
bitious goal to move the current headquarters from New York, Tokyo, and London to
Kuala Lumpur (Rudnyckyj 2018). Arguably, the architects of Islamic finance seek to
replace the problematic debt-​based financing model with an equity-​based one, where
entrepreneurs and financiers share risks rather than seeking opportunities to extract
rent via instruments that mimic the use of interest (Rudnycky 2017, 2018). While there is
a greater call from shariʿa scholars to encourage equity-​based instruments and thereby
finance private entrepreneurs, current evidence reveals that the industry players are
quite sluggish to adopt this as their preferred business model. At least in Malaysia, the
730   Fulya Apaydin

managers of these banks are quite skeptical of this proposal, because it significantly
limits their ability to extend loans on more profitable terms.30
For Islamic banks to add a new impetus to capitalist development, consumer debt-​
driven growth needs to be replaced with one that channels credits into productive
industries, placing the emphasis on small and medium-​sized enterprises. Until recently,
Islamic financial actors have sought legitimacy in international markets by taking con-
ventional finance as the benchmark in developing the industry while claiming to offer
a more sustainable model as an alternative, highlighting the risk-​sharing principle as
their moral imperative. However, as of today, none of the Islamic banks have been able
to fulfil this lofty promise. Consequently, notable critics, such as Kuran (2004, 2018) and
El-​Gamal (2008), accuse the industry players of being deceitful opportunists eager to
scam faithful individuals. Aware of the problems associated with the growing domi-
nance of conventional models in Islamic finance, a group of Islamist economists have
embarked on a search for an alternative legitimacy framework, showing concern for
contemporary challenges such as inequality, climate change, and environmental deg-
radation. Among these, the Responsible Finance Initiative stands out with its ambitious
goals to reroute Islamic finance towards supporting equitable, inclusive, and sustainable
economic development.31 However, current evidence suggests reasons to curb one’s en-
thusiasm in the face of bottom-​up initiatives’ capacity to change the financial landscape
without a clear political commitment by the state.

Acknowledgments
The author would like to thank Melani Cammett, Pauline Jones, Steve Monroe, Jared Rubin,
and Yannis Karagiannis, the participants of The Oxford Handbook of Politics in Muslim Societies
Workshop at the Weatherhead Center for International Affairs, and the participants of the IBEI
research seminar series for very useful comments on earlier versions of this manuscript.

Notes
1. According to some scholars, since the end of the Bretton Woods system, banks play an even
greater role than the state in this regard: printed fiat money and other forms of state-​created
reserves constitute an ever decreasing share in the overall volume that is in circulation
(Ingham 2004; Ryan-​Collins, Greenham, and Werner 2011).
2. While banks may seek collateral or demand the borrower demonstrate future revenue flows
for paying off the loan, they are not limited to the existing amount of savings for extending
credit.
3. Examples of this include Iran, Sudan, and Pakistan.
4. One such case is Turkey between 1984 and 2018. Since the creation of Islamic banks by
Turgut Ozal’s ANAP government, the industry did not have access to government subsidies
and enjoyed limited institutional support to encourage its growth. Consequently, the
majority of the customers—​including the pious—​continue to work with conventional
banks, and the share of Islamic banking in the overall financial sector varies between 3
 Islamic Finance and Development in Malaysia    731

and 5 percent. Other similar cases where Islamic finance lacked state support include
Kyrgyzstan, Egypt, Tunisia, and Morocco, to name a few.
5. Though quite powerful, those occupying the seat of Sheikh-​ul-​Islam were not completely
autonomous in their decision-​making. There have been at least three executions by sultan’s
orders when the Sheikh-​ul-​Islam found himself at odds with the ruler or acted too inde-
pendently. For a detailed account on one of these cases, see Nizri (2014).
6. While Kuran (2018) argues that religious rules that prevented mergers among cash waqfs
prevented large-​scale capital accumulation and stagnated growth, there are reasons to sus-
pect the validity of this claim. Recent findings suggest that the absence of double-​entry
bookkeeping and the use of the stairs method (merdiban technique) in cash waqf ac-
counting partially explains why these funds dwindled over time (Demirhan, Susmus, and
Gonen 2012). Though this technique was also used by the Ilkhanids and Abbasids, there is
no strictly religious basis behind the use of this method in Islamic states.
7. Unlike the British Crown, the Ottomans were very late in borrowing from private lenders;
it was only after the financial difficulties in 1760 that private moneylenders emerged as
critical players to finance Ottoman borrowing needs (Pamuk 2000).
8. Relatedly, Rubin finds that the average real interest rate on loans was much higher—​
around 19 percent in the seventeenth and eighteenth centuries (Rubin 2017, 80).
For the same time period in Britain, interest rates varied between 3 and 5 percent
(Stasavage 2003).
9. The diversity of trade relations is another important dynamic behind a decentralized orga-
nization of the financial marketplace across most Islamic states (Platteau 2017).
10. When they borrowed in the domestic market, the sultans sought to broaden the base of
lenders beyond powerful sarrafs and reached out to small investors (Pamuk 2014).
11. Despite formal limitations on usury, however, borrowers in Muslim societies paid interest
to moneylenders regardless of their religious denomination (Calder 2016; Ercanbrack
2015; Kuran and Rubin 2018; Pamuk 2000; Rubin 2017). According to Ercanbrack (2015),
there were a lot of exceptions to the way financial exchanges operated on the ground. Local
customs were increasingly accommodated as a “go-​to” source to justify decisions that oth-
erwise did not conform to shariʿa.
12. This was the case despite the formal ban on interest under very powerful Muslim clerics,
who arguably delayed the development of organized moneylending and financial services
under banks (Calder 2016; Rubin 2017).
13. As Pamuk also notes, “in their monetary practices, Ottoman governments were well aware
of the limitations of their power. In comparison to goods markets and long-​distance trade,
it was more difficult for governments to control . . . exchange rates and interest rates”
(Pamuk 2000, 226).
14. Majalla is the Ottoman civil code that covers sales, insurance, rental contracts, debt, and
settlements, all informed by the Hanafi jurisprudence.
15. Later, Malaysian success came as a surprise to many, since most experts predicted Gulf
countries—​such as the UAE—​to become global leaders in halal finance, given their histor-
ical role in the emergence of modern Islamic banking.
16. This organization was founded by Ungku Abdul Aziz for pious Muslims to deposit their
savings and take out loans to finance their pilgrimage to Mecca—​with an embedded goal
to increase financial inclusion (Rethel 2015, 120).
17. Interview with an academic expert on Islamic banking and finance, November 29, 2016,
Kuala Lumpur, Malaysia.
732   Fulya Apaydin

18. Interview with former Islamic banking manager, November 22, 2016, Kuala Lumpur,
Malaysia.
19. Interview with a former government member, November 24, 2016, Kuala Lumpur,
Malaysia.
20. Data source: World Bank, Lending interest rate percent, https://​data.worldbank.org/​indi-
cator/​FR.INR.LEND?locations=MY&view=chart, accessed on March 20, 2019.
21. Even today, the Islamic retail banking industry’s biggest share of loans goes to financing
housing and real estate purchases. While they also extend credit to the manufacturing
sector, mortgage credit for buying property constitutes a large share of the business for
these banks. In some cases, their lending rates can be even more competitive than the
conventional interest rates. In that sense, Islamic banking’s largest impact is on the devel-
opment of housing and real estate, because elsewhere they cannot compete with the con-
ventional banks, which are much faster in offering credit.
22. Interview with an Islamic bank manager, November 30, 2016, Kuala Lumpur, Malaysia.
23. See http://​mihas.com.my/​mihas-​over-​the-​years/​, accessed on March 30, 2019. The halal
industry includes production of Islamic halal food, pharmaceuticals, and modest fashion,
as well as tourism.
24. See https://​www.muamalat.com.my/​business-​banking/​investment-​banking/​Investment-​
Account-​Platform.html, accessed on February 15, 2019. Murabaha-​based contracts are the
most commonly used contracts when extending consumer loans.
25. As of 2018, the public sector has issued 34.6 percent of all sukuk in Malaysia, while the
share for private sector stands at 65.4 percent (COMCEC 2018, 2).
26. Employees Provident Fund is a social security institution that manages the savings of
private-​sector employees and provides retirement benefits. KWAP is for public-​sector
workers. As of 2016, 45 percent of EPF’s assets under management are invested in sukuk,
while KWAP has invested 49.7 percent of its total assets in sukuk (COMCEC 2018, 87).
27. Among all sectors corporations that offer sukuk, financial services hold the largest
share and were responsible for nearly 50 percent of all corporate sukuk issued in 2017
(COMCEC 2018).
28. In addition, the Islamic banking industry in Malaysia follows in the footsteps of the con-
ventional bankers in promoting a market-​based solution to improve welfare, and allocates
some of its profits for donations to Islamic charity or halal corporate social responsibility
activities. However, the share of these initiatives is miniscule and their capacity to trigger a
multiplier effect to address socioeconomic inequalities is questionable.
29. In other Muslim-​dominant settings with no similar demand, the development of Islamic
banking and finance has been very limited. Examples include Turkey and Indonesia, where
the volume of Islamic banking transactions constitute around 5 percent of total banking
exchanges. Findings on customer preferences in these contexts reveal that the demand for
shariʿa-​compliant banking services is not very strong and the majority of borrowers feel
comfortable with using conventional banking instruments (Demiralp and Demiralp 2015;
Pepinsky 2013).
30. For example, many Islamic retail banks are also subject to similar liquidity ratio
requirements, as recommended by the Basel III committee, and moving from debt-​
based to equity-​based financing significantly lowers the Tier 1 and Tier 2 capital ratios,
as required by the Central Bank. Elsewhere in the GCC—​including the Emirates—​
there is hardly any intention to abandon debt-​based financing instruments in favor of
 Islamic Finance and Development in Malaysia    733

equity-​based arrangements. Interview with an Islamic bank manager, November 30, 2016,
Kuala Lumpur, Malaysia.
31. See http://​www.rfi-​foundation.org/​, accessed on December 12, 2018. Most RFI board
members and council of advisors have held key positions in Islamic finance industry
in Malaysia, though the initiative includes industry practitioners from the Middle East
as well.

References
Acemoglu, Daron, and James Robinson. 2012. Why Nations Fail. New York: Crown Books.
Agir, Seven. 2017. “The Rise and Demise of Gedik Markets in Istanbul, 1750–​1860.” Economic
History Review 71, no. 1: 133–​156.
Agir, Seven, and Onur Yildirim. 2015. “Gedik: What’s in a Name?” In Bread from the Lion’s
Mouth: Artisans Struggling for a Livelihood in Ottoman Cities,” edited by Suraiya Faroqhi,
217–​236. New York: Berghahn Books.
Apaydin, Fulya. 2018. “Regulating Islamic Banks in Authoritarian Settings: Malaysia and
the United Arab Emirates in Comparative Perspective.” Regulation & Governance 12, no.
4: 466–​485.
Arrighi, Giovanni. 1994. The Long Twentieth Century: Money, Power and the Origins of Our
Times. London: Verso Books.
Aziz, Zeti Akhtar. 2011. “New Platform to Serve the Islamic Financial Industry.” Keynote ad-
dress by Dr Zeti Akhtar Aziz, Governor of the Central Bank of Malaysia, at the launch of
Bloomberg’s Enhanced Islamic Finance Platform, Kuala Lumpur, February 21, 2011.
Broms, Rasmus, and Bo Rothstein. 2020. “Religion and Institutional Quality: Long-​Term
Effects of the Financial Systems in Protestantism and Islam.” Comparative Politics 52, no.
3: 433–​454.
Bulliet, Richard W. 1972. “The Shaikh Al-​Islam and the Evolution of Islamic Society.” Studia
Islamica 35: 53–​67.
Calder, Ryan. 2016. “God’s Technicians: Religious Jurists and the Usury Ban in Judaism,
Christianity, and Islam.” European Journal of Sociology 57, no. 2: 207–​257.
Choudhury, Masudul Alam. 1997. Money in Islam. New York: Routledge.
Clay, Christopher. 1994. “The Origins of Modern Banking in the Levant: The Branch Network
of the Imperial Ottoman Bank, 1890–​1914.” International Journal of Middle East Studies 26,
no. 4: 589–​614.
COMCEC. 2018. The Role of Sukuk in Islamic Capital Markets. Ankara, Turkey: COMCEC.
Demiralp, Seda, and Selva Demiralp. 2015. “The Rational Islamic Actor? Evidence from Islamic
Banking.” New Perspectives on Turkey 52 (May): 3–​27.
Demirhan, Dilek, Turker Susmus, and Seckin Gonen. 2012. “Cash Waqfs and Their Accounting
Applications at the End of the 18th Century in the Ottoman Empire.” Atatürk Üniversitesi
İktisadi Ve İdari Bilimler Dergisi 26, no. 3–​4: 101–​123.
Desan, Christine. 2017. “The Constitutional Approach to Money.” In Money Talks: Explaining
How Money Really Works, edited by Viviana Zelizer, Nina Bandelj, and Frederick Wherry,
109–​130. Princeton, NJ: Princeton University Press.
El-​Gamal, Mahmoud. 2008. Islamic Finance: Law, Economics, and Practice. New York: Cambridge
University Press.
734   Fulya Apaydin

Ercanbrack, Jonathan. 2015. The Transformation of Islamic Law in Global Financial Markets.
New York: Cambridge University Press.
Fleet, Kate. 2003. “Tax-​Farming in the Early Ottoman State.” Medieval History Journal, 6, no.
2: 249–​258.
Hussin, Iza R. 2016. The Politics of Islamic Law: Local Elites, Colonial Authority, and the Making
of the Muslim State. Chicago: University of Chicago Press.
IIFM (International Islamic Financial Market). 2018. IIFM Annual Sukuk Report 2018.
Manama, Bahrain: IIFM.
Ingham, Geoffrey. 2004. The Nature of Money. Cambridge: Polity Press.
Iqbal, Zamir, and Abbas Mirakhor. 2013. Economic Development and Islamic Finance.
Washington, DC: World Bank.
Kuhonta, Erik M. 2011. The Institutional Imperative: The Politics of Equitable Development in
Southeast Asia. Stanford, CA: Stanford University Press.
Kuran, Timur. 2004. Islam and Mammon: The Economic Predicaments of Islamism. Princeton,
NJ: Princeton University Press.
Kuran, Timur. 2011. The Long Divergence: How Islamic Law Held Back the Middle East.
Princeton, NJ: Princeton University Press.
Kuran, Timur. 2018. “Islam and Economic Performance: Historical and Contemporary Links.”
Journal of Economic Literature 56, no. 4: 1292–​1359.
Kuran, Timur, and Jared Rubin. 2018. “The Financial Power of the Powerless: Socioeconomic
Status and Interest Rates under Partial Rule of Law.” Economic Journal 128, no. 609: 758–​796.
Lai, Jikon. 2015. “Industrial Policy and Islamic Finance.” New Political Economy 20, no.
2: 178–​198.
Minsky, Hyman. 1986. Stabilizing an Unstable Economy. New Haven, CT: Yale University Press.
Nizri, Michael. 2014. Ottoman High Politics and the Ulema Household. Basingstoke, UK:
Palgrave.
North, Douglass. 2005. “Institutions and the Process of Economic Change.” Management
International 9, no. 3: 1–​7.
Pamuk, Sevket. 2000. A Monetary History of the Ottoman Empire. New York: Cambridge
University Press.
Pamuk, Sevket. 2014. Türkiye’nin 200 Yıllık İktisadi Tarihi. İstanbul: Turkiye Is Bankasi Kultur
Yayinlari.
Pepinsky, Thomas B. 2013. “Development, Social Change, and Islamic Finance in Contemporary
Indonesia.” World Development 41 (January): 157–​167.
Pettifor, Ann. 2017. The Production of Money. New York: Verso Books.
Platteau, Jean-​ Philippe. 2017. Islam Instrumentalized: Religion and Politics in Historical
Perspective. New York: Cambridge University Press.
Rahman, Syahirah Abdul, Ismail Ertürk, and Julie Froud. 2019. “Financial Citizenship and
Nation-​ Building in Malaysia: Elites’ and Citizens’ Perspectives.” Journal of Economic
Geography 20, no. 1: 225–​248.
Rethel, Lena. 2011. “Whose Legitimacy? Islamic Finance and the Global Financial Order.”
Review of International Political Economy 18, no. 1: 75–​98.
Rethel, Lena. 2015. “Islamic Finance in Malaysia: Global Ambitions, Local Realities.” In The
Everyday Political Economy of Southeast Asia, edited by Juanita Elias and Lena Rethel, 116–​
136. New York: Cambridge University Press.
Rethel, Lena. 2018. “Economic Governance beyond State and Market: Islamic Capital Markets
in Southeast Asia.” Journal of Contemporary Asia 48, no. 2: 301–​321.
 Islamic Finance and Development in Malaysia    735

Rethel, Lena. 2019. “Corporate Islam, Global Capitalism and the Performance of Economic
Moralities.” New Political Economy 24, no. 3: 350–​364.
Rubin, Jared. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the Middle East Did
Not. New York: Cambridge University Press.
Rudnycky, Daromir. 2017. “Subjects of Debt: Financial Subjectification and Collaborative Risk
in Malaysian Islamic Finance.” American Anthropologist 119, no. 2: 269–​283.
Rudnycky, Daromir. 2018. Beyond Debt: Islamic Experiments in Global Finance. Chicago:
University of Chicago Press.
Ryan-​Collins, Josh, Tony Greenham, and Richard Werner. 2011. Where Does Money Come
From? London: New Economics Foundation.
Schenk, Catherine R. 2013. “The Dissolution of a Monetary Union: The Case of Malaysia and
Singapore 1963–​1974.” Journal of Imperial and Commonwealth History 41, no. 3: 496–​522.
Scott, James C. 2009. The Art of Not Being Governed. New Haven, CT: Yale University Press.
Shiozaki, Yuki. 2015. “The Historical Origins of Control over Deviant Groups in Malaysia:
Official Fatwa and Regulation of Interpretation.” Studia Islamika 22: 205–​233.
Singh, Saran. 1986. The Encyclopaedia of the Coins of Malaysia, Singapore, and Brunei, 1400–​1986.
Kuala Lumpur, Malaysia: Malaysia Numismatic Society.
Soh, Jiaming, Amanda Chong, and Kue-​Peng Chuah. 2017. “Household Credit, Growth and
Inequality in Malaysia: Does the Type of Credit Matter?” In Financial Systems and the Real
Economy, 39–​59. BIS Papers 91. Bank for International Settlements. https://​www.bis.org/​
publ/​bppdf/​bispap91c.pdf.
Stasavage, David. 2003. Public Debt and the Birth of the Democratic State: France and Great
Britain 1688–​1789. New York: Cambridge University Press.
Takayasu, Kenichi, and Yosie Yokoe. 1999. “Non-​Performing Loan Issue Crucial to Asia’s
Economic Resurgence.” Japan Research Institute Journal 44. https://​www.jri.co.jp/​english/​
periodical/​rim/​1999/​RIMe199903npl/​.
Wray, Randall. 2015. Why Minsky Matters: An Introduction to the Work of a Maverick Economist.
Princeton, NJ: Princeton University Press.
 Pa rt V I

S O C IA L W E L FA R E
A N D G OV E R NA N C E
 Chapter 36

C ol onial Leg ac i e s a nd
Welfare Prov i si on i n
the M iddle E ast a nd
North Af ri c a

Melani Cammett, Allison Spencer Hartnett,

AND Gabriel Koehler-​D errick

In the decade since the Arab uprisings began, protesters across the Middle East and
North Africa (MENA) region have called upon their governments to address deep-​
seated socioeconomic grievances and long-​ standing regional inequalities. These
grievances mirror the findings of academic analyses of the MENA region as beset by
structural obstacles to economic and political development (Krafft and Alawode 2016;
Malik and Awadallah 2013; Noland and Pack 2007; Rougier 2016).1 Protesters and
analysts alike note the centrality of welfare provision to the well-​being of citizens and the
health of national economies. In public opinion polls, citizens report high levels of dis-
satisfaction with public services and perceptions of rising inequality, and unfair access
to economic opportunities are widespread (Cammett and Salti 2016). As others have
shown, MENA welfare states vary significantly from one another in the targeting and
generosity of services (Eibl 2020). The early design of MENA public welfare, however,
predominantly originated under foreign rule. We propose that colonial configurations
of the welfare state laid the foundations for long-​run patterns of service provision
post-​independence.
In this chapter, we contribute a historical argument that links the development of wel-
fare services under colonialism to postcolonial welfare provision. We seek to explain var-
iation in welfare provision in two core social sectors: health and education. We measure
welfare provision in two ways: physical infrastructure (i.e., hospitals and schools) and
efficacy (i.e., crowding in schools and hospitals). We argue that much of the observed
variation in postcolonial welfare expansion is due to a combination of the degree to
740    Melani Cammett et al.

which welfare provision under colonialism relied on public versus private institutions
and how much the composition of the ruling elite changed post-​independence.2
We argue that post-​independence politics were crucial in determining the scope of
social welfare provision, particularly in the distribution of the expansion of welfare
infrastructure that occurred throughout the region in the decades following inde-
pendence. Under colonialism, European officials designed and implemented limited
welfare services in occupied territories that served political logics of control and co-​
optation, including to educate the nascent bureaucratic class (Hourani 1968) and, where
settler colonialism existed, to benefit European settlers (Owen 1981). Given the limited
scope of welfare provision under colonialism, governments across the region faced
pressure to expand access to social welfare institutions at independence. However, de-
spite popular demand, there was enormous variation in the expansion of social welfare
in the key decades following independence. Drawing on histories of the region, we note
that in the MENA region the expansion of social welfare provision was a top-​down pro-
cess determined at the highest levels of political power. We argue that it was more likely
that governments would expand welfare infrastructure where political elites were more
broadly representative because patronage incentivized politicians to ensure adequate
health and school infrastructure in their “home” regions. The more concentrated the
political elite, the less pressure there was to expand social welfare infrastructure beyond
key urban areas. To measure this, we focus on the relative breadth of ministerial cabinets
as a proxy measure for elite inclusivity.
While increasing access to social welfare was crucial, the decades post-​independence
were characterized by significant variation in the quality of public services across the
MENA region. We argue that the degree to which colonial governments relied on public
versus private institutions in the education and health sectors accounts for much of the
variation in efficacy post-​independence. In some cases, like Tunisia, health and edu-
cation were primarily provided by public facilities under colonialism. In others, like
Jordan, key welfare sectors were dominated by private institutions. Where private
institutions predominated under colonialism, post-​independence states faced higher
costs to train and retain the small pool of teachers and doctors so desperately needed to
meet surging demand for social welfare services. Where public institutions were more
prevalent under colonialism, post-​independence governments faced fewer obstacles to
training and recruiting teachers and doctors even as demand for social welfare services
increased exponentially.
This chapter proceeds as follows. The next section briefly reviews prominent
theories that account for variation in social welfare provision generally, and then
discusses the limitations of these theories for accounting for variation in social wel-
fare in the MENA region. We then outline our theory, introduce our data and key
concepts, and present descriptive evidence supporting our theoretical claims about
the importance of elite coalitions and colonial reliance on public versus private
institutions. We conclude with four brief case studies that substantiate our theoretical
claims using historical evidence.
 Colonial Legacies and Welfare Provision    741

A Historical Framework

There is general consensus on the proximate causes of underdevelopment in the Middle

East, including weak integration in the global economy, low levels of investment, lack of
technology transfer, industrial noncompetitiveness, high levels of government owner-
ship and investment, the low quality of education, and the high costs of doing business
(Noland and Pack 2007, 11). However, these factors are symptoms of deeper causes and
do not entirely account for intraregional variation in economic and social outcomes.3
Theories of the development of welfare states often emphasize factors such as electoral
competition (Niedzwiecki 2018), the role of organized labor (Esping-​Anderson 1990),
or business interests (Mares 2003). Organized labor and business interests in MENA
have traditionally been weak and lacked independence from the state, particularly in the
initial decades after independence (Beinin 2001; Bellin 2011; Cammett 2007).4 Another
dominant strand in the literature concerns the role of ethnic diversity and social wel-
fare outcomes (Alesina and Glaeser 2004). However, other scholars have highlighted
numerous theoretical and conceptual obstacles to assessing this relationship (Baldwin
and Huber 2010; Singh and vom Hau 2016; Wimmer 2016), which helps to explain why
the association between ethnic diversity and social welfare is so highly contested and
inconsistent. In some cases, ethnic differences do appear to have limited social welfare
provision (Franck and Rainer 2012; Miguel and Gugerty 2005; Soifer 2016), but in other
contexts, ethnic groups have transcended these differences (Miguel 2004; Singh 2016;
Tajima, Samphantharak, and Ostwald 2018). Even within the MENA region, where
some parties are closely tied to various ethnic groups, ethnicity alone is an unreliable
predictor of the provision of social welfare services (Cammett 2014).
Recent advances in the literature on nation-​building and welfare under autocracy
provide another set of explanations for examining the historical origins of autocratic
welfare institutions. Scholarship in this vein has shown that contrary to the expecta-
tions of modernization theory, autocrats are frequently more redistributive than their
democratic counterparts (Albertus 2015). Studying the expansion of education over
the last two centuries, Paglayan (2020) finds that most expansion of primary schooling
happened under autocratic regimes. Slater and Fenner (2011) argue that expanding
public services like schooling is a core component of durable regimes because it
increases the state’s infrastructural power by registering and socializing students as well
as making professionals like teachers dependent on state employment. In sum, the liter-
ature leads us to expect that a combination of elite preferences and the historical endow-
ment of welfare infrastructure play an important role in conditioning welfare provision
in autocracies.
We focus on the relationship between institutions and policies, on the one hand,
and welfare provision, on the other, because of the long-​standing emphasis on this
key relationship in the literature on economic development (Acemoglu, Gallego, and
Robinson 2014). We examine the colonial period because prior to colonization, with
742    Melani Cammett et al.

the possible exception of Egypt, no government in the region had attempted to create
a nationwide network of publicly administered schools and hospitals (Gallagher 1983,
1990; Szyliowicz 1973). Furthermore, a large body of literature from across the devel-
oping world suggests that colonial legacies were particularly influential in shaping long-​
run trends in education, and MENA cases stand to contribute to our understanding of
how and why some patterns rooted in colonialism survive beyond independence (Nunn
2014; Valencia Caciedo 2019; Wantchekon, Klasnja, and Novta 2015).
In focusing on the nature of elite coalitions after independence, we draw on a
large literature that suggests that the preferences and strategic considerations of elite
coalitions play a role in shaping a range of policy outcomes (Bueno de Mesquita et al.
2003; Eibl 2020; Haggard and Kaufman 2009; Hodler and Raschky 2014; Khan 2010;
Pepinsky 2009; Ricart-​Huguet 2020; Slater 2010; Teichman 2016; Waldner 1999), as
well as historical accounts that emphasize independence as a critical juncture for ec-
onomic and political development across the MENA region (Issawi 1995; Owen 1981;
Owen and Pamuk 1999). We are, of course, not the first to focus on the role of elite
coalitions and social welfare provision in the region. Eibl (2020) argues that the polit-
ical elite may factor ethnic divisions into their calculus regarding welfare spending, but
only if those divisions are politically salient. While we agree that elite politics are cru-
cial, rather than focusing on the role of external threats for elite cohesion (Eibl 2020;
Slater 2010), we instead examine the relative continuity of political elites after indepen-
dence. Where there was relative continuity, we argue there was much less incentive to
expand social welfare provision.

A Theoretical Framework
for Colonial Legacies and
MENA Development

We contend that two factors shaped welfare provision in the postcolonial Middle
East: the breadth of elite political coalitions post-​independence, and the reliance of
colonial regimes on public versus private institutions. First, we distinguish between
the legacies of colonies that relied heavily on facilities controlled by private, nonstate
interests (i.e., private institutions), and those in which colonial governments expanded
a national network of facilities funded, staffed, and administered by the state (i.e., public
institutions). Variation in colonial reliance on public versus private institutions shaped
welfare promotion post-​independence because it shaped the human capital inherited
at independence and impacted the labor market for scarce white-​collar workers in the
decades after independence. In colonies that relied on public institutions, the post-​
independent state inherited more human capital at independence and faced less com-
petition for the services of doctors and teachers, reducing the costs of training and
recruitment relative to former colonies, which relied on private institutions.
 Colonial Legacies and Welfare Provision    743

Second, the relative breadth of the political elite at independence determined the ex-
tent to which the state would promote welfare in terms of expanding the geographic
reach of services, determining eligibility criteria for public schools and healthcare
facilities, and setting funding levels for public institutions.

Public and Private Institutions and the Welfare State

Even the most ambitious post-​independence plans for welfare expansion faced real
constraints imposed by colonial policy. Although colonial education never reached
more than a small minority of subjects, we argue that its relative breadth proved highly
consequential at independence. There were at least three reasons for this. First, colo-
nies that relied on public institutions generally provided services to larger number
of colonial subjects than colonies that relied on private institutions. Second, public
institutions were typically more geographically dispersed than private ones, which
tended to cluster in a few key urban locales. Third, where private institutions persisted
post-​independence, they often disincentivized investment in the public sector because
they allowed political elites to “opt out” of public institutions and often provided more
attractive terms and working conditions for scarce white collar workers.5
A defining feature of European colonialism across the world was inequality in health
and education outcomes between European citizens and colonial subjects. These
inequalities arose for a variety of reasons, including but not limited to colonizers’ stra-
tegic limitation of education to colonized populations or preferential provision to
European settlers. These dynamics are well documented in historical accounts of colo-
nial education in the Middle East (Allman 1979; Ayachi 2003; Falb Kalisman 2017; Hyde
1978; Salman 1986; Segalla 2009; Szyliowicz 1973) as well as health policy (Amster 2014;
Gallagher 1990, 1983). While inequality characterized access to public institutions be-
tween citizens and subjects in colonies throughout the region, there was significant var-
iation in the ways that colonial governments structured the provision of healthcare and
education. In some countries, like Lebanon, public institutions were limited, and private
institutions, particularly those of a religious character, dominated the provision of edu-
cation and healthcare. In others, like Morocco, the colonial government was hostile to
private institutions, particularly missionary organizations, meaning that with the ex-
ception of a limited number of facilities run by the Catholic Church, the state was solely
responsible for the provision of education and health services.6
Figure 33.1 provides an illustration of variation in public service provision for four
countries: Egypt, Jordan, Morocco, and Tunisia. The plot shows changes in the portion
of public schools at five-​year intervals for all available years between 1900 and 1970.
(The dashed line indicates the year of independence.) We observe significant differences
across all four countries. More than half of schools in Jordan were private prior to inde-
pendence, but this total never exceeded more than 10 percent in Tunisia. We argue that
this variation in the relative influence of private institutions had significant implications
for welfare promotion after independence.
744    Melani Cammett et al.

Public Schools as a Percentage of All Schools

Egypt Jordan
1.00
0.75
0.50
0.25
Percent

Morocco Tunisia
1.00
0.75
0.50
0.25

1910 1920 1930 1940 1950 1960 1970 1910 1920 1930 1940 1950 1960 1970
Year

Public Schools

Figure 33.1: Public schools as a percentage of all schools in Egypt, Jordan, Morocco, and
Tunisia.

Building on classic works emphasizing the essential role of public institutions in the
process of nation-​building and modernization (Anderson 1983; Gellner 1983; Weber
1976), we conceptualize this binary distinction strictly, classifying both traditional
and modern institutions run by nonstate (often religious) authorities as “private.” This
includes the broad network of “traditional” Qurʾanic schools known as kutabs,7 which
often were regulated by the government, but whose funding came from local sources.8
The importance of this binary distinction is shown in Figure 33.1, where the rapid in-
crease in Egyptian public schooling during the 1920s and 1930s reflects not only a pro-
gram of educational expansion that began in 1923, but also the fact that in the mid-​1930s,
the government brought all kutabs under the authority of the Ministry of Education,
making what had been privately run institutions public.

Political Elites and Welfare Outcomes

Our analytical framework combines a focus on the constraints inherited from colonial
rule with an understanding that the political will to expand welfare was a central fea-
ture of postcolonial state-​building. Elite coalitions—​the political elites that constitute
the ruling regime—​shape the institutions and policies that affect subsequent economic
and political policies and, potentially, development outcomes. Put differently, the com-
position of the political elite determines the political motivation to enact more or less
expansive welfare policies.
For example, when post-​independence rulers emerged from a revolutionary break
from the ancien regime, they were more likely to prioritize greater investment in welfare
as part of state-​and nation-​building projects (Cammett and Sasmaz 2016, 8), yielding a
rupture in historical legacies. Since colonial governments typically restricted political
participation to a narrow elite, continuity suggests that postcolonial policies would be
 Colonial Legacies and Welfare Provision    745

less inclusive than those in places that experienced a revolutionary break from the past.
For example, as Waldner (1999) shows, in some newly independent countries, the policy
choices of new ruling elites were driven by their need to launch populist appeals to shore
up their support bases in the face of elite fragmentation.9 Elite coalitions have also been
found to be important in other studies of distributive politics in MENA. Eibl (2020)
uses postcolonial spending data to show that broad and narrow coalitions exert different
levels of welfare effort.10 Hartnett (2019) finds that the inclusivity of colonial and postco-
lonial coalitions conditioned the strength and nature of rising elites’ demands for land
reform, primarily due to the fact that inclusive coalitions engaged in alternative forms
of urban-​rural redistribution while exclusive regimes resisted out-​group demands.
Similarly, in her study of women’s rights and family law in North Africa, Mounira
Charrad (2001) demonstrates that the Moroccan regime’s alliance with conservative
rural notables deterred reform while the Tunisian Neo-​Destour Party’s reliance on a co-
alition of middle-​class professionals brought about changes in the legal status of women,
even in the absence of a grass-​roots feminist movement.
We assume that, like all governments, elite coalitions post-​independence tried to
maintain political power. Unlike previous studies that emphasize divisions among elite
coalitions (Eibl 2020; Waldner 1999), however, we focus specifically on the inclusivity of
the political elite. Elite coalitions were particularly important, we argue, because in the
authoritarian regimes that emerged across the region post-​independence, distributive
politics were personal and parochial. While governments set aspirational goals, political
elites enjoyed tremendous discretion over the allocation and prioritization of resources.
In the absence of electoral accountability, politicians had incentives and opportunity
to direct resources to their home region, where thicker networks of personal contacts
and connections facilitated their control and influence over local implementation and
strengthened their personal and family reputations.11
Following previous scholarship, we focus on the composition of the ministerial cab-
inet to measure elite composition, as this gives us a generalizable measure of the compo-
sition of the highest levels of formal political power (Arriola 2009; Arriola and Johnson
2014; Buehler and Ayari 2018; Nyrup and Bramwell 2020). Although there are many
ways to operationalize elite coalition composition, we follow Ricart-​Huguet (2020)
in adopting a measure of ministers’ geographic diversity as a telling indicator of polit-
ical inclusion. By examining ministers’ birth places, we attempt to capture not only the
individual-​level diversity of ministers, but, more importantly, the potential for diverse
constituencies to be represented at the highest level of politics. Scholars of ministerial
politics in postcolonial sub-​Saharan African countries have noted that a region’s rep-
resentation in the ministerial cabinet often increases the level of distribution to that re-
gion via mechanisms such as elite favoritism, patronage politics, or constituent lobbying
(Koter 2016; Kramon and Posner 2016).
Using data from Hartnett (2020), Figure 33.2 plots the average Herfindahl-​Hirschman
Index (HHI) for postcolonial cabinets in Tunisia,12 Morocco,13 Egypt,14 and Jordan15
from independence until 1970. HHI is a measure of concentration ranging from 0 to
746    Melani Cammett et al.

10,000. Originally used to describe competition in economic markets, a low HHI value
(from zero to 1,500) indicates that the field of competition, in this case geographic rep-
resentation in the ministerial cabinet, is more diverse and not dominated by elites from
any one region. An HHI value ranging between 1,500 and 2,500 signifies a moderately
concentrated cabinet, where geographic representation might draw several ministers
from a few select districts in a given cabinet. Values greater than 2,500 indicate highly
concentrated ministerial geography, where typically one or two regions monopolize
political power.
We observe in Figure 33.2 that no country case falls into the competitive category,
reflecting the propensity toward political inequality in postcolonial and autocratic
regimes. But the plot also shows clear variation between cases. Egypt shows moderate
levels of concentration and is the most geographically inclusive of our cases, with an
HHI of 2,136. Although Tunisia’s HHI of 2,972 falls above the “highly concentrated”
threshold of 2,500, the ministerial elite draw from a larger pool of districts than either
Morocco (HHI = 3,692) or Jordan (HHI = 4,578). Based on these relative differences,
we use this measure of cabinet geographic diversity to categorize the elite coalitions
of Egypt and Tunisia as relatively inclusive, whereas Moroccan and Jordanian elite
coalitions are more exclusive.

5000
4578

4000
Average HHI (Independence to 1970)

3692

3000 2972

2136
2000

1000

0
Egypt Tunisia Morocco Jordan

Competitive (<1500)
Concentrated (>2500)

Figure 33.2: Average cabinet, Herfindahl-​Hirschman Index (HHI), independence to 1970.

Source: Data from Hartnett (2020).
 Colonial Legacies and Welfare Provision    747

A Typology of Postcolonial Welfare Provision

Table 33.1 characterizes our expectations for welfare provision after independence based
on two variables: legacies of public and private welfare provision under colonial rule,
and elite coalition inclusivity.
This typology yields four distinct ideal types of social development regimes in the
postcolonial MENA region. At one end of the spectrum in the typology is the “expan-
sive, effective” pattern, in which post-​independence elites embarked on ambitious state-​
building projects. Breaking with colonial era ruling social coalitions, they aimed to
expand social welfare provision for a broad swath of the population, thereby fostering
broader human development. Building on public institutions built under colonialism,
it was easier to staff these new facilities, both because the absence of a private sector
reduced competition for doctors and teachers, but also because a greater reliance on
public institutions under colonialism increased the pool of future white collar workers.
Political elites in these contexts were more willing to spend significant portions of
their limited budgets on social welfare provision because the expansion of health and
education infrastructure enhanced their local prestige, increased their influence over
an important source of patronage, and helped address soaring demand for services that
were extremely limited under colonialism.
At the other end of the spectrum is the “restrictive, ineffective” pattern, which features
economic and social policies aimed at a narrow subset of the population and continuity
between the colonial and postcolonial ruling social coalitions. In these cases, despite
colonial reliance on public institutions, continuity in the elite coalition pre-​and post-​
independence meant there was little incentive to expand the provision of public goods
beyond rewarding key urban constituencies. This resulted in a publicly oriented social
welfare system, but one shot through with glaring inequalities in both infrastructure
and staffing.
Two other patterns of state investment in welfare provision fall in between these
types. The “restrictive, effective” variant entails policies that are relatively successful in
developing and implementing policies to promote welfare provision. However, given

Table 33.1: Patterns of Welfare Provision in Post-​Independence Polities

Colonial Education and Health

Public Oriented Private Oriented

Elite Coalition Inclusive Expansive, Effective Expansive, Ineffective

Post-​Independence (Broad Elite) (Tunisia) (Egypt)
Exclusive Restrictive, Ineffective Restrictive, Effective
(Narrow Elite) (Morocco) (Jordan)
748    Melani Cammett et al.

the elite social foundations of the post-​independence ruling coalition, which exhibits
relative continuity with the colonial period, economic and social policies involved less
expansion in welfare provision. The “expansive, ineffective” type encompasses regimes
that included a diverse ruling elite, which broke from colonial-​era rulers, and expanded
social welfare infrastructure. However, these regimes often faced huge difficulties in
staffing and training, often because of the continued importance of private institutions
under colonialism and after independence.

Data and Conceptualization

To illustrate our argument, we draw on data from government statistical yearbooks and
archival resources from four countries: Egypt, Tunisia, Morocco, and Jordan. These
data come from the Middle East/​North Africa Historical Data Archive (MENAHDA)
project.16 This database consists primarily of digitized statistical yearbooks, annual
publications that collected statistics from across the bureaucracy in a single bound
volume. Dating back to the colonial period, these yearbooks provide us with valuable
information from a range of different domains, including public health, education, and
infrastructure, as well as economic indicators at both the national and subnational level.
While there is variation in the nature of statistics reported in each country, as well as the
level of aggregation and consistency year-​to-​year, these yearbooks provide a wealth of
demographic, economic, and administrative statistics at the national and subnational
levels. The data below illustrates how these four cases perform based on the expansive-
ness and effectiveness criteria in our theory.

Expansiveness
We measure welfare effectiveness as the number of public schools and hospital per
100,000 citizens.17 As shown in Figure 33.3, there is stark variation across our four cases
in terms of the provision of public schools adjusted by population size. Jordan exhibits
the sharpest increase in public schools post-​independence, followed by a more gradual
increase in Tunisia. In Morocco, there is virtually no change post-​independence, and
public school accessibility actually declined in Egypt, as surging population growth
overcame the expansionary efforts of the Nasser government post-​independence,
discussed in more detail below.
Figure 33.4 shows distinct trends across cases with respect to healthcare infrastruc-
ture. The number of hospitals per capita started to increase in Egypt prior to indepen-
dence, a trend that continued under Nasser. In Jordan, the number of hospitals per
capita increased initially but fell precipitously with the loss of the East Bank in 1967. In
Morocco and Tunisia, trends were essentially flat pre-​and post-​independence.
 Colonial Legacies and Welfare Provision    749

Public Schools Per 100,000 Citizens

Egypt Jordan
60

40
N Per 100K Citizens

20

0
Morocco Tunisia
60

40

20

0
1910 1920 1930 1940 1950 1960 1970 1910 1920 1930 1940 1950 1960 1970
Year

N Schools Per 100K Citizens

Figure 33.3: Education infrastructure in Egypt, Jordan, Morocco, and Tunisia.

Hospitals Per 100,000 Citizens

Egypt Jordan
3

2
N Per Million Citizens

Morocco Tunisia
3

1910 1920 1930 1940 1950 1960 1970 1910 1920 1930 1940 1950 1960 1970
Year
N Hospitals Per 100K Citizens

Figure 33.4: Health infrastructure in Egypt, Jordan, Morocco, and Tunisia.

Effectiveness
What about the relative effectiveness of these distinct patterns of welfare provision?
As previously noted, although governments differed in the degree to which they
encouraged and facilitated the expansion of public education, all governments post-​
independence had to respond to the challenge of a surge in demand for social welfare
services while facing significant constraints in training and staffing these facilities.
Figure 33.5 illustrates this challenge by focusing on trends in student crowding, specifi-
cally the number of students per public school.
750    Melani Cammett et al.

Students Per Public School

Egypt Jordan
750

500
Students Per School

250

Morocco Tunisia
750

500
250

1910 1920 1930 1940 1950 1960 1970 1910 1920 1930 1940 1950 1960 1970
Year
Students Per Public School

Figure 33.5: School crowding in Egypt, Jordan, Morocco, and Tunisia (students per public school).

Doctors Per Hospital

Egypt Jordan
60

40
Doctors Per Hospital

20
0
Morocco Tunisia
60

40
20
0
1910 1920 1930 1940 1950 1960 1970 1910 1920 1930 1940 1950 1960 1970
Year
Doctors Per Hospital

Figure 33.6: Doctors per hospital in Egypt, Jordan, Morocco, and Tunisia.

Independence reduced the barriers to entry to public education across the region,
but governments clearly differed in how quickly and effectively they were able to meet
surging demand for social welfare services. Student crowding surged in Egypt and
Morocco post-​independence, but was much more gradual in Tunisia, a direct result of
a massive expansion of public schools, and increased only modestly in Jordan. While
this was in part due to the role of UNRWA schools in Jordan, as discussed in more detail
below, there is a clear contrast between Jordan and Tunisia and our other cases with re-
spect to school crowding.
Figure 33.6 examines the presence of doctors per hospital. In Morocco and Tunisia,
virtually all hospitals were public, and the vast majority of doctors were French citizens
who left at independence. As a result, hospitals faced a challenge at independence: how
 Colonial Legacies and Welfare Provision    751

to train an adequate number of doctors while simultaneously expanding access to public

healthcare facilities. Despite achieving independence in the same year (1956), Tunisia
clearly weathered this storm much more effectively than Morocco. By the 1960s, there
were roughly twice as many doctors in Tunisian hospitals than peer institutions in
Morocco. Unlike with school crowding, trends were basically flat in both Jordan and
Egypt, reflecting the relatively limited changes pre-​and post-​independence for coun-
tries that were generally more reliant on private institutions under colonialism.

How Colonial Legacies Shape

Post-​Independence Institutions:
Evidence from Four Countries

To explore how patterns of colonial-​era investment in public welfare contribute to post-​

independence welfare regimes, we present brief qualitative case studies of four coun-
tries: Tunisia, Egypt, Jordan, and Morocco. Collectively, these countries illustrate the
four distinct types of welfare regimes theorized in our typology.

Tunisia
Tunisia is the quintessential example of the “expansive, effective” category. Due to a se-
ries of highly contingent events, private institutions played a limited role in the provision
of healthcare and education under colonialism. Access to public institutions was also
relatively broad, as compared to our other cases. Drawing on a much broader group of
elites than under colonialism, Tunisia’s Neo-​Destour Party was ideologically committed
to the continued expansion of social welfare, especially in the countryside. As a result,
Tunisia was able to simultaneously expand access to public facilities while maintaining
relatively high levels of efficiency for decades after independence.

Colonial Education: From Private to Public

Private institutions played an influential role early in the history of the protectorate, es-
pecially among the Italian population, which was initially much larger than the French
settler presence (Lewis 2014). Rather than risk a direct and potentially costly confronta-
tion with the Catholic Church, French colonial officials dramatically expanded public
institutions (Machuel 1900, 3), which were explicitly secular, confident that access to
public goods would increase the colonial government’s influence among the Italian
community.
Education policy in the Tunisian protectorate was exemplary of the campaign to re-
duce the influence of private institutions. The first law governing public schools stated
that neither nationality nor religion would be an obstacle to enrollment (Ayachi 2003,
752    Melani Cammett et al.

23). Tunisian subjects were the unintended beneficiaries of this competition between the
French colonial state and the Italian community. Enrollment in public schools among
Italians citizens was slow, especially prior to World War I, and vacant spots in public
schools were seized by Muslim subjects. By the 1920s, the colonial government was
ready to assert its authority over private schools. Although private schools were allowed
to persist, they were placed under the direct supervision of the Ministry of Education,
giving colonial government oversight of these previously independent institutions
(Macken 1975).

The Emergence of an Educated Middle Class and

the Neo-​Destour Party
Competition for the loyalty of Tunisia’s large Italian population resulted in a relatively
expansive approach to public education under colonialism. Although there were still
huge discrepancies in rates of education between colonial citizens and subjects (Poncet
1949), the expansion of public schools facilitated the emergence of a small but politi-
cally significant middle class outside of Tunis.18 The leadership of the nationalist Neo-​
Destour movement, which successfully agitated against colonialism and would play a
decisive role in post-​independence politics, and the majority of its key leaders were edu-
cated in colonial public schools (Hermassi 1972, 118).

Independence and the Expansion of the Public Sector

Consistent with our theory, the Tunisian government prioritized the expansion of
public education post-​independence (Tarifa 1971, 150). By 1965, 9 percent of the Tunisian
budget was devoted to public health expenditures and 23 percent to public education.
Elementary school enrollment nearly doubled between 1955, a year prior to indepen-
dence, and 1960. Despite this huge increase, a massive campaign of school building
limited overcrowding. These gains were not limited to public education (Allman 1979,
­chapter 3). Despite losing a huge portion of European doctors at independence, the rela-
tive expansion of public education under colonialism meant that many more Tunisians
were able to pursue medical training in France prior to independence. This enabled
Tunisia to maintain relatively high levels of staffing after independence and gradually
began the process of expanding its network of hospitals. Although the Neo-​Destour
did not eliminate regional inequalities in public infrastructure, the expansion of public
facilities, particularly elementary schools, has justifiably been cited as one of the most
impressive dimensions of the Neo-​Destour’s legacy in post-​independence Tunisia
(Micaud 1964).

Egypt
Egypt’s broad elite coalition after independence pursued policies intended to extend
public education and healthcare to all strata of the Egyptian population. After the Free
Officers’ successful coup in 1952, Presidents Naguib and Nasser pursued an expansive,
 Colonial Legacies and Welfare Provision    753

populist approach to welfare reform, but the quality of these services was diminished by
understaffing and overcrowding of public sector institutions. Consequently, although
access to schooling and medical care became increasingly widespread, they were not
entirely effective.

Colonial Education: The Primacy of Private Education in Khedival

and Colonial Egypt
The modern Egyptian state dates to the early nineteenth century. Under Mohammad
Ali Pasha and his dynasty, the Egyptian state was a pioneer in developing one of the
region’s most advanced public welfare systems. By the late nineteenth century, however,
Khedival policies and those of the new British colonial administration curtailed the de-
velopment of public services. Insufficient investment in public services and the explicit
restriction of education to elites during British occupation (1882–​1952) created deficits
in human capital and physical infrastructure.
Under the British occupation, Egyptian education had three tiers. The first, a nongov-
ernmental system supervised by Al-​Azhar, consisted of religious kutabs that comprised
the majority of schools. The national and foreign sectors comprised the second and third
tiers, and were part of the “modern” system and based on European models of education.
Foreign schools were privately run and catered to European and elite Egyptian families.
After the British occupation of Egypt in 1882, Lord Cromer and the colonial govern-
ment explicitly curbed the expansion of national (public) education (Cromer 1908),
abolishing free primary education and reducing funds to the Ministry of Education. The
education system became an arena of political contention between “nationalists who
saw education as means for national improvement and colonial administrators who saw
it as a privilege for those elites they wished to assimilate into the Egyptian bureaucracy”
(Yousef 2012, 56). Distinctions between fee-​paying elite education in national and for-
eign schools and mass education in kutabs were reinforced by linguistic rules and cur-
riculum mismatches that prevented kutab students from pursuing further education.
The British maintained complete control over the education system until 1922, the
first year the Ministry of Education returned to Egyptian control since the occupation
began. The new administration responded to long-​standing Egyptian demands for ed-
ucation expansion and reform. A law passed in 1923 made primary education free and
mandatory although implementation lagged and primary school enrollment in 1950
only amounted to 36 percent of school-​age children (Faksh 1976).
A similar pattern is visible in colonial effects on the Egyptian healthcare system.
Prior to 1882, it was the norm for Egyptian doctors who graduated from the Qasr Al-​
Aini Medical School to enter into government service, while European graduates
monopolized positions in private practice and private hospitals that served foreigners
and the wealthy (Sonbol 1991, 136). The British protectorate marginalized Egyptian
nationals in the Ministry of Health and subordinated the Egyptian medical system
to colonial developmental goals. The number of Egyptian doctors dwindled when
the British instituted high tuition fees and English-​only instruction at Qasr Al-​Aini
(Sonbol 1991, 131). With fewer Egyptian doctors and negligible government support for
754    Melani Cammett et al.

public hospitals, foreign-​oriented, private institutions became the dominant health-

care providers. The historian Amira El-​Azhary Sonbol (1991) describes why: only three
out of the twenty-​one hospitals in Cairo and Alexandria in 1905 were government-​run.
Private care was too expensive for the majority of the native population, and even the
medical fees charged for public hospital visits were too high for most Egyptians to af-
ford. Before colonialism, these services were free. The ultimate effect of British health
administration in Egypt was to create a privately run medical system oriented toward
serving urban Europeans and to undo the foundations of the nascent, but expansive,
public health network that was curtailed after the British occupation in 1882.

A Broad Post-​Independence Elite Coalition

Egyptian independence from Britain progressed in stages. The British declared the
protectorate dissolved in 1922, and although Egypt achieved de jure independence, the
transition to sovereign rule began with the Free Officers’ 1952 coup that overthrew the
monarchy and traditional elite.19 Although the regime was composed of “a charismatic
leader and military cadre” (Hinnebusch 1981, 442), the Free Officers were both ideolog-
ically and socially diverse. Members espoused a range of nationalist and leftist views
(Botman 1986, 350), and the Executive Committee drew its membership from multiple
strata of Egyptian society, notably the rural lower-​middle class (Abou-​El-​Fadl 2018,
104). The early years of the revolutionary transition were also characterized by an alli-
ance between the ruling military officers and modernizing industrial elements from the
old ruling class (Beinin and Goldberg 1982), and they were successful in mobilizing the
middle classes around a modernizing, populist political agenda.

Independence and an Expansive but Struggling Public Sector

The Free Officer’s inclusive coalition prioritized the expansion of government wel-
fare sectors. The regime built on public education policies inaugurated in 1951 by
Education Minister Taha Hussein, who incorporated kutabs into the public primary
school sector (Saleh 2016). Tertiary education rapidly expanded and was tuition-​free
by presidential decree after 1962 (Faksh 1976). But as Saleh (2016) notes, these reforms
had several deficiencies that effected the overall quality of schooling, including minimal
improvements to physical education infrastructure, few changes to curriculum, and
a surge in enrollment that led to immediate overcrowding. Egypt’s student-​to-​school
ratio remained high by regional standards (500 students per school), with only Morocco
performing worse. Not surprisingly, private schooling remains the preference for the
children of the political and economic elite in Egypt to the present day.
The health sector followed a pattern similar to that in education. The post-​1952 regime
built a network of government-​run urban and rural medical facilities that provided free
treatment (Carney 1984, 135). Even so, the pace of infrastructural expansion outstripped
the sector’s ability to provide quality care. The regime and professional doctors’
associations agreed that newly trained doctors would be obliged to work in the public
sector, but also be allowed to treat patients in private practice and hospitals. At the same
time, the government nationalized foreign hospitals and restricted foreign doctors’
 Colonial Legacies and Welfare Provision    755

practices (Hamdy and Bayoumi 2016). In an attempt to increase provision, graduates

were obliged to spend time in public rural clinics post-​graduation, but this early stage
of a new doctor’s career typically lacked supervision by more senior physicians and
featured low compensation rates. The draw of working in both the public and private
sector led to doctors shirking public clinic hours (Carney 1984), contributing to both
overcrowding and a lack of patients’ confidence in public sector clinics. Even into the
2000s, patients reported a preference for the private sector if they could afford treatment
(Rashad and Sharaf 2015, 1167).
The populist politics of Egypt’s inclusive postcolonial elite coalition led to the expan-
sion of public welfare sectors. The British protectorate undermined the development
of public services, instead relying on private institutions to cater to wealthy and foreign
elites. In attempting to quickly address these lingering inequities, the quality of Egyptian
public services continued to suffer long after independence.

Jordan
Jordanian health and education sectors consistently rank among the best in the
MENA region. The relatively high quality of Jordanian welfare does not stem from
higher government expenditure, as these sectors only receive a modest proportion
of the budget, but rather from a hybrid welfare regime that strategically targeted dis-
tinct constituencies. The political logic of government welfare provision historically
prioritized East Bank Jordanians over the large population of Palestinian refugees and
Jordanians of Palestinian descent. The narrow power elite that grew up around the
Mandatory regime remained heavily entrenched after independence in 1946. Waves of
Palestinian refugees after 1948 might have overwhelmed Jordanian capacities, if not for
the large private sector and UNRWA health and education services for the displaced.
This mixed welfare system freed the regime’s hand to continue prioritizing public wel-
fare for the East Bank population in the first decades post-​independence.

Colonial Welfare: Private Institutions in Mandate Transjordan

Prior to the official establishment of the British Mandate in 1921, the East Bank had few
education and medical services. Those that existed were primarily run by local reli-
gious communities or foreign missionaries. Limited, privately run health and educa-
tion infrastructure, combined with British reluctance to overspend in Jordan, resulted
in limited expansion of public services to rural villages and tribal areas. By 1925, only
forty-​six boys’ schools and six girls’ schools existed, and the country’s only secondary
school in Salt prepared male students for university study abroad (Anderson 2005).
When John Glubb assumed command of the Arab Legion in 1939, the military under-
took infrastructure projects like school and clinic building in sedentary and Bedouin
tribal communities across Transjordan (Alon 2007; Massad 2001). By the time Jordan
gained independence in 1946, private schools still outnumbered public, 100 to 73
(Salman 1986).20
756    Melani Cammett et al.

As in the education sector, healthcare provision prior to the Mandate had been a mix
of traditional, local providers and a limited missionary presence in Salt, Karak, Amman,
and later Ajlun (Amadouny 1997). A push to control malaria in the 1930s reflects the
limited state of public health in Transjordan. Amadouny (1997, 467) notes that health
officials estimated a need for 200 kilograms of quinine to control endemic malaria, but
that due to “financial stringency” only 20 kilograms could be used. Throughout the
1940s, the military continued to provide healthcare by building clinics and overseeing
the Desert Medical Unit, which provided free treatment to the Bedouin population
(Massad 2001).

An Exclusive Elite
Jordan’s ruling elite emerged from its close connections to the colonial state. A foreign
Arab bureaucratic elite, hailing mainly from Greater Syria and Palestine, formed the
core of this narrow coalition. As transplants to Jordan via Ottoman and colonial bu-
reaucratic roles, these men allied with the wealthiest tribal sheikhs and merchants to
support the Hashemite regime. Tell (2013, 134) describes the evolution of a distributive
alliance between this insular power elite and the East Bank steppe as the “Hashemite
Compact.” This narrow social pact extended entitlements (including public services) to
East Bank Bedouins and fallahin (peasants) in exchange for political loyalty. In line with
our theory’s expectations, the historical dominance of the private sector and an explicit
focus on targeted provision to the East Banker population produced restricted but effec-
tive social welfare services.

Independence and a Limited but Effective Public Sector

The exclusive elite coalition that developed during the Mandate remained in power after
Jordan’s independence in 1946. The military continued to expand public social welfare
throughout the rural East Bank, with public health and education reaching most of the
Transjordanian population by 1960 (Tell 2013).
The Nakba and resultant resettlement of Palestinian refugees on the East Bank
constituted a serious challenge that might have overwhelmed Jordan’s limited welfare
capacities. After the declaration of Israeli statehood, the forced depopulation of Palestinian
urban centers and villages resulted in the displacement of hundreds of thousands of
Palestinians. A large proportion of these refugees settled in the East Bank. The Jordanian
state also formally annexed the Palestinian West Bank into its territory in 1950.
Although continuity of the Hashemite regime and narrow elite was by no means a cer-
tainty, the bedrock of the Hashemite compact held firm for several decades. Public sector
welfare continued to target East Bank Jordanians. This singular focus was facilitated by
a strong private sector and the support of the United Nations Relief and Works Agency
(UNRWA). UNRWA managed health and education provision for Palestinian refugees
in Jordanian territory, meaning that the state did not have to allocate extra resources to
provide for newcomers. Wealthy and upper-​class Palestinians who relocated to Amman
or other large towns were able to enroll their children in private schools or receive med-
ical treatment at private hospitals. National statistics that show a dramatic increase in
 Colonial Legacies and Welfare Provision    757

public schooling during the 1950s are inflated because they include West Bank public
schools in their tallies. Our data illustrate the data between 1950 and 1965 are largely due
the incorporation of Palestinian and UNRWA schools and hospitals into the Jordanian
system. The visible drop in Jordan’s national-​level service indicators after 1967 is due to
Israel’s annexation of the West Bank during the Six-​Day War (See Figure 33.1). These
schools and services, no longer under Jordanian jurisdiction, were removed from the
national accounts.
In sum, the trajectory of Jordanian welfare shows a long-​term reliance on nonstate and
private-​sector service provision. The Nakba and the 1967 War shaped social welfare ser-
vice provision in several ways. UNRWA’s involvement in the education and health sectors
alleviated some of the demographic pressure facing Jordan’s welfare state. East and West
Bank elites continued to rely on private schools and medical facilities, leaving the expan-
sion of the government sector to cater to the core supporters of the Hashemite regime.

Morocco
Morocco provides an illustration of the “restrictive, ineffective” category. Private
institutions were of limited importance in Morocco under colonialism, but, unlike
Tunisia, access to public institutions for Moroccan subjects was limited throughout the
protectorate period. The absence of a broadened political coalition at independence
meant that there was little pressure or incentive to expand access to public facilities be-
yond the pace deemed acceptable to the palace. While access to public facilities nomi-
nally expanded at independence, glaring inequalities, particularly between urban areas
and the countryside, characterize the provision of education and health services to the
present day. As one observer noted in 2001, “60 percent of rural households have no ac-
cess to paved roads; more than 87 percent have no access to health care; and 77 percent
do not have electricity” (Maghraoui 2001, g. 77).

Colonial Education: A Stunted Public Sector

Morocco’s first Resident Generale, Hubert Lyautey, who governed from 1912 to 1925,
adopted a deliberately conservative approach to public education (Segalla 2009,
­chapter 1). While the sons and daughters of European citizens enjoyed access to primary
and secondary schools that were generally equivalent to those in the metropole, access
to basic education among subject populations was limited almost exclusively to urban
boys, and access to secondary education was virtually nonexistent. The exception to
this generalization was the Ecoles des Fils Notables, a small network of twelve secondary
schools, designed to educate elite Moroccans who would go on to take key roles in the
colonial government. Not only did these schools use an explicitly elitist criteria to de-
termine eligibility for admission, but fees were charged to further dissuade the “wrong”
kinds of students from being admitted (Knibiehler 1994).21
This conservative approach meant that while a network of public schools and hospitals
was gradually developed, these institutions were overwhelmingly concentrated in the
758    Melani Cammett et al.

urban areas with the highest concentrations of Europeans. Despite the fact that the vast
majority of the Moroccan subject population was rural, public provision of basic health
and schooling in these areas was extremely limited until independence (Bidwell 1973,
248–​249).

Elite Continuity
While Morocco’s main political party, the Istiqlal (Independence) Party, was successful
in agitating against the colonial government (Ashford 1961), it was effectively sidelined
by the monarchy shortly after independence (Waterbury 1970). The marginalization
of the Istiqlal ensured the dominance of a narrow coalition of elite families, known in
Morocco collectively as the makhzen (Hermassi 1972, c­ hapter 5), whose importance
persists to the present day (Liddell 2010).22 This narrow elite and the large state bureauc-
racy under its control meant that the elite coalition in Morocco was characterized by
remarkable consistency post-​independence, with a large number of families from the
old imperial capital of Fez holding enormous influence, particularly in the first decades
after independence.

Independence and a Limited Public Sector

The Moroccan government post-​independence could avail itself of a relatively devel-
oped network of brick and mortar institutions, yet public infrastructure was generally
limited to Morocco’s largest cities. While spending levels were relatively high as a per-
centage of the annual budget, demand for public goods far exceeded the government’s
ability to provide them, especially in the countryside. The result was a stunted public
system that failed to meet the needs of the inhabitants of Morocco’s largest cities and
towns, and marginalized entirely those born in rural areas for decades after indepen-
dence (Wagner 1993).
Morocco’s post-​independence experience provides a mirror image to Tunisia’s. Like
Tunisia, Morocco lost the vast majority of its teachers and doctors at independence, as
they returned to Europe. Unlike Tunisia, low levels of public education during the colo-
nial period meant that it took significantly longer to train the white collar professionals
who were essential for meeting the surging demand for social welfare services after in-
dependence. The combination of de jure expansion of access to social welfare services,
surging population growth, and limited expansion of public facilities and staff resulted
in a de facto decline in social welfare provision, as demand for public services far
exceeded supply, even in urban areas.

Discussion
These cases provide descriptive evidence supporting our theory. The end of coloni-
alism provided post-​independence governments with the opportunity to expand the
health and education sectors. Only in cases where the ruling social coalition was rela-
tively inclusive (Egypt, Tunisia) do we see a broad expansion of public sector services;
 Colonial Legacies and Welfare Provision    759

in countries with narrow political elite coalitions (Jordan, Morocco) welfare provision
was limited, particularly in the public sector. A second key insight is that, while the
scope of government intervention in the decade after independence is “sticky,” coun-
tries varied in their ability to provide and maintain public health and education services
to match this scope, in part as a result of the public-​private mix of inherited welfare in-
frastructure. In countries relying primarily on private schools prior to independence,
but where elite coalitions broadened after independence, as in Egypt, the expansion of
public institutions was dramatic but was almost totally overwhelmed three decades after
independence, thanks to rapid population growth and rural to urban migration, which
placed enormous strains on the public system.23 While some of the burden was partially
addressed by the expansion of the private sector, these countries have struggled to main-
tain standards and equity of access to services, given exponential growth in demand for
these services. In other countries, like Tunisia, a broad social coalition and a large net-
work of public infrastructure inherited at independence facilitated dramatic expansions
at independence, while keeping pace with surging population growth and a huge in-
crease in demand for public goods.

Conclusion

Our analysis seeks to understand why welfare provision varies across the MENA region.
We argue that historical processes of state-​building under colonial rule contributed to
post-​independence welfare expansion efforts in two ways. First, the breadth of the po-
litical elite in the post-​independence period as well as colonial legacies of the public-​
private mix of health and education institutions impacted the ability and will of
governments to implement broader welfare policies. Politically, the end of colonialism
provided post-​independence governments with the opportunity to expand or maintain
colonial-​era welfare policies.
Second, our findings suggest that colonial legacies appear to have differential effects
when we disaggregate welfare policies by sector. The efficacy and expansiveness of
healthcare provision remained “sticky” across the transition from the colonial to post-
colonial eras, whereas education outcomes were more variable after independence. This
variation in sectoral legacies clearly needs further exploration. One potential explana-
tion, and an area of future research, could be that education expansion was a central
component of ruling coalitions’ political projects after independence. Education had
mass appeal, was less costly to implement, and brought more citizens into regular con-
tact with the state. Education also gave regimes the opportunity to promulgate new na-
tional narratives, and thus served the interests of ruling elites (Paglayan 2017).
This theory provides a schema for thinking about the political forces that shaped wel-
fare promotion in the crucial decades after independence, but it leaves a number of key
questions unaddressed. First among these is accounting for local variation in welfare
promotion. A growing body of literature notes stark and persistent inequalities across
760    Melani Cammett et al.

the region (Abid, O’Donoghue and Sologon 2016; Belhaj Hassine 2012; Hassine 2015),
particularly in terms of education (Krafft and Alawode 2016), even in “successful”
cases of state-​building like Tunisia (Trabelsi 2013). We test our theory’s implications for
this local variation in ongoing work that examines education outcomes in Jordan and
Tunisia.

Acknowledgments
Authors listed alphabetically. Thanks to all participants at the January OUP book conference
for their feedback. Research assistance for the MENAHDA database project, from which
we draw the data for the chapter, was provided by Yara Abuljadayel, Maha al-​Suwaidi, Steve
Bai, Esat Bayar, Robert Erikson, Amir Hamilton, Elizabeth Herington, Samantha Heyward,
Sean Hwang, Mouad Ibno Bachir, Hossam Mabed, Emily Markowitz, Julia Martin, Shubhan
Nagendra, Talia Orphee, Claire Parker, Brendan Powell, Reshini Premaratne, Heide Rogers,
Lisa Russo, Hannah Sairah, Lia Frances Swiniarski, and Lydia Tahraoui.

Notes
1 An important body of literature examines the relative decline of the region to Europe
and traces this “reversal of fortune” to the Middle Ages (Blaydes and Chaney 2013; Kuran
2011; Kuru 2019; Rubin 2017). While this is an important argument for understanding
differences in cross-​regional variation in development, these arguments are less applicable
to explaining variation within the MENA region, which is our focus.
2 The project focuses initially on the first several decades after most countries in the region
achieved independence, when economic models and welfare regimes were institutional-
ized, and prior to the fiscal crises that most MENA countries faced in the 1980s and 1990s.
We adopt this periodization because the politics of economic retrenchment are arguably
distinct (Hacker 2004; Haggard and Kaufman 2009; Pierson 2000), and because the re-
sponse to these crises depended on the prior establishment of economic and social policies
and institutions.
3 Major overviews of the economic history of the Middle East that cover the colonial period
include: Batatu (1978); Beinin (2001); Hourani (1968); Inalcik (1978); İslamoğlu İnan (1987);
Kazemi and Waterbury (1991); Keyder and Tabak (1991); Lewis (2014); Monroe (1981);
Owen (1981, 1992); Owen and Pamuk (1999); Quataert (2005); Udovitch (1981).
4 An important but distinct literature examines the role of organized labor with respect to
protests and demonstrations especially after 2010 (Beinin 2015; Berman 2020).
5 While the influence of the public sector has declined across the region in response to
shifting economic fortunes, retrenchment, and the influence of the Washington Consensus,
public institutions continue to play a huge role in the provision of social welfare across the
region (Eibl 2020).
6 This reliance was not simply explained by the identity of the colonial power. As the
anecdotes above suggest, the French relied much more heavily on private institutions in
Syria and Lebanon than they did in North Africa.
7 See Saleh (2016) for an account of these institutions in Egypt post-​independence.
 Colonial Legacies and Welfare Provision    761

8 The same holds for missionary schools, which were often nominally regulated by the state
but practically enjoyed significant leeway in their day-​to-​day operations and funding.
9 Initial political settlements may create lock-​in effects with longer-​term consequences.
Since the 1980s and 1990s, when many MENA first experienced fiscal crises, reform
programs have faced serious opposition because popular constituencies for post-​
independence policies had become entrenched.
10 Eibl’s cross-​national comparative study of MENA welfare spending is restricted to the
postcolonial period. We build on his work by operationalizing the character of elite
coalitions quantitatively and allow for historical contingencies by considering the nature
of welfare sectors prior to independence.
11 See Hammoudi (1997) for a classic account in this vein in Morocco.
12 Includes all cabinets from 1956 to 1970.
13 Includes all cabinets from 1955 to 1970.
14 Includes all cabinets from 1952 to 1970. For United Arab Republic cabinets, only Egyptian
ministers are considered.
15 Includes all cabinets from 1946 to 1970. Foreign-​born members of the colonial-​era
Hashemite elite who carry over into the post-​independence period constitute the majority
of ministers in most cabinets well into the 1950s and are coded as being from Amman.
Post-​independence Palestinian ministers who are not from this colonial elite are coded as
being from Palestine.
16 The MENAHDA website, which the co-​authors are developing, will include the original
scans of the statistical yearbooks, host data sets composed of transcribed MENAHDA
data, highlight research that integrates these data, and will be publicly accessible.
17 The post-​independence period was a time of rapid population growth across all of our cases.
However, our cases differed significantly both in terms of the overall size of the population
and rates of growth. Egypt is by far the largest country among our cases. By 1960, Egypt’s
26 million inhabitants were more than double Morocco’s population, six times Tunisia’s pop-
ulation, and more than twenty-​six times Jordan’s population. In order to compare across
cases, we standardize our measures of expansiveness by the overall size of the population.
18 While our focus in this chapter is on primary education, historians justifiably emphasize
the important role played by higher education, like the famous Sadiqi College, which was
established prior to the colonial period (Sraieb 1993). For a discussion of higher education
post-​independence see Micaud (1964, 150).
19 The British did not completely exit from Egyptian affairs until 1956.
20 The public schools served more students than did private schools, however: 9,874 students
attended government-​run schools, while 6,472 attended private schools.
21 Lyautey vehemently opposed missionary schools (Bidwell 1973, 54–​55), successfully lim-
iting private religious schools to a handful of Catholic institutions in Morocco’s largest
cities.
22 While general elections were held in 1963 and 1970, Morocco’s monarch, Hassan II,
dissolved parliament in 1965 and again in 1972, and it was not until the elections of 1984
that Morocco’s elected representatives would actually complete their term in parliament,
and not until 1996 that all members of Morocco’s lower parliament would be directly
elected (Inter-​Parliamentary Union 1970, 1978; White 1997).
23 An inability to meet demand for public goods is not just a challenge for populous countries
like Egypt. Lebanon’s population was comparable to Jordan’s for most of the period under
consideration, but lagged behind Jordan in terms of both health and education outcomes.
762    Melani Cammett et al.

References
Abid, Yosr, Cathal O’Donoghue, and Denisa Sologon. 2016. “Decomposing Welfare Inequality
in Egypt and Tunisia: An Oaxaca-​ Blinder Based Approach.” Working paper. Giza,
Egypt: Economic Research Forum.
Abou-​El-​Fadl, Reem. 2018. Foreign Policy as Nation Making: Turkey and Egypt in the Cold War.
Cambridge: Cambridge University Press.
Acemoglu, Daron, Francisco A. Gallego, and James A. Robinson. 2014. “Institutions, Human
Capital, and Development.” Annual Review of Economics 6, no. 1: 875–​912.
Albertus, Michael. 2015. Autocracy and Redistribution. New York: Cambridge University Press.
Alesina, Alberto, and Edward L. Glaeser. 2004. Fighting Poverty in the US and Europe: A World
of Difference. New York: Oxford University Press.
Allman, James. 1979. Social Mobility, Education and Development in Tunisia. Leiden: Brill.
Alon, Yoav. 2007. The Making of Jordan: Tribes, Colonialism and the Modern State. London: I.B.
Tauris.
Amadouny, Vartan M. 1997. “The Campaign against Malaria in Transjordan, 1926–​1946:
Epidemiology, Geography, and Politics.” Journal of the History of Medicine and Allied
Sciences 52, no. 4: 453–​484.
Amster, Ellen J. 2014. Medicine and the Saints: Science, Islam, and the Colonial Encounter in
Morocco. Austin: University of Texas Press.
Anderson, Benedict. 1983. Imagined Communities: Reflections on the Origin and Spread of
Nationalism. New York: Verso.
Anderson, Betty S. 2005. Nationalist Voices in Jordan: The Street and the State. Austin: University
of Texas Press.
Arriola, Leonardo R. 2009. “Patronage and Political Stability in Africa.” Comparative Political
Studies 42, no. 10: 1339–​1362.
Arriola, Leonardo R., and Martha C. Johnson. 2014. “Ethnic Politics and Women’s Empower­
ment in Africa: Ministerial Appointments to Executive Cabinets.” American Journal of
Political Science 58, no. 2: 495–​510.
Ashford, Douglas Elliott. 1961. Political Change in Morocco. Princeton: Princeton University Press.
Ayachi, Mokhtar. 2003. Ecoles et société en Tunisie. Tunis: Centre de CERES.
Baldwin, Kate, and John D Huber. 2010. “Economic versus Cultural Differences: Forms of
Ethnic Diversity and Public Goods Provision.” American Political Science Review 104, no.
4: 644–​662.
Batatu, Hanna. 1978. The Old Social Classes and the Revolutionary Movements of Iraq: A Study of
Iraq’s Old Landed and Commercial Classes and of Its Communists, Ba’thists, and Free Officers.
Princeton: Princeton University Press.
Beinin, Joel. 2001. Workers and Peasants in the Modern Middle East. New York: Cambridge
University Press.
Beinin, Joel. 2015. Workers and Thieves: Labor Movements and Popular Uprisings in Tunisia and
Egypt. Palo Alto, CA: Stanford University Press.
Beinin, Joel, and Ellis Goldberg. 1982. “Egypt’s Transition under Nasser Reviewed.” MERIP
Reports 107: 23–​26.
Belhaj Hassine, Nadia. 2012. “Inequality of Opportunity in Egypt.” World Bank Economic
Review 26, no. 2: 265–​295.
Bellin, Eva. 2011. Stalled Democracy: Capital, Labor, and the Paradox of State-​Sponsored
Development. Ithaca, NY: Cornell University Press.
 Colonial Legacies and Welfare Provision    763

Berman, Chantal E. 2020. “Policing the Organizational Threat in Morocco: Protest and Public
Violence in Liberal Autocracies.” American Journal of Political Science. Published ahead of
print, December 15, 2020.
Bidwell, Robin. 1973. Morocco under Colonial Rule: Administration of Tribal Areas 1912–​1956.
New York: Cass Publishers.
Blaydes, Lisa, and Eric Chaney. 2013. “The Feudal Revolution and Europe’s Rise: Political
Divergence of the Christian West and the Muslim World before 1500 CE.” American Political
Science Review 107, no. 1: 16–​34.
Botman, Selma. 1986. “Egyptian Communists and the Free Officers: 1950–​54.” Middle Eastern
Studies 22, no. 3: 350–​366.
Buehler, Matt, and Mehdi Ayari. 2018. “The Autocrat’s Advisors: Opening the Black Box of
Ruling Coalitions in Tunisia’s Authoritarian Regime.” Political Research Quarterly 71, no.
2: 330–​346.
Bueno de Mesquita, Bruce, Alastair Smith, Randolph M. Siverson, and James D. Morrow. 2003.
The Logic of Political Survival. Cambridge, MA: MIT Press.
Cammett, Melani. 2007. Globalization and Business Politics in Arab North Africa. New York:
Cambridge University Press.
Cammett, Melani. 2014. Compassionate Communalism: Welfare and Sectarianism in Lebanon.
Ithaca, NY: Cornell University Press.
Cammett, Melani, and Nisreen Salti. 2016. “Popular Grievances in the Arab Region: Evaluating
Explanations for Discontent in the Lead-​Up to the Uprisings.” Middle East Development
Journal 10, no. 1: 64–​96.
Cammett, Melani, and Aytug Sasmaz. 2016. “Social Policy in Developing Countries.” In The
Oxford Handbook of Historical Institutionalism, edited by Orfeo Fioretos, Tulia Gabriela
Falleti, and Adam Sheingate, 239–​255. Oxford: Oxford University Press.
Carney, Kim. 1984. “Health in Egypt.” Journal of Public Health Policy 5, no. 1: 131–​142.
Charrad, Mounira. 2001. States and Women’s Rights: the Making of Postcolonial Tunisia, Algeria
and Morocco. Berkeley: University of California Press.
Cromer, Evelyn Baring. 1908. Modern Egypt. New York: Macmillan.
Eibl, Ferdinand. 2020. Social Dictatorships: The Political Economy of the Welfare State in the
Middle East and North Africa. Oxford: Oxford University Press.
Esping-​Anderson, Gøsta. 1990. The Three Worlds of Welfare Capitalism. Cambridge: Polity Press.
Faksh, M. A. 1976. “An Historical Survey of the Educational System in Egypt.” International
Review of Education 22, no. 2: 234–​244.
Falb Kalisman, Hilary. 2017. “Education Policy in Iraq: Competing Visions of the State.” In State
and Society in Iraq: Citizenship under Occupation, Dictatorship and Democratization, edited
by Benjamin Iskahan, Shamiran Mako and Fadi Dawood, 90–​108. London: I.B. Tauris.
Franck, Raphaël, and Ilia Rainer. 2012. “Does the Leader’s Ethnicity Matter? Ethnic Favoritism,
Education, and Health in Sub-​Saharan Africa.” American Political Science Review 106, no.
2: 294–​325.
Gallagher, Nancy Elizabeth. 1983. Medicine and Power in Tunisia, 1780–​1900. New York:
Cambridge University Press.
Gallagher, Nancy Elizabeth. 1990. Egypt’s Other Wars. Syracuse, NY: Syracuse University Press.
Gellner, Ernest. 1983. Nations and Nationalism. Ithaca, NY: Cornell University Press.
Hacker, Jacob S. 2004. “Privatizing Risk without Privatizing the Welfare State: The Hidden
Politics of Social Policy Retrenchment in the United States.” American Political Science
Review 98, no. 2: 243–​260.
764    Melani Cammett et al.

Haggard, Stephan, and Robert R. Kaufman. 2009. “How Regions Differ.” Journal of Democracy
20, no. 4: 64–​78.
Hamdy, Sherine F., and Soha Bayoumi. 2016. “Egypt’s Popular Uprising and the Stakes of
Medical Neutrality.” Culture, Medicine and Psychiatry 40, no. 2: 223–​241.
Hammoudi, Abdellah. 1997. Master and Disciple: the Cultural Foundations of Moroccan
Authoritarianism. Chicago: University of Chicago Press.
Hartnett, Allison Spencer. 2019. “Redistributive Authoritarianism: Regime Durability and
Land Reform in the Middle East and North Africa.” PhD diss., University of Oxford.
Hartnett, Allison Spencer. 2020. “Selectorate Dynamics in Autocracies: Colonial and Post-​
colonial Minister Appointments in the Middle East and North Africa.” Working paper.
Hassine, Nadia Belhaj. 2015. “Economic Inequality in the Arab Region.” World Development
66: 532–​556.
Hermassi, Elbaki. 1972. Leadership and National Development in North Africa: A Comparative
Perspective. Berkeley: University of California Press.
Hinnebusch, Raymond A. 1981. “Egypt under Sadat: Elites, Power Structure, and Political
Change in a Post-​populist State.” Social Problems 28, no. 4: 442–​464.
Hodler, Roland, and Paul A Raschky. 2014. “Regional Favoritism.” Quarterly Journal of
Economics 129, no. 2: 995–​1033.
Hourani, Albert. 1968. “Ottoman Reform and the Politics of the Notables.” In Beginnings of
Modernization in the Middle East, edited by William Polk and Richard L. Chambers, 41–​68.
Chicago: University of Chicago Press.
Hyde, Georgie D. M. 1978. Education in Modern Egypt: Ideals and Realities. London: Routledge.
Inalcik, Halil. 1978. The Ottoman Empire: Conquest, Organization and Economy. London:
Variorum Reprints.
Inter-​Parliamentary Union. 1970. Morocco, General Election, 1970. Technical report. http://​ar-
chive.ipu.org/​parline-​e/​reports/​arc/​MOROCCO_​1970_​E.PDF
Inter-​Parliamentary Union. 1978. Morocco General Elections, 1977. Technical report. http://​ar-
chive.ipu.org/​parline-​e/​reports/​arc/​MOROCCO_​1977_​E.PDF.
İslamoğlu İnan, Huri, ed. 1987. The Ottoman Empire and the World Economy. New York:
SUNY Press.
Issawi, Charles, ed. 1995. The Middle East Economy: Decline and Recovery. Princeton: Markus
Wiener.
Kazemi, Farhad, and John Waterbury, eds. 1991. Peasants and Politics in the Modern Middle
East. Miami: Florida International University Press.
Keyder, Çağlar, and Faruk Tabak, eds. 1991. Landholding and Commercial Agriculture in the
Middle East. Albany: State University of New York Press.
Khan, Mushtaq. 2010. “Political Settlements and the Governance of Growth-​Enhancing
Institutions.” Research Series on Governance for Growth. London: SOAS, University of
London.
Knibiehler, Yvonne. 1994. “L’enseignement au Maroc pendant le protectorat (1912–​1956): Les
fils de notables.” Revue d’histoire moderne et contemporaine 41, no. 3: 489–​498.
Koter, Dominika. 2016. Beyond Ethnic Politics in Africa. New York: Cambridge University Press.
Krafft, Caroline, and Halimat Alawode. 2016. “Inequality of Opportunity in Higher Education
in the Middle East and North Africa.” International Journal of Educational Development
62: 234–​244.
Kramon, Eric, and Daniel N. Posner. 2016. “Ethnic Favoritism in Education in Kenya.”
Quarterly Journal of Political Science 11, no. 1: 1–​58.
 Colonial Legacies and Welfare Provision    765

Kuran, Timur. 2011. The Long Divergence: How Islamic Law Held Back the Middle East.
Princeton: Princeton University Press.
Kuru, Ahmet T. 2019. Islam, Authoritarianism, and Underdevelopment: A Global and Historical
Comparison. New York: Cambridge University Press.
Lewis, Mary Dewhurst. 2014. Divided Rule: Sovereignty and Empire in French Tunisia. Los
Angeles: University of California Press.
Liddell, James. 2010. “Notables, Clientelism and the Politics of Change in Morocco.” Journal of
North African Studies 15, no. 3: 315–​331.
Machuel, Louis. 1900. L’enseignement public en Tunisie. Tunis: Impremerie Rapide.
Macken, Richard. 1975. “Louis Machuel and Educational Reform in Tunisia during the Early
Years of the French Protectorate.” Revue d’Histoire Maghrebine 3: 45–​55.
Maghraoui, Abdeslam M. 2001. “Monarchy and Political Reform in Morocco.” Journal of
Democracy 12, no. 1: 73–​86.
Malik, Adeel, and Bassem Awadallah. 2013. “The Economics of the Arab Spring.” World
Development 45: 296–​313.
Mares, Isabela. 2003. The Politics of Social Risk: Business and Welfare State Development.
New York: Cambridge University Press.
Massad, Joseph. 2001. Colonial Effects: The Making of Modern Jordan. New York: Columbia
University Press.
Micaud, C. A. 1964. Tunisia: The Politics of Modernization. London: Pall Mall.
Miguel, Edward. 2004. “Tribe or Nation? Nation Building and Public Goods in Kenya versus
Tanzania.” World Politics 56, no. 3: 327–​362.
Miguel, Edward, and Mary Kay Gugerty. 2005. “Ethnic Diversity, Social Sanctions, and Public
Goods in Kenya.” Journal of Public Economics 89, no. 11–​12: 2325–​2368.
Monroe, Elizabeth. 1981. Britain’s Moment in the Middle East, 1914–​1971. Baltimore: Johns
Hopkins University Press.
Niedzwiecki, Sara. 2018. Uneven Social Policies: The Politics of Subnational Variation in Latin
America. New York: Cambridge University Press.
Noland, Marcus, and Howard Pack. 2007. The Arab Economies in a Changing World.
Washington, DC: Peterson Institute for International Economics.
Nunn, Nathan. 2014. “Gender and Missionary Influence in Colonial Africa.” In Africa’s
Development in Historical Perspective, edited by Emmanuel Akyeampong, Robert H. Bates,
Nathan Nunn, and James Robinson, 489–​512. New York: Cambridge University Press.
Nyrup, Jacob, and Stuart Bramwell. 2020. “Who Governs? A New Global Dataset on Members
of Cabinets.” American Political Science Review 114, no. 4: 1366–​1374.
Owen, Roger. 1981. The Middle East in the World Economy: 1800–​1914. New York: Methuen.
Owen, Roger. 1992. State, Power and Politics in the Making of the Modern Middle East.
New York: Routledge.
Owen, Roger, and Sevket Pamuk. 1999. A History of Middle East Economies in the Twentieth
Century. Cambridge, MA: Harvard University Press.
Paglayan, Agustina. 2017. “Political Origins of Public Education Systems.” PhD diss., Stanford
University.
Paglayan, Agustina. 2020. “The Non-​democratic Roots of Mass Education: Evidence from 200
Years.” American Political Science Review 115, no. 1: 179–​198.
Pepinsky, Thomas B. 2009. Economic Crises and the Breakdown of Authoritarian Regimes:
Indonesia and Malaysia in Comparative Perspective. New York: Cambridge University
Press.
766    Melani Cammett et al.

Pierson, Paul. 2000. “Increasing Returns, Path Dependence, and the Study of Politics.”
American Political Science Review 94, no. 2: 251–​267.
Poncet, Jean. 1949. “La scolarisation de la Tunisie et le milieu social.” Régence de Tunis,
Protectorat Français—​Semaine pédagogique, organisée par la direction de l’instruction
publique en Tunisie, Tunis 19–​23, Avril 1949; P 158–​170. Imprimerie officielle, Tunis.
Quataert, Donald. 2005. The Ottoman Empire, 1700–​1922. New York: Cambridge University
Press.
Rashad, Ahmed, and Mesbah Sharaf. 2015. “Who Benefits from Public Healthcare Subsidies in
Egypt?” Social Sciences 4, no. 4: 1162–​1176.
Ricart-​Huguet, Joan. 2020. “Colonial Education, Political Elites, and Regional Political
Inequality in Africa.” Comparative Political Studies, 2021.
Rougier, Eric. 2016. “‘Fire in Cairo’: Authoritarian-​Redistributive Social Contracts, Structural
Change, and the Arab Spring.” World Development 78: 148–​171.
Rubin, Jared. 2017. Rulers, Religion, and Riches: Why the West Got Rich and the Middle East Did
Not. New York: Cambridge University Press.
Saleh, Mohamed. 2016. “Public Mass Modern Education, Religion, and Human Capital in
Twentieth-​Century Egypt.” Journal of Economic History 76, no. 3: 697–​735.
Salman, Amin Mohamed Hasan. 1986. “Some Problems in the Educational System of Jordan.”
PhD diss., University College London.
Segalla, Spencer D. 2009. The Moroccan Soul: French Education, Colonial Ethnology, and
Muslim Resistance, 1912–​1956. Omaha: University of Nebraska Press.
Singh, Prerna. 2016. How Solidarity Works for Welfare: Subnationalism and Social Development
in India. New York: Cambridge University Press.
Singh, Prerna, and Matthias vom Hau. 2016. “Ethnicity in Time: Politics, History, and the
Relationship between Ethnic Diversity and Public Goods Provision.” Comparative Political
Studies 49, no. 10: 1303–​1340.
Slater, Dan. 2010. Ordering Power: Contentious Politics and Authoritarian Leviathans in
Southeast Asia. New York: Cambridge University Press.
Slater, Dan, and Sofia Fenner. 2011. “State Power and Staying Power: Infrastructural Mechanisms
and Authoritarian Durability.” Journal of International Affairs 65, no. 1: 15–​29.
Soifer, Hillel David. 2016. “Regionalism, Ethnic Diversity, and Variation in Public Good
Provision by National States.” Comparative Political Studies 49, no. 10: 1341–​1371.
Sonbol, Amira El-​Azhary. 1991. The Creation of a Medical Profession in Egypt, 1800–​1922.
Syracuse, NY: Syracuse University Press.
Sraieb, Noureddine. 1993. “L’idéologie de l’école en Tunisie coloniale (1881–​1945).” Revue du
monde musulman et de la Méditerranée 68–​69: 239–​254.
Szyliowicz, Joseph S. 1973. Education and Modernization in the Middle East. Ithaca, NY: Cornell
University Press.
Tajima, Yuhki, Krislert Samphantharak, and Kai Ostwald. 2018. “Ethnic Segregation and Public
Goods: Evidence from Indonesia.” American Political Science Review 112, no. 3: 637–​653.
Tarifa, Chadli. 1971. “L’enseignement du 1er et du 2e degré en Tunisie.” Population 26: 149–​180.
Teichman, Judith A. 2016. The Politics of Inclusive Development Policy, State Capacity, and
Coalition Building. New York: Palgrave Macmillan.
Tell, Tariq. 2013. The Social and Economic Origins of Monarchy in Jordan. New York: Palgrave
Macmillan.
Trabelsi, Salwa. 2013. “Regional Inequality of Education in Tunisia: An Evaluation by the Gini
Index.” Region et Developpement 37: 95–​117.
 Colonial Legacies and Welfare Provision    767

Udovitch, Abraham L., ed. 1981. The Islamic Middle East, 700–​1900: Studies in Economic and
Social History. Princeton: Princeton University Press.
Valencia Caciedo, Felipe. 2019. “The Mission: Human Capital Transmission, Economic
Persistence and Culture in South America.” Quarterly Journal of Economics 134, no. 1: 507–​556.
Wagner, Daniel A. 1993. Literacy, Culture and Development: Becoming Literate in Morocco.
New York: Cambridge University Press.
Waldner, David. 1999. State Building and Late Development. Ithaca, NY: Cornell University
Press.
Wantchekon, Leonard, Marko Klasnja, and Natalija Novta. 2015. “Education and Human
Capital Externalities: Evidence from Colonial Benin.” Quarterly Journal of Economics 130,
no. 2: 703–​757.
Waterbury, John. 1970. The Commander of the Faithful: The Moroccan Political Elite; A Study of
Segmented Politics. London: Weidenfeld & Nicolson.
Weber, Eugene. 1976. Peasants into Frenchmen: the Modernization of Rural France, 1870–​1914.
Stanford, CA: Stanford University Press.
White, Gregory. 1997. “The Advent of Electoral Democracy in Morocco? The Referendum of
1996.” Middle East Journal 51, no. 3: 388–​404.
Wimmer, Andreas. 2016. “Is Diversity Detrimental? Ethnic Fractionalization, Public Goods
Provision, and the Historical Legacies of Stateness.” Comparative Political Studies 49, no.
11: 1407–​1445.
Yousef, Hoda A. 2012. “Seeking the Educational Cure: Egypt and European Education,
1805–​1920s.” European Education 44, no. 4: 51–​66.
 Chapter 37

W elfare Stat e s i n
the Middle E ast

Ferdinand Eibl

Social policies and welfare provision have been a cornerstone of government activity
across the Middle East and North Africa (MENA) and have been the focus of a growing
body of literature advanced by area specialists (e.g., Brooke 2017; Cammett and Issar
2010; Cammett 2014a; Loewe 2010).1 Global social policy research, by contrast, has thus
far largely bypassed the region. To cite just one example, there is no entry on the Middle
East in the Oxford Handbook of the Welfare State (Castles et al. 2010), while other re-
gions, such as Asia and Latin America, are covered. This relative neglect is particularly
surprising given the scope and fiscal weight of public welfare provision in the region. On
average, social security systems cover six policy areas across the region, with the largest
programs being health and old-​age pensions, which puts the region roughly on par with
Latin America and only behind one region, post-​Communist Central Asia. And by
committing about 8 percent of GDP annually to public welfare expenditures, the MENA
region counts among the regional high spenders (see Figure 37.1).
One of the reasons why the global welfare state literature has often circumvented
the region may be the tremendous diversity of welfare provision that underlies these
regional averages in the Middle East, which defies easy categorizations in terms of re-
gional trends and makes general statements on welfare provision in MENA analyt-
ically challenging. In a region that includes some of the world’s most affluent (e.g.,
Qatar) and poorest (e.g., Sudan) countries, it is maybe unsurprising that social policies
do not follow a uniform pattern. Indeed, just as Rauch and Kostyshak (2009) speak of
“three Arab worlds” with regard to economic development, three distinct types of wel-
fare states can be found across the region (Eibl 2020):2 (1) Broad and generous welfare
states are characterized by consistently high levels of social spending, providing wel-
fare to a broad cross-​section of the population. Arab Gulf States fall into this category
of authoritarian welfare states, alongside Algeria, Tunisia, and the Islamic Republic of
Iran. For this category to make sense, generous spending needs to be considered in rela-
tion to income levels. In the mid-​2000s, Tunisia and Kuwait spent about the same share
770   Ferdinand Eibl

(a) 8

6
social security areas covered

0
CAC EA LAC MENA SA SEA SSA

10
(b)

8
total social expenditures as % GDP

0
CAC EA LAC MENA SA SEA SSA

Figure 37.1: Areas of social security and social expenditures in regional comparison.
Note: CAC: Central Asia and Caucasus; EA: East Asia; LAC: Latin America and Caribbean;
MENA: Middle East and North Africa; SA: South Asia; SEA: Southeast Asia; SSA: Sub-​Saharan
Africa. Areas of social security refer to distinct risks covered by social security, such as health, old
age, work accidents, etc.
Source: ILO 2014.

of goods and services on social welfare (15 percent of GDP), yet per capita spending
in constant US dollars was more than double in Kuwait than in Tunisia ($2,900 and
$1,400, respectively). Putting both countries in the same category only makes sense then
if we understand generosity in relative, not in absolute, terms. (2) Broad welfare states
 Welfare States in the Middle East     771

provide social policies that are widely, often universally, accessible, but fail to underpin
access with a sufficient stream of financial resources. Egypt and, to some extent, Syria
fit into this pattern. (3) Segmented and/​or minimal welfare states are characterized by
public welfare provision targeted to core constituencies only, with historically low levels
of social spending and an important role for private sector providers. Morocco, Jordan,
and Lebanon, among others, fall into this category. While there has been some conver-
gence in welfare provision across the region, this division has on the whole been highly
persistent.
This chapter seeks to give an overview of these different patterns of welfare provision
and their political economy underpinnings. The next section fleshes out the three types
empirically and highlights the importance of authoritarian coalitions and geostrategic
context as the key drivers of social policies in MENA. The following section looks in
greater detail at the effects of war and conflict on MENA welfare states and introduces
the concept of “cheap social policies,” using Egypt as an example. The conclusion sums
up the main arguments and reflects on future research agendas.

Welfare Provision in MENA and

Its Coalitional Origins

Broadly speaking, the emergence of the three types of welfare states described above
occurred from the early 1950s until the 1970s as a number of countries rapidly expanded
social policies to a growing number of beneficiaries, while others were eager to restrict
the scope of public policies to core constituencies of nascent regimes (e.g., military, bu-
reaucracy), yielding more segmented welfare regimes.
Welfare expansion was underpinned by broad cross-​class authoritarian coalitions,
which were formed in the crucible of large-​scale societal conflicts in postcolonial
societies. With unions often being weak and fragmented, and left-​wing political parties
having limited mobilizational capacity compared to Asia or Latin America (Collier
1999; Slater and Smith 2016),3 the virulent intra-​elite conflicts that pitted rivaling elite
actors against each other in the struggle for postcolonial political dominance were a
major determinant for the formation of cross-​class coalitions (Brownlee 2007; Waldner
1999). Whenever such episodes occurred at the critical juncture of regime formation,
and when conflict among elites was not reconcilable or repressible (Eibl 2020), the for-
mation of broader cross-​class coalitions became viable strategies to ensure political vic-
tory over rivals. A good example is post-​independence Tunisia, where Habib Bourguiba
forged a lasting alliance with the Tunisian General Labour Union (UGTT) to outcom-
pete his rival within the Neo-​Destour Party, Salah Ben Youssef. This alliance afforded
the UGTT unprecedented influence over post-​independence social policies (Eqbal
1967). Where ruling elites remained cohesive, such as in Jordan, bottom-​up threats
emanating from politically organized middle classes, often of the Arab nationalist ilk,
772   Ferdinand Eibl

were successfully fended off, with the result that welfare states remained fragmented.
Where the repressive capacity of one elite group was dramatically higher than the
other side, and the conflict could be ended by coercing elite rivals, such as in Morocco,
coalitions remained small and welfare provision limited. Repression of such threats was
also facilitated by cliency relationships with Western patrons (Yom 2016). Even in revo-
lutionary moments, such as 1979 Iran, where bottom-​up mass mobilization managed to
unseat the incumbent Pahlavi regime, intra-​elite conflicts significantly shaped the na-
ture and scope of the post-​revolutionary welfare state (Harris 2017). The effect of pop-
ular threats on welfare provision in the absence of intra-​elite conflict was thus limited.
Where bottom-​up threats did have a lasting effect was the Arab Gulf States, where leftist
subversive movements, such as the Dhofar rebellion in Oman, threatened conservative
monarchical regimes in the Gulf in the context of an emerging Arab Cold War (Kerr
1965), pitting “progressive” against “reactionary” regimes in the region. These threats,
albeit of varying intensity, provided a major incentive for the nascent oil exporters to
share their resource wealth with their populations (Eibl and Hertog 2018; Bsheer 2018).
Communal cleavages that, whenever salient, worked against the building of
coalitions based on social class were another major determinant of postcolonial political
coalitions. The primary example here is certainly Lebanon, where post-​independence
sectarian leaders (zuʿama) enshrined their status as political patrons in the post-​
independence National Pact that divided power along sectarian lines. This seriously
undermined state-​and welfare state–​building processes (Cammett 2014a). In Jordan,
welfare provision predominantly targeted non-​Palestinian East Bankers, who formed
the backbone of Jordan’s public sector and security apparatus (Baylouny 2010). Other
cases where coalitions became lopsided in ethnic or sectarian terms were Baʿthist Iraq
and Syria, where preferential distribution favoring the rulers’ own communities became
a prominent feature of public welfare provision (Balanche 2006; Blaydes 2018; Drysdale
1981). Intra-​elite conflict, which had a generally broadening effect in the absence of sa-
lient communal cleavages, also tended to activate and reinforce communitarian rivalries
in divided societies, which hampered but did not completely prevent the building of
class-​based coalitions. The mix of universalist welfare policies, coupled with elements
of communal favoritism in Syria, represent the outcome of these conflicting incentives.
The formation of welfare states was importantly also intertwined with geopolitical dy-
namics affecting the region, which conditioned regimes’ abilities for welfare provision.
In the Gulf, the combination of relative peace and accelerating hydrocarbon exports,
which by the time of the first oil boom had turned all Gulf States bar Yemen into ren-
tier states (e.g., Beblawi and Luciani 1987; Hertog 2010), meant that fiscal constraints
were not a major issue. In fact, resource abundance abetted the emergence of a specific
Gulf variety of welfare states, in which welfare provision not only comprises generous
social policies in traditional fields, such as health and education, but also the employ-
ment of the vast majority of the national labor force in the public sector. And while
large public sectors have been described as a key feature of an Arab variety of capitalism
(Hertog 2020), the extent of public employment sets these countries apart from other
(welfare) states in the region. By contrast, non-​Gulf regimes faced important budgetary
 Welfare States in the Middle East     773

trade-​offs, and, in the absence of resource endowments, only those unaffected by violent
interstate conflict—​most notably the Arab-​Israeli conflict—​had the ability to provide
welfare broadly and generously. For cross-​class regimes at the epicenter of the con-
flict (e.g., Egypt, Syria), the ambition to build authoritarian welfare states was severely
hampered as a result, yielding underfunded policies with low barriers to access. War and
resource endowments thus represented the key conditioning variables for welfare state
development in the Middle East. Having adumbrated the coalitional origins of MENA
welfare states, I will now present trends in education, health, and social protection in
greater detail.

Education
Education began to expand significantly across the region in the 1950s as nascent
regimes (1) made education compulsory and tore down existing barriers to entry in pri-
mary, secondary, and, later, tertiary education; (2) invested in educational infrastructure
in a hitherto unprecedented scale; and (3) either abolished or disempowered religious
institutions as main providers of education. That educational attainment did not im-
prove as rapidly in any other world region as it did in the Middle East is testimony to the
scope of this process. Structurally, educational systems were built along a continental
European model with a high level of tracking into academic and vocational streams at
the secondary level. That said, differences between the three types of welfare states are
visible regarding both the speed and the scope of educational expansion.
In Tunisia, for example, the flagship educational reform under Minister of Education
Mahmoud Messadi virtually achieved universal primary enrolment in 1968, compared
to 35 percent enrolment on the eve of independence. Algeria—​which until the na-
tionalization of oil in 1971 did not receive significant oil rents—​tripled educational
expenditures to nearly 30 percent of the budget in the first decade after independence in
1963, achieving universal primary enrolment in 1978. The share of educational spending
similarly doubled in post-​revolutionary Iran relative to Pahlavi-​era levels, reaching a
fifth of the budget. The post-​1979 regime made a particular effort to bring education
to the countryside, which had been hitherto neglected (Harris 2017; Messkoub 2006).
As for the Gulf States, the expansion of education unfolded even more rapidly, since it
occurred at primary and secondary levels at the same time. In the United Arab Emirates,
for example, secondary enrolment was practically nonexistent in 1967 (0.33 percent),
but ten years later it stood at 31 percent.
Broad (yet not generous) welfare states, such as Egypt and Syria, saw their educa-
tional ambitions seriously hampered by successive military conflicts with Israel, which
tended to crowd out welfare spending (see next section). Primary enrolment, for in-
stance, stagnated in Egypt between 1960 and 1970 and only increased by about 10 per-
cent in Syria. Segmented-​minimal welfare states, such as Morocco, Jordan, or Pahlavi
Iran, exhibited a noticeable bias in their educational policies toward secondary and
tertiary education, which played into the hands of key constituencies from which the
774   Ferdinand Eibl

bureaucratic and military elites were recruited, while undermining the large-​scale ex-
pansion that occurred in authoritarian welfare states (Eilers 1978). By the mid-​1970s,
Morocco’s primary enrolment barely exceeded 50 percent. Jordan, in turn, managed to
outsource the education of Palestinian refugees to UNRWA (the United Nations Relief
and Works Agency for Palestine Refugees), sparing precious fiscal resources. And
Lebanon historically developed an educational system with a significant share of private
institutions, which reflects sectarian divides and affords political elites sources of pa-
tronage (Cammett 2011; Cammett 2014b).
Trends toward a limited convergence of educational policies have been observable
since the mid-​1970s, driven by fiscal difficulties in educational high spenders, a global
trend to universalize secondary education, and a limited peace dividend in the wake
of the Middle East peace process. The latter was particularly significant in Jordan,
where a reduction in defense spending freed up more money for education. In Egypt,
by contrast, any potential peace dividend was swallowed up by the mounting debt ser-
vice, which at its height consumed about a quarter of public spending (Wahid 2009).
Legacies of historical welfare state divergence remain clearly visible, though, in partic-
ular with regard to the role of private education, which remains either banned or seri-
ously restricted in authoritarian welfare states outside the Gulf, whereas private tertiary
institutions capture 17 and 25 percent of the student population in Egypt and Jordan,
respectively (Kohstall 2012; World Bank 2008). In the Gulf, the implantation of private
universities is not driven by a waning commitment to public education, but rather by
expatriate demand and brand-​seeking of aspiring Gulf citizens.
Legacies of welfare state development are also noticeable when it comes to educa-
tional inequalities. While data are only available for a limited number of countries, for
children from the most vulnerable households, the probability of not entering school
at all were 2.5 percent in Tunisia, 6 percent in Jordan, and 21 percent in Egypt (Assaad,
Salehi-​Isfahani, and Hendy 2014).4 Finally, across countries, educational systems suffer
from a serious problem of mismatching between the skills taught and the demands
of the labor market, as public education was historically geared toward recruiting fu-
ture civil servants, and vocational degrees do not provide certified skills with a strong
signaling effect on the job market (World Bank 2013).

Health
Postcolonial healthcare systems have differed significantly with regard to their acces-
sibility, funding, and institutional structure. Regarding the latter, a number of coun-
tries in the region opted for a mixed healthcare system, combining a tax-​funded public
healthcare sector with an insurance-​based system granting access to both public and,
later, private facilities. This is the case in, among other countries, Tunisia, Algeria,
Libya, and Egypt. Such mixed systems resulted from the requirement of having to take
over insurance systems previously established for colonial civil servants while trying
to expand access to healthcare to the wider population. Other countries, such as Syria
 Welfare States in the Middle East     775

or the oil-​rich Gulf States, established purely tax-​based healthcare systems. A third
group of generally segmented-​minimal welfare states built uniquely insurance-​based
systems, frequently with a significant role for private sector providers, as in the case of
Lebanon (Cammett 2014a).
Regarding accessibility, it is clear that regimes based on cross-​class coalitions sought
to provide healthcare universally. Tunisia, for example, created health insurance funds
for the public and private sectors—​which were merged in 2004—​while simultaneously
granting uninsured patients virtually free access to healthcare through the publicly
funded healthcare system (Camau, Zaïem, and Bahri 1990). In post-​revolutionary Iran,
the regime made a concerted effort to universalize healthcare by expanding the public
health insurance scheme to the entire population in 1994 (Messkoub 2006). This had
been preceded by significant hikes in health spending that particularly sought to en-
hance provision in rural areas (Harris 2010).
Similar efforts were visible in Egypt and Syria, which were, however, seriously
undermined by the regimes’ conflicting commitments to welfare and warfare. Egypt
established a tax-​financed pillar of its public healthcare system with free-​of-​charge ac-
cess, yet due to notorious underfunding these facilities tend to be avoided even by the
lowest income groups, as quality is considered wanting. Out-​of-​pocket expenditures
of 72 percent are a negative side effect of this situation (World Bank 2015). Better-​
equipped facilities exist but cater uniquely to those covered by public health insurance.
Throughout the 2000s, the Mubarak regime sought to universalize public health insur-
ance but eventually balked at the prospect of having to subsidize the premiums of up
to 40 percent of Egypt’s population.5 In Syria, the universalization of public healthcare
occurred from the 1960s by gradually expanding access to public health cards, which
granted free access to healthcare based on a means test. This method was later aban-
doned in favor of the introduction of a fully free-​of-​charge national health service. This
process was lengthy and haphazard, though, leaving up to a third of Syria’s population
completely uncovered for a long period of time (Drysdale 1981).
Segmented-​minimal welfare states made the least effort to expand healthcare cov-
erage. A good example is Morocco, where public health insurance was for a long time
limited to urban, white-​collar middle classes in the public sector. Granted, to counter-
vail negative distributive effects, the lowest income brackets have had, in theory, ac-
cess to subsidized healthcare, subject to a means test. In reality, many poor Moroccans
have been unable to afford even the reduced service charges (Loewe 2010). To address
the lack of coverage, Morocco adopted a system of subsidized healthcare provision
for low-​income groups in 2009 (régime d’assistance médicale, or RAMED), which un-
fortunately still leaves about a third of the population uncovered. Jordan’s healthcare
system is truly fragmented, providing high-​quality care to public sector workers and
members of the armed forces—​historical East Banker constituencies—​while the rest
of the population has had to pay a subsidized fee to access public healthcare (Knowles
2005). In Lebanon, healthcare provision has been predominantly private and
represents a major playground for patronage politics along sectarian lines (Cammett
and Şaşmaz 2017).
776   Ferdinand Eibl

Social Protection
Social protection in MENA has consisted of two main pillars: (1) social security insur-
ance in areas such as old age, maternity, and disability, funded through payroll taxes
and, in some countries, voluntary contributions of self-​employed; and (2) noncontrib-
utory social assistance, with the subsidization of basic commodities and energy being
the dominant type of transfer (World Bank 2012). While differences across welfare state
types are less pronounced in this area, they are visible nonetheless.
This is particularly the case in the area of social security, where the pattern of coverage
across broad and generous, broad, and segmented-​minimal welfare states has followed
a very similar pattern as health insurance. On the eve of the Arab Spring, pension cov-
erage outside the Gulf, for example, stood at 66–​80 percent of the workforce in Tunisia,
Algeria, and Iran; 35–​55 percent in Egypt and Syria; and 22–​50 percent in Morocco and
Jordan (Eibl 2020). Jordan’s relatively high level of coverage is concentrated among the
East Banker population, which has historically benefited from employment in the public
sector. While the rapid increase of informal employment after the Arab Spring (e.g.,
Gallien 2017) has leveled these differences to some extent, the social security system re-
mains generally more accessible in historical cross-​class regimes. Broad and generous
welfare states, in particular, have made consistent efforts to enlarge their social secu-
rity scheme to hitherto excluded segments of the population, such as the self-​employed
or agricultural workers (e.g., Chaabane 2002). These efforts notwithstanding, social
security systems across the entire region have reinforced relatively sharp boundaries
between labor market insiders and outsiders in the formal and informal sectors, while
noncontributory social safety nets, such as cash grants known from the Latin American
cases, are underdeveloped (World Bank 2012).
The only major noncontributory social assistance schemes are food and energy
subsidies, which have been used to varying degrees by all countries in the region. With
subsidies absorbing considerable fiscal resources and exposing MENA countries to
adverse shocks as a result of fluctuating commodity prices—​all MENA countries are
net food importers—​governments across the region have been under pressure by the
International Monetary Fund (IMF) and World Bank to reduce subsidy spending
for some time (IMF 2014). Reforms have had variable success, however, and been
accompanied by backlashes from urban middle classes, which constitute the major
beneficiaries of distributionally regressive subsidies (Sadiki 2000). As of 2019, energy
subsidies for petrol have only been fully faded out in Morocco. Jordan and Egypt have
established a managed price adjustment system, which is aligned to the world market
price but subject to discretionary decisions. Tunisia is considering the introduction of
similar systems. Liquefied petroleum gas, which is particularly important for lower-​
income households, remains widely subsidized, similar to basic commodities such as
bread, sugar, or milk. In the Gulf countries, where excessive subsidies have led to wide-
spread overconsumption by households, the drop of the oil price since 2014 has equally
necessitated a reduction of subsidies (Boersma and Griffiths 2019).
 Welfare States in the Middle East     777

War and Welfare State Development

The analytical framework in this chapter emphasizes war-​making and war preparation,
as well as resource endowments, as crucial conditional variables of welfare state devel-
opment. This section corroborates this claim by presenting evidence on the crowding-​
out effect of defense spending on welfare expenditures in MENA. Using the case of
Egypt, I also highlight how regimes with a strong incentive but a lack of ability to dis-
tribute welfare have used what I call “cheap social policies” to square the circle.

MENA’s “Butter or Guns” Trade-​Off

Although the metaphor of a “butter versus guns” trade-​off has intuitive appeal, empir-
ical findings at the global level have actually been very mixed (Gupta et al. 2004; Hess
and Mullan 1988; Narizny 2003; Rasler and Thompson 1985). Whitten and Williams
(2011, p. 117) even claim that “governments do not face a trade-​off between ‘butter’ or
‘guns.’” Similar to European patterns (Tilly 1985), war-​making has generally increased
fiscal capacity in the Global South (Thies 2004), although this association seems ab-
sent in the MENA region (Lu and Thies 2012). That said, extant research (Cammett et al.
2015; see also Richards and Waterbury 2008, p. 350) has found little systematic evidence
for a negative effect of defense spending on welfare.
To test the empirical association between war and welfare provision, I present the
results of a fixed-​effects, dynamic panel regression of defense expenditures on welfare
expenditures, drawing on the Global State Revenues and Expenditures data set (Lucas
and Richter 2016), supplemented with data from the IMF’s Government Financial
Statistics (IMF 2015a), as well as statistical yearbooks. Included countries all have long
and consistent time-​series data available between 1960 (or the earliest year after inde-
pendence) and 2005. Welfare expenditures are measured as a percentage of the budget,
although findings are substantively identical with other measures. The model includes
standard controls from the welfare state literature, such as GDP per capita, trade, and
the dependency ratio. The regression also includes a measure of resource rents per
capita taken from the Ross (2019) data set. To test the countervailing effect of resources,
this variable is interacted in a second model with defense expenditures. I refer the reader
to Eibl (2020, chap. 4) for further details and regression tables, and instead would like to
zoom in here on the main findings shown below.
Figures 37.2 and 37.3 display the simulated predicted effects of, respectively, a short-​
term and a permanent increase in defense spending. The short-​term hike represents a
change from the mean to the maximal value of defense spending found in the data (17.8
and 57.4 percent, respectively). Permanent changes are based on a scenario where de-
fense remains at the maximum value indefinitely. Although these values seem drastic,
they have historically been observed in times of war and in situations of severe external
 60

50

40

30

20

10
0 2 4 6 8
Year

welfare spending/budget defense spending/budget

60

50

40

30

20

10
0 2 4 6 8
Year

defense spending/budget resource-poor resource-rich

Figure 37.2: Short-​term defense shocks on welfare spending.

Note: Shaded areas represent 95 percent confidence interval. Year refers to the number of years
after the increase in defense spending. The left-​hand axis indicate percentages.
Source: IMF 2015a; Lucas and Richter 2016; and various statistical yearbooks.
 60

50

40

30

20

10

0
0 5 10 15
Year
welfare spending/budget defense spending/budget

60

50

40

30

20

10

0
0 5 10 15
Year
defense spending/budget resource-poor resource-rich

Figure 37.3: Long-​term defense shocks on welfare spending.

Note: Shaded areas represent 95 percent confidence interval. Year refers to the number of years
after the increase in defense spending. The left-​hand axis indicate percentages.
Source: IMF 2015a; Lucas and Richter 2016; and various statistical yearbooks.
780   Ferdinand Eibl

threats without ongoing hostilities (e.g., Jordan in the early 1960s). The scenarios for the
interaction models are identical, with the addition that the effect is shown for resource-​
poor and resource-​abundant countries. The graphs yield two important take-​away
messages. First, there has been a significant trade-​off between welfare and warfare in
MENA. Second, large resource endowments allow regimes to eschew, or at least alle-
viate, the dilemma between “butter or guns” by allowing governments to shield social
welfare from the negative impact of external conflict. While social spending eventu-
ally declines in both resource-​rich and resource-​poor countries in our, arguably, rather
extreme scenario, social spending cuts are much more pronounced in resource-​poor
countries. Given that the MENA region to date has one of the highest levels of defense
spending in the world, these findings suggest that war-​making and war preparation con-
tinue to significantly undermine welfare provision across the region.

Cheap Social Policies

Regimes with the incentives but, due to external threats, lacking capacity to distribute
welfare face a particular dilemma regarding how to honor coalitional commitments
while avoiding fiscal collapse. Post-​1952 Egypt under Nasser is a good example of such as
regime. Strong intra-​elite conflict within the ruling Revolutionary Command Council
(RCC), which came to a head in the wake of President Mohamed Naguib’s ouster in 1954,
prompted Gamal Abdel Nasser to reach out to working and middle-​class constituencies
to outcompete his rival (for a similar argument in the context of party regimes, see
Brownlee 2007). In the words of Beattie (1994, p. 98), the events of 1954 “crystallized
Nasser’s thoughts regarding real and prospective sources of regime support.” The events
of 1954 thus stood at the origin of Egypt’s broad-​based regime coalition.
As the nascent regime soon came to realize, though, the ambition for large-​scale wel-
fare distribution needed to be reconciled with ensuring survival in the face of strong
external threats. The Arab-​Israeli conflict, and foreign policy more generally, did not
feature prominently among the Free Officers’ concerns when they seized power in 1952.
In fact, wanting neither peace nor war, the Free Officers exhibited a certain ambivalence
vis-​à-​vis Israel (Barnett 1992, 84). A key indicator of this low level of threat perception
was the “heavy cuts in expenditure on defence” (IMF 1954, p. 15) carried out between
1952 and 1954. The Gaza raid of February 1955 carried out by Israel and the Suez War
in late 1956 fundamentally changed the regime’s risk perception, however, and shifted
its spending commitments from socioeconomic development to defense. Defense
expenditures increased by 116 percent between 1955 and 1960 in absolute terms, whereas
social expenditure only rose by 39 percent in the same period. Putting these outlays into
broader perspective, defense expenditures consumed yearly about two-​thirds of what
was spent in the first five-​year plan altogether (Al-​Jiritli 1974, p. 42). With Egypt lacking
a natural resource endowment, social policies thus had to compete, from the beginning,
with security-​related expenditures, which greatly undermined the regime’s distributive
capacity.
 Welfare States in the Middle East     781

Within this particular context of strong incentives to distribute, but little fiscal ca-
pacity to do so, Egypt resorted to what I have called “cheap social policies” (Eibl 2020),
comprising at least one of the following features: (1) a heavy reliance on “free” social
policies distributing rent-​generating statutory rights to lower and middle classes;
(2) the systematic use of windfalls to finance social expenditures; and (3) devising social
policies that alleviate the regime’s shortage of resources. In the following section, I will
focus in particular on the latter feature and demonstrate how it has shaped Egyptian so-
cial policies.
Designing social policies to alleviate the shortage of resources arguably puts the logic
of distributive social policies on its head, as social policies devised in such a way become
a mechanism of resource generation instead of resource distribution. Given conditions
of resource scarcity and the need to cater to a large support base, these social policies
become extremely advantageous, since they enable the generation of additional revenue
while bolstering domestic legitimacy. A main area where revenue-​generating social
policies has been implemented is social insurance. To illustrate the general logic behind
revenue-​generating cheap social policies, I rely on the correspondence between the
Egyptian Ministry of Social Affairs and the Council of Ministers in the preparation of
social security legislation in the early 1950s, which established the cornerstone of Egypt’s
current social security system.
When presenting the draft law to the Council of Ministers in November 1952, the min-
ister of finance pointed to the “copious resources” that the funds would make available
to the government, and the extent to which this would support the national economy
(Egyptian National Archives 1953b). In fact, the explanatory note contains no reference
to any social motives whatsoever. In a second note, dating from June 1953, the Ministry
of Finance and Economics urged the Council of Ministers to extend social insurance
to all new state employees (Egyptian National Archives 1953a). Underlining the “pre-
sent conditions of austerity,” the minister highlighted the “vast funding opportunities”
(Egyptian National Archives 1953a) that would become available if social insurance
were to be extended. In turn, this would help the government finance its development
projects.
The same arguments for social legislation appear again in the context of the 1955 law,
which extended social insurance to the private sector and thus considerably broadened
its coverage. The archival notes strongly suggest that economic factors played a major
role in the government’s motivation to pass the law (Egyptian National Archives 1955).
Particularly insightful is an explanatory note from the Ministry of Social Affairs to
the Council of Ministers from July 1955, laying out two key motives behind the draft
law. First, social insurance was meant to “calm down” workers and to make them more
productive, so social considerations, albeit in a rather paternalistic sense, did indeed
motivate the adoption of the 1955 law. Second, and importantly, the generation of re-
sources was again a primary argument presented in favor of the law: “Given the enor-
mous amount of money that the social insurance system will accumulate after a while,
the social insurance fund will become a powerful source of money that the government
can rely on to finance social reforms and economic development projects” (Egyptian
782   Ferdinand Eibl

National Archives 1955). Taken together, both notes demonstrate that the idea of
implementing social policies to fund social policies was not something the regime “dis-
covered” along the way. On the contrary, the generation of resources was a central moti-
vation behind the social insurance legislation from the outset.
Ample borrowing from Egypt’s social insurance funds by the state since 1952
demonstrates the importance of social policies as an income-​generating mechanism.
At the height of war preparation in 1971–​1972, the government covered over two-​thirds
of all expenditures from the social security fund. The average between 1960 and 1976
was 42 percent. Borrowing continued through the 1980s and 1990s, when state-​owned
enterprises started to use the social insurance funds to access cheap loans. Since 1993,
surpluses from the work accident insurance have been used to cover resulting deficits in
Egypt’s health insurance fund. Crucially, until the late 1990s, borrowing occurred at in-
terest rates consistently below inflation, which effectively amounted to an expropriation
of savings.
While it may seem that revenue-​generating social policies are specific only to the
period of welfare state formation in Egypt, there is evidence to suggest that cheap so-
cial policies of this nature matter until the present period—​and could well be at play in
other countries in the region. Specifically, by increasing domestic saving and curbing
consumption, Egyptian governments have systematically used coverage extensions of
the social security scheme to counterbalance current account problems. As Figure 37.4
demonstrates (for further details on the underlying regression model, see Eibl 2020,
pp. 238–​234), coverage extensions of the Egypt’s social security system have been sys-
tematically associated with falling foreign currency reserves. The figure highlights the
growing likelihood of coverage extension as reserves start to shrink, which is indicated
by the increasing slope of the graph left of the zero line. It is interesting to note that
the pursued extension of Egypt’s health insurance by President Sisi occurs in a similar
context.
Conversely, coverage extensions of social security that are not revenue-​generating,
at least in the short term, have been very difficult to implement. The primary example
in this context is the failed introduction of a universal health insurance scheme under
Mubarak in the late 2000s. As many lower-​income groups would have been unable
to pay their premiums, the Egyptian state would have had to shoulder an additional
cost of at least 17 billion EGP, nearly twice the size of the country’s health budget of
10 billion EGP in the late Mubarak period. According to former Health Minister Tag
Eddin, “money was the key issue in the failure of the health insurance reform. [ . . . ]
Our budget was never enough and our spending lower than needed. [ . . . ] In the
drafting of the budget, the Ministry of Finance would only ever give us the ‘leftovers’”
(personal interview, Cairo 2013). Given these historical patterns, and the fact that the
structural constraints weighing on President Sisi are essentially the same as those of
his predecessors, the decision to introduce universal health insurance in 2018 should
be viewed with caution. The envisaged implementation period of fifteen years, gover-
norate by governorate, is extremely long, allowing for policy slippage and backtracking
along the way.
 Welfare States in the Middle East     783

0.7
Predicted probability of coverage extension

0.6

0.5

0.4

0.3

0.2

0.1

0.0

–100 –50 0 50 100 150

Change of reserves in %

Figure 37.4: Balance of payments crises and social security extension.

Note: Shaded areas represent 95 percent confidence interval.
Source: IMF 2015b; own data on extensions.

In sum, the concept of cheap social policies offers a theoretical mechanism to under-
stand how interstate war and external threats have shaped social policies in the Middle
East and other regimes where the fiscal base for redistributive taxation cannot be easily
extended due to institutional constraints and the durable “conscription of wealth”
(Scheve and Stasavage 2010) has remained elusive. The concept could also be easily
extended to take into account other forms of “cheap” social policies. For example, for-
bearance or non-​enforcement (Holland 2017) in the context of social policies could be
another mechanism through which social policies, albeit meaningful and beneficial for
target groups, do not aggravate the fiscal burden on the state budget.

Conclusion

Since the late 1950s, social policies and welfare distribution have been major areas
of government activity across the MENA region. Shaped by foundational political
struggles at the moment of regime formation and conditioned by geopolitical context,
MENA welfare states have appeared in three distinct forms: (1) broad and generous wel-
fare states have been characterized by a high level of social expenditures and predom-
inantly universal, open-​access social policies (countries such as Tunisia or the GCC
[Gulf Cooperation Council] countries fall into this category); (2) broad welfare states
are similar in their ambition to guarantee, as much as possible, universal access but
784   Ferdinand Eibl

have been hampered in their scope by geopolitical challenges, mostly stemming from
external threats undermining generous social outlays—​this has led, as the example of
Egypt above illustrates, to the emergence of “cheap social policies”; (iii) segmented and/​
or minimalist welfare states have not only spent much less on social welfare, but also put
in place important barriers of access to social policies, limiting their scope to key po-
litical constituencies of the regime. They are generally characterized by a much greater
importance of private sector providers.
While the political economy work on welfare provision in the MENA region has made
great strides in recent years (e.g., Brooke 2017; Cammett and Issar 2010; Cammett 2014a;
Loewe 2010), many important questions remain unsolved. First, the political economy
of social policy reform has lagged behind the study of economic reforms in MENA.
Key constituencies of reform, their dynamics of mobilization, and how they affect re-
form processes—​such as the above-​mentioned health insurance reform in Egypt—​
have so far been understudied. Second, more systematic attention should be given to
the link between electoral politics—​in autocratic and democratic contexts—​and social
policies. While important work has been done in the context of Lebanon, Egypt, and
Iran (Brooke 2019; Cammett and Issar 2010; Harris 2017), we know very little about the
electoral cycle of social policymaking in other parts of the region. Third, data related
to social policies (e.g., expenditures, coverage) should be gathered more systematically
across the region, including in the Gulf countries, where budgetary and labor market
data are more spotty. Only by tracing spending patterns back to critical junctures can
we better understand the political origins—​and the likely future trajectory—​of welfare
states in the Middle East and North Africa.

Notes
1. For the purpose of this analysis, MENA includes all Muslim-​majority countries in the
Middle East and North Africa as defined by the World Bank. As a result, Turkey is not in-
cluded in this analysis. See Eder (2010) and Buǧra and Keyder (2006) for a comparative
perspective on Turkey. To simplify diction, I use the terms “Middle East” and “Middle East
and North Africa” interchangeably.
2. This typology is admittedly highly state-​centric and analytically less ambitious than alterna-
tive typologies (e.g., Wood and Gough 2006), which are systematically anchored in a nor-
mative framework. My main aim is to evaluate the role of the state in providing welfare,
rather than looking at the whole policy mix of private and public welfare provision. Welfare
provision by nonstate actors is analyzed elsewhere in this volume.
3. The notable exception here is the Tunisian General Labour Union (UGTT) (Bellin 2002;
Hartshorn 2019).
4. The figures are based on the averages between boys and girls.
5. This project was resumed under President Sisi and a gradual expansion of the health in-
surance system to the entire population has in principle been agreed upon. That said, the
roll-​out period (into the 2030s) means that full implementation of the reform remains
uncertain.
 Welfare States in the Middle East     785

References
Al-​Jiritli, Ali. 1974. Al-​Tarikh al-​iqtisadi lil-​thawrah, 1952–​1966 [The economic history of the
revolution, 1952–​1966]. Cairo: Dār al-​Ma’arif bi Misr.
Assaad, Ragui, Djavad Salehi-​Isfahani, and Rana Hendy. 2014. “Inequality of Opportunity
in Educational Attainment in Middle East and North Africa: Evidence from Household
Surveys.” ERF Working Paper 834. Cairo, Egypt: Economic Research Forum.
Balanche, Fabrice. 2006. La région alaouite et le pouvoir syrien. Paris: Karthala.
Barnett, Michael N. 1992. Confronting the Costs of War: Military Power, State, and Society in
Egypt and Israel. Princeton: Princeton University Press.
Baylouny, Anne Marie. 2010. Privatizing Welfare in the Middle East: Kin Mutual Aid
Associations in Jordan and Lebanon. Bloomington: Indiana University Press.
Beattie, Kirk J. 1994. Egypt during the Nasser Years: Ideology, Politics, and Civil Society. Boulder,
CO: Westview Press.
Beblawi, Hazem, and Giacomo Luciani, eds. 1987. The Rentier State. London: Croom Helm.
Bellin, Eva. 2002. Stalled Democracy: Capital, Labor, and the Paradox of State-​Sponsored
Development. Ithaca, NY: Cornell University Press.
Blaydes, Lisa. 2018. State of Repression: Iraq under Saddam Hussein. Princeton, NJ: Princeton
University Press.
Boersma, Tim, and Steve Griffiths. 2019. Reforming Energy Subsidies: Initial Lessons from the
United Arab Emirates. Washington, DC: Brookings Institution. https://​www.brookings.
edu/ ​ r esearch/ ​ r eforming- ​ e nergy- ​ s ubsidies- ​ i nitial- ​ l essons- ​ f rom- ​ t he- ​ u nited- ​ a rab-​
emirates/​.
Brooke, Steven. 2017. “From Medicine to Mobilization: Social Service Provision and the
Islamist Reputational Advantage.” Perspectives on Politics 15, no. 1: 42–​61.
Brooke, Steven. 2019. Winning Heats and Votes: Social Service and the Islamist Political
Advantage. Ithaca, NY: Cornell University Press.
Brownlee, Jason. 2007. Authoritarianism in an Age of Democratization. Cambridge: Cambridge
University Press.
Bsheer, Rosie. 2018. “A Counter-​Revolutionary State: Popular Movements and the Making of
Saudi Arabia.” Past & Present 238, no. 1: 233–​277.
Buǧra, Ayşe, and Çaǧlar Keyder. 2006. “The Turkish Welfare Regime in Transformation.”
Journal of European Social Policy 16, no. 3: 211–​228.
Camau, Michel, Hédi Zaïem, and Hajer Bahri. 1990. État de santé: Besoin médical et enjeux
politiques en Tunisie. Paris: Editions du CNRS.
Cammett, Melani. 2011. “Partisan Activism and Access to Welfare in Lebanon.” Studies in
Comparative International Development 46, no. 1: 70–​97.
Cammett, Melani. 2014a. Compassionate Communalism: Welfare and Sectarianism in Lebanon.
Ithaca, NY: Cornell University Press.
Cammett, Melani. 2014b. “Sectarian Politics and Social Welfare: Non-​state Provision in
Lebanon.” In The Politics of Non-​state Social Welfare, edited by Melani Cammett and Lauren
M. MacLean, 137–​155. Ithaca, NY: Cornell University Press.
Cammett, Melani, Ishac Diwan, Alan Richards, and John Waterbury. 2015. A Political Economy
of the Middle East. Boulder, CO: Westview Press.
Cammett, Melani, and Sukriti Issar. 2010. “Bricks and Mortar Clientelism: Sectarianism and
the Logics of Welfare Allocation in Lebanon.” World Politics 62, no. 3: 381–​421.
786   Ferdinand Eibl

Cammett, Melani, and Aytuğ Şaşmaz. 2017. “Political Context, Organizational Mission, and the
Quality of Social Services: Insights from the Health Sector in Lebanon.” World Development
98: 120–​132.
Castles, Francis G., Stephan Leibfried, Jane Lewis, Herbert Obinger, and Christopher Pierson,
eds. 2010. The Oxford Handbook of the Welfare State. Oxford: Oxford University Press.
Chaabane, Mohamed. 2002. “Towards the Universalization of Social Security: The Experience
of Tunisia.” ESS Paper no. 4. Geneva: International Labour Organization.
Collier, Ruth Berins. 1999. Paths toward Democracy: the Working Class and Elites in Western
Europe and South America. Cambridge: Cambridge University Press.
Drysdale, Alasdair. 1981. “The Regional Equalization of Health Care and Education in Syria
Since the Ba’thi Revolution.” International Journal of Middle East Studies 13, no. 1: 93–​111.
Eder, Mine. 2010. “Retreating State? Political Economy of Welfare Regime Change in Turkey.”
Middle East Law and Governance 2, no. 2: 152–​184.
Egyptian National Archives. 1953a. 0081-​097940 Juz Awal qawanin wa murasim wa makatabat
bi-​sh’an sunduq al-​ta’min wa sunduq al-​idkhar wa al-​mu’ashat li-​muwadhafi al-​hukuma al-​
madaniin [First part of Laws, Decrees, and Correspondences regarding the Insurance Fund
and the Savings and Pensions]. Cairo: Egyptian National Archives.
Egyptian National Archives. 1953b. “Session of the Council of Ministers, 14 October 1953.”
Cairo: Egyptian National Archives.
Egyptian National Archives. 1955. Insurance and Savings for Workers (Dec. 52–​Dec. 55). 0081-​
053914/​1952-​55. Cairo: Egyptian National Archives.
Eibl, Ferdinand. 2020. Social Dictatorships: The Political Economy of the Welfare State in the
Middle East and North Africa. Oxford: Oxford University Press.
Eibl, Ferdinand, and Steffen Hertog. 2018. When Do Oil Dictators Share the Wealth? Evidence
from the Gulf Monarchies. Unpublished manuscript.
Eilers, Wilhelm. 1978. “Educational and Cultural Development in Iran during the Pahlavi Era.”
In Iran under the Pahlavis, edited by George Lenczowski, 303–​331. Stanford, CA: Hoover
Institution Press.
Eqbal, Ahmad. 1967. “Politics and Labor in Tunisia.” PhD diss., Princeton University.
Gallien, Max. 2017. Seven Years On: Tunisia’s Legacies of Neglect. Middle East Eye, December 18,
2017. https://​www.middleeasteye.net/​opinion/​seven-​years-​tunisias-​legacies-​neglect.
Gupta, Sanjeev, Benedict Clements, Rina Bhattacharya, and Shamit Chakravarti. 2004. “Fiscal
Consequences of Armed Conflict and Terrorism in Low-​and Middle-​Income Countries.”
European Journal of Political Economy 20, no. 2: 403–​421.
Harris, Kevan. 2010. “Lineages of the Iranian Welfare State: Dual Institutionalism and
Social Policy in the Islamic Republic of Iran.” Social Policy & Administration 44, no.
6: 727–​745.
Harris, Kevan. 2017. A Social Revolution: Politics and the Welfare State in Iran. Berkeley:
University of California Press.
Hartshorn, Ian M. 2019. Labor Politics in North Africa: After the Uprisings in Egypt and Tunisia.
Cambridge: Cambridge University Press.
Hertog, Steffen. 2010. Princes, Brokers, and Bureaucrats: Oil and the State in Saudi Arabia.
Ithaca, NY: Cornell University Press.
Hertog, Steffen. 2020. “Segmented Market Economies in the Arab World: The Political
Economy of Insider-​Outsider Divisions.” Socio-​Economic Review, Online First, April 13, 2020.
doi:10.1093/​ser/​mwaa016.
 Welfare States in the Middle East     787

Hess, Peter, and Brendan Mullan. 1988. “The Military Burden and Public Education Expenditures
in Contemporary Developing Nations: Is There a Trade-​off?” Journal of Developing Areas 22,
no. 4: 497–​514.
Holland, Alisha C. 2017. Forbearance as Redistribution: The Politics of Informal Welfare in Latin
America. Cambridge: Cambridge University Press.
ILO (International Labour Organization). 2014. World Social Protection Report 2014–​15.
Geneva: ILO.
IMF (International Monetary Fund). 1954. 1953 Consultations–​Egypt. SM-​54-​1. Washington,
DC: IMF.
IMF (International Monetary Fund). 2014. Subsidy Reform in the Middle East and North Africa.
Washington, DC: IMF.
IMF (International Monetary Fund). 2015a. Government Financial Statistics Yearbook 2015.
Washington, DC: IMF. https://​www.elibrary.imf.org/​view/​IMF043/​23060-​9781513559766/​
23060-​9781513559766/​23060-​9781513559766?rskey=2GLYdZ&result=18.
IMF (International Monetary Fund). 2015b. International Financial Statistics (IFS). Washington,
DC: IMF. https://​www.elibrary.imf.org/​search?t_​7=urn:Series/​041&type_​0=booksandjournals
&browsePage=countries&sort=datedescending.
Kerr, Malcolm H. 1965. The Arab Cold War, 1958–​1964: A Study of Ideology in Politics.
Oxford: Oxford University Press.
Knowles, Warwick. 2005. Jordan since 1989: A Study in Political Economy. London: I.B. Tauris.
Kohstall, Florian. 2012. “Free Transfer, Limited Mobility: A Decade of Higher Education Reform
in Egypt and Morocco.” Revue des mondes musulmans et de la Méditerraneé 131: 91–​109.
Loewe, Markus. 2010. Soziale Sicherung in den arabischen Ländern. Baden-​Baden, Germany:
Nomos.
Lu, Lingyu, and Cameron G. Thies. 2012. “War, Rivalry, and State Building in the Middle East.”
Political Research Quarterly 66, no. 2: 239–​253.
Lucas, Viola, and Thomas Richter. 2016. “State Hydrocarbon Rents, Authoritarian Survival
and the Onset of Democracy: Evidence from a New Dataset.” Research & Politics 3, no. 3.
doi:10.1177/​2053168016666110.
Messkoub, Mahmood. 2006. “Social Policy in Iran in the Twentieth Century.” Iranian Studies
39, no. 2: 227–​252.
Narizny, Kevin. 2003. “Both Guns and Butter, or Neither: Class Interests in the Political.”
American Political Science Review 97, no. 2: 203–​220.
Rasler, Karen A., and William R. Thompson. 1985. “War Making and State Making:
Governmental Expenditures, Tax Revenues, and Global Wars.” American Political Science
Review 79, no. 2: 491–​507.
Rauch, James E., and Scott Kostyshak. 2009. “The Three Arab Worlds.” Journal of Economic
Perspectives 23, no. 3: 165–​188.
Richards, Alan, and John Waterbury. 2008. A Political Economy of the Middle East. Boulder,
CO: Westview Press.
Ross, Michael L. 2019. Oil and Gas Data, 1932–​2011. Harvard Dataverse, V2. http://​hdl.handle.
net/​1902.1/​20369.
Sadiki, Larbi. 2000. “Popular Uprisings and Arab Democratization.” International Journal of
Middle Eastern Studies 32: 71–​95.
Scheve, Kenneth, and David Stasavage. 2010. “The Conscription of Wealth: Mass Warfare and
the Demand for Progressive Taxation.” International Organization 64, no. 4: 529–​561.
788   Ferdinand Eibl

Slater, Dan, and Nicholas Rush Smith. 2016. “The Power of Counterrevolution: Elitist Origins
of Political Order in Postcolonial Asia and Africa.” American Journal of Sociology 121,
no. 5: 1472–​1516.
Thies, Cameron G. 2004. “State Building, Interstate and Intrastate Rivalry: A Study of Post-​
colonial Developing Country Extractive Efforts, 1975–​2000.”International Studies Quarterly
48, no. 1: 53–​72.
Tilly, Charles. 1985. “War Making and State Making as Organized Crime.” In Bringing the State
Back In, edited by Peter B. Evans, Dietrich Rueschemeyer, and Theda Skocpol, 169–​187.
Cambridge: Cambridge University Press.
Wahid, Latif. 2009. Military Expenditure and Economic Growth in the Middle East. London:
Palgrave Macmillan.
Waldner, David. 1999. State Building and Late Development. Ithaca, NY: Cornell University Press.
Whitten, Guy D., and Laron K. Williams. 2011. “Buttery Guns and Welfare Hawks: The Politics
of Defense Spending in Advanced Industrial Democracies.” American Journal of Political
Science 55, no. 1: 117–​134.
Wood, Geof, and Ian Gough. 2006. “A Comparative Welfare Regime Approach to Global Social
Policy.” World Development 34, no.10: 1696–​1712.
World Bank. 2008. The Road Not Traveled: Education Reform in the Middle East and North
Africa. Washington, DC: World Bank.
World Bank. 2012. Inclusion and Resilience: The Way Forward for Social Safety Nets in the
Middle East and North Africa. Washington, DC: World Bank.
World Bank. 2013. Jobs for Shared Prosperity: Time for Action in the Middle East and North
Africa. Washington, DC: World Bank.
World Bank. 2015. A Roadmap to Achieve Social Justice in Health Care in Egypt. Washington,
DC: World Bank.
Yom, Sean. 2016. From Resilience to Revolution: How Foreign Interventions Destabilize the
Middle East. New York: Columbia University Press.
 Chapter 38

Isl am ist Orga ni z at i ons

and the Prov i si on of
So cial Se rv i c e s

Steven Thomas Brooke

Many governments in the Islamic world struggle to extend social protections across
their populations. Alongside these faltering public systems, a variety of nonstate
providers, including Islamist organizations, have assumed a larger and larger role in fur-
nishing schooling, healthcare, unemployment assistance, and financial aid to millions
of people each year. Purely in terms its influence on the lived daily experience of ordi-
nary people, Islamist groups’ social service activism constitutes an important field for
research. At the same time, recent scholarship has built an empirically rich and theoret-
ically specified account of how this provision influences, and is influenced by, politics.
This chapter focuses on this work, with an eye toward highlighting research advances
for the field of comparative politics more broadly.
This chapter proceeds as follows: I first conceptualize Islamist social service provi-
sion, focusing on identifying precisely what distinguishes it from other types of nonstate
provision, including by non-​Islamist, and nonreligious, actors. The next section situates
Islamist social service provision as an outcome to be explained: given the purported ef-
fectiveness of this provision in generating support for provider organizations, and the
dominant status of Islamist groups as members of the political opposition, why has
it flourished in regimes that devote extensive efforts to suppressing their opponents?
Cross-​national variation in the existence of these networks suggests the importance of
structural factors, specifically the reaction of authoritarian regimes to economic crises.
The final section probes the potential of Islamist social service provision to explain po-
litical mobilization. In contrast to classic, exchange-​based theories of political support
or ideological transformation, a major finding is that social service provision influences
how recipients perceive the values of the provider organization. While this finding
provides new insight into voting behavior in the Islamic world, it also requires better
theorizing of the role of religion as a potential “recipient side” factor. Returning to the
790   Steven Thomas Brooke

opening theme of the chapter, this fact in particular seems to distinguish Islamist social
service provision from other types.

Islamist Social Service Provision

Among the multitude of social service providers are Islamist groups like the Muslim
Brotherhood. As “Islamists,” these actors seek not only to make Islam the basis of public
and private life, but also to use the coercive power of the state—​through legislation
and enforcement—​to do so. This chapter focuses largely on nonviolent Islamists, spe-
cifically those who have chosen to pursue this power through peaceful participation
in electoral politics (although at times violent Islamist groups, as well as Islamic or-
ganizations in general, will be used to illustrate important comparisons). As discussed
at greater length below, the fact that Islamist organizations often find themselves ex-
pressly at odds with ruling regimes marks out an important additional characteristic of
this phenomenon.
There is a general assumption that Islamist groups’ religious identity creates a spe-
cial category of nonstate providers. In one popular argument, Islamist groups benefit
from a dense and hierarchical organizational structure, which scholars often connect to
their particular interpretation of Islam (al-​Anani 2016; Kandil 2014). To the extent that
Islamist groups can count on ideologically driven cadres to do the work of social ser-
vice provision, it confers on these groups a notable advantage over their competitors in
the realm of social service provision by reducing costs (Bano 2009; Cammett and Jones
Luong 2014). Hamayotsu describes the development of the Islamist PKS in Indonesia,
stating, “Given limitations on financial resources and on access to state bureaucracies
and privileges, PKS relies heavily on cadres’ service for extensive and effective operation
of its welfare programs” (2015, 985). If Islamist parties draw on this reservoir of cheap
organizational support, that may help account for the “Islamist Advantage” over non-​
Islamist providers (Cammett and Jones Luong 2014).
The evidence that this distinctly advantages Islamist parties is mixed, however. In
other contexts, ethnic parties have similarly been able to convert the labor of move-
ment activists into political gains (Anria 2018; Van Cott 2005). Other religious—​but not
Islamic—​organizations have also shown the ability to capitalize on the ideological de-
votion of adherents to reduce the costs of social service provision (Thachil 2014a). And,
of course, there seems little reason why leftist groups could not also draw on the “free
labor” of an ideologically committed cadre just as Islamist groups can.
Returning to the specific case of Islamists further suggests that the advantage might
not necessarily flow through infrastructure or organization. Other Islamist providers,
such as the Salafi party Hizb al-​Nour, managed to provide extensive social services
without a Muslim Brotherhood–​style formal organization. Even in the case of the
Muslim Brotherhood, the group’s “brick and mortar” medical networks (Cammett
and Issar 2010) did not run on free activist labor. Not only was every employee paid
 Islamist Organizations and the Provision of Social Services    791

a competitive salary, but their attitudes toward the Muslim Brotherhood ranged from
sympathetic to indifferent, and the movement went to great lengths to separate itself
from the party (Brooke 2019). This brief examination suggests that doctrinal factors
alone do not seem sufficient to explain Islamist proficiency, at least to the extent that
they manifest in a particular type of organizational structure.
A second factor that may mark Islamist provision as a unique phenomenon is the
extensive variation in type, style, and target of their social services. This is even the
case within the same organization. For example, two authors describe how the Muslim
Brotherhood operated “a massive, decentralized, society-​wide network of clinics and
hospitals, Islamic banks . . . textile factories, food banks, youth clubs, social welfare
agencies, unemployment services, discount grocery stores, legal aid agencies, and much
more” (Davis and Robinson 2012, 48). Empirically, however, much of the research on
this provision is limited to health and, to a lesser extent, education (Azra, Afrianty,
and Hefner 2007; Brooke 2017a, 2017b, 2019; Cammett 2011, 2014; Cammett and Issar
2010; Challand 2008; Clark 2004; Hefner 2008; Morsy 1988; Sciortino, Ridarineni, and
Marjadi 2010). This web of interlinked, cradle-​to-​grave services—​many of which have
been operational for decades—​seems to lack parallel in other providers.
This extensive array of services suggests the analytical utility of disaggregating the
broader category of “nonstate services.” As a starting point, Tugal (2013, 150) identifies
a helpful four-​part typology based on the method of delivery: (1) provision through
quasi-​independent organizations, (2) syndicates (effectively white-​ collar unions),
(3) direct provision by individuals, and (4) through infiltration of independent organ-
izations. One could plausibly add here a fifth dimension, particularly in weaker states,
based around Islamists’ ability to access state resources and redirect them toward their
constituents (Cammett 2014). While this fifth approach resembles classic brokerage
more than “nonstate” provision, to the extent that Islamists engage in “credit claiming”
for state resources that they have redirected, it may approximate the effects of non-
state provision (Bueno 2018). For example, Masoud (2014, 90–​92) notes how Muslim
Brotherhood candidates seamlessly integrated into their list of campaign achievements
the services for neighborhood residents that they successfully extracted from the state.
And as Jenny White identified when studying the local service provision networks of the
Turkish Justice and Development Party (AKP), “whether or not the activists provided
services in the name of the party, it was known that they were party members and credit
for their neighborly assistance was informally attributed to the party” (2012, 61). Other
researchers have similarly noted the AKP’s adeptness at using public goods to grow their
own constituencies (Aytaç 2014; Marschall, Aydogan, and Bulut 2016).
An explanation for this variation potentially lies in the institutional context,
and in particular in a given regime’s ability to monitor and control these providers
(Wiktorowicz 2000; Tadros 2011). Mubarak-​era Egypt retained a significant level of
control over the civic and social sectors, which meant that operating social services on
an informal or below-​the-​radar capacity was not sustainable, and in fact was actively
dangerous. Instead, the safer choice was to operate legally, generally by forming offi-
cial service-​providing NGOs that kept obvious partisan activism at arm’s length. As
792   Steven Thomas Brooke

Hamid puts it, Islamists in Egypt made a calculation that “legal status, while offering no
guarantees, at least offers some protection” (2014, 50). The Jamaʿat-​i Islami in Pakistan
seemingly made a similar judgement. As one member told Masooda Bano, formal sep-
aration granted the social activism some protection: “Whenever the military came into
power, it put restrictions on our political work; this also restricted our welfare work.
That is why we thought it is best to make our welfare wings completely separate” (2009,
90). This reinforces the importance of institutional context, and particularly regime-​
level relationships and constraints on how these organizations operate (Cammett and
MacLean 2014).1
In addition to the institutional context, scholars also highlight how local-​level dy-
namics of competition with other providers, the emergence of specific opportunity
structures, and particular availability of resources influence the style and type of provi-
sion. For example, in a case study of the Indonesian Islamic movement Muhammadiyah,
Mohamed Fuad (2002, 140–​141) shows how the precise form the movement’s social
service provision takes is determined by opportunity, political dynamics, inter-​and
intra-​religious rivalries, and an assessment of existing services. Similarly, in Pakistan
and Bangladesh, Islamist parties decide what social service to provide (and how to pro-
vide it) “mainly in response to the demands of the context” (Bano 2009, 25). Empirically
assessing why certain types of services are delivered in certain ways will require sys-
tematically disaggregating to the local level of Islamist politics, which has rarely been
systematically studied (Baker, et al. 2017). But doing so will likely help connect national-​
level variables such as economic crisis—​see below—​to subnational patterns of public
goods provision and political competition. For example, based on a spatial analysis of
mosque activism and infrastructure density across greater Cairo in the interwar pe-
riod, Neil Ketchley and Steven Brooke (2018) find that activists affiliated to the Muslim
Brotherhood were more likely to appear in mosques in areas distal to government
healthcare services. They then find that the Brotherhood went on to establish formal
(physical) medical services in those areas.
While the above points reveal some ways that Islamist social service provision is a
distinct type of nonstate activism, there are also important tie-​ins to broader research
agendas in comparative politics. For example, disaggregating Islamist social service pro-
vision by type is potentially useful in light of the recent interest in relaxing the “single
good” assumptions of clientelism. Albertus (2012), for example, frames a provider’s
choice between durable goods (such as land transfers) and more contingent types of
aid. He finds that providers target staunch supporters with contingent aid, while swing
supporters are furnished with more permanent transfers. Dynamics related to the
Muslim Brotherhood’s provision of healthcare in Egypt suggest the reverse: the group
disproportionately sited “brick and mortar” facilities (Cammett and Issar 2010) in
neighborhoods where their middle-​class core was predominant, while using intermit-
tent “medical caravans” to extend provision to poorer voters living in swing districts
(Brooke 2019). As this example implies, different types of social services—​even coming
from the same organization—​may generate different types of “linkages” between the
 Islamist Organizations and the Provision of Social Services    793

provider organization and the recipient (Kitschelt 2000), a point to which the final sec-
tion of this chapter returns.

Institutional Origins of Islamist

Social Service Provision

Islamist groups provide an extensive array of services that tangibly connect them
to millions of citizens. Simultaneously, in many cases, Islamist groups are political
opponents of extant authoritarian regimes. This frames an important research ques-
tion: Why do regimes that restrict the political space allow opposition movements to en-
gage in activism that reaches millions of citizens per year? Under what conditions does
Islamist social service provision emerge?
Prior researchers have outlined a type of “democratic bargain” that emerges in times
of economic stress, where political liberalization compensates citizens for economic
hardship (Brumberg 1992). It is also the case that these economic cuts do not always
occur evenly. For example, Huber, Mustillo, and Stephens (2008, 432) note how social
spending is differentially exposed to political and economic pressures, with social se-
curity spending being fairly insulated while the health and education sectors are more
exposed. These moments of state retrenchment and consequent reshuffling of bases of
support appear to open the space for Islamists to expand social service networks, despite
regimes that remain wary of their operations.
Egypt exemplifies this dynamic. Amid a dangerous series of economic reforms,
highlighted by the eruption of nationwide “Bread Riots” in 1977, Anwar el-​Sadat began
to encourage nonstate actors to expand their social service operations, including both
Christian and Muslim providers. The Muslim Brotherhood also benefitted from this
policy, seizing on the opportunity to establish what would become their largest and
oldest social service institution, the Islamic Medical Association. While this activism
posed a risk to el-​Sadat, such a broad-​based expansion of nonstate providers did help to
ease some of the social consequences of economic reform, and thus reduced the imme-
diate risk of further regime-​threatening uprisings (Brooke 2019).
While a full investigation is outside the scope of this chapter, other cases from the
region similarly suggest that economic crises play in important role in the emergence
of Islamic providers (Vandewalle 1992, 110). In Jordan, following a series of economic
reforms that began in the late 1980s, Baylouny argues that “Islamic NGOs spread into
new geographic areas and began to bridge the urban-​rural divide, employing urban
funds and mobilizing bases to do so. During this period Islamist charities increased
their presence in social services, providing higher quality and less-​expensive services
that the private sector or other philanthropic charities” (2006, 112–​113). The Algerian
case also helps tentatively substantiate a linkage between financial crisis, political
794   Steven Thomas Brooke

liberalization, and the expansion of Islamist social services. “Through an extensive net-
work of mosques, the [Islamist] FIS dispensed religious and socialization programs as
well as welfare and social services rarely provided by the government,” Imad Shahin
writes. “With more than three million followers, the party involved its members in vol-
untary activities, including collecting garbage, tutoring high school students, and of-
fering medical care services for needy patients” (1998, 138). Even pre–​Arab Spring Syria,
despite a hegemonic single-​party regime and repressive state apparatus historically hos-
tile to Islamic activism, the restructuring of the “Ba’thist Social Contract” gave Islamic
providers room to operate (Ruiz de Elvira and Zintl 2014). As Thomas Pierret and Kjetil
Selvik note in their study of the Zayd movement, “Starting in the late 1990s, the [Syrian]
government progressively relaxed its restrictions on charitable associations, which have
provided a useful—​although still relatively modest—​contribution through the growth
of private-​sector-​funded welfare schemes . . . . the growth of the Islamic charitable
sector speaks of a détente between the Ba’th regime and the Islamic trend that began in
the mid-​1990s” (2009, 596). There is, of course, a chasm between the emergence of such
providers and the mobilization of political support, especially in the Syrian context, but
it is telling that the pressures of economic reform incentivize initial steps toward liberal-
ization of the nonstate sector even in highly repressive contexts.
The Gulf monarchies offer one relevant comparison set. In those cases, the exist-
ence of natural resource rents and the codification of a robust social safety net as part
of the rentier bargain effectively foreclosed the emergence of nonstate Islamist service
providers (Freer 2018).2 However, to better trace mechanisms, it is helpful to expand
the scope of the investigation beyond the Arab world. In addition to its sheer size and
long period of developmental authoritarian rule, the Indonesian case is particularly
relevant because of the presence of Islamic social movements that reasonably approx-
imate those in the Arab countries. Nahdlatul Ulama (NU) and Muhammadiyah boast
long histories, millions of members, and the type of local networks that could plausibly
have mobilized electoral opposition to the New Order regime (Noer 1973). To the ex-
tent that these organizations could have used social service provision to generate oppo-
sition mobilization, but did not, they are important comparisons for theory building
and testing.
As Thomas Pepinsky (2014a, 2014b) suggests, an important reason for the “failure”
of these organizations to mount a social service–​based challenge seems to be the rela-
tively higher capacity that the New Order regime maintained throughout its lifecycle.
Higher regime capacity seemingly forestalls the development of opposition providers
through two mechanisms. First, it enables the regime to ensure that political mobili-
zation is channeled through regime institutions and takes place purely on the regime’s
terms. In Indonesia the key instrument for this strategy was the capacious Golkar, which
monopolized politics down to the local level and forced Islamic providers to accom-
modate to its rules or face severe sanctions (King and Rasjid 1988; Pepinsky 2014b).
Muhammadiyah’s relationship with the state exhibited similar dependencies, where
social service provision was allowed, but was also depoliticized to avoid interference
with potential opposition parties (Syamsuddin 1995; Fuad 2002). As Mitsuo Nakamura
 Islamist Organizations and the Provision of Social Services    795

describes Muhammadiyah’s relationship to the regime early in the New Order era, “so
long as the Muhammadiyah kept its activities to religious propagation, education, and
social welfare, the local government authorities did not intervene” (2012, 363).
Beyond policing the political arena, higher-​capacity regimes can use access to funding
streams and positions in the state bureaucracy to co-​opt these providers.3 As Porter
describes the relationship for Indonesia’s other large Islamic organization, “Nahdlatul
Ulama’s policies were accommodative to secular political authority in exchange for pe-
ripheral influence in the cabinet, control of the ministry of religion and the thousands
of patronage positions within it, and economic advantages given to its leaders in their
business ventures” (2004, 30–​31).4 Any movement toward opposition would risk access
to these resources, which Nahdlatul Ulama discovered in the 1970s. At that time, amid
increasing criticism of the New Order regime, Suharto took the Ministry of Religion out
of their hands, which sapped revenue for their local social service networks. In the mid-​
1980s, however, a new generation of NU leaders revised their criticisms and returned to
New Order’s good graces (Bush 2009, 70–​81).
By co-​opting these challengers, the New Order regime was able prevent the emer-
gence of the type of independent provision that could fuel an alternative electoral
challenge. However, following the collapse of the New Order regime in the late 1990s,
political constituencies became scrambled (Pepinsky 2009). This opened up an opportu-
nity for service provision as a strategy of electoral mobilization, particularly because the
financial crisis that ended the regime also damaged welfare provision and created space
for new providers (Sciortino et al. 2010; Kusujiarti 2011). In particular, Islamic organiza-
tions became less tied to the state (for instance by reorienting around local fundraising),
and thus more available as potential bases that political activists potentially could capi-
talize upon (Fuad 2002). Indeed, post-​democratization, one Islamist political party, the
Prosperous Justice Party (Partai Keadilan Sejahtera, or PKS), used social service pro-
vision to break through strong social networks and expand its support, particularly in
poorer and more remote areas (Hamayotsu 2011, 2015). These developments align the
Indonesian case with other young and unconsolidated democracies, whereby social
services provide a valuable tool for parties to use to extend their appeal beyond core ide-
ological or programmatic supporters (Cammett 2014; Thachil 2014a; Brooke 2019).
While the Indonesian case helps to identify how regime capacity limits opposition
social service activism, Malaysia suggests an alternative pathway. In some ways, it
resembles the Middle Eastern cases rather than the Indonesian one, because an eco-
nomic crisis punctuated, but did not end, the regime (Pepinsky 2009). This generates
an expectation that Islamist parties would be able to exploit the loosening of constraints
to expand their provision, and thus challenge the United Malay National Organization
(UMNO), the dominant party (Mueller 2014, 16–​ 24). Yet this has generally not
happened. Instead, Islamist parties, with the Malaysian Islamic Party (Parti Islam Se-​
Malaysia, or PAS) as the most prominent example, have effectively remained regional
parties with limited appeal. What, then, differentiates the Malaysian case from the Arab
ones? Why did the Malaysian regime lose fiscal capacity yet not face more of a challenge
from Islamist providers?
796   Steven Thomas Brooke

The likely reason lies in Malaysia’s ethno-​religious cleavages. In the early 1970s, the
Malaysian regime began to devote increasing amounts of resources to the Malay (Muslim)
population at the expense of non-​Malays, including Chinese. One aspect of this policy
change was Islamization of the state bureaucracy and, in general, a consolidation of the
regime along religious lines (Liow 2009). This tightly linked UMNO to a Muslim core
and created a dominant party that, in James C. Scott’s words, was “a well-​organized and
well-​financed political machine providing both individual and collective blandishments
that reach into every Malay village” (1985, 57).5 Beyond eliminating the opportunity for
alternative providers (such as PAS), state-​level Islamization became a source of consid-
erable patronage for local Islamic service providers (Shamsul 1983; Hamayotsu 2004).
Thus, even after the financial crisis, potential opposition providers in the Malay constitu-
ency remained tied in to the regime’s largesse, rather than defecting toward the opposition
(Hamayotsu 2015). As Hassan describes, “these Islamic NGOs therefore seldom if ever
publicly express critical or evaluative opinions on government policies touching upon
Islam and the Muslim Community” (2002, 102). Indeed, UMNO’s material hegemony
over the Malay population has caused the PAS, which might be expected to use social ser-
vice provision to challenge the regime, to basically abandon the field of distributional pol-
itics and instead attempt to “outbid” the regime over questions of Islamic identity (Liow
2009).6 To the extent that the Malaysian case offers a generally different set of cleavages
than exist in the Arab world, it highlights important additional factors to consider.
Even in the most repressive regimes, religious social service providers can find
the space to operate. But a brief investigation of majority-​Muslim countries in the
Middle East and Southeast Asia suggests that the main constraint on these providers
transforming their activism into political gains is the capacity of the regime. In the Arab
world, the onset of financial crisis in the 1970s and 1980s forced nondemocracies to
cede space to Islamist providers, while, in contrast, Indonesia’s fiscal health allowed it
to co-​opt potential challengers until the collapse of the New Order. Yet, as the Malaysian
case suggests, an ethno-​religious cleavage that overlaps with regime constituencies also
influences how social service provision unfolds.

Islamist Social Service Provision and

Party-​Voter Linkages

Social service provision attracts considerable scholarly attention for how it can influ-
ence the various beliefs and behaviors of recipients, up to and including generating po-
litical support for the provider organization. This is perhaps the area where the recent
study of Islamist social service provision has made the strongest contributions, both in
terms of empirically documenting how receipt of Islamist-​provided services influences
behavior, but also using these cases to build theories relevant to scholars of comparative
political behavior more broadly.
 Islamist Organizations and the Provision of Social Services    797

One prominent argument is that Islamic identity functions as a more-​or-​less rational

mechanism that helps citizens learn more about the policies political parties stand for
(Pepinsky, Liddle, and Mujani 2012). Tarek Masoud (2014) extends this to Islamists’
social service provision, showing how individuals embedded in Islamic institutions,
including social service ones, are more convinced that Islamist parties are economi-
cally redistributive. Other authors deemphasize these information-​based arguments
to identify how social service provision leads to ideological, and often religious, attitu-
dinal change. For instance, Carrie Wickham, in her study of Islamic social organizing
in Egypt, suggests that continual exposure to a “parallel Islamic sector” of businesses,
mosques, and social service organizations can “change the preferences of educated
youth” (2002, 148). A third mechanism centers on how provision conveys broadly de-
sirable traits about providers separate from programmatic stances. Bano hypothesizes
about the Jamaʿat-​i Islami, stating that “if an individual builds credibility through good
social service work and gradually becomes known to be engaged with a specific party, it
is expected that he can lend that party credibility” (2009, 33). For Szekely (2015), Hamas’s
social service activism is a form of “political advertising” that communicates informa-
tion about bureaucratic competence to constituents. Melani Cammett and Pauline
Jones Luong (2014) suggest similar emotional and sentiment-​based roles for social ser-
vice provision. And through a survey experiment, I show that individuals indeed use
their experiences with Islamist social service providers to form impressions about the
traits of Islamist candidates, including their honesty, competence, and modesty (2017a).
Regardless of the mechanisms connecting social service provision to political
outcomes, the ways in which these outcomes are conditional on various individual-​level
characteristics of the recipient are not well understood. For example, many of the above-​
mentioned studies are built on a general assumption that experience with or exposure
to these types of social service providers is more or less randomly distributed across the
population. Because the citizens who opt in to these providers are not systematically
different from those who opt out, the subsequent effects of this provision are produced
in some way by exposure to the provision itself. This can be better substantiated (or
challenged) by more careful empirical work. Fully accounting for the ways that various
effects are jointly produced by exposure to social service networks and individual recip-
ient characteristics is an important task for future research.7
Among potential differential effects, social class and religiosity are worth considering
in more depth. There is a general assumption that Islamist social service networks ex-
tend among the poor, often with the assumption that these efforts exist as a subtype of
charity. Mishal and Sela, for example, discuss the Palestinian Islamist group Hamas’s so-
cial service network, arguing that “the Brotherhood’s traditional practice of applying the
Islamic duty of charity to the poor” is a reason why that sector tends to support the group
(2006, 20). Yet the empirical record on this count is mixed. Whatever impulse may drive
providers to focus on a particular segment of the population, many have found that
pressures for sustainability force provision mainly toward a class of customers who can
afford to pay for what they receive (Brooke 2019). In her pioneering study of Islamic
NGOs, Janine Clark (2004) found that quotidian needs to receive permits, hire staff,
798   Steven Thomas Brooke

obtain materials, and ensure funding meant that Islamic providers operated by and for
the middle class. Indeed, she refers to these as “Islamic Commercial Institutions.” One
World Bank study found that, in Indonesia, Muhammadiyah has increasingly oriented
its services toward a middle-​class audience to cope with limited financial resources.
As the authors put it, “Aware that . . . it cannot compete with high-​ranking, for profit
hospitals for upper class clientele, [Muhammadiyah] has identified as its “market seg-
ment” people with lower-​middle and middle incomes” (Sciortino et al. 2010, 29). The
Jamaʿat-​i Islami’s initiatives in Pakistan were not free, but “less than the market rate and
the respondents considered that there was a better guarantee of the quality of the serv-
ices than available from other providers, especially the public sector, where the quality
of services is doubtful . . . . The same was the case in Bangladesh, where most services are
provided at twenty-​five percent below the market price . . . the Jamaʿat prefers to pro-
vide subsidized rather than free services because . . . it is conscious of the need to ensure
sustainability” (Bano 2009, 20). While in some cases a robust social movement affiliate,
or access to outside resources, can sustain provision targeted at an indigent population,
most religious providers seem forced to accommodate doctrinal mandates with mun-
dane material considerations.
While the ability of structural (class-​based) factors to condition the effect of Islamist
provision on political outcomes is relatively more developed in the literature, the po-
tential influence of religion is far less so. For example, much of the literature on party-​
voter linkages outside the Islamic world emphasizes monitoring (Stokes 2005). In
parallel, however, scholars have theorized how norms, such as reciprocity and fair play,
likely also ensure provision translates into political support (Auyero 1999; Finan and
Schechter 2012; Lawson and Greene 2014). This research agenda can likely benefit from
a closer engagement with the specific case of Islamist social service providers, specifi-
cally considering how their religious identity may trigger such informal mechanisms.
For example, Hamayotsu (2015, 162) identifies how the Islamist PKS’s social service
provision is embedded in a larger web of religious solidarity and activities, which in
turn influences a sense of reciprocity and feeling of community. By extension, because
nonreligious providers cannot rely on the webs of social and physical infrastructure
available to religious groups, they may be forced to make the exchange more explicit.
Rebecca Weitz-​Shapiro (2014) shows how, among certain audiences, clientelism
produces both a moral revulsion and a suspicion of administrative incompetence.
Islamist groups may have an advantage in escaping these connotations because their
social service provision simultaneously exists as a manifestation of religious activism
as well as political ambition. As one scholar of Hizbullah puts it, “the activities of
Hizbullah’s Social Unit emanate from a strong ideological commitment because so-
cial service is a fundamental tenet of the faith” (Hamzeh 2004, 53). Another scholar
describes the Jordanian Muslim Brotherhood’s social service provision as “one of the
basic principles of the Muhammadan mission . . . a product of this Islamic culture and
the values and ethics that are derived from the Islamic Sharia and the Arab-​Islamic her-
itage” (Hammad 1997, 187). Bano found similar motivations for the Jamaʿat-​i Islami in
Pakistan and Bangladesh: “it was clear that spiritual rewards were the main incentives
 Islamist Organizations and the Provision of Social Services    799

motivating many members to volunteer to manage the Jamaʿat welfare organization”

(2012, 89). And members of the Egyptian Muslim Brotherhood describe their social ser-
vice activism similarly. Amr Darrag, a former minister in Mohammed Morsi’s govern-
ment, explains that the group’s interests in social service provision “are not traditionally
the most expedient or the most utilitarian, but rather are principled and faith-​based”
(2017, 222).
On precisely this point, Marie Vannetzel (2016) identifies a “symbolic economy of
disinterestedness” that shapes social service provision by Islamist actors in Egypt. As she
shows, the public (and to some extent performative) disavowal of political motives for
service provision in favor of reference to religious motivations is, paradoxically, part of
the reason why it is effective at generating political support for the organization. Put dif-
ferently, because Islamist actors are expected to provide services because of their religious
identity, it simultaneously helps them to escape the taint of clientelism (and potentially
help their recipients cognitively disassociate from the exchange as well). The implication
is that because non-​Islamist providers cannot do this, they similarly struggle to distance
themselves from the negative connotations of traditional exchange.
Provider religiosity likely also functions subtly, conditioning the set of expectations
that individuals bring to specific interactions (Pepinsky et al. 2012, 587). In this con-
struct, customers may perceive the service offered by Islamists to be better, or believe
that it says something about Islamists that it does not say about non-​Islamists, thus
generating a unique political effect for Islamist providers despite technical similarities
(or differences) across different provider organizations. As Asef Bayat concludes,
“many private schools and clinics chose Islamic tags even though they differed little
from non-​Islamic institutions of the same type” (2007a, 149). Suheir Morsy calls this
“the placebo effect provided by religious symbolism” (1988, 366). Latief hints at this
in his study of health providers in Indonesia, noting how, “despite the preservation of
Islamic identity, there is no particular distinction between Islamic clinics and other
‘secular’ private or government clinics” (2010, 508). One implication of this argu-
ment is the need to isolate the particular characteristics of the social service provision
that is responsible for generating the effect of interest, separating technical quality
from atmosphere by comparing religious and nonreligious providers (Cammett and
Şaşmaz 2017).
To extend this argument, when recipients form judgments about particular services
received, and are thus motivated toward some behavioral or attitudinal outcome, they
are almost certainly doing so based on an implicit comparison with another provider.
In many cases, this provider is the struggling public sector. As Soliman quips, “the word
public in Egypt today is synonymous with mediocrity” (2011, xiv).8 Or, in his aforemen-
tioned study of the Indian BJP, Tariq Thachil finds that the “very basic standard of serv-
ices offered by the VKA [the BJP’s movement affiliate] is that such services are appealing
only in areas where even basic public provisioning is absent, and even then only to those
voters who cannot pay to obtain these services privately” (2011, 446). Given that a de-
cline in state capacity is an important factor in the emergence of opposition providers,
this is a natural comparison set. However, it also raises the possibility that an opposition
800   Steven Thomas Brooke

organization’s ineffectiveness at converting social service provision to political capital

is not only due to the high-​capacity regime’s ability to depoliticize and co-​opt that pro-
vider, but also because the provider’s efforts simply do not stand out when set against a
well-​functioning public sector.
Given the importance of declining state capacity and the consequent decay in public
services, in many cases a plausible alternative provider could be the state itself. Yet this
raises an additional question: If public services are so poor, why is it that Islamist so-
cial services seem to stand out? In Indonesia, for instance, Hamayotsu suggests that
providers tied to the PKS benefit because “their reputation and impact are in stark con-
trast with some of the other similar institutions and programs that are almost bankrupt
or known for corruption and inefficient management” (2015, 163). But if multiple pro-
vider organizations exist, why are religious providers able to offer services that seem-
ingly stand out for their high quality, lack of corruption, or efficient operation? One
way to find out, presumably, would be to not only establish more rigorous comparisons
against the state, but against alternative nonstate providers as well.
The emphasis on comparisons and, in particular, individual-​level factors provides im-
portant insight into the question that motivated the outset of this chapter—​what makes
social service provision by Islamist organizations unique? The evidence presented implies
that Islamist provision is different perhaps because recipients come to it with different
perceptions and assumptions than they do with non-​Islamist organizations. To move be-
yond anecdotal evidence, and toward systematic theorization, will require rigorous com-
parison against competitors across both organizational and individual level factors.

Conclusion

Across the Islamic world a variety of religious organizations, from political parties to
social movements, provide an array of social services to millions of citizens per year.
This chapter reviews this growing literature and makes three major arguments. First, it
attempts to define Islamist social service provision and identify what makes it a unique
case of nonstate provision. While religious factors do not seem to be the source of
Islamists’ purported advantage, the variation in type and method of social service de-
livery is notable. Islamist providers are also somewhat unique in that they exist under au-
thoritarian regimes that deliberately constrict the political space for the opposition. The
second contribution of this chapter is to theorize why this is so, in the process making
a broader contribution to the study of the political economy of authoritarian durability.
Finally, the chapter identifies how and why the receipt of Islamist social services shifts
the beliefs and behaviors of recipients. While this provides new insights into the role of
perceptions and expectations as part of a broader research agenda in comparative polit-
ical behavior, it also suggests that the demand side—​in terms of the assumptions ordi-
nary people hold about Islamist providers—​is likely a key factor distinguishing Islamist
provision from that engaged in by non-​Islamist actors.
 Islamist Organizations and the Provision of Social Services    801

Two areas emerge as particularly relevant for additional theoretical development and
empirical documentation. First, the conditions under which Islamists provide one type
of social service, and deliver it in a particular way, are likely relevant for better under-
standing the political effects of such provision, but are themselves poorly understood.
Increasing focus on the local dynamics of Islamist activism will likely yield rele-
vant explanations, including spatial and temporal variation inside the same provider.
Second, provision is not only filtered through recipients’ perceptions about the service
they receive (and the provider itself), but also based on an implicit comparison between
the Islamic provider and the plausible alternative. To the extent that there is a consider-
able role for subjectivity in the political manifestations of Islamic groups’ provision of
social services, it stands out as a considerable gap in our knowledge.

Notes
1. For example, in the realm of health, the Brotherhood’s Islamic Medical Association
was a formally registered, independent organization with dozens of brick-​and-​mortar
facilities around the country (Brooke 2019). And while they too were “brick and mortar”
institutions, the Brotherhood’s schools operated on a mostly different basis than the med-
ical facilities: instead of being registered under the umbrella of a formal, legal NGO, the
Brotherhood’s schools were mostly operated as private businesses of specific individuals.
Interestingly, these divergent modes of legal operation complicated the state’s crackdown
on the group after 2013 (Brooke 2017a).
2. Although, in some cases, Islamists still managed to influence social service delivery inside
the state itself, particularly in the realm of education.
3. State capacity is one variable that influences the extent to which alternative providers
can use their provision to generate mobilization. But it is not fully satisfying; despite de-
clining fiscal capacity, the Egyptian regime remained able to co-​opt a prominent Islamic
provider, the sprawling Islamic organization al-​Gamʿiyya al-​Sharʿiyya. As with the NU in
Indonesia, successive Egyptian regimes enlisted al-​Gamʿiyya al-​Sharʿiyya to help operate
the country’s religious sector, including the administration of mosques and oversight of
charity funds (zakat). In return, al-​Gamʿiyya al-​Sharʿiyya enrolled their networks of mostly
poor beneficiaries on behalf of the regime. As Sarah Ben Nefissa, puts it, the group provided
“a base of support for the political regime and, especially, the party in power, the NDP”
(2002, 149).
4. NU leaders, including Abdurrahman Wahid, recognized this lack of financial indepen-
dence as a key constraint on their ability to operate their network of religious schools
(Porter 2004, 112).
5. As he goes on to note, the Islamist PAS “cannot offer any of the material incentives that
UMNO’s control of the government provides” (Scott 1985, 58).
6. The Malaysian case is also interesting in light of Corstange’s (2016) work on ethnicity and
clientelism. As he argues, ethnic groups dominated by a single provider generally receive
lower levels of goods provision than those that are internally competitive. To the extent that
UMNO dominates the Muslim community, it would suggest that alternative providers may
have the opportunity to mobilize support among UMNO’s dissatisfied base. This, appar-
ently, has not happened.
802   Steven Thomas Brooke

7. Better identifying these interactions is also highly important given potential issues of
endogeneity whereby receipt of social services is an effect, rather than cause, of political al-
legiance. Cammett, for example, finds that in Lebanon, political partisans are more likely to
receive services than nonpartisans, and also that the baskets of benefits that they receive are
comparatively greater. As she summarizes, “access to diverse packages of social assistance
entail distinct outlays of political activism” (2011, 73).
8. The historic Egyptian case suggests that the relationship may in fact be endogenous, in that
when opposition candidates win, the Mubarak government was known to withhold serv-
ices from those districts (El-​Borai 1996, 36–​43; Blaydes 2011, chap. 4).

References
al-​Anani, Khalil. 2016. Inside the Muslim Brotherhood: Religion, Identity, Politics. New York:
Oxford University Press.
Albertus, Michael. 2012. “Vote Buying with Multiple Distributive Goods.” Comparative
Political Studies 46, no. 9: 1082–​1111.
Anria, Santiago. 2018. When Movements Become Parties: The Bolivian MAS in Comparative
Perspective. New York: Cambridge University Press.
Auyero, Javier. 1999. “From the Client’s Point(s) of View: How Poor People Perceive and
Evaluate Political Clientelism.” Theory and Society 28, no. 2: 297–​334.
Aytaç, S. Erdem. 2014. “Distributive Politics in a Multiparty System: The Conditional Cash
Transfer Program in Turkey.” Comparative Political Studies 47, no. 9: 1211–​1237.
Azra, Azyumardi, Dina Afrianty, and Robert W. Hefner. 2007. “Pesantren and Madrasa: Muslim
Schools and National Ideals in Indonesia.” In Schooling Islam: The Culture and Politics of
Modern Muslim Education, edited by Robert W. Hefner and Muhammad Qasim Zaman.
Princeton, NJ: Princeton University Press.
Baker, Lauren, Kristen Kao, Marc Lynch, and Ellen Lust, eds. 2017. Local Politics and Islamist
Movements. POMEPS Studies no. 27. Washington, DC: Project on Middle East Political
Science. https://​pomeps.org/​wp-​content/​uploads/​2017/​09/​POMEPS_​Studies_​27_​Web.pdf.
Bano, Masooda. 2009. “Market of Identity: Religious Political Parties and Welfare Work—​the
Case of Jama’at-​i-​Islami in Pakistan and Bangladesh.” Religions and Development Research
Programme Working Paper 34.
Bano, Masooda. 2012. “Welfare Work and the Politics of Jamaʿat-​i-​Islami in Pakistan and
Bangladesh.” Economic and Political Weekly 47, no. 1: 86–​93.
Baylouny, Anne Marie. 2006. “Creating Kin: New Family Associations as Welfare Providers in
Liberalizing Jordan.” International Journal of Middle East Studies 38, no. 3: 349–​368.
Ben Nefissa, Sarah. 2002. “Citoyennete Morale En Egypte: Une Association Entre Etat et Freres
Musulmans (Moral Citizenship in Egypt: An Association between the State and the Muslim
Brothers).” in Pouvoirs et Associations Dans Le Monde Arabe, edited by Sarah Ben Nefissa,
147–​179. Paris: CNRS Editions.
Blaydes, Lisa. 2011. Elections and Distributive Politics in Mubarak’s Egypt. Cambridge: Cambridge
University Press.
Brooke, Steven. 2017a. “From Medicine to Mobilization: Social Service Provision and the
Islamist Reputational Advantage.” Perspectives on Politics 15, no. 1: 42–​61.
Brooke, Steven. 2017b. “Egypt.” In Rethinking Political Islam, edited by Shadi Hamid and Will
McCants, 17–​31. New York: Oxford.
 Islamist Organizations and the Provision of Social Services    803

Brooke, Steven. 2019. Winning Hearts and Votes: Social Services and the Islamist Political
Advantage. Ithaca, NY: Cornell University Press.
Brumberg, Daniel. 1992. “Survival Strategies vs. Democratic Bargains: The Politics of Economic
Reform in Contemporary Egypt.” In The Politics of Economic Reform in the Middle East, ed-
ited by Henri J. Barkey, 73–​104. New York: St. Martin’s Press.
Bueno, Natalia Salgado. 2018. “Bypassing the Enemy: Distributive Politics, Credit Claiming,
and Nonstate Organizations in Brazil.” Comparative Political Studies 51, no. 3: 304–​340.
Bush, Robin. 2009 Nahdlatul Ulama and the Struggle for Power within Islam and Politics in
Indonesia. Singapore: ISEAS.
Cammett, Melani. 2011. “Partisan Activism and Access to Welfare in Lebanon.” Studies in
Comparative International Development 46, no. 1: 70–​97.
Cammett, Melani. 2014. Compassionate Communalism: Welfare and Sectarianism in Lebanon.
Ithaca, NY: Cornell University Press.
Cammett, Melani, and Sukriti Issar. 2010. “Bricks and Mortar Clientelism: Sectarianism and
the Logics of Welfare Allocation in Lebanon.” World Politics 62, no. 3: 381–​421.
Cammett, Melani, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17: 187–​206.
Cammett, Melani, and Aytuğ Şaşmaz. 2017. “Political Context, Organizational Mission, and the
Quality of Social Services: Insights from the Health Sector in Lebanon.” World Development
98: 120–​132.
Challand, Benoît. 2008. “A Nahda of Charitable Organizations? Health Service Provision and
the Politics of Aid in Palestine.” International Journal of Middle East Studies 40, no. 2: 227–​247.
Clark, Janine A. 2004. Islam, Charity, and Activism: Middle Class Networks and Social Welfare
in Egypt, Jordan, and Yemen. Bloomington: Indiana University Press.
Corstange, Daniel. 2016. The Price of a Vote in the Middle East. New York: Cambridge
University Press.
Darrag, Amr. 2017. “Politics or Piety? Why the Muslim Brotherhood Engages in Social Service
Provision.” In Rethinking Political Islam, edited by Shadi Hamid and Will McCants, 218–​230.
Oxford: Oxford University Press.
Davis, Nancy, and Robert V. Robinson. 2012. Claiming Society for God: Religious Movements
and Social Welfare. Bloomington: Indiana University Press.
El-​Borai, Negad, ed. 1996. Democracy Jeopardized: Nobody “Passed” the Elections. Cairo: Egyptian
Organization for Human Rights.
Fealy, Greg, and Robin Bush. 2014. “The Political Decline of Traditional Ulama in Indonesia.”
Asian Journal of Social Science. 42, no. 5: 536–​560.
Finan, Frederico, and Laura Schechter. 2012. “Vote-​Buying and Reciprocity.” Econometrica 80,
no. 2: 863–​881.
Freer, Courtney. 2018. Rentier Islamism: The Influence of the Muslim Brotherhood in Gulf
Monarchies. New York: Cambridge University Press.
Fuad, Muhammad. 2002. “Civil Society in Indonesia: The Potential and Limits of Muhammadiyah.”
Sojourn: A Journal of Social Issues in Southeast Asia. 17, no. 2: 133–​163.
Hamayotsu, Kikoue. 2004. “Islamisation, Patronage, and Political Ascendancy: The Politics
and Business of Islam in Malaysia.” In The State of Malaysia: Ethnicity, Equity and Reform,
edited by Terrence Edmund Gomez, 229–​252. London: Routledge.
Hamayotsu, Kikoue. 2011. “The Political Rise of the Prosperous Justice Party in Post-​
Authoritarian Indonesia: Examining the Political Economy of Islamist Mobilization in a
Muslim Democracy.” Asian Survey. 51, no. 5: 971–​992.
804   Steven Thomas Brooke

Hamayotsu, Kikoue. 2015. “Patronage, Welfare Provisions, and State-​Society Relations: Lessons
from Muslim-​Dominant Regimes in Southeast Asia (Indonesia and Malaysia).” In Religion
and the Politics of Development, edited by Philip Fountain, Robin Bush, and R. Michael
Feener, 155–​176. London: Palgrave Macmillan.
Hamid, Shadi. 2014. Temptations of Power: Islamists and Illiberal Democracy in a New Middle
East. Oxford: Oxford University Press.
Hammad, Walid. 1997. “‘Islamists and Charitable Work in Jordan: The Muslim Brotherhood
as a Model.’” In Islamic Movements in Jordan, edited by Hani Hourani and Jillian Schwedler,
169–​194. Amman, Jordan: Markaz al-​Urdun al-​Jadid.
Hamzeh, Ahmad Nizar. 2004. In the Path of Hizbullah. Syracuse, NY: Syracuse University Press.
Hassan, Saliha. 2002. “Islamic Non-​governmental Organizations.” In Social Movements in
Malaysia: From Moral Communities to NGOs, edited by Saliha Hassan and Meredith Weiss,
97–​114. London: Routledge.
Hefner, Robert W. 2008. “Islamic Schools, Social Movements, and Democracy in Indonesia.”
in Making Modern Muslims: The Politics of Islamic Education in Southeast Asia, edited by
Robert W. Hefner, 55–​105. Honolulu: University of Hawai‘i Press.
Huber, Evelyne, Thomas Mustillo, and John D. Stephens. 2008. “Politics and Social Spending in
Latin America.” Journal of Politics 70, no. 2: 420–​436.
Kandil, Hazim. 2014. Inside the Brotherhood. London: Polity. 2014.
Ketchley, Neil, and Steven Brooke. 2018. “Islamist Politicization of Religious Infrastructure:
Evidence from Interwar Cairo.” Paper presented at the 2018 Annual Meeting of the American
Political Science Association.
King, Dwight Y., and M. Ryass Rasjid. 1988. “The Golkar Landslide in the 1987 Indonesian
Elections: The Case of Aceh.” Asian Survey 28, no. 9: 916–​925.
Kitschelt, Herbert. 2000. “Linkages between Citizens and Politicians in Democratic Polities.”
Comparative Political Studies 33, no. 6–​7: 845–​879.
Kusujiarti, Siti. 2011. “Pluralistic and Informal Welfare Regime: The Roles of Islamic Institutions
in the Indonesian Welfare Regime.” In The Sociology of Islam: Secularism, Economy, and
Politics, edited by Tugrul Keskin, 419–​452. Reading, UK: Ithaca Press.
Latief, Hilman. 2010. “Health Provision for the Poor: Islamic Aid and the Rise of Charitable
Clinics in Indonesia.” South East Asia Research 18, no. 3: 503–​553.
Lawson, Chappell, and Kenneth F. Greene. 2014. “Making Clientelism Work: How Norms of
Reciprocity Increase Voter Compliance.” Comparative Politics 47, no. 1: 61–​85.
Liow, Joseph. 2009. Piety and Politics: Islamism in Contemporary Malaysia. New York: Oxford
University Press.
Marschall, Melissa, Abdullah Aydogan, and Alper Bulut. 2016. “Does Housing Create Votes?
Explaining the Electoral Success of the AKP in Turkey.” Electoral Studies 42: 201–​212.
Masoud, Tarek. 2014. Counting Islam: Religion, Class, and Elections in Egypt. New York:
Cambridge University Press.
Mishal, Shaul, and Avraham Sela. 2006. The Palestinian Hamas: Vision, Violence, and Coexistence.
New York: Columbia University Press.
Morsy, Soheir A. 1988. “Islamic Clinics in Egypt: The Cultural Elaboration of Biomedical
Hegemony.” Medical Anthropology Quarterly 2, no. 4: 355–​369.
Mueller, Dominik M. 2014. Islam, Politics, and Youth in Malaysia: The Pop-​Islamist Reinvention
of PAS. London: Routledge.
Nakamura, Mitsuo. 2012. The Crescent Arises over the Banyan Tree: A Study of the Muhammadiyah
Movement in a Central Javanese Town, c. 1910s–​2010. 2nd ed. Singapore: ISEAS.
 Islamist Organizations and the Provision of Social Services    805

Noer, Deliar. 1973. The Modernist Muslim Movement in Indonesia, 1900–​1942. Singapore: Oxford
University Press.
Pepinsky, Thomas B. 2009. Economic Crises and the Breakdown of Authoritarian Regimes:
Indonesia and Malaysia in Comparative Perspective. Cambridge: Cambridge University Press.
Pepinsky, Thomas. 2014a. “Political Islam and the Limits of the Indonesian Model.” Taiwan
Journal of Democracy 10, no. 1: 105–​121.
Pepinsky, Thomas. 2014b. “Social Services and Political Islam in Southeast Asia: Two Failures.”
Paper presented at the Islamist Social Services workshop, September 23, 2014. https://​
pomeps.org/​social-​services-​and-​political-​islam-​in-​southeast-​asia-​two-​failures.
Pepinsky, Thomas B., R. William Liddle, and Saiful Mujani. 2012. “Testing Islam’s Political
Advantage: Evidence from Indonesia.” American Journal of Political Science 56, no.
3: 584–​600.
Porter, Donald J. 2004. Managing Politics and Islam in Indonesia. London: Routledge.
Ruiz de Elvira, Laura, and Tina Zintl. 2014. “The End of the Ba’thist Social Contract in Bashar
al-​Asad’s Syria: Reading Sociopolitical Transformations through Charities and Broader
Benevolent Activism.” International Journal of Middle East Studies 46: 329–​349.
Sciortino, Rosalia, Neni Ridarineni, and Brahmaputra Marjadi. 2010. “Caught between Social
and Market Considerations: A Case Study of Muhammadiyah Charitable Health Services.”
Reproductive Health Matters 18, no. 36: 25–​34.
Scott, James C. 1985. Weapons of the Weak. New Haven, CT: Yale University Press.
Shahin, Emad. 1998. Political Ascent: Contemporary Islamic Movements in North Africa.
Boulder, CO: Westview.
Shamsul, A.B. 1983. “The Politics of Poverty Eradication: The Implementation of Development
Projects in a Malaysian District.” Pacific Affairs 56, no. 3: 455–​476.
Soliman, Samer. 2011. The Autumn of Dictatorship: Fiscal Crisis and Political Change in Egypt
under Mubarak. Palo Alto, CA: Stanford University Press.
Stokes, Susan C. 2005. “Perverse Accountability: A Formal Model of Machine Politics with
Evidence from Argentina.” American Political Science Review 99, no. 3: 315–​325.
Syamsuddin, M. Din. 1995. “The Muhammadiyah Da‘wah and Allocative Politics in New Order
Indonesia.” Studika Islamika 2, no. 2. http://​journal.uinjkt.ac.id/​index.php/​studia-​islamika/​
article/​view/​834.
Szekely, Ora. 2015. “Doing Well by Doing Good: Understanding Hamas’s Social Services as
Political Advertising.” Studies in Conflict and Terrorism 38, no. 4: 275–​292.
Tadros, Mariz. 2011. “The Securitisation of Civil Society: A Case Study of NGOs–​State Security
Investigations (SSI) Relations in Egypt.” Conflict, Security & Development 11, no. 1: 79–​103.
Thachil, Tariq. 2011. “Embedded Mobilization.” World Politics 63, no. 3: 434–​469.
Thachil, Tariq. 2014a. “Elite Parties and Poor Voters: Theory and Evidence from India.”
American Political Science Review 108, no. 2: 454–​477.
Thachil, Tariq. 2014b. Elite Parties, Poor Voters: How Social Services Win Votes in India.
Cambridge: Cambridge University Press.
Tugal, Cihan. 2013. “Contesting Benevolence: Market Orientations among Muslim Aid
Providers in Egypt.” Journal of Qualitative Sociology 36: 141–​159.
Van Cott, Donna Lee. 2005. From Movements to Parties in Latin America: The Evolution of
Ethnic Politics. Cambridge: Cambridge University Press.
Vandewalle, Dirk. 1992. “Ben Ali’s New Era: Pluralism and Economic Privatization in Tunisia.”
In The Politics of Economic Reform in the Middle East, edited by Henri J. Barkey, 105–​126.
New York: St. Martin’s Press.
806   Steven Thomas Brooke

Vannetzel, Marie. 2016. “Confronting the Transition to Legality.” In Egypt’s Revolutions: Politics,
Religion, and Social Movements, edited by Bernard Rougier and Stéphane Lacroix, 41–​61.
Basingstoke, UK: Palgrave Macmillan.
Weitz-​Shapiro, Rebecca. 2014. Curbing Clientelism in Argentina: Politics, Poverty, and Social
Policy. Cambridge: Cambridge University Press.
White, Jenny B. 2012. “Islamist Social Networks and Social Welfare Services in Turkey.” In
Islamist Politics in the Middle East: Movements and Change, edited by Samer Shehata, 59–​67.
London: Routledge.
Wickham, Carrie. 2002. Mobilizing Islam. New York: Columbia University Press.
Wiktorowicz, Quintan. 2000. “Civil Society as Social Control: State Power in Jordan.”
Comparative Politics 33, no. 1: 43–​61.
 Chapter 39

Expl oring t h e Rol e

of Isl am i n Ma l i
Service Provision, Citizenship, and Governance

Jaimie Bleck and Alex Thurston

Why would clerics who demonstrate an explicit interest in politics, sit atop formi-
dable infrastructures of alternative service provision, and repeatedly mobilize followers
around political demands refrain from directly seeking elected office or other means
of capturing the state? In this chapter, we examine clerics’ roles in politics and govern-
ance in Mali, where the country’s leading clerics have pursued both competition and
complementarity with the state. We argue that even at the height of contention between
mainstream Muslim clerics and the state, both clerics and elected politicians share an
underlying understanding about the limits of clerics’ ambitions. For clerics, refraining
from office-​seeking allows them to influence policy and gain personal power without
sacrificing their reputational advantage (Cammett and Jones Luong 2014) and popular
appeal by risking a total immersion into politics. However, the arrival of jihadist groups,
whose repertoires of service provision explicitly challenge the state, is a strong depar-
ture from these existing patterns of political engagement. We conclude with a discussion
speculating on the ways that jihadist presence could challenge patterns of accommoda-
tion between religious leaders and the state.
In many other Muslim-​majority contexts in the Middle East, North Africa, and South
Asia, Islamic service provision has been used to fuel electoral support for Islamic parties
or Islamic candidates running as independents. If governments underperform in vital
sectors such as public health or education, religious providers can fill the void and dem-
onstrate their comparative ability to reach the population with essential goods and serv-
ices (Brooke 2019, 2021; Cammett and Jones Luong, 2014; Cammett and MacLean 2011;
Cammett and Issar 2010; Harrigan and el-​Said 2009). Quality service provision can
build groups’ reputational advantage at the ballot box, offering insight into the compas-
sionate way that these groups would govern (Brooke 2017; Cammett and Jones Luong
2014); some groups leverage this advantage to reinforce a perception that Islamists will
808    Jaimie Bleck and Alex Thurston

perform better on economic and welfare policy in a context where voters are uncer-
tain about party platforms or performance (Pepinsky, Liddle, and Mujani 2012). Often
Islamist organizations draw on alternative funding, including diasporas, donor coun-
tries, or illicit networks that balance against government-​funded clientelist networks
(Wickham 2002), thus enhancing Islamists’ power and ability to act as opposition
players independent from the state (Arriola 2013). Service provision has also proved ef-
fective for insurgent and jihadist groups seeking popular support (Cammett and Issar
2010; Flanigan 2008) and/​or information about the territory that they control (Wagstaff
and Jung 2020).
In both settings, service provision can help nonstate actors to outperform states, but
also affords an opportunity to reshape political imagination about the spheres of au-
thority and legitimacy in society. For example, schools provide venues for indoctrina-
tion into religious ideology, but also forums to connect with parents and enlist their
allegiance in alternative forms of political authority. Similarly, Islamic courts, if they
perform well, offer the promise of more equitable provision of justice in contrast to
perceived corruption within bureaucratic states (Kendhammer 2013).
In much of West Africa, Islamic service providers play key roles in providing edu-
cation, justice provision and mediation, and alternative clientelist welfare networks to
the channels afforded by the state (Kang 2015; Kendhammer 2013, 2016; Villalón 2010;
Wing 2009; Schulz 2008, Kuenzi 2006).1 However, this vast network of service provision
has not translated into many Islamist parties in the region. Part of the explanation is the
structure of West African political systems. Like several other Muslim-​majority coun-
tries in Francophone West Africa, such as Senegal and Niger, Mali has a constitutionally
laïc state, following the French model of secularism that seeks to protect the state and
public life from religious influence, rather than the American model that protects the
church from the state. Successive constitutions have forbidden the establishment of reli-
gious parties (Bogaards et al. 2010).
Since the overthrow of a long-​running authoritarian regime in 1991, Mali has held
regular multiparty elections. Yet Mali, despite being over 90 percent Muslim, lacks a
mass-​based, bureaucratically organized Islamist movement comparable to the Muslim
Brotherhood. This absence reflects not just the formal prohibitions on Islamist party
formation, but also several other factors, including the enduring power of Sufism,
the relatively small size of the kind of middle-​class base that has populated Islamist
movements elsewhere, and the particular histories of the democratization era (Thurston
2020b). Under these conditions, Mali’s foremost Muslim religious leaders have not, so
far, competed directly in elections, even as independents or under the banner of estab-
lished secular parties—​as Islamists did in neighboring Mauritania prior to the legaliza-
tion of Islamist parties there in 2007, or as Egyptian Muslim Brotherhood candidates did
at various points prior to the Arab Spring.
However, the absence of Islamic parties from the electoral terrain understates the
tremendous role that religious leaders and their followers play in politics. In Mali, re-
ligious authority and religio-​political activism largely emanate from individual clerics
and their associated followings. Malian clerics as well as peers in neighboring countries
 Exploring the Role of Islam in Mali     809

back candidates, mobilize followers to engage in protests, and/​or lobby government

(Bleck and Patel 2021; Villalón 2010). As such, clerics’ forays into service provision and
the “parallel sector” tend to complement rather than challenge state authority. Islamic
actors have become an enduring informal institution of governance.
In contrast to these patterns of accommodation, complementarity, and competition
between clerical networks and the state, the emergence of jihadist actors in Mali and
their repertoires of service provision poses a challenge to these older patterns of ac-
commodation between religious constituencies and the state. Insurgent actors use ser-
vice provision strategies to outperform, delegitimize, and displace the secular state. We
conclude by speculating how their emergence in the country’s periphery might chal-
lenge the existing status quo relationship between the most powerful clerics and poli-
tics in the capital.

Clerics and the State

Malian religious leaders jockey for power and influence over elected leaders, while
also asserting their own independence from government to maintain credibility with
their constituencies. Less formally, religious authorities shape policy, mediate state-​
society relations, and even, in some senses and especially in the country’s peripheries,
govern in tandem with bureaucratic and elected officials. For many citizens with lim-
ited contact with the bureaucratic state, religious and traditional leaders continue to
be the de facto authority. Mali, following Senegal, boasts the second-​highest rating of
traditional authorities on the continent. Malians put trust in traditional leaders higher
than the president, members of the national assembly, municipal government or the
courts (Logan 2009, 119). Seventy-​six percent of Malians favor increasing the power of
traditional authority (Logan 2013, 364). The elected elite routinely show deference to
and consult the Muslim elite in this laïc state, which lifts Islam to the status of state re-
ligion in practice (Amselle 2018). In the eyes of much of the population, clerics serve
as a moralizing “check” on secular power—​almost as a branch of government, or at
least a counterweight to the executive office (Bleck 2021). Focus groups with over 3302
members of informal, tea-​drinking clubs ages 18–​45 in Bamako, Sevare, and Mopti re-
vealed that most respondents value religious pluralism, and they also favor religious
leaders’ involvement in governance (Bleck 2021).
In independent Mali, clerics’ visibility in politics has grown over time. Under the so-
cialist regime of Modibo Keïta (in power 1960–​1968) and the military-​turned-​civilian
regime of Moussa Traoré (in power 1968–​1991), authoritarianism discouraged overt
expressions of politicized religiosity, especially religious-​based dissent. Although laïc,
the state attempted to control clerics. Starting in 1980, the year after Traoré formally
transitioned his regime from a military to a civilian one, religious leaders were techni-
cally grouped under the state-​sponsored Association malienne pour l’unité et le progrès
de l’Islam (Malian Association for the Unity and Progress of Islam, AMUPI), which “was
810    Jaimie Bleck and Alex Thurston

meant to parallel, in the religious domain, the ideological functions of Mali’s single po-
litical party” (Brenner 1993, 73). Some religious actors created independent movements,
notably Ousmane Madani Haïdara’s Ançar Dine (Supporters of the Religion),3 which
is now one of the most popular religious networks in the country (Holder 2012; Schulz
2012). Yet such movements remained partly underground until Traoré was overthrown.
His fall accelerated liberalization and enabled a transition to multiparty democracy.
These changes created greater space for Islamic organizations, media outlets, and
NGOs. The early 1990s saw some incipient, would-​be Islamist parties, but they proved
short-​lived (Brenner 2001; Soares 2006).
In 2002, Malian president Alpha Oumar Konaré created a new state-​backed Muslim
umbrella body, the Haut Conseil Islamique (High Islamic Council, HCI), which partly
supplanted and partly absorbed AMUPI. Via the HCI, Konaré sought to domesticate
and control Mali’s proliferating Islamic associations in the new, post-​9/​11 era (Idrissa
2017, 179). The HCI has greater independence from the government than AMUPI had,
and the HCI has been a platform that helps clerics mobilize supporters in order to check
government policies, personnel, and legislation that they view unfavorably. Not all HCI
leaders and members support greater Islamization of Malian society—​nor do they agree
on what such Islamization should look like—​but the HCI has at times been one vehicle
for a more assertive Islamic politics in Mali.
Amid the rise of new Islamic associations, prominent new clerics have emerged,
while some of the country’s established religious lineages have maintained their influ-
ence (Lebovich 2019). The most prominent of the new clerics include Haïdara, a populist
preacher who borrows from Sufi styles without belonging to a Sufi order (Soares 2006,
86), and who took over as head of the HCI in 2019. Another major figure is Mahmoud
Dicko, the foremost Malian Salafi. Dicko’s blend of traditionalist learning in the Sahara
and studies in Saudi Arabia helped bring him to prominence as an imam in Bamako in
the 1980s and 1990s. Dicko served two terms as president of the HCI from 2008 to 2019,
having been initially denied that position when the HCI was created (Idrissa 2017, 179–​
180). In 2019 Dicko stepped down as HCI president and decided not to put forth a clear,
Salafi successor, bowing to pressure from Sufis and other non-​Salafis within the organi-
zation, as well as pressure from President Ibrahim Boubacar Keïta (in office 2013–​2020),
who all demanded new, non-​Salafi leadership for the HCI in the person of Haïdara.
Some Sufis and traditionalists have remained influential on the religious scene. The
most important Sufi leader is Mohamed Ould Cheickna Hamaullah, better known as
the Chérif of Nioro du Sahel. The Chérif is the son of Ahmad Hamaullah (d. 1943), who
led a popular and controversial branch of the Tijaniyya Sufi order, the Hamawiyya, and
had repeated conflicts with the French colonial authorities, culminating in his deporta-
tion to France (Soares 2005). The Tijaniyya is one of the most widespread Sufi organi-
zations across Africa, and the Tijaniyya Hamawiyya has a substantial following within
Mali and in other parts of West Africa.
Despite the diversity of their religious orientations and their differences on politics,
particularly in the era of President Keïta, these leaders and their peers adopt a common
foundational strategy of working within the framework of the existing Malian state. Of
 Exploring the Role of Islam in Mali     811

the three clerics mentioned above, Dicko is the most politically outspoken and the most
willing to challenge presidential authority, while Haïdara has been the most quietist—​
although still willing to participate in organizing mass demonstrations on social issues
when he chooses to. The Chérif, although not often physically present at mass rallies in
Bamako, also intervenes directly in politics by endorsing candidates, making political
statements, and, in recent years, publicly backing some of Dicko’s political maneuvers.
All of these clerics supply a range of services to citizens. This role in turn positions
clerics as power brokers and the heads of patronage networks. For example, as Dorothea
Schulz (2012, 104–​105) writes, “Villagers use their ties to [Haïdara] in Bamako when-
ever they need support in local conflicts over business transactions or with representa-
tives of the administration. Ties of religious and commercial-​entrepreneurial patronage
that link rural followers to Haïdara and his followers in Bamako are facilitated by
intermediaries, most of them merchants and teachers at local medersas.” Such networks
give the clerics social and political weight that elected politicians cannot ignore.
Meanwhile, the weakness of the Malian state expands the space available to clerics as
political entrepreneurs and alternative service providers. To cite a harsh but well-​argued
assessment, Mali “never quite met the prerequisites for functional statehood” (Craven-​
Matthews and Englebert 2018). In multiple domains, vacuums left by state weakness
have created opportunities for alternative service providers to fill gaps. Independence-​
era socialism fell short of its ambitions to remake state and society, and since the 1970s,
NGOs have had a major role in governance across the Sahel (Mann 2015), including in
remote areas of northern Mali (Gutelius 2007; Lecocq and Schrijver 2007). More in-
formal religious networks have also filled gaps in governance, or, in some cases, never
surrendered the roles that they had long played.
Weak state capacity, as well as the limited strength of political parties, leaves ample
opportunity for religious leaders to demonstrate their comparative ability to influ-
ence policy and mobilize their constituencies. Unlike in more restrictive authoritarian
contexts, contentious politics, including engagement in wide-​scale protests, has proven
very effective in achieving policy objectives. For instance, Haïdara and Dicko criticized
the adoption of a Family Code in 2009, which they claimed contradicted Islamic values.
Articles in the code proposed raising the legal age of marriage, but also required that
couples have a civil ceremony—​drawing power and resources away from religious elites
who conducted marriages. Religious leaders organized nationwide protests with tens of
thousands of participants, which led to a significant revision of the contested provisions
(Soares 2011). More recently, Dicko was the face of mass protests in Bamako in summer
2020, with tens of thousands of demonstrators, which eventually culminated a coup that
removed President Keïta from office. At the time of writing in December 2020, Mali
has passed from a short-​lived military junta to a roughly eighteen-​month, transitional
government, with religious leaders continuing to jockey for influence over rapidly
changing government structures. Presidential and legislative elections are scheduled for
spring 2022.
The political roles of religious leaders have expanded not just because of liberalization
and because of many Malians’ perceptions that elected politicians have failed, but also
812    Jaimie Bleck and Alex Thurston

because of the country’s long-​running security crisis. In 2012, Mali suffered a multifac-
eted rupture that involved a separatist rebellion in the North, a junior officers’ coup in
the South, a jihadist force that wrested control of the North from the separatists, and a
humanitarian emergency. A French-​led military intervention in January 2013 and pres-
idential elections in summer 2013 ended the immediate crisis, but international security
deployments and a resumption of electoral politics have not prevented the continuation
and even escalation of multidirectional violence in northern and central Mali.
As Mali attempted to transition out of the crisis in 2013, the presidential election saw
religious figures openly backing presidential candidates. Dicko helped to create an ex-
plicitly “Islamic” movement called Sabati 2012 to back Ibrahim Boubacar Keïta, who
won the election; Sabati followed an earlier, pro-​Keïta effort organized by Dicko in
the 2002 election (Elischer 2019, 211). The Chérif openly encouraged followers to vote
for Keïta in 2013 (Bleck 2021). Dicko and the Chérif fell out with Keïta during his first
term in office. In the lead-​up to the 2018 elections, the Chérif joined Dicko in opposing
Keïta, although the president and his surrogates continue to court the Chérif ’s support
(Coulibaly 2019). Meanwhile, Keïta made lavish gestures of patronage to Haïdara, no-
tably by gifting him 150 hectares of land in 2017 and ensuring that a paved road reached
the cleric’s hometown of Tamani, in the Segou region. This did not elicit a formal en-
dorsement from Haïdara for Keïta (Koné 2018). Yet amid Dicko and the Chérif ’s antag-
onism, Haïdara’s tacit support for Keïta appears to have earned him material rewards,
access to the presidency, and the presidency’s own tacit support, as Haïdara became head
of the HCI in 2019. Amid these leading clerics and the president, numerous alignments
and realignments occur; any actor who comes to be seen by the others as too powerful
will elicit a counter-​reaction and various realignments.

Justice and Schooling as Key Sectors

of Cooperation and Competition

The judicial and education sectors exemplify two areas where the state has clear
shortcomings and where religious actors provide influential alternatives. These two
domains are characterized by compartmentalization, coordination, and competition in
service provision
Malians trust the judiciary less than any other state institution (Little and Logan
2009). A 2014 poll found that 14 percent of Malians living in Bamako reported paying a
bribe to a police officer, judge, or civil servant in the previous month (Bleck, Dembele,
and Guindo 2016). And in this predominately rural country, many courts sit in regional
centers—​far from villages—​making courts inaccessible and unaccountable to many
rural inhabitants (Doungon and Sangaré 2016).
In the face of corrupt or unresponsive bureaucratic legal apparatus, many Malians
turn to religious authorities to settle conflicts (Soares 2005, 83). This type of Islamic
 Exploring the Role of Islam in Mali     813

justice, as well as arbitrage by village elders, has gained ground, or in some cases
maintained its hegemony, in a context where many rural citizens view government
courts with skepticism, believing that judgments will reflect the interest of those with
the most money (Doungon and Sangaré 2016). The state justice system can also be
opaque for some citizens because the language of the courts is French, which a small mi-
nority of the Malian population speaks fluently (Remane 2010). Religious or traditional
justice provisions can offer local communities more legible, transparent, and predict-
able alternatives.
Schooling is another important arena of service provision, not just because of the mul-
tifaceted impacts that education has on citizens’ economic prospects and overall quality
of life, but also because of its ability to draw people into engagement with the state. Bleck
(2015) finds that parents with children enrolled in public schools are more likely to vote,
campaign, and hold voter identification cards than citizens without children enrolled.
In contrast, Bleck finds that parents who have their children in modern, Islamic schools
were less likely to vote in the 2009 communal elections or to claim having voted in 2007,
as compared to Malians without a child enrolled in school. This research suggests that,
even amid skepticism about government performance, access to state services might
foster participation in formal politics. Meanwhile, engagement with nonstate services
or membership in related communities might discourage formal participation or orient
claims-​making toward other venues.
The strong history of Islamic education in what is now Mali, as well as the association
of post-​independence education with the French colonial state, contributes to enduring
tension around schooling. In Mali and the surrounding region, Islamic schooling pro-
vision precedes “Western education” by centuries (Hall and Stewart 2010; Nyanchoga
2014; Sanankoua 1985; Ware 2014). When French colonial authorities encountered a
substantial Muslim educational presence in the north of the country, they sought to co-​
opt schools for the colonial mission, recruit students to forced labor, or delimit schools’
curricula (Brenner 2007; Saidou 2012). Malian Muslims often feared that French schools
would turn their children into unbelievers (Brenner 2007; Gérard 1997). This belief
continues to be held by segments of the Malian population. Meanwhile, despite French
intentions to contain Islam, the colonial era saw widespread conversions to Islam in
southern Mali, where French policies and their impacts (emancipation, labor migra-
tion, conscription, etc.) created conditions that prompted conversions, often due to
new or strengthened linkages between Islam and different forms of economic and so-
cial belonging (Peterson 2011). The wave of conversions not only made Mali a Muslim-​
majority society, but also bound new converts into networks of patronage and belonging
that existed outside of the state (Launay and Soares 1999; Soares 2005). As of the early
1950s, meanwhile, only 4 percent of the school-​aged population was attending French
schools (Bolibaugh 1972, 139).
The Malian state inherited the weak colonial education legacy, and the independent
state continued the colonial state’s contentious relationship with Islamic schools. The
regime maintained a relatively antagonistic relationship with Islamic schools, which did
not receive comparable government subsidies that Catholic Schools received (Lange and
814    Jaimie Bleck and Alex Thurston

Diarra 1999). Yet madrasas—​“modernized” Islamic schools with a broader curriculum

than Qurʾanic schools—​have become popular. As the public education system struggled
in the 1970s and 1980s, enrollment in madrasas likely increased at a rate greater than en-
rollment in public schools (Brenner 2001, 171–​172). With democratization, the Malian
state embarked on an educational expansion, which included an attempt to integrate
madrasas as part of the broader educational infrastructure. By counting students
enrolled in madrasas toward the total number of children enrolled in school, the gov-
ernment moved closer to goals of universal primary enrollment, while simultaneously
appealing to Malian constituencies that preferred Islamic schooling (Bleck 2015, 89–​90).
Most madrasas are private schools, making this one area where religious leaders and
networks have a substantial role in service provision. Educational entrepreneurs can
draw on substantial resources: if they studied abroad in North Africa or the Arab world,
personal networks can help fund projects back home in Mali. Roy (2012, 154) argues that
these personal connections are far more powerful than any state-​sponsored financing of
madrasas. This creates a diverse, and often fragmented, Islamic landscape in Bamako,
whereas each association has ties to different foreign funders. Many madrasas offer
subsidized fees for the poorest learners as well as scholarship opportunities abroad.
While these opportunities are dwindling, there remain channels for this type of edu-
cational trajectory (137). Madrasas represent another sphere of power, education, and
influence linked to North Africa and the Middle East. Roy argues that graduates of
madrasas gained a “madrasa consciousness”—​a specific type of Muslim identity (101).
She stresses that many parents saw this as an alternative, rather than competing, net-
work to state channels; by enrolling some of their children in Islamic schools—​families
could diversify their clientelistic networks. Similarly, Sanankoua and Brenner (1991) dis-
cuss the social prestige associated with a madrasa education. By some metrics, madrasas
offer higher-​quality education than contemporary public schools. In 2005–​2006, for
example, more than eighty-​one teachers in madrasas held a four-​year diploma—​more
than three times the number of comparable degree holders in public schools (Bleck
2015, 81–​82).
Islamic schools form key parts of different clerics’ and associations’ webs of service
provision. Although sometimes stereotyped as a Salafi project, madrasas are run by all
of the clerical networks (Roy 2012, 177–​118; Roy 2012 86–​88). Schooling represents one
ingredient in religious networks’ ability to channel funding from abroad, particularly
from the Middle East and North Africa. Meanwhile, there is interplay between clerical
bodies and service provision networks; some of the foremost service providers also hold
senior positions within bodies such as the HCI (Watling and Raymond 2015).
Qurʾanic schools, in contrast, have remained isolated from the formal education
system. Qurʾanic schools have more limited curricula than madrasas, as students
focus on memorizing the Qurʾan and, at higher levels, pursuing esoteric knowledge.
Since students beg for their own school fees and, in some instances, lodge with their
religious instructor, these schools are affordable options for many Malian parents.
Qurʾanic teachers feel as though they are providing welfare—​including housing and
food—​for some of the most vulnerable children in Mali, yet do not receive any financial
 Exploring the Role of Islam in Mali     815

support from the state (Bleck 2015, 75). Qurʾanic schools play powerful roles, espe-
cially throughout rural Mali—​particularly in the Kayes, Mopti, and Timbuktu regions
(Autoportrait du Mali, 2015). Many of these schools are decentralized and linked to spe-
cific religious leaders, including adherents of Sufi orders.
Islamic schools continue to offer a complement or alternative to state schooling and
have varying degrees of engagement with the Malian state. Given a high degree of dissat-
isfaction with the quality of secular education or the job prospects it creates, madrasas
offer an acceptable alternative to many parents, who can diversify their portfolio or
choose a type of schooling that reaffirms a Muslim identity and inserts their children in
a new network of power and resources.
In sum, religious leaders and networks have a powerful role in the judicial and edu-
cation sectors specifically, and in service provision more broadly. This has not, however,
translated into Islamist mobilization in the style of the Egyptian Muslim Brotherhood or
other well-​known Islamist movements from the Arab world. In all cases, these religious
leaders have maintained their position as formidable welfare providers and key polit-
ical actors without formally entering the “dirty” realm of electoral politics.4 By inserting
themselves as central actors in elite political negotiations, but keeping enough distance
to preserve autonomy and independence, they are able to maintain “reputational” credi-
bility,5 which could be damaged with an explicit entry into electoral politics. Instead their
repertoire of contentious politics and backroom negotiation proves lucrative for their
own personal power and serves an effective policymaking instrument without needing to
engage in campaigning or partisan politics that many Malians hold in disdain.

Jihadist Service Provision:

A Challenge to Patterns
of Accommodation

The current conflict between jihadists and the state—​wherein the former, although they
are not usually portrayed this way, could be considered armed, coercive religious ser-
vice providers—​raises questions about the limits of Islamic service provision as comple-
mentary to the Malian state. In many ways, the state’s inability to respond to citizens in
their direst time of need has reinforced state impotence in the minds of many rural citi-
zens (Bleck and Michelitch 2015). When presented with a better-​performing alternative
(e.g., jihadists, whose governance styles do appeal to some segments of the population
even as jihadists kill, intimidate, and marginalize others), this competition may force
reflection on the state performance, making citizens question their allegiance to the
state. Jihadist service provision can proceed alongside and interact with jihadist service
restriction. When they exercise sway in a given locale, jihadists often shut down some
public institutions; allow other services, such as health provision, to function under sur-
veillance; and create new structures of their own.
816    Jaimie Bleck and Alex Thurston

Since the early 2000s, jihadist groups have had a presence in Mali, beginning in the
far north. It is important to note that al-​Qaʿida in the Islamic Maghreb (AQIM) used
welfare provision during its initial entry into northern Mali in the late 1990s. They dis-
tributed money, medicine, phone credit, and healthcare to local populations and mar-
ried women from less-​powerful families to demonstrate solidarity with “poor local
lineage” (Bøås and Torheim 2013, 1287). Jihadists controlled northern Mali’s major cities
for much of 2012–​2013, before a French-​led military intervention expelled them. During
the occupation, they continued to use security provision—​including a hotline for those
experiencing harassment or extortion by bandits or other insurgent groups, as well as
other service provision tactics in Timbuktu.
Since 2017, the most important jihadist formation in Mali has been Jamaʿat Nusrat al-​
Islam wa-​l-​Muslimin (the Group for Supporting Islam and Muslims, JNIM). Formed as
a coalition of local jihadist actors and the Mali-​based units of AQIM, JNIM’s area of op-
erations extends from Mali’s northern border with Algeria south to the Mopti and Ségou
regions in the center of Mali. JNIM also conducts attacks in Burkina Faso, while AQIM
units have struck as far south as Cote d’Ivoire. In addition to JNIM/​AQIM, since 2015
Mali has also been troubled by the Islamic State in the Greater Sahara (ISGS). This outfit
operates in the borderlands of eastern Mali, western Niger, and northern and eastern
Burkina Faso.
Part of JNIM’s challenge involves thinking through the lessons of 2012–​2013, in-
cluding the ways that jihadists’ uncompromising rule alienated potential allies. AQIM’s
own leader criticized the manner in which his field commanders behaved in Mali during
that period. Although JNIM’s leaders have rarely acknowledged any past mistakes, they
do now invoke concepts such as “the popular embrace” (al-​hadina al-​shaʿbiyya) to de-
scribe their search for support (Al-​Masra 2017). Throughout the areas where they are
present, JNIM’s field commanders are enmeshed in hyper-​local politics (Thurston
2019). Jihadists have relationships not just with young male fighters, but also with var-
ious politicians, clan leaders, businesspersons, religious leaders, and other power
brokers (Sandor and Campana, 2020). Such relationships allow some figures to occupy a
gray area between overt jihadism and mainstream life; by 2018, for example, the former
jihadist judge Houka Houka ag Alhousseini had allied himself with Timbuktu-​based
militias and based himself in Zoura, a town “well known in the Timbuktu region for
its weekly market days (foire), frequently used by AQIM as a platform to intimidate
the population by displaying decapitated heads and distributing threatening leaflets”
(United Nations Panel of Experts 2018, 5). In this type of shadow governance by AQIM/​
JNIM, however, threats and coercion are accompanied by forms of outreach that take
advantage of the very insecurity that jihadists help produce. If in 2012 cutting off hands
seemed largely calculated to demonstrate jihadists’ power and produce a media spec-
tacle, by 2019 jihadists were presenting such acts as public safety measures. For example,
a jihadist flier posted in 2019 in Acharane, a village in the Timbuktu region, called upon
thieves to repent and return stolen goods; “if not,” the flier continued, “he exposes him-
self to prosecution and accountability in accordance with the Law of God, Glorious and
Exalted be He.”6
 Exploring the Role of Islam in Mali     817

Since 2013, two important trends in jihadist governance and service provision have
occurred in Mali. First, jihadists have moved away from their bid for urban control,
which in retrospect appears anomalous, to a mode of rural-​based shadow governance
and political influence (Jezequel and Foucher 2017). Second, jihadists have partly shifted
from the shock-​and-​awe tactics of stoning accused adulterers and destroying Sufi
shrines to a more encompassing effort to present shariʿa as an alternative moral order
and a mode of equitable justice that meets community needs. Established clerics, in this
context, have often remained relatively quiet, reflecting both the fear of retaliation by
jihadists (numerous imams have been assassinated in central Mali) and the fact that, in
certain areas, established clerics’ prestige had been challenged even before the current
wave of conflict began (Thurston, 2020a).
As context for why some citizens would welcome the changes that jihadists bring, re-
call that many Malians distrust the judiciary. Further, Malian communities’ interactions
with the police or the army are often predatory. This pushes certain populations toward
armed actors, including jihadist groups, if they do not feel protected by the state (Nievas
and Sangaré 2016). For instance, rent-​seeking by the state has pushed some pastoralists
into the arms of jihadists who promise more equitable treatment (Benjaminsen and Ba
2019). In the context of the current conflict, most Malians living in the North and center
of the country feel exposed to criminal and insurgent actors, and many have formed
self-​defense groups and militias in the vacuum left by the state (Bleck, Boisvert, and
Sangaré 2018). In some areas, jihadist groups have stepped in to offer protection from
criminal actors and championed fair treatment for all members of the population as a
way to distinguish themselves from the state, but also from other insurgent actors. The
current wave of conflict has, moreover, in many cases weakened the power of traditional
and religious elites, especially in the North. A 2019 study by the Clingendael Institute
found that “the increased fragmentation of conflict has limited traditional authorities’
scope of authority and their legitimacy. As communities become divided along ethnic,
resource-​based and clan-​based fault lines, so do their traditional authority structures”
(Molenaar et al., 2019: 6). Armed groups have in some cases supplanted the historical
roles of hereditary authorities in terms of conflict mediation and service provision.
Much of what jihadists offer, in terms of governance and services, comes back to the
notion of equitable justice. Jihadists present the shariʿa as an alternative moral order,
an ideal of justice that they contrast with the allegedly corrupt, Francophile state. In
a widely discussed August 2017 recording, for example, the central Malian jihadist
leader Amadou Kouffa said, “Our only ‘fault’ is that we defend Islam; if not for that, one
couldn’t reproach us for anything. We haven’t stolen anyone’s goat . . . . We the jihadists
only attack our known adversaries, the allies of France” (Kouffa 2017). In central Mali,
Kouffa’s forces have reconfigured the management of pastures, placed restrictions on
the dowries expected of bridegrooms, instituted the collection of what they term zakat
(obligatory alms in Islam), and expelled many state agents (Sangaré 2016; Thiam 2017).
Such maneuvers have won jihadists some support, however grudging, among locals
who had suffered for years under corrupt judges and gendarmes as well as under the
“rackets” practiced by village-​level oligarchies.
818    Jaimie Bleck and Alex Thurston

Consistent with other insurgent strategies that focus on less capital-​intensive sectors
(Wagstaff and Jung 2020), since 2016 jihadists have intervened dramatically in the do-
main of schooling, one of the foremost sectors where religious actors and the Malian
state compete and sometimes collaborate. Yet in contrast to patterns of accommodation
involving nonviolent religious actors, jihadist strategies to close government schools
and remove bureaucratic officials have generated a zero-​sum struggle between the
Malian state and new forms of religious provision. By the spring of 2017, 23 percent of
schools in Mali and 100 percent of schools in eight communes were closed; in areas with
the most concentrated jihadist activity, teachers, decentralized state administrators, and
health officials were all targeted by threats and harassment (Bleck, Boisvert, and Sangaré
2021, 16). A wider set of schools face a jihadist assault against schooling as a symbol of
the secular state, including threats against administrators, calls for schools to turn into
Qurʾanic teaching centers, or burning of material and closures. Parallel to the closing
of government schools has been the expansion of Qurʾanic schools in Mopti and Segou
(Adam, Golovko, and Sangaré 2017).
These campaigns take place in the broader context of the Ministry of Education’s in-
ability to manage its own schools. In December 2019, 895 schools were close and, as a
result, 333,900 children were out of school (UNICEF 2019). These school closures over
the past few years are due to insecurity and targeting, but also to teacher strikes. The
current educational crisis in Mali—​the perceived decline of educational quality over the
last twenty years (Ba 2009; Diakité 2000), as well as the impact of the recent political
crisis on the education sector (Bleck, Boisvert, and Sangaré 2021)—​may force a more
explicit reevaluation of political allegiance as some actors seek to differentiate them-
selves as being in opposition the secular state. The jihadist efforts to generate services
and the state’s incapacity in the face of threat is likely to inform other rural citizens’ own
strategies to cope with the ongoing insecurity.

Conclusion

This paper has described the complementarities and competition between religious
service provision and state-​directed governance in Mali. We have described clerics
as key political actors who primarily rely on contentious politics, endorsement (or
denouncements) of existing candidates, and elite lobbying to heighten their personal
power and achieve their policy goals. We argue that these strategies, rather than overt
electoral participation, prove very effective in achieving policy goals and fueling their
personal power, while allowing them to maintain reputational advantage (Brooke
2017; Cammett and Jones 2014) and keep the moral high ground as compared to other
political elites.
These relationships are shaped by Malian history, the nature of its laïc political system,
and the way that popular attitudes appear to condition clerics’ engagement in poli-
tics and government. The preceding section, however, described the severe challenge
 Exploring the Role of Islam in Mali     819

that jihadist activity—​including zero-​sum competition in service provision between

jihadists and the state—​represents to this model of collaboration and nonviolent com-
petition. What, then, might come next for Mali, particularly in terms of Islam’s role in
governance?
Theoretically, religious leaders are supposed to be part of the solution to the conflicts
in northern and central Mali. The 2015 Algiers Accord, meant to bring peace to the
North, includes among its provisions a call for “revalorizing the role of Qadis [Islamic
judges] in the administration of justice, notably in terms of civil mediation in a manner
that takes into account cultural, religious, and customary specificities” (Accord pour la
paix et la reconciliation au Mali 2015, 13). The scope that Islamic law will have in the
postconflict North, however, has been a point of debate between northern military-​
actors and the government in Bamako. In early 2019 the ex-​rebel bloc the Coordination
of Azawad Movements (French acronym CMA) unilaterally declared the application
of a version of shariʿa in the northern region of Kidal—​but their decision was swiftly
criticized by the national government and by the monitoring committee for the Algiers
Accord, and the CMA formally backed down while nevertheless retaining de facto con-
trol over Kidal and, indeed, many areas of northern Mali continue to apply aspects of
shariʿa. If Malians want a greater role for traditional and religious authorities in the jus-
tice sector, politicians have not yet agreed on the exact modalities of that role.
One question is how the explicit competition between jihadist and the state could
change or inform religious leaders’ strategies in the capital. On one hand, the three most
influential religious leaders are highly networked to state power. While they challenge
the state at times, they largely operate in a complementary sphere. While the govern-
ment often appeals to them, the president is not entirely dependent on religious leaders.
President Keïta was able to win re-​election without heavy religious backing in 2018,
which demonstrates the electoral resilience of the incumbent president. On the other
hand, the jihadist strategies in the center as well as efforts to adopt shariʿa in the North,
may open a political opportunity for some leaders to challenge to laicité and might buoy
the reputational advantage (Cammett and Jones Luong 2014) of religious leaders associ-
ated with the secular state into question.
The question of religion, service provision, and the state is largely subordinate to
questions of Malian state presence in the geographic periphery. Given the recent edu-
cation crisis (more than 800 schools closed last year), as well as the state’s inability to
thwart expanding interethnic violence, many Malians in peripheral areas may look for
other forms of protection or service provision. If jihadists can execute justice more equi-
tably, they can likely win adherents independent of broader ideology. Greater questions
remain about how religious leaders in the periphery will respond to jihadist rule in the
longer term. Those who are incorporated into the networks of Haïdara, Dicko, and the
Chérif have incentives to keep the status quo relationship to the state. However, others
who operate without the protection or even any visible signs of state presence will likely
cede to the most powerful brokers in their areas or flee as a matter of pragmatism.
Lastly, there are broader questions about the role of religion and the state in the long
term. Many religious leaders have adopted a populist rhetoric that critiques the state
820    Jaimie Bleck and Alex Thurston

and offers better governance through Islam. As jihadists have a chance to rule over a
longer period of time, citizens will have more evidence to evaluate the effectiveness of
these new models. They may also grow unsatisfied if a regime is unable to generate jobs
or economic growth, or if it shows vulnerability to corruption (Kendhammer 2016), and
they might seek other forms of governance. In contrast, better long-​term performance
by these actors may raise questions about the applicability of this alternative governance
scheme in other parts of the country.

Acknowledgments
We thank Callie Whelan for research assistance on this project. Christiana Parreira and Pierre
Englebert, as well as the editors, gave very useful comments on previous drafts.

Notes
1. As Platas writes in this volume, in many of Muslim-​ majority states, independent
governments inherited particularly weak bureaucratic and public goods infrastructure, and
as a result, many Islamic providers preceded the state or operated in its void.
2. The vast majority of respondents were male. We note that respondents were much more
divided about the role of religion in electoral politics and direct governance by religious
leaders.
3. Not to be confused with the more recently created, Northern-​based jihadist movement
Ansar al-​Din.
4. This is consistent with the strategies Brooke describes, which underscore the importance of
groups distancing themselves from the “dirty river of politics,” which can help to “capture
diffuse (reputational) benefits of the provision . . . but retain the image of being above the
fray of politics” (2017, 45).
5. Cammett and Jones Luong (2014) argue that Islamist advantage largely comes from this
reputational advantage, thus a central concern of Islamic leaders should be maintaining the
relative reputational credibility—​especially vis-​à-​vis the other political elite.
6. This flier is available at https://​mobile.twitter.com/​walid6519/​status/​1142368408341766145.

References
Accord pour la paix et la reconciliation au Mali. 2015. Available at https://​photos.state.gov/​
libraries/​mali/​328671/​peace-​accord-​translations/​1-​accord-​paix-​et-​reconciliation-​francais.pdf.
Adam, Nadia, Ekaterina Golovko, and Bouboucar Sangaré. 2017. Terrorism Puts School
on Hold in Mali. Pretoria: Institute for Security Studies. https://​issafrica.org/​iss-​today/​
terrorism-​puts-​education-​on-​hold-​in-​mali
Amselle, Jean-​Loup. “L’excision et l’homosexualité, enjeux politiques au Mali.” Les Temps
Modernes 2 (2018): 3–​19.
Al-​Masra 2017. “‘Al-​Masra’ Tuhawir al-​Shaykh Aba al-​Fadl Iyad Ghali.” Al-​Masra 45, 3 April.
 Exploring the Role of Islam in Mali     821

Arriola, Leonardo. 2013. Multi-​Ethnic Coalitions in Africa: Business Financing of Opposition

Election Campaigns. Cambridge: Cambridge University Press.
Autoportrait du Mali: Les Obstacles à la Paix. 2015. Bamako: Institut Malien de Recherche
Action Pour la Paix.
Ba, Oumar Issiaka. 2009. Une histoire de l’enseignement au Mali: Entre réforme et Réticences.
Paris: L’Harmattan.
Benjaminsen, Tor A., and Boubacar Ba. 2019. “Why Do Pastoralists in Mali Join Jihadist
Groups? A Political Ecological Explanation.” Journal of Peasant Studies 46, no. 1: 1–​20.
Bleck, Jaimie. 2015. Education and Empowered Citizenship in Mali. Baltimore: Johns Hopkins
University Press.
Bleck, Jaimie. 2021. Social Capital and Social Clubs in Mali. Unpublished Book Manuscript.
Bleck, Jaimie, Marc-​ André Boisvert, and Boukary Sangaré 2018. “I Joined to Save my
People: Children and Non-​State Armed Groups in Mali.” In Cradled by Conflict: Child
Involvement with Armed Groups in Contemporary Conflict, edited by Siobhan O’Neil and
Kato van Broeckhoven, 142–​175. Tokyo: United Nations University.
Bleck, Jaimie, Marc-​André Boisvert, and Boukary Sangaré. 2021. “Mali’s Education Crisis.”
Unpublished working Paper.
Bleck, Jaimie, Abdoulaye Dembele, and Sidiki Guindo. 2016. “Malian Crisis and the Lingering
Problem of Good Governance.” Stability 5, no. 1. https://​www.stabilityjournal.org/​articles/​
10.5334/​sta.457/​.
Bleck, Jaimie, and Kristin Michelitch. 2015. The 2012 Crisis in Mali: Ongoing Empirical State
Failure. African Affairs 114, no. 457: 598–​623.
Bleck, Jaimie, and David Patel. 2021. “Out of Africa: Electoral Failure and the Future of Political
Islam in West Africa.” Working paper.
Bøås, Morten, and Liv Elin Torheim. 2013. “The Trouble in Mali—​Corruption, Collusion,
Resistance.” Third World Quarterly 34, no. 7: 1279–​1292.
Bogaards, Matthijs, Matthias Basedau, and Christof Hartmann. 2010. “Ethnic Party Bans in
Africa: An Introduction.” Democratization 17, no. 4: 599–​617.
Bolibaugh, Jerry. 1972. “Educational Development in Guinea, Mali, Senegal, and Ivory Coast.”
DHEW Publication. Washington, DC: US Department of Health, Education and Welfare.
Brenner, Louis. 1993. “Constructing Muslim Identity in Mali.” In Muslim Identity and Social
Change in Sub-​Saharan Africa, edited by Louis Brenner, 59–​78. Bloomington: Indiana
University Press.
Brenner, Louis 2001. Controlling Knowledge: Religion, Power, and Schooling in a West African
Muslim Society. Bloomington: Indiana University Press.
Brenner, Louis. 2007. “The Transformation of Muslim Schooling in Mali: The Madrasa as an
Institution of Social and Religious Mediation.” In Schooling Islam: The Culture and Politics of
Modern Muslim Education, edited by Robert Hefner and Muhammad Qasim Zaman, 199–​
223. Princeton, NJ: Princeton University Press.
Brooke, Steven. 2017. “From Medicine to Mobilization: Social Service Provision and the
Islamist Reputational Advantage.” Perspectives on Politics 15, no. 1: 42–​61.
Brooke, Steven. 2019. Winning Hearts and Votes: Social Services and the Islamist Political
Advantage. Ithaca, NY: Cornell University Press.
Brooke, Steven. 2021. “Islamist Organizations and the Provision of Social Services.” In The
Handbook of Politics in Muslim Societies, edited by Melani Cammett and Pauline Jones
Luong. Oxford: Oxford University Press.
822    Jaimie Bleck and Alex Thurston

Cammett, Melani, and Sukriti Issar. 2010. “Bricks and Mortar Clientelism: Sectarianism and
the Logics of Welfare Allocation in Lebanon.” World Politics 62, no. 3: 381–​421.
Cammett, Melani, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17: 187–​206.
Cammett, Melani, and Lauren M. MacLean. 2011. “Introduction: the Political Consequences
of Non-​state Social Welfare in the Global South.” Studies in Comparative International
Development 46, no. 1: 1–​21.
Coulibaly, Cheick. 2019. “Révélation du Chérif de Nioro sur la visite d’IBK: ‘J’étais contre le
pouvoir et je demeure dans cette même posture.’” Maliweb.net, June 12. https://​www.
maliweb.net/​societe/​revelation-​du-​cherif-​de-​nioro-​sur-​la-​visite-​dibk-​%E2%80%89jetais-​
contre-​le-​pouvoir-​et-​je-​demeure-​dans-​cette-​meme-​posture-​2828401.html.
Craven-​Matthews, C., and P. Englebert. 2018. “A Potemkin State in the Sahel? The Empirical
and the Fictional in Malian State Reconstruction.” African Security 11, no. 1: 1–​31.
Diakité, Drissa. 2000. “La crise scolaire au Mali.” Nordic Journal of African Studies 9, no. 3: 6–​28.
Doungon, Isaie, and Boukary Sangaré. 2016. “Local Conflict and the Call for Justice.” In
Perception Study on the Drivers of Insecurity and Violent Extremism in the Border Regions of
the Sahel. UNDP and Centre for Humanitarian Dialogue, 2016. New York: United Nations
Development Programme.
Elischer, Sebastian. 2019. “Governing the Faithful: State Management of Salafi Activity in the
Francophone Sahel.” Comparative Politics 51, no. 2: 199–​218.
Flanigan, Shawn Teresa. 2008. “Nonprofit Service Provision by Insurgent Organizations: The
Cases of Hizballah and the Tamil Tigers.” Studies in Conflict & Terrorism, 31:499–​519,
Gérard, Étienne. 1997. La tentation du savoir en Afrique: Politiques, mythes, et strategies
d’éducation au Mali. Paris: Ostrom.
Gutelius, David. 2007. “Islam in Northern Mali and the War on Terror.” Journal of Contemporary
African Studies 25, no. 1: 59–​76.
Hall, Bruce and Charles Stewart. 2010. “The Historic ‘Core Curriculum’ and the Book Market
in Islamic West Africa.” In The Trans-​Saharan Book Trade: Manuscript Culture, Arabic
Literacy and Intellectual History in Muslim Africa, edited by Graziano Krätli and Ghislaine
Lydon, 109–​174. Leiden: Brill.
Harrigan, Jane, and Hamed el-​ Said. 2009. Economic Liberalisation, Social Capital, and
IslamicWelfare Provision. London: Palgrave Macmillan.
Holder, Gilles. 2012. “Chérif Ousmane Madani Haïdara et l’association islamique Ançar
Dine. Un réformisme malien populaire en quête d’autonomie.” Cahiers d’études africaines
206–​207: 389–​425.
Idrissa, Rahmane. 2017. The Politics of Islam in the Sahel: Between Persuasion and Violence.
New York: Routledge.
Jezequel, Jean-​ Hervé, and Vincent Foucher. 2017. “Forced Out of Towns in the Sahel,
Africa’s Jihadists Go Rural.” Brussels: International Crisis Group, January 11. https://​www.
crisisgroup.org/​africa/​west-​africa/​mali/​forced-​out-​towns-​sahel-​africas-​jihadists-​go-​rural.
Kang, Alice J. 2015. Bargaining for Women’s Rights: Activism in an Aspiring Muslim Democracy.
Minneapolis: University of Minnesota Press.
Kendhammer, Brandon. 2013. “The Sharia Controversy in Northern Nigeria and the Politics
of Islamic Law in New and Uncertain Democracies.” Comparative Politics 45, no. 3: 291–​311.
Kendhammer, Brandon. 2016. Muslims Talking Politics: Framing Islam, Democracy, and Law in
Northern Nigeria. Chicago: University of Chicago Press.
 Exploring the Role of Islam in Mali     823

Koné, Issouf. 2018. “Le Chérif Ousmane Madani Haïdara évite le piège d’IBK.” Benbere, July 21,
2018. https://​benbere.org/​politique/​cherif-​madani-​haidara-​ibk-​ousmane-​evite-​piege/​.
Kouffa, Amadou. 2017. Recorded statement translated from Fulfulde to French by Oumar Sow,
and from French to English by the authors.
Kuenzi, Michelle T. 2006. “Nonformal Education, Political Participation, and
Democracy: Findings from Senegal.” Political Behavior 28, no. 1: 1–​31.
Lange, Marie-​France, and Sékou Oumar Diarra. 1999. “Ecole et démocratie: L’explosion
scolaire sous la IIe République du Mali.” Politique Africaine 76: 164–​176.
Launay, Robert, and Benjamin Soares. 1999. “The Formation of an ‘Islamic Sphere’ in French
Colonial West Africa.” Economy and Society 28, no. 4: 497–​519.
Lebovich, Andrew. 2019. “Sacred struggles: How Islam shapes politics in Mali,” European
Council on Foreign Relations, Policy Brief 1, November 2019: https://​ecfr.eu/​publication/​
secular_​stagnation_​malis_​relationship_​religion/​
Lecocq, Baz, and Paul Schrijver. 2007. “The War on Terror in a Haze of Dust: Potholes and
Pitfalls on the Saharan Front.” Journal of Contemporary African Studies 25, no. 1: 141–​166.
Little, Eric, and Carolyn Logan. 2009. “The Quality of Democracy and Governance in Africa: New
Results from Afrobarometer Round 4—​A Compendium of Public Opinion Findings from
19 African Countries, 2008.” Working Paper No. 108, Afrobarometer Network. http://​
afrobarometer.org/​sites/​default/​files/​publications/​Working%20paper/​AfropaperNo108.pdf.
Logan, Carolyn. 2009. “Selected Chiefs, Elected Councilors and Hybrid Democrats: Popular
Perspectives on the Co-​existence of Democracy and Traditional Authority.” Journal of
Modern African Studies 47, no. 1: 101–​128.
Logan, Carolyn. 2013. “The Roots of Resilience: Exploring Popular Support for African
Traditional Authorities.” African Affairs112, no. 448: 353–​376.
Mann, Gregory. 2015. From Empires to NGOs in the West African Sahel. Cambridge: Cambridge
University Press.
Molenaar, Fransje, Jonathan Tossell, Anna Schmauder, Rahmane Idrissa, and Rida Lyammouri.
2019. “The Status Quo Defied The legitimacy of traditional authorities in areas of lim-
ited statehood in Mali, Niger and Libya.”Cligendael: Netherlands Institute of International
Relations, CRU Report September 2019: https://​www.clingendael.org/​sites/​default/​files/​
2019-​09/​legitimacy_​traditional_​authorities_​mali_​niger_​libya.pdf
Nievas Bullejos, David, and Boukary Sangare. 2016. “Social and Territorial Control of Northern
Mali by Jihadism in a Context of Crisis.” Policia y Seguridad Publica 6, no. 1: 29–​81.
Nyanchoga, S. A. (2014). "Politics of knowledge production in Africa: A critical 41 reflection
on the idea of an African university in sustainable development." European Law and Politics
Journal, no. 1: 37–​55.
Pepinsky, Thomas B., R. William Liddle, and Saiful Mujani. 2012. “Testing Islam’s Political
Advantage: Evidence from Indonesia.” American Journal of Political Science 56, no.
3: 584–​600.
Peterson, Brian. 2011. Islamization from Below: The Making of Muslim Communities in Rural
French Sudan, 1880–​1960. New Haven, CT: Yale University Press.
Remane, Katja. 2010. “Le Mali, le pays le moins francophone d’Afrique.” Swiss Info, October
12, 2010. https://​www.swissinfo.ch/​fre/​le-​mali-​-​le-​pays-​le-​moins-​francophone-​d-​afrique/​
28515390.
Roy, Émilie. 2012. Educating Pious Citizens: Local Politics, International Funding, and
Democracy in Bamako’s Islamic Schools. PhD diss., MacMaster University.
824    Jaimie Bleck and Alex Thurston

Saidou, Abdoul Kader. 2012. “Le projet indépendantiste touareg: Quête de reconnaissance ou
quête de pouvoir ?” In La crise au Mali: Recueil de contributions, edited by Mamadou Lamine
Doumbia. Cahier spécial, Décembre 2012. Société Malienne des Sciences Appliquées
(MSAS).
Sanankoua, Bintou, and Louis Brenner. 1991. L’Enseignement Islamique au Mali. Bamako:
Jamana.
Sanankoua, Diarah Bintou. 1985. “Les écoles ‘Coraniques’ au Mali: Problèmes actuels.”
Canadian Journal of African Studies 19, no. 2: 359–​367.
Sandor, Adam, and Aurelie Campana. 2020. “Les groups djihadistes au Mali, entre violence,
recherche de legitimite et politiques locales.” Canadian Journal of African Studies 53, no.
3: 415–​430.
Sangaré, Boukary. 2016. “Le Centre du Mali: épicentre du djihadisme?” GRIP, 20 May. https://​
grip.org/​le-​centre-​du-​mali-​epicentre-​du-​djihadisme/​.
Schulz, Dorothea. 2008. “(Re)Turning to Proper Muslim Practice: Islamic Moral Renewal
and Women’s Conflicting Assertions of Sunni Identity in Urban Mali.” Africa Today 54, no.
4:21–​43.
Schulz, Dorothea. 2012. “Dis/​ Embodying Authority: Female Radio” Preachers” and the
Ambivalences of Mass-​Mediated Speech in Mali.” International Journal of Middle East
Studies 44, no. 1: 23–​43.
Soares, Benjamin. 2005. Islam and the Prayer Economy: History and Authority in a Malian
Town. Ann Arbor: University of Michigan Press.
Soares, Benjamin. 2006. “Islam in Mali in the Neo-​Liberal Era.” African Affairs 105: 77–​95.
Soares, Benjamin. 2011. “Family Law Reform in Mali: Contentious Debates and Elusive
Outcomes.” In Gender and Islam in Africa: Rights, Sexuality, and Law, edited by Margot
Badran, 263–​290. Washington, DC: Woodrow Wilson Center Press.
Thiam, Adam. 2017. “Centre du Mali: Enjeux et Dangers d’une crise négligée.” Humanitarian
Dialogue/​ Institut du Macina. https://​www.hdcentre.org/​wp-​content/​uploads/​2017/​03/​
Centre-​du-​Mali-​Enjeux-​et-​dangers-​dune-​crise-​n%C3%A9glig%C3%A9e.pdf.
Thurston, Alexander. 2019. “Timbuktu: A Laboratory for Jihadists Experiment with Politics.”
War on the Rocks, January 23, 2019. https://​warontherocks.com/​2019/​01/​timbuktu-​a-​
laboratory-​for-​jihadists-​experimenting-​with-​politics/​.
Thurston, Alexander. 2020a. Jihadists of North Africa and the Sahel: Local Politics and Rebel
Groups. Cambridge: Cambridge University Press.
Thurston, Alexander. 2020b. “Why Are There Few Islamist Parties South of the Sahara?”
Project on Middle East Political Science. Washington, DC: Elliot School of International
Affairs. https://​pomeps.org/​why-​are-​there-​few-​islamist-​parties-​south-​of-​the-​sahara.
United Nations Panel of Experts on Mali. 2018. Report. http://​www.securitycouncilreport.org/​
atf/​cf/​%7B65BFCF9B-​6D27-​4E9C-​8CD3-​CF6E4FF96FF9%7D/​s_​2018_​581.pdf.
Country Office Humanitarian Situation Report Number 8: January 1 –​December 31, 2019.
UNICEF: https://​reliefweb.int/​sites/​reliefweb.int/​files/​resources/​UNICEF%20Mali%20
Humanitarian%20Situation%20Report%20No.%208%20-​%201%20January%20-​%2031%20
December%202019.pdf
Villalón, Leonardo A. 2010. “From Argument to Negotiation: Constructing Democracy in
African Muslim Contexts.” Comparative Politics 42, no. 4: 375–​393.
Wagstaff, William, and Danielle F. Jung. 2020. “Competing for Constituents: Trends in
Terrorist Service Provision.” Terrorism and Political Violence, 32, no. 2: 293–​324.
 Exploring the Role of Islam in Mali     825

Ware, Rudolph, III. 2014. The Walking Qur’an: Islamic Education, Embodied Knowledge, and
History in West Africa. Chapel Hill, NC: University of North Carolina Press.
Watling, Jack, and Paul Raymond. 2015. “The Struggle for Mali.” Guardian, November 25, 2015.
https://​www.theguardian.com/​world/​2015/​nov/​25/​the-​struggle-​for-​mali.
Wickham, Carrie Rosefsky. 2002. Mobilizing Islam: Religion, Activism, and Political Change in
Egypt. New York: Columbia University Press
Wing, Susanna. 2009. “Family Law Reform in Francophone Africa: Women’s Rights and Legal
Pluralism.” Paper presented at Rethinking Development: Societal Transformations and the
Challenges of Governance in Africa and the Middle East, Conference at Yale University.
 Chapter 40

Isl am ist Part i e s a nd

Women ’s Repre se ntat i on
in Moro c c o
Taking One for the Team

Lindsay J. Benstead

When it comes to political representation, women face major obstacles in Arab and
Muslim countries. Female candidates encounter biases at the polls and in many coun-
tries, and ordinary women are less likely to enjoy clientelistic networks with politicians,
who are often men (Bauer and Burnet 2013; Beck 2003; Benstead 2016; Bjarnegård 2013;
Tripp 2001). Even when women win elections, they may not be able—​or even inclined—​
to promote liberal gender laws in weak and co-​opted legislatures.
The electoral success of Islamist parties and its implication for women has received
substantial attention and spanned fears that the electoral success of Islamist parties in
places like Turkey, Morocco, Egypt, and Tunisia will threaten women’s rights. Islamist
parties may incorporate women’s rights into their platforms, some argue, but they may
not genuinely promote gender equality (Clark and Schwedler 2003).1
Yet the relationship between gender and political representation in authoritarian
regimes is not straightforward. A growing number of studies of Islam, gender, and
governance suggest that women may be better represented in districts with Islamist
governments due to Islamist parties’ efforts to create effective social service institutions.
For instance, Meyersson (2014) finds that the Refah (Welfare) Party in Turkey, which
won several mayoral elections, improved poor women’s access to education. Henderson
and Kuncoro (2011) found that Islamist parties decreased corruption, while Blaydes
(2014) showed that women living in Egyptian localities controlled by radical Islamists
experienced better healthcare. Abdel-​Samad and Benstead (2016) find that in Tunisia,
women living in districts with Islamist and female legislators are more likely to interact
with elected officials than those living in electoral districts without female legislators,
due to female legislators’ access to sex-​segregated spaces, which increases ordinary
828   Lindsay J. Benstead

women’s interaction with deputies (especially with female deputies), and Islamist
parties’ desire to serve marginalized populations including women.2
The suggestion that Islamist parties’ institutionalization has positive consequences for
female citizens is compelling. Yet it has been tested in only a few cases. This chapter draws
on original data to test the Islamic mandate effect in Morocco, where women gained 35
of 325 seats (11%) due to a reserved seat quota in the 2002 elections when the Party of
Justice and Development (PJD) won 42 of 325 (13%) of all seats in parliament. Using data
from a survey of 112 parliamentarians and qualitative interviews in parliamentary offices
in Tangiers, a city in northern Morocco, this chapter tests the impact of electing Islamist
parties and women on female citizens’ access to networks with politicians, which they
use to seek help with constituency service requests.3 Expanding the Islamic mandate
effect (Abdel-​Samad and Benstead 2016), it makes a singular claim: the PJD’s more in-
ternally democratic party candidate recruitment fosters incentives to cultivate the party
reputation over personal reputation and fosters an organizational culture of teamwork,
resulting in more responsive constituency service to women.
The findings are striking. Islamist deputies receive more requests from women, as do
female deputies of any party. Secular male MPs interacted least with female citizens,
while Islamist male, Islamist female, and non-​Islamist female deputies had similar, high
levels of service-​oriented interactions with women.4
By highlighting the gendered impacts of Islamist governance, this study builds on an
already growing literature theorizing about the role that party institutionalization plays
in supporting Islamist parties’ electoral success (Cammett and Jones Luong 2014). It
also builds on gender and politics research on Islamist females’ work in sex-​segregated
environments by showing that the gender gap in interactions between parliamentarians
and citizens is also narrowed mixed-​sex teamwork in the context of constituency offices.

Conceptualizing Clientelistic
Responsiveness

This chapter focuses on state-​citizen interactions in clientelistic settings—​authoritarian

regimes with weak rule of law in which personal connections are needed to resolve social
challenges that citizens face. Clientelism is an “informal relationship between two actors
of asymmetrical socioeconomic power where the patron . . . controls . . . resources . . .
clients pursue but often cannot receive otherwise” (Manzetti and Wilson 2007, 953).
Citizens frequently contact intermediaries, including elected officials, for help with per-
sonal problems such as accessing healthcare, education, or jobs, which functions as a
form of constituency service when leaders are willing and able to assist them, a form
of service responsiveness.5 Because access to it is based on preferential networks, not
on fair treatment of citizens under the law, service responsiveness should be referred
to as “clientelistic responsiveness” (Benstead 2016). Yet although service responsiveness
 Islamist Parties and Women’s Representation in Morocco    829

takes place in clientelistic settings, it nevertheless serves as a vital link to the state that
helps citizens navigate the myriad bottlenecks of corrupt and inefficient bureaucracy.
Interactions between legislators and citizens may promote different forms of respon-
siveness,6 including by bringing women’s concerns into the policymaking sphere (Arat
2005, 113; Kudva 2003), an outcome akin to policy responsiveness. When citizens come
in contact with legislators and other representatives of the state, they may also ask for
other non-​exclusionary goods such as a road or other service for a whole community
(i.e., allocation responsiveness). Moreover, these interactions also promote citizens’
trust in their representatives, which is a form of symbolic responsiveness. In reality, the
dimensions of responsiveness overlap (Schwindt-​Bayer and Mishler 2005), forming a
seamless process of responsiveness to citizens’ needs (Celis et al. 2008).

The Case of Morocco

Morocco is a useful case for examining the effect of Islamist party orientation and
gender on women’s access to networks with elected officials. As shown in Table 40.1,
Morocco’s population is 99% Muslim. While many Moroccans identify as Arab, the in-
digenous Amazigh population accounts for 40% of its population. Social norms in many
rural areas are strongly traditional, with patrilineal tribes playing a strong role (Charrad
2001).7 Recent family code reform in 2003 (Waltz and Benstead 2006) and the imple-
mentation of a parliamentary gender quota in 2002 contributed to the improvements
in gender equality. However, women from lower socioeconomic classes and those in
rural areas can face particularly severe social and political discrimination in addition to
experiencing poverty.
Gender Inequality Index scores of .44 illustrate a high level of gender discrimination
in politics, healthcare, and economic life. Women’s low workforce participation (26%)
suggests that they could be relatively more excluded from formal political networks
(Ross 2008). Morocco has low historical levels of women’s representation. The first
woman was elected to the Moroccan parliament in 1993 (Gribaa 2008). At the time of
the study, thirty-​five (11%) of the Moroccan Chamber of Representatives were women;
thirty were elected to the national list as a result of a reserved seat quota negotiated by
the political parties in 2002, and five, including two from the PJD, were elected in reg-
ular constituencies (Darhour and Dahlerup 2013).
With a gross domestic product (per capita PPP) of $4,600, Morocco is also a devel-
oping country in which citizens encounter many social and economic challenges that
require them to seek help from friends, the government, and other sources. Due to un-
employment, unresponsive government, and corruption, Moroccans’ needs for housing,
jobs, and public services abound. According to the Arab Barometer (2015), many citi-
zens find services difficult to obtain: 34% found getting an identity document difficult,
75% stated that obtaining medical services is difficult, and 26% found getting help from
police difficult. While women did not have less access to these services, according to the
830   Lindsay J. Benstead

Table 40.1: Moroccan Case

Social Population (millions)i 32.0

ii
Percent Muslim (official) 99%
Percent Amazighiii 40–​50%
iv
Family law Conservative until 2004
Gender Inequality Indexv .444
Proportion women, House of 35 (10.7%)
Representatives (2002)vi
Political Regime typevii Lynchpin monarchy
viii
Parliamentary Powers Index .31
Freedom House (2012)ix Partly free
x
Electoral system (2002) Closed-​List PR, 325-​member Chamber of
Representatives, upper house
District magnitude 2–​5 seats, 95 geographical constituencies/​
30-​seat national constituency for women
Gender quota Reserved seat
xi
Economic GDP per capita (US$) $4,600
Oil dependencyxii Low
xiii
Percent women in workforce 26%

Source:
i
World Bank (2014);
ii
CIA, World Factbook;
iii
Ennaji (2005); Benstead and Reif (2013);
iv
Charrad (2001);
v
UNDP (2013);
vi
Darhour and Dahlerup (2013), Inter-​Parliamentary Union (IPU);
vii Jamal and Lust-​Okar (2002);
viii Fish and Kroenig (2009);
ix Freedom House (2012);
x Benstead (2008);
xi World Bank (2014);
xii
Proportion export earnings from oil/​natural gas: 0–​33% (low)/​34–​65% (moderate)/​66–​100% (high),
Cammett (2011, 101);
xiii World Bank (2014).

Arab Barometer, females were significantly more likely to have never tried, suggesting
patriarchal norms limit women’s access to public spaces (Benstead 2016; Sadiqi 2008).
Politically, Morocco is an authoritarian regime that is considered partly free by
Freedom House (Freedom House 2012), a result of decades of political liberalization
 Islamist Parties and Women’s Representation in Morocco    831

under the country’s linchpin monarchy, headed by King Mohammed VI, who ascended
the throne in 1999. As in other Arab countries, when the regime faced pressure to lib-
eralize politically in the 1980s, it feared that its grip on power might be challenged by
rising Islamist parties. The Moroccan PJD emerged several decades earlier, but it was
not until the 1990s that it made significant strides in the parliamentary elections. In the
1960s, Abdelkrim al-​Khatib founded the National Popular Movement, a royalist party
with significant support in the rural south. His party would later split, leading to the
emergence of the Islamist movement, the Movement for Unity and Reform (MUR),
from which he later founded the Party of Justice and Development (PJD).
After al-​Khatib’s initial experience in the 1977 parliamentary elections, the MUR
participated in elections in 1997 and was named the PJD in 1998. By 2002 the PJD
was still considered an opposition party, winning 13% of the seats in the Chamber of
Representatives. Yet the PJD achieved 13% of the seats after fielding candidates only in
some districts. Based on interviews that I conducted with journalists and civil society
actors in Morocco in 2007, I found two narratives explaining why the PJD did not run
nationwide. One view held that the PJD did not field candidates in all of the districts
to avoid being “too successful” and facing state repression, such as that meted out to
Islamist parties in Algeria and Egypt (Masoud 2014). Another view was that the PJD
was not able to break into the regime’s solid grasp on patronage networks in rural areas,
which traditionally benefited from electoral gerrymandering and ties with the palace.
As a party with a base among middle-​class, educated Moroccans, the PJD may have
avoided fielding candidates in rural areas where it did not expect to succeed electorally.
It was from this position in the opposition that the PJD deputies answered the survey.
The Islamist PJD has become a major political force in the country, rising from a small
opposition party in parliament in the 1990s to a party that enjoyed a plurality in par-
liament and the prime minister’s office in 2011. The regime tried to weaken the PJD be-
tween 2007 and 2010, only to see it become a major winner after the Arab uprisings. By
2011, in the wake of the Arab Spring, the PJD won a plurality (26%) of seats in the legisla-
ture and its leader was for the first time named as the country’s prime minister, though a
shadow cabinet still ran important government affairs (Buehler 2015).
Although the founders of the PJD were men, women joined in large numbers, and
within a generation they constituted one-​fourth of the party’s membership. As early
as 1999, the PJD’s General Secretariat enacted a quota of 15% female delegates at the
party congress (Wegner 2011, 54). In addition, like the Ennahda movement in Tunisia,
the PJD also never practiced gender segregation within its party structures (Wegner
2011), and it took steps early on to increase women in its internal party structures and
as candidates. In 2000, King Mohammed VI made several high-​level female political
appointments, and in 2002 political parties signed a “gentleman’s agreement,” reserving
thirty seats in the Chamber of Representatives for women (Sadiqi 2008). But the PJD is
not the only party in Morocco to do so. The socialist USFP, for instance, also has a strong
showing of female parliamentarians (Benstead 2021). This raises questions about how
and why electing women generally, and Islamist parties in particular, shapes women’s
representation.
832   Lindsay J. Benstead

Theoretical Framework:
Existing Literature

Applied to political life, network homosociality helps explain why electing women
supports greater gender balance in political life, whether at the elite or citizen levels.
Examining Thailand, Bjarnegård (2013) argues that women face barriers to elected office
because they lack ties to powerful individuals, who are usually male, which are needed
to access resources and succeed in elections. A network-​based approach suggests that
it is not gender stereotypes that reproduce women’s marginalization from politics—​
though stereotypes may contribute to the problem—​but men’s numerical dominance
in the political sphere. Bjarnegård argues that men have advantages accruing homoso-
cial capital—​that is, resources in same-​sex friendships and working relationships such
as political ties—​which is made up of two components. Homosocial capital consists of
instrumental resources that flow through political networks (“instrumental resources”)
and relational similarities (“expressive resources”). Expressive resources are relational
similarities (e.g., ease relating, fewer conflicts), which stem from traditional norms de-
fining public space as male and private space as female (Sadiqi 2008) and encourage sex
segregation, while instrumental resources are resources needed to reach goals, such as
winning office or attracting patronage. Since males are numerically dominant, instru-
mental and expressive resources are not in opposition to one another for men, helping
to explain why men are more likely to be recruited for and succeed in elections.
The concept of homosocial capital is also useful for understanding why women’s ac-
cess to linkages with politicians differs from that of men (Benstead 2016). Male leaders
have more men in their networks and thus are more likely to serve male citizens.
Electing female legislators increases women’s access to networks with elected officials
because women tend to have more heterosocial networks. Women have more females
in their network than men, but female legislators must engage with males to gain the re-
sources (i.e., instrumental resources) they need to be electorally successful.
Benstead (2016) offers evidence from several countries that women are marginalized
from linkages with elected officials. She finds in Morocco and Algeria that electing
women, especially through quotas, improves women’s access to national legislators. In
a nationally representative survey of 1,200 Libyans,8 Benstead (2015) found that women
were significantly less likely to contact parliamentarians than men. But electing women
reduced the gender gap in access to political networks because female constituents were
more likely than male citizens to contact female representatives. Female constituents
reported that the gender of the MP they contact was female 17% of the time, while male
citizens stated that the gender of the MP they contacted was never female. When male
constituents contacted an MP, it was never a female member.9 The same was true for
Tunisia, where citizens more often contacted same-​gendered local councilors, partic-
ularly when making large requests for club goods, which require a close relationship
(Benstead 2019).
 Islamist Parties and Women’s Representation in Morocco    833

The Islamic Mandate Effect

The Islamic mandate theory is helpful for explaining why electing Islamist parties pro-
mote female citizens’ access to networks with legislators. Abdel-​Samad and Benstead
(2016) argue that this is due to the Islamic mandate effect, which has both a “party” and
a “gendered” mechanism, leading to improved service responsiveness toward women
when Islamist parties are elected.10 They argue that Islamist parties strive to serve
marginalized populations, including women, through the institutionalization of con-
stituency service and levering female members to reach out to women in private and
public spaces (party component). This results in more interaction with women and
members of other marginalized groups.

Islamic Mandate Effect: Party Component

First, Islamists engage in more interactions with female citizens because they institu-
tionalize service provision by establishing direct contact with citizens through a net-
work of party and local offices, mosques, social service institutions, and schools
(Aldrich 1995; Keefer and Vlaicu 2008; Roháč 2013, 256). Moreover, they emphasize sup-
port from the needy in their communities (Bayat 2002; Catusse and Zaki 2009; Esposito
1999; Hamzeh 2001; Harik 1994; Ismail 2001; Milton-​Edwards 2007; Troudi 2011; Walsh
2006; Wiktorowicz 2002), which also support greater access to networks with legislators
by female citizens.11 Because Islamist parties use service provision to promote party rep-
utation for fairness and social responsibility (Cammett and Jones Luong 2014), they
are different from other non-​Islamist parties in authoritarian regimes, which, while
they may not all be corrupt, tend to operate in a highly clientelistic way (Abdel-​Samad
2009; Benstead 2008; Brumberg 2002; Cammett and Jones Luong 2014; Esposito 1999;
Wickham 2002), in which individual members seek to foster their own reputation over
that of the party. Brooke (2019) argues that in Egypt, the Muslim Brotherhood was
successful not through proselytizing or electoral mobilization or reaching out to the
poor, but by providing high-​quality, paid medical services to the middle classes, which
improved the group’s reputation. As such, the Brotherhood sought to build a good rep-
utation more than to engage in proselytizing or electoral mobilization.12 Because of
this approach, Islamist parties stand out from others in different ways, including by
serving more female citizens than other more clientelistic parties (Abdel-​Samad and
Benstead 2016).

Islamic Mandate Effect: Gendered Component

Second, Islamists engage in more interactions with female citizens because they involve
female members in reaching out to women citizens in sex-​segregated environments,
leading to an increase in women’s access to networks with parliamentarians. This is the
second, “gendered” component of their theory of Islamist representation of women.
Female members not only mobilize voters, but they also establish new social networks
(Clark 2004b) by meeting women who are not engaged in political networks through
834   Lindsay J. Benstead

social functions like homes, weddings, and social teas (Arat 2005; Blaydes and Tarouty
2009). By emphasizing piety, Islamist parties foster “sacred” spaces that ensure that
women can safely enter political offices and participate in meetings while being free
from harassment and negative impacts on their reputation. Islamist parties in several
countries utilize quotas to increase the number of women not only on party lists but also
within their organization (Clark and Schweder 2003), and often have women’s sections
in their party offices and women’s committees (White 2002; Yadav 2010). Direct contact
with constituents allows Islamist women to access female citizens who are practicing
Muslims and may be reluctant to approach male legislators, and to talk to someone who
understands the obstacles they face.

Extending the Islamic Mandate Effect: The Role

of Mixed-​Sex Teamwork
Abdel-​Samad and Benstead (2016) theorized that female deputies’ promoted women’s
representation through fostering interactions with women in sex-​segregated spaces and
by creating spaces that allow gender mixing while respecting piety. Yet neither the MUR
nor the PJD practiced sex segregation within their party structures. Rather, teamwork
as a party involving party members and legislators at all levels has characterized their
work. This does not mean that sex segregation does not occur within the PJD or any
other party. Indeed, homosociality is a feature of human interactions in many settings
(Bjarnegård 2013). Rather, an extension of the Islamic mandate theory holds that mixed-​
sex teamwork, in addition to interactions with citizens in sex-​segregated settings,
contributes to the PJD’s ability to establish networks with female citizens. I argue that
the PJD’s stronger party institutionalization enhances legislators’ electoral incentives
to work in mixed-​sex teams and build their party’s reputation, leading to Islamist
parliamentarians’ more frequent service interactions with female citizens.

Measuring Incentives to Promote the Party Reputation

Islamist MPs do in fact perceive greater incentives to cultivate their party’s reputa-
tion over their personal reputation than do members of other parties (Benstead 2016).
I asked MPs to tell me about their electoral strategy by responding to the following
survey question: “In your opinion, what is the best strategy for seeking reelection to the
Chamber of Representatives? Build a strong party image (=1); build a strong party and
personal image, but focus on the party image (=2); build a strong party and personal
image equally (=3); build a strong party and personal image, but focus on the personal
image (=4); and, build a strong personal image (=5).” I found in Morocco that the PJD
was the most party-​focused, scoring 1.6 on average compared to 3.2 for all other parties
(p<.001).13 That is, PJD deputies were more likely than other MPs to believe that the
pathway to re-​election depending on what they did for their party as a unit, rather than
their efforts to establish their personal reputation in the district.
 Islamist Parties and Women’s Representation in Morocco    835

Incentives to Promote the Party Reputation and Teamwork

I also found evidence for a link between these incentives and the tendency toward
teamwork involving party members and legislators of both sexes. I conducted field
research throughout Morocco, including in the northern city of Tangiers (Benstead
2008). During my visit to Tangiers, I visited the offices of the four parliamentarians who
represented one of the city’s electoral districts. The type of office that each deputy had
and its organizational culture differed markedly, with profound implications for citi-
zens’ access.

Office 1
The first office that I visited belonged to a socialist deputy—​a lawyer—​with whom I had
arranged a meeting. His parliamentary office doubled as his legal practice and was
located on an upper floor of a commercial building with a doorbell and receptionist. The
MP was friendly and approachable. He detailed the types of requests that he received
from constituents. Yet no constituents were present at the time of the visit. This deputy,
while working conscientiously as a lawyer and member of the House of Representatives,
did not have a constituency office that was set up to foster direct contact with citizens.

Office 2
The second office that I visited was that of a pro-​regime party MP who utilized his home
as his office. Although I rang the bell, there was no answer. A young girl was sitting be-
side the deputy’s gate in a wheelchair asking passersby for money. This deputy, if he had
a constituency office at all, utilized his home and was not on this occasion available to
visitors, although he may have been at other times.

Office 3
The third office I visited was that of two Islamist deputies from the district—​a man and
a woman—​working together in a party office. Upon reaching the neighborhood, I asked
a man standing beside his truck for help locating the office. He said, “I know them. They
are the only MPs around here who work hard. I’ll take you there.”14 Then he took me to
an office where a female receptionist welcomed me and took me into the reception area,
where the two parliamentarians were seated with two other party members. Together,
the male and female MPs were engaged in talking with their constituents about an issue
that concerned them. Anyone in the constituency could directly enter the office without
making a prior appointment.
In this example, the PJD was the only party that had elected two representatives in the
district, and thus it was more natural that they would work together. Yet this teamwork
and direct contact with citizens was aided by the presence of a constituency office with a
staff that was dedicated to constituent relations rather than to another function, as was
the case with the lawyer’s office. While I visited other non-​Islamist offices in which many
citizens were present, this PJD office from Tangiers was the only one in which multiple
members were present working together with citizens.
836   Lindsay J. Benstead

This team orientation and institutionalization of constituency services was also evi-
dent in the qualitative interviews. When I asked parliamentarians to tell me the average
number of requests that they receive per month, many Islamist deputies clarified this,
stating, “It is the party that takes care of citizens who ask for help.”15 The PJD, in con-
trast to other parties, indicated that the party works together as a unit to resolve citizens’
needs and build the party’s reputation.
The PJD’s candidate recruitment practices also contrast with those of non-​Islamist parties,
creating a strong incentive for teamwork, based upon my field research in Morocco. The PJD
was the only party that noted in interviews that they conduct internal votes among party
members to select candidates. When I asked parliamentarians whether they planned to run
again, all MPs could respond, but PJD MPs’ response was often, “That’s not up to me. It’s the
party that decides.”16 No other socialist or pro-​regime party members stated that it was the
party that decides who would be a candidate in the future, and party-​switching is common.
One female deputy from a socialist party shared the internal power struggles regarding who
would be selected to head the party’s list for the female seats.17 She stated that party members
traded favors in an effort to jockey to be a candidate on the party’s electoral list. Other MPs
noted that they had changed parties when they were not able to be successful.
Thus, to sum up, expanding on the Islamic mandate effect (Abdel-​Samad and
Benstead 2016), I argue that the PJD’s more internally democratic candidate recruit-
ment practices foster incentives to cultivate the party over candidates’ personal reputa-
tion and an organizational culture of mixed-​sex teamwork in addition to women’s work
in sex-​segregated settings, resulting in more responsive constituency service to women.

Hypotheses

Accordingly, I propose several hypotheses. First, I expect that, due to network

homosociality, female parliamentarians from any party will provide more constitu-
ency services to female constituents than male parliamentarians, helping to reduce the
gender gap in access to connections with parliamentarians (H1).

Homosociality Hypothesis
H1: Female parliamentarians will engage in more constituency contacts
with female constituents than male parliamentarians.

Second, because Islamist parties more than others use voting by party members to foster
more internally democratic candidate selection and foster incentives to promote the
party image than do other parties, Islamist male parliamentarians will work in mixed-​
sex teams and provide more constituency services to female constituents and other
marginalized citizens than non-​Islamist male parliamentarians (H2).
 Islamist Parties and Women’s Representation in Morocco    837

Islamic Mandate Hypotheses: Party Mechanism

H2: Islamist male parliamentarians will engage in more constituency
contacts with female constituents than non-​Islamist male parliamentarians.

Finally, I hypothesize that Islamist female parliamentarians will provide more con-
stituency services to female constituents than non-​Islamist female parliamentarians
(H3). I argue that this occurs because female Islamists meet female constituents in sex-​
segregated environments more than non-​Islamist females to adhere to Islamic values of
piety. In addition I argue that Islamist parties’ use of teamwork in mixed-​sex teams also
promotes female citizens’ interactions with parliamentarians.

The Islamic Mandate Hypotheses: Gendered Mechanism

H3: Islamist female parliamentarians will engage in more constitu-
ency interactions with female constituents than non-​
Islamist female
parliamentarians.

Data and Methods

The data used to test the homosociality (H1) and Islamic mandate hypotheses
(H2–​H3) are from an original paper-​and-​pencil survey of 112 Moroccan Chamber
of Representatives members conducted between September 2006 and May 2007.
I used two-​stage probabilistic sampling to draw a representative sample of depu-
ties to whom I administered the questionnaire, either at the parliament or in their
district.18 The survey focused on constituency activities and had forty questions in
French and Arabic. When possible, I also conducted semi-​structured interviews
with deputies when I administered the paper-​and-​pencil survey.19 While many
interviews were conducted at the national parliament, others took place in the
electoral constituency itself, including in Tangiers, a city in northern Morocco
that is the particular focus of this paper. Through the survey and qualitative in-
terview, I learned more about the constituency service practices of the legislators.
I conducted these interviews in district offices and the parliament. The survey had a
response rate of 57%.20

Measurement of Dependent and Independent Variables

To measure the dependent variable, interactions with female constituents, I use this
item: “What percent of these requests come directly from women? Less than 10% (=1),
838   Lindsay J. Benstead

Table 40.2: Moroccan House of Representatives (2002–​2007)

Total Number of Percent seats
Parliamentary group seats females elected held by female

Parliamentary groupsa
Movement Group (MP) 72 5 6.9
Independence Group of Unity and Equality (Istiqlal) 60 6 10.0
Group of the National Rally of Independents (RNI) 39 4 10.3
Group of the Constitutional Democrat Union (CD) 28 2 7.1
Socialist Group (USFP) 48 5 10.4
Group of the Socialist Alliance 21 2 9.5
Justice and Development Group (PJD) 42 6 14.3
Small Parties and Independents
Deputies of the Democratic Forces Front (FFD) 8 2 25.0
Deputies of the Unified Socialist Left (GSU) 3 0 0
Deputies of the Alliance of Freedoms 1 0 0
Deputies without Party Affiliation 3 0 0
Total 325 32 9.8

a Parliamentary group composition at the survey, 2006–​2007.

10–​19% (=2), 20–​29% (=3), 30–​39% (=4), 40–​49% (=5), 50–​59% (=6), 60–​69% (=7), 70%
or more (=8).” The mean response in the full sample of MPs is 3.87.
To measure the key independent variables, deputy gender, and party affiliation,
I used the gender and party affiliation reported on the survey. As shown in Table 40.2,
twenty-​six parties contested the elections in 2002 in ninety-​five districts with two to
five seats each, as well as the national list, for a total of 325 seats. The Socialist Union
of Socialist and Popular Forces (USFP) won 50 seats; the country’s oldest party, the
Istiqlal (Independence) Party, received 48 seats; the Islamist PJD gained 42 seats; the
conservative party of the makhzen, the National Rally of Independents (RNI), took 41
seats; the People’s Movement and the National People’s Movement, which subsequently
formed a single Popular Movement (MP), received 27 and 18 seats, respectively; and,
the Constitutional Union (UC) had 16 seats. The remaining fifteen parties received 83
seats. These parties formed parliamentary groups to establish offices in the parliament.
Nearly all parties elected female members, most to the 30 reserved seats for women.
Five women were elected in geographical constituencies as well. A few small parties
elected no women, while most parties elected between 7 and 25 percent women. The
PJD elected 14% women, which was greater than any other party except the Democratic
Forces Front (FFD).
 Islamist Parties and Women’s Representation in Morocco    839

The key explanatory variable is MP gender and party affiliation. The Islamist Party of
Justice and Development (PJD) is the only Islamist party, while other parties are coded
as non-​Islamist: the Koutla (governing coalition)—​Socialist Union of Popular Forces
(USFP), Istiqlal, Party of Progress and Socialism (PPS), and the Unified Socialist Party
(GSU)—​were moderately oppositional and may not differ from members of pro-​palace
parties, including the National Rally of Independents (RNI). In the sample as a whole,
74% are male non-​Islamist deputies, 9% are female non-​Islamist deputies, 13% are male
Islamist deputies, and 3% are female Islamist deputies. Thus, the conclusions that can be
drawn from comparing Islamist female and non-​Islamist female deputies is limited by
the small n.

Descriptive Statistics
In support of H2, the PJD interacts with women more than any other party. As shown
in Figure 40.1, the PJD was the only party that achieved gender parity (5.7) in its con-
stituency service provision. All other parties scored 4.0 or lower on the scale of in-
teraction with female constituents, and the difference across the parties is statistically
significant (p < .01).
To be sure, this effect could partially be driven by the relatively high number of
women in the PJD party, at 14% (Table 40.2). However, one other party, the FFD,
had more women, at 25% of its MPs. As shown in Figure 40.2 and in support of the
homosociality hypothesis (H1), female members interact with female citizens more
than male members. Male members serve nearly 30–​39% women (3.68 on an 8-​point
scale), while female members serve 50–​59% on average (5.57 on an 8-​point scale).

6.0 5.7

5.0
3.9 3.8 4.0
4.0 3.6 3.5 3.4
3.0
3.0
2.0
2.0

1.0

0.0
t

lal

FP

CD

I
RN
en

ist
FF

PJ
GS
US
iq
em

ial
Ist

oc
ov

eS
M
ar

nc
l

lia
pu

Al
Po

Figure 40.1: Percentage of requests received from female constituents, by party.

840   Lindsay J. Benstead

7.00

6.00 6.00
6.00 5.78 5.67

5.00

3.87
4.00 3.68
3.27
3.00

2.00

1.00

0.00
All MPs All male MPs All female MPs Non-Islamist Non-Islamist Islamist male Islamist female
male MPs female MPs MPs MPs

Figure 40.2: Percentage of requests received from female constituents, by MP identity.

Question wording: “What percent of these requests come directly from women? Less than 10%
(=1), 10-​19% (=2), 20-​29% (=3), 30-​39% (=4), 40-​49% (=5), 50-​59% (=6), 60-​69% (=7), 70% or
more (=8).” The average among all MPs is 3.87.

Yet the comparisons between male and female deputies from Islamist and non-​
Islamist tendencies also support the Islamic mandate hypothesis (H2). Female
deputies serve more citizens than male deputies, but male Islamist deputies are as
responsive to females as their female colleagues from any party. The average non-​
Islamist male deputy reported a score of 3.27 on a scale of 1–​8 (about 20–​29% of
requests from women), while an average Islamist male deputy has a score of 5.67
(almost 50–​59% on an 8-​point scale). This suggests that Islamists generally institu-
tionalize constituency service in a way that helps them reach out to citizens from
different backgrounds, especially those who are more marginalized from formal
networks.
Yet unlike in the Tunisian case (Abdel-​Samad and Benstead 2016), there is no ev-
idence for the Islamic mandate hypothesis’s gendered component, which expects
that Islamist female MPs are more prone to serving citizens in sex-​segregated
environments than are non-​Islamist female MPs. Inconsistent with H3, female
parliamentarians do not differ from one another in their responsiveness to female
constituents. Non-​Islamist female MPs and Islamist female MPs receive about 50–​
60% requests from women. This could be due to the small sample size. Only 3% of
MPs who responded were Islamist females, due to their small number within the
Chamber of Representatives.
 Islamist Parties and Women’s Representation in Morocco    841

Results and Discussion

I use ordinary least squares (OLS) regression to test the impact of the deputy party iden-
tification and gender on responsiveness to female citizens. The overall n is low, but the
regressions are significant and robust to multiple specifications. Model 1, which includes
only parliamentarian gender, explains 30% of the variation in responsiveness to fe-
male constituents, while Model 2, which includes the intersection of parliamentarians’
gender and Islamist orientation, explains 35% of the variation in the dependent variable
in Morocco. Breusch-​Pagan/​Cook-​Weisberg heteroskedasticity tests indicate the con-
stant variance assumption is not violated.

Measurement of Control Variables

In the regressions, I include three measures of electoral incentives: a proxy for dis-
trict population, perceived incentives to provide clientelism, and political ambition
(Table 40.3). Women elected in larger districts may be more responsive to females,
due to weaker incentives to cultivate a personal vote and greater freedom to serve
less politically influential constituencies; a high density of civil society organizations;
and urban, employed female constituents. Women elected in small districts—​such as
Moroccan districts, with 2–​5 seats—​may have strong incentives to cultivate a personal
vote among female and male constituents and lower responsiveness to women. District
population, measured with five levels, is included as a control variable.21
Because numerous factors, including electoral competitiveness, can affect how in-
clusive legislators’ networks are of women, I control for perceived importance of pro-
viding services, measured as agreement on a four-​point Likert scale: “Serving citizens is
important for maintaining electoral support.” Political ambition could affect inclusive-
ness. Thus, a measure of whether the respondent plans to seek higher office is included.
Deputies who plan to run again or perceive electoral benefits of service provision may
be less likely to serve females because patriarchally organized constituencies, such as
tribes or Sufi brotherhoods, are more advantageous to serve.
Second, I control for the extent to which deputies have networks helpful for solving
requests: education, tapping workforce participation, and political experience.
Education has six levels, ranging from high school or less to doctorate. In Morocco,
members may hold concurrent functions, such as mayor. Previous political experience
is measured as a dummy variable equal to one if the respondent indicated a past public
role, such as local councilor. The models also include a dummy variable equal to one if
the member served at least one previous parliamentary term. Those with experience
should have denser networks and be more likely to serve a broad constituency, including
men and women.
842   Lindsay J. Benstead

Table 40.3: Measurement of Control Variables

District population:
Morocco: “What is the best description of your electoral zone? Large city, more than 200,000=4/​
100,000–​200,000=3/​less than 100–​000=2/​rural and small villages=1/​national list=5.”
Electoral incentives:
Electoral basis: “Serving citizens is important for maintaining electoral support: Strongly agree=4/​
agree=3/​disagree=2/​strongly disagree=1.”
Political ambition: “Will you try to attain a higher political position in the future? Definitely=2/​
probably=1/​no=0.”
Network density:
Education: “Please indicate your highest level of education: High school or less=1/​baccalaureat=2/​short
cycle university diploma=3/​long cycle university diploma=4/​master’s=5/​doctorate=6.”
Previous political experience, 1 for any of the following: “What was your occupation prior to being
elected?”/​“How many years have you served as an elected official a local council?”/​“How many years
have you served as an elected official in a regional council?”/​“Which of the following positions do you
hold or have you held?”/​“Do you have a second public function?”
Previous mandate: “How many mandates have you served as a deputy? Current only=0/​two,
including current=1/​three or more, including current=1.”
Controls:
Age: Less than 35 years=1/​35–​39=2/​40–​44=3/​45–​49=4/​50–​54=5/​55–​60=6/​61 or more=7.

Third, I control for age and district socioeconomic level. Age has six levels, from
35–​39 years to 61 or more. In rural and less developed areas, women should be less likely
to approach members, leading to less inclusiveness. Development level is measured as
the log of hospital beds/​100,000 people.

Homosociality Hypothesis (H1)

As hypothesized (H1) and consistent with existing research (Benstead 2016), female
members are more responsive to women than are males, but the model is significant
only at the p<.10 level (Model 1). Responsiveness is 2.69 higher (p<.05), on average,
on an 8-​point scale for female than male deputies, suggesting that electing women
decreases the gender gap. But given the small number of observations, this difference
does not reach conventional significance levels.

Islamic Mandate Effect: Party Component (H2a and H2b)

Responsiveness is also higher among Islamist male deputies than non-​Islamist male
deputies (Model 2), as hypothesized by the Islamic mandate effect party component
(H2). Islamist male deputies are 1.80 units more responsive than non-​Islamist male
 Islamist Parties and Women’s Representation in Morocco    843

deputies, all else being equal (p<.05). In Model 2, female non-​Islamist deputies are 3.68
units more responsive than non-​Islamist male deputies (p<.05). But due to the small size
of the sample, female non-​Islamist deputies are not significantly more responsive to fe-
male constituents than Islamist male deputies or Islamist female deputies.

Islamic Mandate Effect: Gendered Component (H3)

Contrary to the Islamic mandate effect’s gendered component (H3), the evidence does
not support the claim that, in Morocco, Islamist female deputies are more likely to re-
spond to women than non-​Islamist female deputies due to the possibility of greater con-
stituency service in sex-​segregated environments. This is somewhat surprising, given
female Islamist deputies were particularly likely to serve women in Tunisia. The lack of
evidence in support of H3 could be due to the small number of Islamist females.

Control Variables
Responsiveness to women is higher in urban districts with populations greater than
200,000. In Model 1 (Table 40.4), the average responsiveness is 1.63 units higher, on
average, than in rural areas (p<.05), while in Model 2 it is 1.29 units higher (p<.05).
Perceptions of the importance of providing services or the desire to run again in the
future do not predict responsiveness for female citizens. Neither deputy political experi-
ence, age, and education, nor the development level of the district were significant.

Conclusions and Implications

Theories of homosocial capital offer fertile ground for examining the impact of Islamist
parties on women’s access to constituency service. In the study, nearly 30–​39% of
requests to legislators in Morocco come from women. Yet, consistent with previous
work, electing women improves the potential for their service and symbolic representa-
tion by increasing women’s interaction with elected officials. Contrary to conventional
wisdom, electing Islamist parties also contributes markedly to closing the gender gap.
Islamist males were as likely as women from any party to serve female citizens. This
supports the claim that Islamist parties foster women’s representation by working in
mixed-​sex teams in order to support their party’s reputation over their own.
Yet the results did not support the hypothesis that electing female Islamists has a par-
ticularly strong impact on improving responsiveness to women. This may be due to the
small number of female Islamists in the study. The number of women elected—​thirty-​
five—​is too small to speak definitely about how partisan differences across these women
may impact their representative styles. It may also mean that the improvements in
Table 40.4: Determinants of Service Responsiveness to Women
Model 1 Model 2

OLS coefficients

Electoral incentives
Urban (Less than 100,000)1 .89(.79) .96(.77)
Urban (100,000-​200,000)1 1.26(.88) 1.09(.87)
Urban (More than 200,000)1 1.63(.63)* 1.29(.64)*
National list .80(1.47) .37(1.49)
Electoral basis (Agree)2 -​.47(1.02) -​.81(1.07)
2
Electoral basis (Strongly agree) -​1.15(1.02) -​.95(1.02)
Political ambition (Probably)3 -​.11(.69) -​.36(.71)
3
Political ambition (Certainly) -​.31(.69) -​.47(.73)
Network density
Higher education .03(.16) .06(.16)
Previous/​current political experience .45(.55) .66(.56)
Previous parliamentary mandate .10(.51) .25(.50)
Controls
Higher age -​.16(.19) -​.22(.20)
Higher socioeconomic level (hospital beds) -​.43(.36) -​.50(.35)
Parliamentary traits
Female MPs 2.69(1.38)†
Female non-​Islamist MPs4 -​ 3.68(1.52)*
Male Islamist MPs4 -​ 1.80(.89)*
Female Islamist MPs4 -​ 2.56(1.91)
Constant 7.27(3.10)* 7.85(3.05)*
Test of linear hypothesis:
H0=b (Female non-​Islamist)–​b (Male Islamist)=0 -​ F(1, 65)=1.27
H0=b (Female non-​Islamist)–​b (Female Islamist)=0 -​ F(1, 65)=.33
H0=b (Male slamist)–​b (Female Islamist)=0 -​ F(1, 65)=.16
N 82 82
F 2.03 2.16
Prob.>F .1513 .0156
R2 .2980 .3468

† p <.10
* p <.05 ** p <.01 *** p <.001 two-​tailed test. Standard errors in parentheses.
Comparison categories:
1 Rural/​villages.
2 2=Disagree.
3 No.
4 Male non-​Islamist MP.
 Islamist Parties and Women’s Representation in Morocco    845

women’s access to Islamist MPs are due more to party institutionalization and mixed-​
gender teamwork than to female Islamists’ efforts in sex-​segregated environments like
women’s organizations or mosques. I contend that both of these processes contribute to
Islamist parties’ unique engagement with female citizens.
To be sure, interaction and the ability to make a request is not the same thing as that
request being favorably addressed or citizens developing feelings of trust toward their
representatives. Yet women’s access to members is the first step in a potential chain of re-
sponsiveness that can lead to symbolic responsiveness (trust, political engagement) and
substantive outcomes, including receiving help with a personal or community problem.
The findings have important implications for gender and governance research by
showing that parties that institutionalize and become less clientelistic are more respon-
sive to women. Islamist parties do not completely reject clientelistic practices, but when
it comes to service provision, they are more internally democratic than other parties.
This positively impacts marginalized groups, including women. This is consistent with
the argument that clientelism and informality, rather than Islam or something else,
inhibits women’s equality in political life (Sung 2003).
Existing literature on Islamist parties illustrates the integral role that party reputation
plays in supporting Islamist parties electoral success (Brooke 2019; Cammett and Jones
Luong 2014). This chapter extends this work by providing new evidence that Islamists
MPs believe that promoting their party’s reputation is central to their electoral success.
And it also shows the implications of this orientation for mixed-​sex teamwork and fe-
male citizens’ access to networks with legislators.
The findings also extend work on Islam, gender, and politics by suggesting that party
institutionalization and the creation of mixed-​sex constituency service environments is
as important as engagement in sex-​segregated environments for ensuring that elected
officials engage with citizens from more diverse social backgrounds.
The findings concerning Islamist parties are likely to be broadly applicable to clien-
telistic societies in the Arab and Muslim world, where Islamist parties have the oppor-
tunity to mobilize supporters through the openness to engage with them about their
basic needs that are unmet by the state.22 The impact of electing women is likely to
shape women’s representation in a broader swath of countries, given male advantages
in accessing patronage networks in Africa (Tripp 2001), Asia (Bjarnegård 2013), and
the Middle East (Benstead 2016). Because Morocco has achieved some expansion of
women’s rights in the form of the 2003 reform of the Moudawana, the effects of electing
women and Islamists are likely to have a greater impact than in more conservative na-
tions such as Yemen.
Nevertheless, this must be tested. A comparative framework should take into ac-
count not only how conservative a society is, but also the number of women engaged
in the Islamist party as elected officials and as party members. New research should
also directly test the role of mixed-​sex teamwork and gender segregation in improving
women’s access to elected leaders. More specifically, this research must focus on insti-
tutional differences in how women are engaged in Islamist parties, whether in a more
gender-​segregated way, as in the case of the Islamic Action Front (IAF) in Jordan (Clark
846   Lindsay J. Benstead

and Schwedler 2003), or without segregation, as has been the case for the Moroccan PJD
and the Tunisian Ennahda.
Other citizen and member identities must also be considered in future research.
When parliamentarians and citizens share a family or tribe, they are more likely to
have a network with an MP (Benstead 2015). For instance, I find in Jordan that men and
women are equally likely to enjoy a network with an MP. This may be due to tribes’ po-
litical influence, which improves women’s service responsiveness, because women can
leverage an intersecting identity (i.e., co-​ethnicity) to access services (Benstead 2015).
Co-​ethnicity has also been found to be an important factor facilitating women’s political
engagement and electoral success (Kao and Benstead 2021).
Future surveys should include questions on citizens’ interactions with parlia­
mentarians, and, if possible, surveys of parliamentarians should also be conducted in
other countries. In critical cases like Egypt, this is impossible, due to the dissolution of
the Muslim Brotherhood parliament. New research should also consider different types
of Islamist parties, including those that are more authoritarian and seek to establish a
theocratic state. Future studies should also assess other intersectional aspects of parlia-
mentarian and citizen identity, such as ethnicity and class, and look more closely at the
extent to which women and men who are helped by Islamist and other parties turn out
to vote for them. Through these extensions, scholars will have a better understanding of
how electoral politics impact responsiveness to women and other marginalized citizens.
Acknowledgments
I would like to thank the funders of the survey of Moroccan parliamentarians, which in-
clude the Charles Cannell Fund in Survey Methodology, the William Davidson Institute, and
the Nonprofit and Public Management Center at the University of Michigan, and the Social
Science and Humanities Research Council, Canada. Special thanks to Adil Fala and Alae Serra,
who provided me with invaluable assistance with the survey.

Notes
1 The PJD is not a militant wing but a political party that always focused on electoral par-
ticipation. Nor did it have social service institutions or directly participate in da’wa
(proselytizing). Instead, it participated in elections and pushed for social and economic
concerns more than religious ones (Brown and Hamzawy 2010).
2 In this chapter, marginalized citizens are those who have been excluded from traditional
political networks. Studies from the Middle East indicate that in many countries, women
are less able to access political networks than men. See Benstead 2016.
3 The measure of service responsiveness taps whether citizens have asked for help, not
whether the service was actually received. This is an important avenue of citizen engagement
in clientelistic settings. The survey is based on self-​reported data from parliamentarians. Yet
there is no reason to believe that male Islamists would be more likely due to social desira-
bility to report interacting with females.
4 Unlike my previous work in Tunisia with Abdel-​Samad, I found that Islamist female
deputies in Morocco are no more likely to interact with female citizens in their districts
than women from socialist or other parties. This may be due to the small number of
 Islamist Parties and Women’s Representation in Morocco    847

observations or it may suggest that all female parliamentarians tend to interact equally
often with female constituents in different types of settings (e.g., constituency offices, civil
society organizations).
5 According to my research, the most common requests made to Moroccan
parliamentarians involve “social problems such as medicine, hospital beds,” “work,” “local
development, security, parks”; “job transfers for teachers, family reunification,” and “edu-
cation” (Benstead 2016).
6 Building on Pitkin’s (1967) conceptualization of representation as responsiveness rather
than policy congruence, Eulau and Karps (1977) identified four dimensions of respon-
siveness: symbolic, service, policy, and allocation (i.e., district projects) responsiveness.
Symbolic responsiveness refers to interactions with constituents (Eulau and Karps 1977),
which create conditions for service, allocation (i.e., club goods), and policy (i.e., legisla-
tive) responsiveness.
7 Tribes in Morocco are strong and supportive of the monarchy; patronage flows to rural
notables, strengthening tribes (Charrad 2001).
8 2013 National Democratic Institute poll of 1,200 Libyans. Diwan Research. Benstead, Lust,
and JMW Consulting. Transitional Governance Project (2021).
9 In Jordan, men and women are equally likely to ask a female for help. This may be due to
tribes’ political influence, which improves women’s service responsiveness because women
can leverage an intersecting identity (i.e., co-​ethnicity) to access services (Benstead 2015).
10 Abdel-​Samad and Benstead (2016) adapt this term from Franceschet and Piscopo (2008),
who argued that quota-​elected women’s perceive a mandate, or “obligation to act on behalf
of women,” due to their underrepresentation (394–​395). Along similar lines, research on
race in the United States finds black state legislators were more likely to respond to letters
from black constituents, even those living outside their district, suggesting an intrinsic
motivation to serve members of one’s own, marginalized group (Broockman 2013).
11 Some theorists claim this constitutes an exchange of favors for votes (Alterman 2000;
Flanigan 2008; Hamzeh 2001). Others argue that these activities are different from tra-
ditional clientelistic networks because they are based on “horizontal networks” between
group members working to provide services to citizens to promote the party’s goals (Clark
2004a; Roháč 2013, 267). Islamist parties focus on fairness and responsiveness, but may
also utilize some traditional patronage networks or establish new ones. The concept of cli-
entelistic responsiveness (Benstead 2016) suggests that these exchanges and interactions
are inherently clientelistic. This is due to the context in which they are embedded, but this
does not diminish their importance for citizens needing support for their needs or their
consequences for improving trust their representatives.
12 Yet, unlike the Muslim Brotherhood in Egypt, the PJD does not have a service network. It
is also not the only political party in Morocco with a history of electing women.
13 The Socialist Alliance (3.6) Istiqlal (3.3), RNI (3.3), USFP (3.1), Popular Movement (MP)
(3.1), GSU (3.0), CD (2.9), and FFD (2.7).
14 Interview with male citizen, 25–​40 years old, Tangiers, Morocco (2008).
15 Interview with male MP, 41–​55 years old, Rabat, Morocco (2008).
16 Interview with male MP, 41–​55 years old, Rabat, Morocco (2008).
17 Interview with female MP, 41–​55 years old, Casablanca, Morocco (2008).
18 First, random stratified sample with parliamentary group as strata; females and members
of parliamentary groups with fewer than 8% of seats over-​sampled. Second, random strati-
fied sample of electoral districts with region and population as strata.
848   Lindsay J. Benstead

19 I also conducted qualitative interviews with 10% of the MPs who responded to the survey
and agreed to a secondary qualitative interview. No differences in item response rates
were noted.
20 American Association for Public Opinion Research (AAPOR), Response Rate 3 (RR3).
21 Electoral districts in Morocco do not correspond to administrative areas. Gerrymandering
increases the importance of rural areas, where palace support is strong. District popula-
tion is difficult to obtain and estimates rendered regressions insignificant. MP age was not
statistically significant and not included in the regressions.
22 Effects may not generalize where Islamist parties establish an authoritarian, theocratic
government, or when few women are engaged in the party, whether as elected officials or
party militants.

References
Abdel-​Samad, Mounah. 2009. “Exchanging Favors: The Predominance of Casework in
Parliamentarians’ Behavior in Jordan and Lebanon.” Journal of Legislative Studies 15, no.
4: 420–​438.
Abdel-​Samad, Mounah, and Lindsay J. Benstead. 2016. “Why Do Islamist Parties Reduce
the Gender Gap in Service Provision?” Paper presented at the “After the Uprisings: Public
Opinion, Gender, and Conflict in the Middle East” workshop, Kansas State University, May
5, 2016.
Aldrich, John H. 1995. Why Parties? The Origin and Transformation of Political Parties in
America. Chicago: University of Chicago Press.
Alterman, Jon B. 2000. “Egypt: Stable, but for How Long?” Washington Quarterly 23, no.
4: 107–​118.
American Association for Public Opinion Research (AAPOR). 2021. https://​www.aapor.org/​
Education-​Resources/​For-​Researchers/​Poll-​Survey-​FAQ/​Response-​Rates-​An-​Overview.
aspx.
Arab Barometer. 2015. https://​www.arabbarometer.org/​.
Arat, Yesim. 2005. Rethinking Islam and Liberal Democracy: Islamist Women in Turkish Politics.
Albany: State University of New York Press.
Bauer, Gretchen, and Jennie E. Burnet. 2013. “Gender Quotas, Democracy, and Women’s
Representation in Africa: Some Insights from Democratic Botswana and Autocratic
Rwanda.” Women’s Studies International Forum 41: 103–​112.
Bayat, Asef. 2002. “Activism and Social Development in the Middle East.” International Journal
of Middle East Studies 34, no. 1: 1–​28.
Beck, Linda J. 2003. “Democratization and the Hidden Public: The Impact of Patronage
Networks on Senegalese Women.” Comparative Politics 35, no. 2: 147–​169.
Benstead, Lindsay J. 2008. “Does Casework Build Support for a Strong Parliament? Legislative
Representation and Public Opinion in Morocco and Algeria.” PhD diss., University of
Michigan.
Benstead, Lindsay J. 2015. “Women’s Access to Services: Homophily and the Paradox of
Tribe.” Paper presented at the American Political Science Association Annual Meeting, San
Francisco, California, September 3–​6.
Benstead, Lindsay J. 2016. “Why Quotas Are Needed to Improve Women’s Access to Services in
Clientelistic Regimes.” Governance 29, no. 2: 185–​205.
 Islamist Parties and Women’s Representation in Morocco    849

Benstead, Lindsay J. 2019. “Do Female Local Councillors Improve Women’s Representation?”
Journal of the Middle East and Africa 10, no. 2: 95–​119.
Benstead, Lindsay J. 2021. “The Role of Women in Arab Political Parties.” In Routledge
Handbook on Political Parties in the Middle East and North Africa, edited by Francesco
Cavatorta, Lise Strom, and Valeria Resta, ­chapter 20. New York: Routledge.
Benstead, Lindsay J., and Megan Reif. 2013. “Polarization or Pluralism? Language, Identity, and
Attitudes toward American Culture among Algeria’s Youth.” Middle East Journal of Culture
and Communication 6, no. 1: 75–​106.
Bjarnegård, Elin. 2013. Gender, Informal Institutions and Political Recruitment: Explaining Male
Dominance in Parliamentary Representation. Basingstoke, UK: Palgrave Macmillan.
Blaydes, Lisa. 2014. “How Does Islamist Local Governance Affect the Lives of Women?”
Governance 27, no. 3: 489–​509.
Blaydes, Lisa, and Safinaz El Tarouty. 2009. “Women’s Electoral Participation in Egypt: The
Implications of Gender for Voter Recruitment and Mobilization.” Middle East Journal 63, no.
3: 364–​380.
Broockman, David E. 2013. “Black Politicians Are More Intrinsically Motivated to Advance
Blacks’ Interests: A Field Experiment Manipulating Political Incentives.” American Journal
of Political Science 57, no. 3: 521–​536.
Brooke, Steven T. 2019. Winning Hearts and Votes: Social Services and the Islamist Political
Advantage. Ithaca, NY: Cornell University Press.
Brown, Nathan J., and Amr Hamzawy. 2010. Between Religion and Politics. Washington,
DC: Carnegie Endowment for International Peace.
Brumberg, Daniel. 2002. “Islamists and the Politics of Consensus.” Journal of Democracy 13, no.
3: 109–​115.
Buehler, Matt. 2015. “Continuity through Co-​optation: Rural Politics and Regime Resilience in
Morocco and Mauritania.” Mediterranean Politics 20, no. 3: 364–​385.
Cammett, Melani. 2011. “The Political Economy of Development in the Middle East.” In The
Middle East, edited by Ellen Lust. 12th ed., 99–​142. Washington, DC: CQ Press.
Cammett, Melanie, and Pauline Jones Luong. 2014. “Is There an Islamist Political Advantage?”
Annual Review of Political Science 17: 187–​206.
Catusse, Myriam, and Lamia Zaki. 2009. “Gestion communale et clientelisme moral au
Maroc: Les politiques du Parti de la justice et du développement.” Critique international
42: 73–​91.
Celis, Karen, Sarah Childs, Johanna Kantola, and Mona Lena Krook. 2008. “Rethinking
Women’s Substantive Representation.” Representation 44, no. 2: 99–​110.
Central Intelligence Agency (CIA). 2015. “The World Factbook.” <https://​www.cia.gov/​library/​
publications/​the-​world-​factbook/​> (accessed July 11, 2015) https://​www.cia.gov/​the-​world-​
factbook/​.
Charrad, Mounira M. 2001. States and Women’s Rights: The Making of Postcolonial Tunisia,
Algeria, and Morocco. Berkeley: University of California Press.
Clark, Janine A. 2004a. Islam, Charity, and Activism: Middle-​Class Networks and Social Welfare
in Egypt, Jordan, and Yemen. Bloomington: Indiana University Press.
Clark, Janine A. 2004b. “Social Movement Theory and Patron-​Clientelism: Islamic Social
Institution and the Middle Class in Egypt, Jordan and Yemen.” Comparative Political Studies
37, no. 8: 941–​968.
Clark, Janine A., and Jillian Schwedler. 2003. “Who Opened the Window? Women’s Activism
in Islamist Parties.” Comparative Politics 35, no. 3: 293–​312.
850   Lindsay J. Benstead

Darhour, Hanane, and Drude Dahlerup. 2013. “Sustainable Representation of Women through
Gender Quotas: A Decade’s Experience in Morocco.” Women’s Studies International Forum
41, no. 2: 132–​142.
Ennaji, Moha. 2005. Multilingualism, Cultural Identity, and Education in Morocco. New York:
Springer.
Esposito, John L. 1999. The Islamic Threat: Myth or Reality? Oxford: Oxford University Press.
Eulau, Heinz, and Paul D. Karps. 1977. “The Puzzle of Representation: Specifying Components
of Responsiveness.” Legislative Studies Quarterly 2, no. 3: 233–​254.
Fish, Steven M., and Matthew Kroenig. 2009. The Handbook of National Legislatures: A Global
Survey. Cambridge: Cambridge University Press.
Freedom House. 2012. Freedom in the World 2012: The Arab Uprisings and Their Global
Repercussions. Washington, DC: Freedom House. https://​freedomhouse.org/​report/​freedom-​
world/​freedom-​world-​2012#.VOEjw_​nF_​9s.
Flanigan, Shawn T. 2008. “Nonprofit Service Provision by Insurgent Organizations: The Cases
of Hizbalah and the Tamil Tigers.” Studies in Conflict and Terrorism 31: 499–​519.
Franceschet, Susan, and Jennifer M. Piscopo. 2008. “Gender Quotas and Women’s Substantive
Representation: Lessons from Argentina.” Politics & Gender 4, no. 3: 393–​426.
Gribaa, Boutheina. 2008. “Strengthening Women’s Leadership and Participation in Politics
and Decision Making Process in Algeria, Morocco and Tunisia: Mapping of the Situation.”
Hamzeh, Ahmad Nizar. 2001. “Clientelism, Lebanon: Roots and Trends.” Middle East Studies
37, no. 3: 167–​178.
Harik, Judith Palmer. 1994. The Public and Social Services of the Lebanese Militias. Oxford:
Centre for Lebanese Studies.
Henderson, J. Vernon, and Ari Kuncoro. 2011. “Corruption and Local Democratization in
Indonesia: The Role of Islamic Parties.” Journal for Development Economics 94, no. 2: 164–​180.
Inter-​Parliamentary Union (IPU). 2019. “Women in National Parliaments.” http://​www.
ipu.org/​wmn-​e/​classif-​arc.htm (accessed June 16, 2016) http://​archive.ipu.org/​wmn-​e/​
world.htm.
Ismail, Salwa. 2001. “The Paradox of Islamist Politics.” Middle East Report 221: 34–​39.
Jamal, Amaney, and Ellen Lust-​Okar. 2002. “Rulers and Rules: Reassessing the Influence of
Regime Type on Electoral Law Formation.” Comparative Political Studies 35, no. 3: 337–​366.
Kao, Kristen, and Lindsay J. Benstead. 2021. “Female Electability in the Arab World: The
Advantages of Intersectionality.” Comparative Politics 53, no. 3: 427–​464.
Keefer, Philip, and Razvan Vlaicu. 2008. “Democracy, Credibility and Clientelism.” Journal of
Law, Economics and Organization 24: 371–​406.
Kudva, Neema. 2003. “Engineering Elections: The Experiences of Women in Panchayati Raj
in Karnataka, India.” International Journal of Politics, Culture, and Society 16, no. 3: 445–​463.
Manzetti, Luigi, and Carole Wilson. 2007. “Why Do Corrupt Governments Maintain Public
Support?” Comparative Political Studies 40, no. 8: 949–​970.
Masoud, Tarek. 2014. Counting Islam: Religion, Class, and Elections in Egypt. Cambridge:
Cambridge University Press.
Meyersson, Erik. 2014. “Islamic Rule and the Empowerment of the Poor and Pious.”
Econometrica 82, no. 1: 229–​269.
Milton-​Edwards, Beverley. 2007. “Hamas: Victory with Ballots and Bullets.” Global Change,
Peace & Security 19: 301–​316.
Pitkin, Hanna Fenichel. 1967. The Concept of Representation. Berkeley: University of California
Press.
 Islamist Parties and Women’s Representation in Morocco    851

Roháč, Dalibor. 2013. “Religion as a Commitment Device: The Economics of Political Islam.”
Kyklos 66, no. 2: 256–​274.
Ross, Michael. 2008. “Oil, Islam, and Women.” American Political Science Review 102, no.
1: 107–​123.
Sadiqi, Fatima. 2008. “The Central Role of the Family Law in the Moroccan Feminist
Movement.” British Journal of Middle Eastern Studies 35, no. 3: 325–​337.
Schwindt-​Bayer, Leskue A., and William Mishler. 2005. “An Integrated Model of Women’s
Representation.” Journal of Politics 67, no. 2: 407–​428.
Sung, Hung-​En. 2003. “Fairer Sex or Fairer System? Gender and Corruption Revisited.” Social
Forces 82, no. 2: 703–​723.
Transitional Governance Project. 2021. http://​transitionalgovernanceproject.org (accessed
May 5, 2021).
Tripp, Aili Mari. 2001. “The Politics of Autonomy and Cooptation in Africa: The Case of the
Ugandan Women’s Movement.” Journal of Modern African Studies 39, no. 1: 101–​128.
Troudi, Mohammed Fadhel. 2011. “La révolution tunisienne entre attente et déception, fragilité
et maturité.” Géostratégiques 32: 163–​182.
United Nations Development Programme. 2013. “Gender Inequality Index.” <http://​hdr.undp.
org/​en/​statistics/​gii>.
Walsh, John. 2006. “Egypt’s Muslim Brotherhood.” Harvard International Review 24, no.
4: 32–​36.
Waltz, Susan, and Lindsay Benstead. 2006. “When the Time Is Ripe: The Struggle to Create
an Institutional Culture of Human Rights in Morocco.” In Human Rights in the Arab
World: Independent Voices, edited by Anthony Chase and Amr Hamzawy, 174–​ 196.
Philadelphia: University of Pennsylvania Press.
Wegner, Eva. 2011. Islamist Opposition in Authoritarian Regimes: The Party of Justice and
Development in Morocco. Syracuse, NY: Syracuse University Press.
White, Jenny B. 2002. Islamist Mobilization in Turkey: A Study in Vernacular Politics.
Seattle: University of Washington Press.
Wickham, Carrie Rosefsky. 2002. Mobilizing Islam: Religion, Activism, and Political Change in
Egypt. New York: Columbia University Press.
Wiktorowicz, Quintan. 2002. “Islamist Activism in Jordan.” In Everyday Life in the
Muslim Middle East, edited by Donna Lee Bowen and Evelyn A. Early, 227–​245. 2nd ed.
Bloomington: Indiana University Press.
World Bank. 2014. “World Development Indicators.” https://​datatopics.worldbank.org/​world-​
development-​indicators/​.
Yadav, Stacey Philbrick. 2010. “Segmented Publics and Islamist Women in Yemen: Rethinking
Space and Activism.” Journal of Middle East Women’s Studies 6, no. 2: 1–​30.
 Chapter 41

The Isl amic Stat e as a

Revolu tiona ry Re be l
Grou p
IS’s Governance and Violence in Historical Context

Megan A. Stewart

At the Islamic State’s (IS) height, the rebel group controlled over 100,000 square
kilometers of territory in Iraq and Syria (Jones et al. 2017, 20), captivating both adherents
and adversaries worldwide. In the process of taking control of a territorial space that
spanned the sovereign borders of two nation-​states, the IS unleashed spectacular and
gruesome violence. At the same time, it began to co-​opt, construct, and consolidate an
interconnected web of institutions that not only governed millions of people (Jones et al.
2017, 20), but frequently did so in ways that some Iraqis perceived to be superior to the
Iraqi state itself (Revkin 2020a). Throughout the process, the Islamic State’s specific reli-
gious perspective appeared to undergird a number of its behavioral choices. What then
explains the Islamic State’s warmaking and governance strategies?
In this essay, I argue that the Islamic State’s overarching strategy with respect to
warmaking and governance is not derived from its religious ideology, but rather
represents a jihadist adaptation of a strategy for revolutionary warfare first coherently ar-
ticulated, and then especially propagated, by Mao Zedong and the Chinese Communist
Party (CCP) during the Chinese Civil War. This strategy includes a military doctrine that
integrates both conventional and guerrilla warfare alongside other violent tactics, simulta-
neously coupled with intensive sociopolitical restructuring through governance. Not only
is the Islamic State’s strategy not new, it is also not the only other jihadist organization to
adopt the dual-​front revolutionary warfare approach. Instead, religious ideology tends to
influence the nature of governing institutions (e.g., religious taxes and shariʿa courts), the
primary beneficiaries of governance and the targets of violence. The strategic template for
854   Megan A. Stewart

the construction of a specific portfolio of governance institutions alongside specifically

patterned violence, however, is unrelated to religious ideology.
Importantly, I am not the first to argue that the Islamic State is a revolutionary organi-
zation and that its military behaviors reflect its revolutionary character (Whiteside 2016a,
2016b; Kalyvas 2018). Rather, my contribution is to first articulate the nature and origins
of a single strategy for rebel groups that combines military and governance elements, to
explain that this strategy is most frequently adopted by rebel groups with revolutionary
goals, to show how this strategy transcends beyond discrete ideologies, to illustrate that
this strategy appears to shape both the Islamic State’s military and governing behaviors in
practice, and to identify the role religious ideology plays, if any, in the Islamic State’s gov-
ernance and violence. By approaching the Islamic State through a framework embedded
within a broader historical context, the Islamic State is correctly understood as not fun-
damentally unique compared to other rebel groups—​jihadist or otherwise—​and presents
the organization’s strategy and behaviors as akin to any other rebel group with similar
ambitions (revolution), regardless of ideology. In so doing, by distinguishing the aspects
of the Islamic State’s behavior attributable to religious ideology and those that are not,
we are better able to understand the conditions under which religious ideology plays a
role in conflict, and in what ways, thereby avoiding essentialist arguments that root the
Islamic State’s behaviors in Islam. Finally, one of the primary contributions is to illustrate
how ideas and practices are learned by rebel groups, and the conditions under which
these ideas spread, even across ideologically opposed organizations.
This essay is structured as follows. First, I explain a strategy for the conduct of civil war
implemented, but especially propagated, by Mao Zedong and the Chinese Communist
Party (CCP). This strategy consciously combines political (meaning governance) and
military elements to realize social and political transformation during civil war. Then,
I explain how the CCP’s strategy has diffused across ideological lines such that it has
been adapted and incorporated into jihadist rebel groups with similarly revolutionary
ambitions. Third, I review how the IS puts into practice both the violence and govern-
ance components of revolutionary war. Finally, I conclude with a discussion about how
the IS’s violent and governing behaviors are directly linked to, and an echo of, commu-
nist rebel groups of the mid-​twentieth century and are neither new nor particularly
Islamic, but I illustrate how and when an Islamic ideology affects rebel organizational
behavior with respect to governance.

The Origins of the Islamic

State’s Strategies

The Islamic State’s behaviors with respect to governance and violence are not new or
original emanations of the IS’s religious ideology. Rather, the strategic origins begin
with Mao Zedong and the CCP during the Chinese Civil War that began in the 1920s.
 The Islamic State as a Revolutionary Rebel Group    855

Prior to the Chinese Civil War, many revolutions and civil wars unfolded as periods
of state contestation, followed by social and political restructuring and consolidation
after this period (Skocpol 1979, 163–​4; Mondlane 1983, 163; Huntington 2006, 266). For
instance, the Bolsheviks in Russia and the Jacobins in France first seized the reins of state
power before undertaking massive socioeconomic and political restructuring. Likewise,
this step-​wise trajectory was similar for many rebel groups: although some rebel groups
governed during war, many preferred to save technically complex reforms not directly
related to the military struggle until the end of conflict.
Unlike many other previous revolutionaries and rebel groups, the Chinese
Communist Party undertook a concerted effort to integrate the typically sequen-
tial processes of contestation and consolidation. Rather than pursue military victory
followed by socioeconomic and political restructuring, the CCP pursued social trans-
formation alongside and simultaneous with military confrontations against the state or
other rebel organizations (Chen 1986; Selden 1995; Pepper 1999; Opper 2018; Mao 2011,
[1937] 1989). To Mao, the political and the military aspects of civil war are inherently
intertwined (Mao [1937] 1989, 42–​44).
On the military side, although the CCP is best known for its use of guerrilla warfare—​
defined as relatively mobile bands of combatants that use the element of surprise to
attack much more powerful enemies (Mao [1937] 1989, 96–​98)—​this was not the sole el-
ement with which the CCP pursued military victory. Instead, the CCP understood civil
war as a dynamic process characterized by three phases. In the first phase, frequently
covert and cellular rebel elements would lay the groundwork for more overt operations.
In the second phase, these cells would eventually become mobile guerrilla bands, which
would eventually transition to, in the third phase, conventional military forces that
would encircle major cities and finally seize power through conventional forces (21–​22).
The dynamic process was not expected to proceed in a linear and step-​wise fashion: Mao
envisioned regression, progression, and combination across different military stages in
pursuit of final victory (see, e.g., Mao 2011, loc. 43-​75).
At the same time, the CCP also undertook a set of political behaviors intended to
accelerate the transformation of societies and individuals during conflict. This basket
of political activities formed a concerted propaganda (psychological) campaign (Mao
[1937] 1989, 85, 103), with appeals to certain constituencies (see, e.g., Snow 1968; Larkin
1973; Friedman 2015; Lovell 2019) within the international system and the creation of an
elaborate system of governance (Seybolt 1971; Minden 1979; Yung-​Fa 1986; Huang 1995;
Pepper 1999; Griffin 2015; Opper 2018). The governance apparatus the CCP constructed
included new educational (Seybolt 1971) and healthcare systems (Minden 1979, 302),
utilities and electricity response teams (Pepper 1999, 392), land redistribution (Pepper
1999, 307) and regulatory economic policies (Huang 1995, 119), the “liberation” of
women (Seybolt 1971, 646), the introduction of elections (Selden 1995, 103–​109), and a
new judicial structure (Griffin 2015, 97–​100).
Most critically, however, was that the CCP packaged together these behaviors as part
of a coherent and unified strategy, and then propagated this strategy globally. Indeed, the
CCP even understood its experience—​the multidimensional approach to conflict—​as
856   Megan A. Stewart

a “model” to be emulated by other rebel groups the world over (Larkin 1973, 148). Lu
Dingyi, one of the CCP’s chief theorists, argued that the Chinese Civil War provided a
“model for the ‘colonial and semi-​colonial countries’” in pursuing their own revolution
(Friedman 2015, 27). For instance, during the civil war, the CCP was keen to highlight its
electoral institutions to British and American missionaries in China, who were quick to
praise the communist rebels for adopting these governance forms upon witnessing the
CCP’s elections take place (Selden 1995, 103). Mao and the CCP cadres also composed
and disseminated numerous texts relating to the civil war and Maoist thought that were
published globally (e.g., Mao [1937] 1989, 2011). In what was perhaps most important to
the propaganda campaign, the CCP also carefully orchestrated a visit from a Western
journalist, Edward Snow, to tour its liberated areas (Lovell 2019, 61–​62). Snow then
published his favorable account of life in the CCP’s liberated areas, titled Red Star over
China, a text that captured the attention of people and partisans around the world. Even
anti-​Nazi partisans in Europe read these highly favorable accounts of the CCP (Lovell
2019, 62, 470).
Beyond just spreading information during the Chinese Civil War, the CCP also
trained many would-​be rebels, none of whom were more famous than Ho Chi Minh,
who would lead the Viet Minh against the French in the First Indochinese War (Wilson
Center, n.d.). Although Ho’s training took place in the heat of battle during the Chinese
Civil War itself, once the CCP secured victory over the Kuomintang in 1949, the CCP
was able to institutionalize a training course for rebel groups. For rebellions already un-
derway, the CCP offered political and material aid (Van Ness 1970, 82, 87), and even
offered to provide military support if cadres of rebel groups agreed to complete a
training course in China (Crosby 1968).
Over time, more and more rebel groups with similar long-​ term objectives of
revolution—​meaning goals for the creation of a new state and society, (and not goals
for autonomy, reform, or the preservation of culture)—​but not necessarily the same
ideologies as the CCP, learned from and decided to imitate the CCP’s template for
how to execute revolution during their own civil wars. As an increasing number of
rebel groups with revolutionary goals adopted the CCP’s strategy, the CCP’s template
became delinked from the CCP’s experience specifically (Stewart 2021). Instead, it ul-
timately became a general framework for the pursuit of revolutionary warfare—​a spe-
cific portfolio of governance institutions alongside a strategy for how to contest the
state and civilian enemies with violence. Importantly, as I discuss more in the next sec-
tion, this institutional and warmaking framework was ideologically flexible and adapt-
able to local conditions. For instance, the revolutionary warfare template included the
construction of schools, but the nature of the schools, the content of the curriculum,
and the beneficiaries of such education were left to the interpretation and adaptation
of later rebel groups. Likewise, although discriminate violence is a feature of nearly all
civil wars to varying degrees, the targets of discriminate violence may vary from loca-
tion to location and depend upon ideology. In other words, the strategy of revolutionary
warfare provided a general framework for rebels, which could then be adapted to local
conditions and ideologies as necessary.
 The Islamic State as a Revolutionary Rebel Group    857

Furthermore, this revolutionary warfare approach came to dominate over alter-

native strategies for pursuing revolution during civil war. For instance, although the
Russian Revolution preceded the Chinese Civil War by about a decade, the Bolsheviks
achieved their revolution through a putsch, not civil war. As a result, even the Chinese
Communist Party came to cast the Russian Revolution as a model for revolution in in-
dustrial countries, but not in colonial or semi-​colonial countries (Friedman 2015, 27).
Likewise, because the CCP had begun propagating its experience as a template during
its civil war in the early half of the twentieth century, it had a several-​decades head start
before the innovation of other models. Even if one considers Che Guevara’s approach to
guerrilla warfare to fully constitute an alternative and nonderivative strategy,1 Guevara
only published his seminal text for guerrilla fighters in 1961, years after other rebel
leaders were trained or had their cadres trained in China. Ultimately, the subsequent
adaptation of the CCP’s strategy by later revolutionary rebel groups limited the space
for innovation for alternative approaches and strategies for the conduct of revolution
during civil war.
To summarize, the CCP integrated sociopolitical transformation alongside warfare
during civil conflict, packaged this strategy together, and propagated it globally. Other
rebel groups with goals similar to the CCP increasingly adopted this strategy, such that
it became delinked from the CCP’s experience specifically, and generally became known
as revolutionary warfare. Furthermore, the commonplace implementation of this
strategy among rebels meant that there was limited space for innovation or alternative
strategic models for pursuing revolution during civil war. In the next section, I illustrate
the flexibility of revolutionary warfare with respect to ideological backgrounds.

Revolutionary Warfare Spreads

to Jihadist Rebel Groups

The adaptation and implementation of non-​Islamic strategic approaches and texts by

jihadist organizations is not unusual (see, e.g., Lia and Hegghammer 2004; Brooke
2008) and the CCP’s revolutionary warfare strategy is not immune to cross-​ideolog-
ical adaptation. For instance, in its infancy, Hezbollah leaders learned from the CCP’s
revolutionary warfare model and even explicitly referenced Mao Zedong and Ho Chi
Minh as paragon examples worthy of emulation (al-​Hassan 2012; Daana 2013; Al Sayegh
2014). Hezbollah’s mimicry of leftist groups was not simply a function of its leaders’
lack of knowledge about other organizational forms or strategies: Hezbollah’s leaders
consciously determined to avoid organizational models such as that of the fellow Shiʿa
Daʿwa party in Iraq (AbuKhalil 1991, 392), and instead decided to learn from and emu-
late leftist revolutionary rebel groups that achieved success just a decade or two earlier
(Kahveci 1998; Khashan and Mousawi 2007, 5–​7; Daana 2013; Qassim 2017). In addi-
tion, Hezbollah’s organizational structure closely mirrors the organizational structure
858   Megan A. Stewart

of leftist revolutionary rebels (AbuKhalil 1991, 397), and commentators on Hezbollah’s

own website even note that Hezbollah implements a strategy of revolutionary warfare
that contains asymmetric violence alongside governance, that this strategy was created
by the CCP, and that the organization should be understood and considered within the
tradition of Mao Zedong and Che Guevara (Kahveci 1998; Qassim 2017).
Hezbollah is far from the only jihadist organization to learn from the CCP’s revolu-
tionary warfare strategy. In fact, several al-​Qaʿida strategists have learned directly from
the Chinese model. One such example is Abu Musab al-​Suri, a jihadist writer who is
acutely aware of his intellectual tradition among leftist revolutionary warfare thinkers,
including Mao (Ryan 2013, 231), and even explicitly cites these communist thinkers and
practitioners to justify his own claims (Ryan 2013, 233–​234). Al-​Suri’s revolutionary war
strategy, like Mao’s, includes three phases of war—​exhaustion, balance, and decision or
liberation (Ryan 2013, 230)—​and culminates in establishing a political entity with a basis
in shariʿa law during war itself (Ryan 2013, 234).
A second al-​Qaʿida strategist, Abu Ubayd al-​Qurashi, is similarly explicit in his ad-
vocacy for jihadist rebels adopting the Chinese Communist Party’s model of revolu-
tionary war as a template for the relatively “weak against the strong” (Ryan 2013, 86).
He repeatedly and blatantly cites the CCP as a model of “Revolutionary War” to emu-
late, specifically emphasizing Mao’s classic three-​phased military campaign, as well as
the CCP’s governing and propaganda components. According to al-​Qurashi, “[m]‌ost
revolutionary movements spend some time to set up not only the local administration
but also parallel governmental institutions. . . . Thus the Chinese ‘people’s war’ (for ex-
ample) witnessed the establishment of a state within a state in regions controlled by the
communist forces in China” (Ryan 2013, 88–​89). Al-​Qurashi’s acknowledgement of the
influence of the CCP’s experience on his own work is clear, and, like the CCP, Qurashi
advocates for the implementation of both the military and political aspects of revolu-
tionary warfare.
Al-​Qurashi’s work would go on to inspire other al-​Qaʿida strategists, including Abu
Bakr Naji, author of the influential jihadist text The Management of Savagery (Naji
2006; see also Ryan 2013, 199). Although Naji does not explicitly reference Mao in The
Management of Savagery, he nevertheless encourages his readers to turn to al-​Qurashi,
who does reference Mao (Naji 2006, 30),2 tells readers to review “[g]‌eneral books on
the art of war, especially guerrilla wars, as long as the student is able to correct the
Sharia mistakes that are in them” (30), and specifically references other rebel groups
that adopted and implemented the revolutionary warfare strategy, including describing
leftist organizations, as examples to emulate (14).
Though not explicitly stated, the similarities between Naji and Mao are unmistakable.
For instance, both outline a three-​phased, multidimensional military approach: Naji
advocates for a period of violence that forces state or enemy forces to withdraw, followed
by the control of territory whereby the initial institutions of an Islamic state adminis-
tration are created, leading to a final decisive victory whereby an Islamic state is estab-
lished (Ryan 2013, 168–​169). Like Mao, Naji also describes the importance of small and
mobile guerrilla warfare tactics that surprise the enemy through waves of consistent
 The Islamic State as a Revolutionary Rebel Group    859

attacks, thereby militarily and economically draining him, and, like Mao, Naji advocates
for initiating attacks in depopulated areas before contesting spaces where the state has
greater control (Naji 2006, 29–​30).
In pursuit of achieving the political, governance aspects of revolutionary warfare,
Naji also left clear guidance as well. To facilitate this transformation from violent chaos
to administered order, Naji lists twelve basic steps of governance procedures that ought
to be completed during the course of war. This guidance includes:

• Spreading internal security and preserving it in every region that is managed

• Providing food and medical treatment
• Securing the region of savagery from the invasions of enemies by setting up defen-
sive fortifications and developing fighting capacities
• Establishing Sharia justice among the people who live in the regions of savagery
• Raising the level of faith and combat efficiency during the training of the youth of
the region of savagery and establishing a fighting society at all levels and among
all individuals by means of making them aware of its importance. But it must be
made clear that it is an obligatory duty . . . which does not mean that every indi-
vidual member of society must practice fighting; rather, only a part or portion of the
fighting ranks (must practice it) in the form which the society knows best and needs.
• Working for the spread of legal, Sharia science (putting the most important aspects
before those of lesser importance) and worldly science (putting the most important
aspects before those of lesser importance).
• Disseminating spies and seeking to complete the construction of a minimal intelli-
gence agency.
• Uniting the hearts of the people by means of money and uniting the world by
Sharia governance and (compliance with) rules which are at least exemplified by
individuals in the administration.
• ...
• To these we add a future goal, which is the advancement of managerial groups to-
ward the attainment of the power of establishment and readiness for plucking the
fruit (of their efforts) and establishing the state. (Naji 2006, 15–​18)

Like Mao, the political element of revolutionary warfare, meaning governance to

construct an ideal state during war, was an essential component of Naji’s vision for
executing jihad.
In the same way that Mao propagated his military and governance achievements, Naji
is explicit about the use of propaganda through military and governance actions to at-
tract a broad (if not global) coalition of followers. For instance, Naji notes that the “rate
of operations escalates in order to send a living, practical message to the people, the
masses, and the enemy’s low-​ranking troops that the power of the mujahids is on the
rise . . . . This encourages the masses and revives hope within them and facilitates per-
manent support for the movement and the automatic escalation of the movement” (Naji
2006, 29). At the same time, Naji insists on developing a media plan for the “masses” that
860   Megan A. Stewart

focuses on governance and administration so that “the masses fly to the regions which
we manage, particularly the youth after news of (our) transparency and truthfulness
reaches them so that they may be fully aware of the loss of money, people, and worldly
gains” (21). Like Mao, Naji emphasizes the importance of propagating both the govern-
ance and military behaviors of jihadist groups, even beyond the theater of conflict.
To summarize, revolutionary warfare is not simply a leftist or communist strategy,
and many jihadist rebels—​ostensibly ideologically opposed to communism—​adopted
the package of military and political behaviors that make up revolutionary warfare. For
instance, in the 1980s, Hezbollah adopted revolutionary warfare to the context of the
Shiʿa in Lebanon. Likewise, al-​Qaʿida strategists adapted and spread Mao’s framework
to other jihadist rebel groups globally.

The Islamic State as a Revolutionary

Organization

Although the Islamic State’s ideology diverges from the communist, leftist, and national
liberation rebel movements that preceded it decades earlier, the IS shares these rebel
groups’ goals of fundamentally transforming state and social institutions, a key deter-
minant of revolutionary goals (Ryan 2014; Salih 2015; Whiteside 2016a, 2016b; Kalyvas
2018; Stewart 2021). Like rebel groups with revolutionary ambitions before them, the
Islamic State implemented a template for pursuing revolution during civil war that
closely mirrors the Chinese model, but the implementation thereof represents a jihadist
interpretation of revolutionary warfare. This revolutionary warfare strategy is articu-
lated throughout several jihadist texts, but especially in The Management of Savagery,
which has in turn become one of the touchstone strategic works for the Islamic State’s
pursuit of its goals (Ryan 2014; Salih 2015; Al-​Arab 2016; Whiteside 2016a, 2016b).
In this section, I summarize the Islamic State’s strategic foundation in revolutionary
warfare. I then detail how elements of IS’s revolutionary warfare strategy, both violent
and political, has shaped the group’s behaviors in practice. As a result, because of the
IS’s ties to revolutionary warfare, its behaviors can be understood as neither particu-
larly new nor inherently Islamic. I conclude with a discussion of what the Islamic State’s
adaptation of revolutionary warfare means for both the organization’s future, as well as
future jihadist movements.

The Islamic State’s Strategic Foundation

As already noted, The Management of Savagery represents the evolution of revolu-
tionary warfare, and is a jihadist interpretation of what is, at its heart, a relic of leftist
revolutionaries of the Cold War. Despite its non-​jihadist (even atheist) origins, most
 The Islamic State as a Revolutionary Rebel Group    861

experts agree that the IS relies on the guidance contained in its pages (Ryan 2014; Hassan
2015; Weiss and Hassan 2016; Whiteside 2016b). According to Hassan (2015) and Weiss
and Hassan (2016, 41–​6, 218) the IS uses the text as a central part of its training cur-
riculum. Additionally, because of the strict ideological interpretation and worldview of
the IS, the organization has banned many texts that its cadres can read and consume.
One of the few nonreligious texts that its leaders can read, however, is The Management
of Savagery (see Weiss and Hassan 2016, 218), underscoring its importance among the
leadership class and intellectual underpinnings of the Islamic State.
Even when the Islamic State itself is cautious not to overemphasize the roll of The
Management of Savagery in the strategic guidance it issues to its cadres, the group
admitted that “[a]‌s for the concise but beneficial 100 page book titled [The Management
of Savagery] by an unknown author who only went by the penname Abu Bakr Naji, then
when Shaykh al Zarqawi read this book he commented, ‘it is as if the author knows what
I am planning.’ Note: Although Naji’s book describes very precisely the overall strategy
of the mujahidin, Naji fell into some errors in his discussions on issues related to the
takfir of parties who forcefully resist the Shariah and its laws” (Islamic State 2015b, 39, fn.
2; see also Whiteside 2016b, 10). Thus, the IS admits to using the text in its curriculum,
allows leaders to read it, and claims that it reflects the overall strategic choices that the
group makes.3

Violence
From an operational level, the Islamic State’s pattern of violence closely mirrors Naji’s
three-​phase strategy of revolutionary war, and thus, in turn, Mao’s three phases of rev-
olutionary war. In its incipient stage (around the early 2000s), the IS launched a rela-
tively anemic campaign of urban suicide terrorism combined with improvised explosive
devices (IEDs) that left the organization almost totally defeated (Hashim 2019, 26). By
2011, however, the organization had regrouped and began to diversify the types of vio-
lence the organization used. The IS used truck bombs to free imprisoned jihadist and
to destroy soft targets, while also assassinating key opposition figures (Hashim 2019,
27). Around the time the IS entered into Syria, the group had, “conduct[ed] robberies,
assassinations, acts of terror, sabotage, and other attacks termed ‘war out of the dark’”
(Whiteside 2016a, 745). Importantly, IS members explicitly pointed to certain passages
as influential in justifying some of the violent tactics they have used. For instance, the
group rationalizes public beheading through appeals to The Management of Savagery
(Hassan 2015).
When the Islamic State entered the Syrian war and the second phase of the revolu-
tionary warfare struggle, it relied on mobile guerrilla tactics to push out existing state
forces or rival rebel organizations (Hashim 2019, 27). As the organization captured and
controlled territory, it was able to seize and repurpose heavy artillery and weapons left
by the fleeing forces. The IS amassed a large battalion of tanks, trucks, and anti-​tank
missiles used in both offensive and defensive campaigns. It even developed an elite
862   Megan A. Stewart

conventional combat unit and had cultivated a relatively high degree of command and
control (Maurer 2018, 232). The result was that by its height, the IS relied on an array
of violent behaviors that included conventional military attacks (Maurer 2018; Hashim
2019, 27–​28). The use of this form of violence coincides with Mao’s and Naji’s second
phase of civil war and approached its third phase—​guerrilla warfare to realize the con-
struction of administrative structures, culminating in the final decisive blow to achieve
victory with conventional warfare.
As an international coalition against the rebel movement coalesced and the Islamic
State’s territorial grasp slipped from its hands, the group’s ability to conduct conven-
tional attacks dwindled and disappeared. Increasingly, the IS resembles the organization
it once was before it captured vast swaths of territory throughout Iraq and the Levant
(Hashim 2019, 28). This regression away from conventional warfare—​though certainly
undesirable by the Islamic State—​is completely consistent with Mao’s own writing,
which notes that progress in war is not linear, and a rebel group could move through and
away from multiple phases of the guerrilla war campaign (Mao [1937] 1989, 21–​22). Even
Mao and the CCP progressed and regressed through the three stages of civil war, facing
severe and almost insurmountable setbacks as well as realizing successes, before finally
securing victory after nearly two decades of fighting (see, e.g., Mao 2011; Opper 2018).

Governance
As I and other scholars have explained, the Islamic State’s violent behaviors closely
mirror the revolutionary warfare that once characterized leftist and anticolonial rebel
groups of the Cold War (Whiteside 2016a; Kalyvas 2018). But revolutionary warfare it-
self is not simply a military strategy; indeed, what makes revolutionary warfare revolu-
tionary is that politics are central and indistinguishable from the military struggle (Mao
[1937] 1989, 42–​44). For the IS, the political aspects of revolutionary warfare entailed the
creation of an Islamic state and a set of administrative structures that would underscore
a wholly new Islamist society.
This portfolio of IS institutions mirrored many of the same institutional forms that
the CCP created during war, although of course these institutions clearly differed along
ideological grounds. For instance, while both the CCP and the Islamic State set up court
systems, a central institutional form in the construction of a state, the CCP’s included
People’s Tribunals and trials for counter-​revolutionaries (Griffin 2015, 98–​100), while
the Islamic State’s were premised on shariʿa law (Robinson et al. 2017, 167).
The similarities in the Islamic State’s and the CCP’s institutional forms do not simply
end with the creation of a new judicial structure. Both organizations opened and
transformed schools and hospitals (Robinson et al. 2017, xxi, xiv) under their control, al-
beit in different ways. Both organizations “nationalized” certain institutions (Robinson
et al. 2017, 167), made land more easily available for its central constituency (Rukmini
2018),4 and regulated the marketplace (Robinson et al. 2017, xxvi). Both focused their
efforts to rehabilitate infrastructure and utilities (Robinson et al. 2017, xiv), and both
 The Islamic State as a Revolutionary Rebel Group    863

changed the political institutions within the places they controlled (Reuter 2015). Both
organizations sought to restructure the lives of women to better comport with ideolog-
ical ideals (Islamic State 2015a, see, for example, specimens M, 1F, 1G, 1I, 1Y). As a whole,
the Islamic State and the CCP created almost the same set of institutional forms, though
the content and nature of these institutions varied in accordance with their divergent
ideologies (see also Robinson et al. 2017, 80; Revkin 2020b).
In the same way that the Islamic State used violence to propagate itself and its deeds,
the organization also propagated itself through the governance institutions that it
created, just as the CCP had done. In its magazine, the IS called on all educated per-
sons with technical expertise to serve as administrators within the caliphate (Islamic
State 2014, 11). The magazine also featured a full page filled with quotes from its enemies,
describing the IS as a “de facto state,” in order to attract the backing and attention of po-
tential supporters (Islamic State 2014, 32–​33). Finally, a study of the Islamic State’s propa-
ganda found that over half of the material it produced covered its governance operations
and the daily life of civilians living in its territory (Osborne 2015), again conforming
to Naji’s guidance while mirroring Mao’s behaviors by propagating the governance
institutions created within the territories the rebel group controlled.

Discussion and Role

of Religious Ideology

In the previous section, I discussed how the Islamic State’s violence and governance
originate from a single strategy of revolutionary warfare. This strategy has its roots in
the communist insurgencies of the Cold War that date back almost a century. Not only
has the revolutionary warfare strategy been adapted by jihadist organizations and stra-
tegic thinkers, but the Islamic State learned from these jihadist thinkers, and in prac-
tice it unmistakably realizes almost the same governance portfolio and violence profile
as the CCP and other revolutionary rebel groups. At the same time, like the CCP, the
Islamic State has gone to great pains to highlight globally its violent and governing
behaviors, highlighting both its violence alongside its simultaneous realization of its so-
cioeconomic and political goals. Given the historical origins of the Islamic State’s stra-
tegic behaviors, it would be incorrect to understand the Islamic State or its behaviors as
something that is fundamentally new.
If the Islamic State’s strategic framework is neither strictly jihadist nor an emanation
of Islam, what role does religious ideology play in its behavior? Although a religious
ideology did not appear to influence the decision to provide the portfolio of governance
institutions the Islamic State has or the Islamic State’s conduct of violence heretofore,
religious ideology still influences some aspects of the Islamic State’s behavior in ways
that represent a point of departure from previous leftist and some jihadist organizations.
Specifically, the Islamic State’s religious ideology influences the nature, content, and
864   Megan A. Stewart

beneficiaries of governance institutions, while also influencing the targets of violence

and the primary recruits for the organization.
In terms of governance, religious ideology and cultural history have played a role
in shaping the nature of the Islamic State’s institutions, though not necessarily the de-
cision to build institutions in the first place, or which specific institutions to build
during war. For instance, although the Islamic State organized schools, just like the
CCP, the schools’ curriculum is devoid of secular topics and is instead closely tied
to its religious ideologies (Khalil 2014). Likewise, where the CCP held elections for
positions to administrate certain CCP programs (Selden 1995, 103–​109), the Islamic
State appoints an emir (Reuter 2015). The Islamic State’s courts are also rooted in
shariʿa law, while the CCP’s courts were not. Finally, while both the CCP and IS sought
to alter the roles of women (Spencer 2016), they held different expectations for what
the role of women ought to be.
Similarly, Gutiérrez-​Sanín and Wood (2014, 215) note that central to any ideology is
a referent group around which grievances are articulated, as well as a plan of action to
address these grievances. For the Islamic State, Sunni Muslims are the referent group,
whereas for leftist rebel groups these referent classes are typically peasants or colonized
persons. Thus, in the distribution of governance, the Islamic State prioritizes Sunni
Muslims, whereas communists typically prioritized peasants. For instance, Christians
are charged a religious tax that Sunni Muslims do not have to pay (Revkin 2020b),
thereby implementing policies that favor Sunni Muslims over other religious and ethnic
groups.
Likewise, while the specific forms of violence the Islamic State uses are not new (e.g.,
guerrilla warfare, discriminate lethal violence), the targets of this violence could be
shaped by its religious ideology. For instance, the Islamic State’s violence against certain
religious or minority sects is a reflection of the group’s ethno-​sectarian referent group
of Sunni Muslims. Likewise, the CCP’s violence against members of a certain class
(see, e.g., Opper 2018) is again partially a reflection of its ideological priorities. In other
words, the Islamic State and the CCP might rely on the same forms of violence, but the
targets of this violence are often shaped by their differing ideologies.
Because the revolutionary warfare strategy is not wedded to a particular ideology, the
Islamic State’s religious ideology does not make the group inherently unable to adapt the
revolutionary warfare strategy. However, whereas members of other revolutionary rebel
groups, including organizations like Hezbollah, were quick to explicitly compare them-
selves to the CCP, the Viet Minh, or other adapters of the Chinese model (Al Sayegh
2014), the strict religiosity of the Islamic State’s ideology may prevent members from
making direct comparisons between the two organizations. For instance, even though
The Management of Savagery encourages readers to examine leftist revolutionary rebel
groups in South America, it warns readers to correct for their un-​Islamic tendencies
(Naji 2006, 14). Thus, the Islamic State might be careful to distance itself from leftist
rebel groups.
 The Islamic State as a Revolutionary Rebel Group    865

Alternatively, the consistent application of the CCP’s strategy meant that the model
became delinked from the CCP’s experience specifically. Over time, it became a ge-
neral strategic framework for pursuing revolution that could be learned from any other
organization that adopted the CCP’s approach. As a result, the Islamic State could be
unaware of its communist roots and could have learned its strategies from a study of
Hezbollah or some other jihadist rebel group that adopted the revolutionary warfare
framework. Either way, the Islamic State need not know about the CCP specifically to
adopt the strategy of revolutionary warfare that the CCP propagated.

Conclusion

Ultimately, in this essay, I argued that the Islamic State’s strategy is not a unique product
of its religious ideology. Rather, its strategy has its origins in the Chinese Communist
Party’s approach first propagated by the CCP almost a century ago. This strategy is ide-
ologically flexible, however, allowing for local and non-​leftist adaptations of the ap-
proach. As a result, although the Islamic State’s strategic approach is not a product of
religion or Islam, its interpretation and application of the approach does reflect its par-
ticular worldview and the ideas and practices that constitute this worldview.
Today, the Islamic State is but a shadow of its once formidable self. Reduced to al-
most no territorial control and holdings, the organization appears to be all but defeated.
Although the Islamic State pales in comparison to itself just a few years ago, rebel regres-
sion from once great heights of power and control is built into the genetic makeup of the
revolutionary warfare strategy. Not only did Mao, the initial propagator of revolutionary
warfare, assume that rebel groups would face ebbs and flows in the pursuit of victory, he
experienced them too. In the 1930s, the Chinese Communist Party introduced govern-
ance initiatives that were so radical and extreme that it ended up triggering waves of ci-
vilian defections in support of its rival, the Kuomintang. With the help of civilian forces,
the Kuomintang forced the CCP out of its foothold in southern China, precipitating a
debilitating retreat to the CCP’s northern base. Despite, these setbacks, Mao was never-
theless able to return and succeed a little more than a decade later. Because the CCP and
the Islamic State rely on similar strategies, perhaps the Islamic State’s current military
regression and near defeat are more closely akin to Mao’s Long March, meaning that
total defeat is far from certain, and the Islamic State could survive to fight another day.
Even if the Islamic State never recovers and is duly defeated, the Islamic State’s ability
to capture global attention through its use of a revolutionary warfare strategy may serve
to perpetuate the strategy, now nearly a century old. Furthermore, the application of the
revolutionary warfare strategy by the Islamic State may serve as another template for
how jihadist groups could integrate the CCP’s strategies into a non-​leftist framework
that might be used by jihadist organizations at a later time.
866   Megan A. Stewart

Notes
1. According to Wickham-​Crowley (1987, 483–​489) the Cuban revolutionaries engaged in
some forms of governance and similarly undertook guerrilla warfare, so the degree of sub-
stantial difference between the CCP’s approach and Guevara’s remains unclear.
2. The translation of The Management of Savagery contains page numbers for the translation
and notes, but also keeps the original page numbers of the untranslated work in brackets.
I cite the original page numbers of the text reported in brackets.
3. Although the work of al-​Qaʿida strategist al-​Suri was highly influential, the Islamic State’s
leadership does not recommend this work (Islamic State 2015b, 39, fn. 2; Whiteside 2016b,
10). As a result, I primarily focus on the Islamic State’s imitation and implementation of The
Management of Savagery.
4. For instance, the CCP redistributed land after a class analysis of peasants, whereas the
Islamic State confiscated land often owned by Christians and then rented it back to poorer
Muslims.

References
AbuKhalil, As’ad. 1991. “Ideology and Practice of Hizballah in Lebanon: Islamization of
Leninist Organizational Principles.” Middle Eastern Studies 27, no. 3: 390–​403.
Al-​Arab. 2016. “The Book The Management of Savagery: The Holy Constitution of ISIS on the
Internet.” Translated by Mollie Todd. Al-​Arab Online, March 29, 2016. https://​i.alarab.co.uk/​
pdf/​2016/​03/​29-​03/​p1000.pdf#page=19.
Brooke, Steven. 2008. “Jihadist Strategic Debates before 9/​11.” Studies in Conflict & Terrorism
31, no. 3: 201–​226.
Chen, Yung-​fa. 1986. Making Revolution: The Communist Movement in Eastern and Central
China. Vol. 1, 1937–​1945. Berkeley: University of California Press.
Crosby, J. M. 1968. “Reference.” F0 1043/​7 1, 1968. Document 15. Archived Letter issued July 25,
1968. Accessed September 13, 2018, at the UK National Archives in Kew, United Kingdom.
Daana, Saif. 2013. “Hezbollah in the July War: The Engineering Miracle.” Translated by Nada
Mousa and Mollie Todd. Al-​Manar, July 22, 2013. http://​archive.almanar.com.lb/​article.
php?id=543497.
Friedman, Jeremy. 2015. Shadow Cold War: The Sino-​Soviet Competition for the Third World.
Chapel Hill: University of North Carolina Press.
Griffin, Patricia E. 2015. The Chinese Communist Treatment of Counterrevolutionaries, 1924–​1949.
Princeton, NJ: Princeton University Press.
Gutiérrez-​Sanín, Francisco, and Elisabeth Jean Wood. 2014. “Ideology in Civil War: Instrumental
Adoption and Beyond.” Journal of Peace Research 51, no. 2: 213–​226.
Hashim, Ahmed S. 2019. “The Islamic State’s Way of War in Iraq and Syria: From Its
Origins to the Post Caliphate Era.” Perspectives on Terrorism 12, no. 1: 23–​32. https://​www.
universiteitleiden.nl/​binaries/​content/​assets/​customsites/​perspectives-​on-​terrorism/​2019/​
issue-​1/​hashim.pdf.
Hassan, Hassan. 2015. “Isis Has Reached New Depths of Depravity. But There Is a Brutal Logic
Behind It.” Guardian, February 7, 2015. https://​www.theguardian.com/​world/​2015/​feb/​08/​
isis-​islamic-​state-​ideology-​sharia-​syria-​iraq-​jordan-​pilot.
 The Islamic State as a Revolutionary Rebel Group    867

al-​Hassan, Sair. 2012. “‘Ayoub’ a New Slap to Classical Wars.” Al-​Manar, November 8, 2012.
http://​archive.almanar.com.lb/​article.php?id=343051.
Huang, Philip C. C. 1995. “Rural Class Struggle in the Chinese Revolution: Representational
and Objective Realities from the Land Reform to the Cultural Revolution.” Modern China 21,
no. 1: 105–​143.
Huntington, Samuel P. 2006. Political Order in Changing Societies. New Haven, CT: Yale
University Press.
Islamic State. 2014. “Dabiq: The Return of the Khalifah.” The Islamic State lists its publication
according to the Islamic calendar and the publication for this title is Ramadan 1435, meaning
approximately June/​ July 2014. https://​jihadology.net/​wp-​content/​uploads/​_​pda/​2014/​07/​
islamic-​state-​22dc481biq-​magazine-​122.pdf.
Islamic State. 2015a. Archive of Islamic State Administrative Documents. Translated by
Aymenn Jawad al-​Tamini. http://​www.aymennjawad.org/​2015/​01/​archive-​of-​islamic-​state-​
administrative-​documents.
Islamic State. 2015b. “Dabiq: The Revival of Jihad in Bengal.” The Islamic State lists its publica-
tion according to the Islamic calendar and the publication for this title is Safar 1437, meaning
approximately November/​ December 2015. https://​jihadology.net/​wp-​content/​uploads/​_​
pda/​2015/​11/​the-​islamic-​state-​e2809cdc481biq-​magazine-​12e280b3.pdf.
Jones, Seth G., James Dobbins, Daniel Byman, Christopher S. Chivvis, Ben Connable,
Jeffrey Martini, Eric Robinson, and Nathan Chandler. 2017. Rolling Back the Islamic State.
Santa Monica, CA: Rand Corporation. https://​www.rand.org/​pubs/​research_​reports/​
RR1912.html.
Kahveci, Riad. 1998. “Hezbollah and the Strategy of Fatigue and Options, ‘Israel.’” Translated by
Mollie Todd. Al-​Moqawama, December 9, 1998. https://​www.moqawama.org/​essaydetailsf.
php?eid=7281&fid=29.
Kalyvas, Stathis N. 2018. “Jihadi Rebels in Civil War.” Dædalus 147, no. 1: 36–​47.
Khalil, Ahmad. 2014. “A Teacher in Raqqa, Living under ISIS Rule.” News Deeply, Syria Deeply,
November 18, 2014. https://​deeply.thenewhumanitarian.org/​syria/​articles/​2014/​11/​18/​
a-​teacher-​in-​raqqa-​living-​under-​isis-​rule.
Khashan, Hilal, and Ibrahim Mousawi. 2007. “Hizbullah’s Jihad Concept.” Journal of Religion &
Society 9: 1–​19.
Larkin, Bruce D. 1973. China and Africa, 1949–​1970: The Foreign Policy of the People’s Republic of
China. Berkeley: University of California Press.
Lia, Brynjar, and Thomas Hegghammer. 2004. “Jihadi Strategic Studies: The Alleged al Qaida
Policy Study Preceding the Madrid Bombings.” Studies in Conflict & Terrorism 27, no.
5: 355–​375.
Lovell, Julia. 2019. Maoism: A Global History. London: The Bodley Head.
Mao Zedong. (1937) 1989. On Guerrilla Warfare. Translated by Brigadier General Samuel
B. Griffith. U.S. Marine Corps, Document FMFRP 12–​18. https://​www.marines.mil/​Portals/​
1/​Publications/​FMFRP%2012-​25%20%20The%20Guerrilla%20and%20How%20to%20
Fight%20Him.pdf.
Mao Zedong. 2011. Selected Military Writings of Mao Tse-​Tung. Kindle ed. Barnsley, UK:
Praetorian Press.
Maurer, Thomas. 2018. “ISIS’s Warfare Functions: A Systematized Review of a Proto-​
state’s Conventional Conduct of Combat Operations.” Small Wars & Insurgencies 29, no.
2: 229–​2 44.
868   Megan A. Stewart

Minden, Karen. 1979. “The Development of Early Chinese Communist Health Policy: Health
Care in the Border Region, 1936–​1949.” American Journal of Chinese Medicine 7, no.
4: 299–​315.
Mondlane, Eduardo. 1983 The Struggle for Mozambique. London: Zed Books.
Naji, Abu Bakr. 2006. The Management of Savagery: The Most Critical Stage Through Which the
Umma Will Pass. Translated by Will McCants, with financial support from the John M. Olin
Institute for Strategic Studies at Harvard University. https://​archive.org/​stream/​TheManag
ementOfBarbarismAbuBakrNaji/​The+Management+of+Barbarism+-​+Abu+Bakr+Naji_​
djvu.txt.
Opper, Marc. 2018. “Revolution Defeated: The Collapse of the Chinese Soviet Republic.”
Twentieth-​Century China 43, no. 1: 45–​66.
Osborne, Samuel. 2015. “Isis Propaganda Shifts to Focus on State-​Building Rather than
Glorified Acts of Violence, Study Suggests.” The Independent, October 6, 2015, 12:55. http://​
www.independent.co.uk/​news/​world/​middle-​east/​isis-​propaganda-​shifts-​to-​focus-​on-​
state-​building-​rather-​/​than-​glorified-​acts-​of-​violence/​-​study-​a6681856.html.
Pepper, Suzanne. 1999. Civil War in China: the Political Struggle 1945–​1949. Lanham, MD:
Rowman & Littlefield.
Qassim, Abdul Satar. 2017. “We Will Not Copy the Experience of Hezbollah”. Translated
by Mollie Todd. Rai al-​Youm, October 19, 2017. https://​www.raialyoum.com/​index.
php/ ​ % D9%84%D9%86- ​ % D9%86%D8%B3%D8%AA%D9%86%D8%B3%D8%AE-​
%D8%AA%D8%AC%D8%B1%D8%A8%D8%A9- ​ % D8%AD%D8%B2%D8%A8-​
%D8%A7%D9%84%D9%84%D9%87/​.
Reuter, Christoph. 2015. “Secret Files Reveal the Structure of Islamic State.” Der Spiegel, April
18, 2015. https://​www.spiegel.de/​international/​world/​islamic-​state-​files-​show-​structure-​of-​
islamist-​terror-​group-​a-​1029274.html.
Revkin, Mara. 2020a. “Competitive Governance and Displacement Decisions under Rebel
Rule: Evidence from the Islamic State in Iraq.” Journal of Conflict Resolution 65, no.1: 46–​80.
Revkin, Mara. 2020b. “What Explains Taxation by Resource-​Rich Rebels? Evidence from the
Islamic State in Syria.” Journal of Politics 82, no. 2: 757–​764.
Robinson, Eric, Daniel Egel, Patrick B. Johnston, Sean Mann, Alexander D. Rothenberg, and
David Stebbins. 2017. When the Islamic State Comes to Town. Santa Monica. CA: RAND
Corporation. https://​www.rand.org/​pubs/​research_​reports/​RR1970.html.
Rukmini, Callimachi. 2018. “The ISIS Files: When Terrorists Run City Hall.” New York Times, April
4, 2018. https://​www.nytimes.com/​interactive/​2018/​04/​04/​world/​middleeast/​isis-​documents-​
mosul-​iraq.html.
Ryan, Michael W. S. 2013. Decoding Al-​Qaeda’s Strategy: The Deep Battle against America.
New York: Columbia University Press.
Ryan, Michael W. S. 2014. “From Theory to Action: The Rationale behind the Re-​establishment
of the Caliphate.” Terrorism Monitor 12: 8–​11. https://​jamestown.org/​program/​from-​theory-​
to-​action-​the-​rationale-​behind-​the-​re-​establishment-​of-​the-​caliphate/​.
Salih, Mu’ad Jabar Hassan. 2015. “Reading the Deaesh Ideology Depending on ‘The
Management of Barabarism’ Book.” Translated by Mollie Todd. Journal of Kerbala University
13, no. 4: 1–​21. https://​www.iasj.net/​iasj?func=fulltext&aId=107539.
Al Sayegh, Nasri. 2014. “From Ho Chi Minh to Sayeed Hassan Nasrallah.” Translated by Nada
Mousa. Al-​Manar, July 17, 2014. http://​archive.almanar.com.lb/​article.php?id=904385.
Selden, Mark. 1995. China in Revolution: The Yenan Way Revisited. Armonk, NY: ME Sharpe.
Seybolt, Peter J. 1971. “The Yenan Revolution in Mass Education.” China Quarterly 48: 641–​669.
 The Islamic State as a Revolutionary Rebel Group    869

Skocpol, Theda. 1979. States and Social Revolutions: A Comparative Analysis of France, Russia
and China. Cambridge: Cambridge University Press.
Snow, Edgar. 1968. Red Star over China: The Classic Account of the Birth of Chinese Communism.
New York: Grove Press.
Spencer, Amanda N. 2016. “The Hidden Face of Terrorism: An Analysis of the Women in
Islamic State.” Journal of Strategic Security 9, no. 3: 74–​98.
Stewart, Megan A. 2021. “Governing for Revolution: Statebuilding in Civil War. Cambridge:
Cambridge University Press.
Van Ness, Peter. 1970. Revolution and Chinese Foreign Policy: Peking’s Support for Wars of
National Liberation. Berkeley: University of California Press.
Weiss, Michael, and Hassan Hassan. 2016. ISIS: Inside the Army of Terror. Updated edition.
New York: Simon & Schuster.
Whiteside, Craig. 2016a. “The Islamic State and the Return of Revolutionary Warfare.” Small
Wars & Insurgencies 27, no. 5: 743–​776.
Whiteside, Craig. 2016b. “New Masters of Revolutionary Warfare: The Islamic State Movement
(2002–​2016).” Perspectives on Terrorism 10.4. http://​www.terrorismanalysts.com/​pt/​index.
php/​pot/​article/​view/​523.
Wickham-​Crowley, Timothy P. 1987. “The Rise (and Sometimes Fall) of Guerrilla Governments
in Latin America.” Sociological Forum 2.3: 473–​499.
Wilson Center. n.d.“Biographies: Ho Chi Minh.” Washington, DC: Wilson Center. https://​
digitalarchive.wilsoncenter.org/​resource/​cold-​war-​history/​ho-​chi-​minh/​biography.
 Index

Tables and figures are indicated by t and f following the page number

A Ali Khan Mahmudabad, 284

Abbasid Caliphate, 37–​39, 619, 627–​28 The Alliance (Malaysia), 261
‘Abd al-​Karim Mouti, 489 Alliance of the Forces of Progress (AFP), 307
Abdou, Sheikh Mohamed, 485f Anavatan Partisi (Motherland Party, ANAP),
ʿAbdulḥamīd II (sultan of Ottoman Empire), 146–​47, 760n4
55–​60 Ançar Dine movement, 810
ʿAbdülmecīd Efendi (sultan of Ottoman Andijan uprising (Uzbekistan, 2005), 423
Empire), 65 Anjuman-​i-​Sipah-​i-​Sahaba of Pakistan
Abdul Rahman, Tunku, 87 (ASSP), 176
Adityanath, 283 Ansar al-​Sharia
al-​ʿAdl wal-​Ihsan (JCG), 110, 484, 490–​93, 495, Arab Spring uprising and, 586–​89
605, 609 demise of, 23, 582, 591–​92
Advani, Lal Krishna, 281 Ennahda and, 589–​92
al-​Afghani, Djamal ad-​Din, 485 Salafism and, 23, 581–​82, 586–​89, 591–​93
Afghanistan, 119, 273, 414, 417, 423–​25, 434 US Consulate protests (2012) and, 591
Africa. See Sub-​Saharan Africa; specific Ansar-​Beit al-​Maqdis, 573–​74
countries Anwar Ibrahim, 87
Ahle Hadith sect, 434 Arab-​Israeli wars, 432–​33, 563, 756–​57, 773, 780
Ahmad Hamaullah, 810 Arab Spring uprisings
Ahmadinejad, Mahmoud, 253 Egypt and, 22, 101, 197, 561, 566–​67, 597–​98,
Ahmadis, 120–​22, 135, 172–​73, 176, 273, 600–​3, 608–​13
286–​87, 701 Islamists and, 22, 597–​613
AIMIM Party (India), 285 Libya and, 198
AIUDF Party (India), 285 military responses to, 193, 197–​98, 205
Akaev, Askar, 420 Morocco and, 22, 491–​92, 597–​98, 600–​1,
Aḳseki, Ḥamdī, 67 604–​6, 609–​12, 831
Alawites, 522–​26, 528–​32 Syria and, 10, 197, 524–​31, 534
al-​Alfi, Mohamed Bey, 659–​60 Tunisia and, 22, 197, 204, 391, 581, 585–​89,
Algeria 597–​98, 600–​1, 603, 606–​12
coup (1992) in, 109 Yemen and, 198
education in, 773 Arba’een pilgrimage (Karbala, Iraq), 355, 358
French colonial era in, 650 al-​Arefe, Mohamed, 571
healthcare in, 774 Arınç, Bülent, 147–​48, 150
Islamism in, 109–​11, 793–​94, 831 Armenians in the Middle East, 637, 640, 644,
pension system in, 776 647
welfare service provision in, 769 al-​Aroor, Adnan, 532
women legislators in, 832 Arslan, Chakib, 485
872   Index

Ashanti kingdom, 711 economic performance under, 128

Asian Financial Crisis (1998), 181, 262, 727 elections under, 118, 122–​23
al-​Assad, Bashar Islamism and, 173–​74
Arab Spring uprising (2011) and, 529–​31 modernizing religious reform and, 126, 132
ascent to power (2000) of, 524 resignation (1969) of, 126, 131, 133, 174
sectarian political appeals by, 525–​26, 529, 534 shrine familes in Pakistan and, 688–​89
Syrian civil war (2011-​) and, 531–​32 Azerbaijan, 116
al-​Assad, Hafez, 109, 522–​24 al-​Azhar University, 486, 564, 659, 662, 753
Association malienne pour l’unité et le progrès Aziz, K. K., 678–​79, 684
de l’Islam (AMUPI), 809–​10
Association of Independent Industrialists and B
Businessmen (MUSIAD), 152 Badawi, Abdullah Ahmad, 728
al-​Aswani, Alaa, 108 Baghdad (Iraq), 555, 619–​21
Atatürk, Kemal, 53, 65–​66, 82, 337 al Baghdadi, Abu Bakr, 469–​70
At-​Tawhid wa al-​Islah, 489 Bahrain, 197
authoritarianism Bahujan Samaj Party, 285
coups d’état as means of ending, 195 Bakhtiar, Makhdoom Khusro, 693
courts and, 129–​30 Bamba, Cheikh Amadou, 296–​97
economic shocks and, 127–​28 Bangladesh, 133, 174–​75, 284, 435, 792, 798
in Egypt, 21, 101–​5, 130, 214–​15 Bank Islam Malaysia Berhad, 726–​27
elections as means of ending, 195 Bank Negara Malaysia (Malaysian Central
electoral authoritarianism and, 142–​43, 260, Bank), 725–​26, 728
481–​83, 495–​96 al-​Banna, Hassan, 602
foreign aid and, 201–​3 al-​Bardisi, Uthman Bey, 659–​60
in Iran, 215, 365 Barelvi sect, 288
military as political actor and, 21, 193–​96, 205 Barisan Nasional (BN), 261–​62, 264
Muslim-​majority societies’ overall record bayʿah, 245
regarding, 193–​98 Baʿth Party, 522–​23, 794
in Pakistan, 117, 126–​34, 170 Bellin, Eva, 132, 195, 391–​92
popular uprising as a means of Ben Ali, Zine el-​Abidine
overturning, 195 Arab Spring uprising (2011) and ouster of,
pre-​Islamic empires and, 38 603, 606–​9
regime-​friendly political parties and, coup (1987) by, 583, 607
128–​29 Ennahda and, 109, 584
social media and, 575–​77 Islamists suppressed by, 22, 220, 584
in Syria, 522–​24 political parties’ relationships with, 494
threat perception and, 103 Ben Barka, Mehdi, 488, 490
in Tunisia, 214, 494, 583–​84 Bengal, 121, 132, 170, 275, 284–​85
in Turkey, 53–​54, 130, 143–​60 Benkirane, Abdelilah, 489, 491–​92, 605
typologies of, 117–​18 Ben Shitrit, Lihi, 503, 516
urban protests and, 130–​32 Berat Albayrak, 150
Awami League, 174 Bharatiya Janata Party (BJP)
Ayodhya religious site controversy (India, affirmative action programs criticized by, 279
1989-​1992), 280–​81, 283 Ayodhya religious site controversy (India,
Ayub Khan, Muhammad 1989-​1992), 280–​81, 283
authoritarian nature of rule of, 173, 286 emergence as a national party of, 271–​72,
Convention Muslim League and, 129 281–​85
 Index   873

Hindu nationalism and, 22, 271, 281–​84, 289 Christian Democratic parties (Europe), 216,
social service provision and, 281, 799–​800 253
technocratic political appeals made by, 22, clientelism
272, 281 constituency services compared to, 319–​20,
Bhutto, Benazir, 124, 126, 133 325–​29
Bhutto, Zulfikar Ali, 129, 131, 133, 174–​75, 272–​ definition of, 318–​19, 828
73, 286–​87 in Lebanon, 320–​24
Bierstedt, Robert, 688–​89 Muslim versus non-​Muslim societies and,
Bint Muhammad Hassan, Sara, 512, 516 23, 315–​18
Blaydes, Lisa, 18, 622, 625, 633, 656 perverse accountability created by, 314, 329
Bloc Démocratique Sénégalais (BDS), 299 poorer voters and, 322–​24
Bolkiah, Hassanal, 89–​92 responsiveness under, 828–​29
Bolsheviks, 855, 857 in Yemen, 320–​24
Bosnia, 541 Cold War, 195, 772, 862–​63
Bouabid, Abbderrahim, 488 Committee for Union and Progress (CUP), 55,
Bouazizi, Mohamed, 391 58–​62
Bourguiba, Habib, 214, 217, 220, 583, 607, 771 Congress for the Republic Party (CPR), 222–​24,
British Empire 226, 230
Brunei and, 90–​91 Congress Party, 179, 184, 272–​83, 289. See also
Egypt and, 56, 650, 667–​68, 753–​55 Indian National Congress
India and, 274–​75 Constitutional Union (UC), 838
Malaysia and, 85–​86, 724 Convention Muslim League (CML), 129
mandate system in Middle East and, 650, Coordination of Azawad Movements (CMA),
755–​56 819
Pakistan and, 170, 677–​80 Coptic Christians
property rights regimes and, 19 conversions to Islam during medieval era
United Provinces (India) and, 170–​7 1, 275 among, 642, 647–​49
Brunei, 75, 78, 80, 89–​93 education levels among, 644, 648
al-​Bukhari, Muhammad, 628 health outcomes among, 651
Bulut, Yigit, 150 poll taxes during medieval and early
Burkina Faso, 256, 438–​40, 700, 702–​3, 707–​8, modern era on, 9, 638–​39, 646–​49
816 role of saints and martyrs in the faith
Burundi, 540, 700, 702–​3, 706, 708 of, 643
rural concentration of, 641, 644
C socioeconomic status of, 9, 637–​50
Cameroon, 700–​3, 706–​8 Cote d’Ivoire, 308, 700, 702–​3, 816
Catholics, 77, 306–​7, 551, 640, 643 Council of Malay Rulers, 90
Cevdet Paşa, Aḥmed, 57 Crescent Star Party (PBB), 260
Chad, 700–​3, 706, 708 Cromer, Lord, 649, 753
Chaney, Eric, 18, 622, 624–​25, 630, 633, 676
Chaudhri Zafrullah Khan, 172 D
Chaudhry, Mohammad Iftikhar, 126, 130 Daesh. See Islamic State (ISIS)
Chérif of Nioro du Sahel (Mohamed Ould Dahl, Robert, 78, 117
Cheickna Hamaullah), 810–​12, 819 Darling, Malcolm, 678
Chinese Communist Party (CCP), 853–​58, Darrag, Amr, 799
862–​65 Davutoglu, Ahmet, 150
Christian Association of Nigeria, 548 Daʿwa Party, 104, 109, 857
874   Index

democracy Ecoles des Fils Notables, 757

definitional aspects of, 115–​16 economic development
democratic backsliding and, 143 colonialism and, 637–​38, 644–​46, 650
economic development and, 335, 709 democracy and, 335, 709
Egypt and, 221–​22 Islamic banks and finance and, 720–​21
hybrid democradura regimes and, 122, 126 Islamic law and, 622–​23, 630–​31
hybrid dictablanda regimes and, 126, 130 madrassa system and, 630
India and, 167, 182–​86 non-​Muslim minorities in the Middle East
Indonesia and, 80–​81, 83–​85, 93, 127, 167, and, 637–​51
180–​81, 245, 256 poll taxes and, 649–​50
Iran and, 356 religious legitimacy claims and, 622–​23, 633,
Iraq and, 356–​57, 360, 362–​64 722
legacy of medieval Arab conquest and, 43–​44 state institutions and, 655–​56, 668
Malaysia and, 80, 88–​89, 93, 245, 256 Sub-​Saharan Africa and, 697–​7 15
Muslim-​majority societies’ overall record Eddin, Tag, 782
regarding, 194–​95, 335, 709 education
Pakistan and, 13–​14, 115–​19, 122, 125–​26, 131–​ in Algeria, 773
36, 167, 172–​73, 177, 430, 435–​36, 440–​41 in Egypt, 646, 710, 748–​50, 752–​54, 773–​74
pluralism and, 13–​14 in Jordan, 748–​50, 755–​57, 773–​74
populism and, 346–​51 madrasas and, 40, 42, 45, 62–​63, 288, 630,
religiosity and, 13–​14, 167, 336, 345–​46, 430, 633, 814–​15
434–​36, 439–​41, 444 in Mali, 813–​15, 818
secularism and, 75–​77 in Morocco, 748–​50, 757–​58, 773–​74, 841
Tunisia and, 204, 221–​22, 245, 581, 589 Ottoman state secular schools and, 56–​57
Turkey and, 13–​14, 53–​54, 69–​70, 156, 159, political participation levels and, 397–​98, 401
256, 346–​51 in Sub-​Saharan Africa, 698, 701–​7, 710–​11,
typologies of, 117–​18, 122 713–​15
Democratic Action Party (DAP), 261–​62 in Tunisia, 748–​52, 773–​74
Democratic Forces Front (FFD), 838–​39 in Turkey, 61–​63, 339–​40
Democratic Party (Turkey), 146, 149, 337–​38 Egypt
Democrat Party (Indonesia), 259–​60 Advisory Council of Delegates in, 664–​67
Demokratik Sol Parti (Democratic Left Party, Arab Conquest ( 641 CE) of, 638–​39
DSP), 147 Arab-​Israeli wars and, 563, 773, 780
Deobandi sect, 288, 434 Arab Spring uprising (2011) in, 22, 101, 197,
Devasher, Madhavi, 280 561, 566–​67, 597–​98, 600–​3, 608–​13
Dhofar Rebellion (Oman), 772 authoritarianism in, 21, 101–​5, 130, 214–​15
Dicko, Mahmoud, 810–​12, 819 Bread Riots (1977) in, 793
Diola ethnic group, 296 British colonial era in, 56, 650, 667–​68,
Diouf, Abdou, 299, 301–​5, 307 753–​55
Djibouti, 700–​1, 703 Caisse de la Dette in, 665
Doğru Yol Partisi (DYP), 146–​49 Consultative Legislative Council in, 667
Dufferin, Lord, 667 Coptic Christians in, 9, 637–​50
Durkheim, Emile, 4, 6–​7, 372 Council of Consultation in, 663
Dutch Republic, 623, 632 coup (2013) in, 101, 106–​10, 568–​69
defense spending in, 732
E democracy in, 221–​22
Eastern Turkestan Islamic Movement, 421 education in, 646, 710, 748–​50, 752–​54,
Ebussuud Efendi, 722 773–​74
 Index   875

election (2012) in, 567–​68 religious beliefs among supporters of,

food and energy subsidies in, 776 227–​30
Free Officers’ coup (1952) in, 752, 754, 780 secularism and, 589
French occupation (1798-​1801) of, 657, 659 shar’ia law and, 601
healthcare in, 664, 748–​55, 774–​75, 782, 792, suppression of, 22, 109–​10, 221, 584
827 “Turkish model” and, 145
intelligence agencies and secret police women’s participation in, 831, 846
forces in, 105, 195 Erbakan, Necmettin, 68–​69, 146–​48, 339
Islamism in, 16, 22, 101–​11, 213, 215–​22, 225–​31, Erdoğan, Recep Tayyip
256, 563–​77, 597–​98, 601–​3, 608–​10, 613, attempted coup (2016) against, 158–​60
790–​93, 797, 799, 827, 833 consolidation of power by, 20, 143–​45, 156,
Mamluks as administrators during 159–​60, 345
nineteenth century in, 657–​60, 662–​63, Egypt and, 571
668 Islamism and, 145, 148–​49, 159
Mamluk Sultanate in, 656–​57 Israel and, 157
Ministry of Social Affairs in, 781 Justice and Development Party’s rise to
Muqabala Law (Law for Exchange) in, power and, 147–​48, 339
664–​65 Kotku and, 68
Nile River flooding in, 624, 658, 676 pro-​government businesses and, 154
Ottoman viceroys (pashas) in, 657–​58, 660 Welfare Party and, 146
pension system in, 776 Yarbay and, 150
poll taxes during medieval and early Ettakatol Party, 223–​26, 229–​30
modern era in, 9, 638–​39, 646–​49
postcolonial era cabinets in, 745–​46 F
Rabaa Massacre (2013) in, 107–​9, 569, Facebook, 101, 561, 566–​73, 575–​76, 608
571–​72 al-​Fasi, Allal, 485, 488, 493
shariʿa law and, 610–​11 February 20th movement (Morocco), 491, 493,
social media in, 561–​62, 565–​77 601, 604–​6, 609
state centralization and government Federation of Malaya, 261
spending during nineteenth century in, Felicity Party (SP), 146, 149
19, 656, 659–​68 Fidan, Hakan, 157–​58
state of emergency (1987-​2011) in, 201 Fırat, Dengir Mir Mehmet, 150
Tahrir Square protests (2011) in, 561, 602 First Congress of Indonesian Youth, 178
United States and, 201 France
welfare service provision in, 710, 743–​44, Algeria and, 650
747–​55, 758–​59, 771, 773–​75, 780–​82, 784 Catholic Church and, 751
England, 623, 632 Egypt occupied (1799-​1801) by, 657, 659
Ennahda Lebanon and, 650
Ansar al-​Sharia and, 589–​92 Mali and, 812–​13
Arab Spring uprising and, 585–​86, 597, 601, Morocco and, 485–​87, 757
606–​7 revolution (1789) in, 855
democratization and, 589 Senegal and, 297, 299, 304
economic platform of, 225–​27 Syria and, 522, 650
electoral success of, 213, 607, 612–​13 Tunisia and, 56, 650, 751–​52
al-​Fajr newspaper and, 607 Franz, Erica, 131–​32
government suppression of, 607, 613 Freedom and Justice Party (FJP), 213, 217,
organizational strength of, 222–​24 222–​25, 229, 565, 570, 602
political mobilization strategies of, 581 Free Syrian Army, 531–​32
876   Index

French Socialist Party (SFIO), 299 Halkların Demokratik Partisi (Peoples’

Front Islamique du Salut (FIS), 109–​10, 794 Democratic Party, HDP), 148, 151
Hamas, 253, 797
G Hama uprising and massacre (Syria, 1982),
Galata bankers (sarrafs), 722 109–​10, 522–​23
Gandhi, Indira, 184–​85 Hamawiyya order, 810
Gandhi, Mahatma, 183, 272, 275 al-​Haqq Party, 321
Gandhi, Rahul, 283, 285 Hashemite Dynasty, 756–​57
Geddes, Barbara, 131–​32 Hassan II (king of Morocco), 488, 490, 493, 604
Genc Parti (Young Party, GP), 153 Hatta, Mohammad, 179
Gerth, Hans, 688 Hausa ethnic group, 545, 549–​51
Gezi Park protests (Istanbul, 2013), 399–​401, Haut Conseil Islamique (HCI), 810, 812, 814
406 headscarves. See hijab wearing
Ghana, 700–​3, 706–​8, 711 healthcare
Ghannouchi, Rached, 221, 583, 607 in Algeria, 774
Ghulam Ishaq Khan, 123–​24, 133 colonial regimes and, 743
Gibb, Hamilton Alexander Rosskeen, 40 in Egypt, 664, 748–​55, 774–​75, 782, 792, 827
Gilani, Yousuf Raza, 126, 134, 690 in Jordan, 748–​51, 755–​57
Gilani Syeds of Multan, 690 in Morocco, 748–​51, 758, 775
Glubb, John, 755 in Syria, 774–​75
Godhra riots (India, 2002), 282 in Tunisia, 748–​51, 774–​75
Golden Age of Islam, 36, 622–​23, 630 Heck, Paul L., 675
Golkar Party, 259–​60, 264, 794 hijab wearing
Great Indonesia Movement Party, 259–​60 Kyrgyzstan and, 370, 374–​75, 377–​79, 385,
Greek Orthodox Christians, 637, 640, 644, 647 387–​88, 390–​91
Grief, Avner, 693 political mobilization regarding rights for,
Grubman, Nathan, 494 16, 370, 377–​79, 385, 387–​91
Guevara, Che, 857–​58 state prohibitions regarding, 16, 335, 340
Guèye, Lamine, 299, 303–​4 Tunisia and, 335
Gujarat (India), 281–​83 Turkey and, 335, 340
Gül, Abdullah, 147–​50 Hinduism
Gülen, Fethullah caste system and, 277–​80
coup (1980) and, 68 Hindu-​Muslim violence and, 282–​84
influence in Turkish government of Hindu nationalism and, 22, 185–​86, 271–​72,
followers of, 68–​69, 148, 158–​59 281–​85, 289
Israel and, 157 Hindu Mahasabha Party, 275
Justice and Development Party and, 156–​58, Hizb al-​Istiqlal (Independence Party), 486,
160 488, 758
pious bourgeoisie among followers of, 151 Hizb al-​Nour Party, 789
purge of the followers of, 70, 156, 158–​60 Hizbullah, 15, 321, 798, 857–​58, 860, 864–​65
Gulf Cooperation Council (GCC), 467, 783 Hizb ut-​Tahrir, 422
Gulf States, 532, 769, 772–​73, 775 Ho Chi Minh, 856–​57
Gupta, Pranav, 280, 282t homosocial capital, 832, 836–​37, 842–​43
Houka Houka ag Alhousseini, 816
H Hussein (imam of Karbala), 355, 358
Haïdara, Ousmane Madani, 810–​12, 819 Hussein, Saddam, 104, 109, 524
Halkın Sesi Partisi (HAS Parti), 149 Hussein, Taha, 754
 Index   877

I Indonesia
Ibn Khaldun, 37–​38, 48n24 anti-​colonial movement and nationalism
Ibrahim Bey, 659 in, 14, 81, 168–​69, 177–​79
Ibrahim Pasha, 663 Asian Financial Crisis (1997-​1998) and, 181
Igbo ethnic group, 545, 550 Chinese population in, 177–​81
ijtihād (independent reasoning), 627–​30 civil rights and religious freedoms in, 80–​81,
Ila al-​Amam Party, 488, 490 83–​84, 93, 179
illicit economies democracy in, 80–​81, 83–​85, 93, 127, 167,
Afghanistan and, 414, 417, 423–​25 180–​81, 245, 256
Central Asia and, 417, 419–​25 Dutch colonial era in, 81, 86, 177–​78
organized crime and, 412, 417–​18 federalism in, 84
political elites’ participation in, 21, 417, 422, Guided Democracy era in, 180, 258–​60
424 Islam and national identity in, 14, 81, 167,
terrorism and, 412, 416–​25 169, 178, 187
Imran Khan, 125, 134, 287–​88, 692 Islamism in, 15, 81–​85, 260, 263–​64, 790,
India 792, 794–​95, 798, 800
anti-​colonial movement and nationalism land reforms (1960) in, 676
in, 14, 168–​69, 182–​83 military as political actor in, 82, 180–​81, 205
Ayodhya religious site controversy (India, New Order Era in, 180, 258–​60, 264, 794–​96
1989-​1992), 280–​81, 283 Pancasila ideology in, 82–​84, 179–​81, 187
Bangladesh and, 175 party competition in, 252, 257–​60, 262–​65
British colonial era in, 274–​75 pre-​Islamic societal structures in, 43
China’s border conflict with, 289 religiosity in, 438–​39
Citizenship Amendment Act in, 284 secularism and, 75, 80–​85, 93
Constitution of 1950 in, 272, 276 shariʿa law, 440
democracy in, 167, 182–​86 state-​building in, 81–​83
Emergency Period (1975-​1977) in, 185 ulama in, 78, 81, 85
Godhra riots (2002) in, 282 United States and, 180, 201
Government of India Act (1935) and, 274 World War II and, 179
Hindu-​Muslim conflict in, 183 Indonesian Communist Party (PKI), 180,
Hindu nationalism in, 22, 185–​86, 271–​72, 259–​60
281–​85, 289 Indonesian Democratic Party (PDI), 180,
Lahore Declaration (1999) and, 124 259–​60
military culture in, 204 Indonesian Democratic Party of Struggle
Muslims’ socioeconomic status (PDIP), 84, 259, 263
in, 277–​79 Indonesian Nationalist Party (PNI), 177–​80,
nonviolent strategies in anti-​colonial 259–​60
movement in, 183–​85, 196–​97, 200, 204 International Monetary Fund (IMF), 776–​80,
partition of, 184, 272–​73, 276, 430, 436, 783
683–​84 iqta system (agricultural taxation), 18, 40
religion and national identity in, 167, 169, Iran
182–​87 authoritarianism in, 215, 365
religious political parties in, 15, 22, 271–​73, constitution in, 119
277–​85 democracy and, 356
Sufism and political authority in, 693 education in, 773
Indian National Congress, 170, 182–​84. See elections in, 256
also Congress Party healthcare in, 775
878   Index

Iran (cont.) Islamic Medical Association, 793

Islamic Revolution (1979) in, 774 Islamic Movement of Uzbekistan (IMU), 414,
Pahlavi Dynasty in, 772–​73 420–​25
pension system in, 776 Islamic State (ISIS)
Sasanian Era in, 38 demographic profile of supporters of, 456,
Saudi Arabia and, 434 459–​60, 468
Shiʿi Muslims and, 355–​66 foreign fighters in, 412, 449
Syria and, 524, 530 governance strategies of, 23, 862–​63
theocracy in, 79, 356–​57, 360, 362, 365–​66 guerrilla warfare and, 853, 861–​62, 865
United States and, 201 in Iraq, 466, 470, 853, 862
welfare service provision in, 769, 772, 784 in Libya, 450, 466
Iraq political religiosity among supporters of,
Abbasid Caliphate and, 619 452–​55, 459–​60, 468
British mandate during interwar period political trust levels among supporters of,
in, 650 457–​61, 468
civil war (2000s) in, 524 public opinion in Muslim-​majority
democracy and, 356–​57, 360, 362–​64 societies regarding, 17, 449–​70
Dujail massacre (1982) in, 109–​10 religious ideology and, 863–​65
Gulf War (1991) and, 490 as revolutionary organization, 854, 860–​63
Islamic State in, 466, 470, 853, 862 shar’ia law and, 864
Islamism in, 104, 109–​11 Shiʿi Muslims and, 455, 469
sectarian politics in, 360, 362–​64 social media and, 574
Shiʿi Muslims in, 9, 355–​66 in Syria, 450, 466, 470, 532, 853, 861
welfare service provision in, 772 terrorist attacks by, 861
Islah Party, 15, 321 in Tunisia, 470
Islamic Action Front (IAF), 109–​10, 845 women and, 864
Islamic banks and finance in Yemen, 450, 466
economic development and, 720–​21 Islamic State in the Greater Sahara (ISGS), 816
household debt levels and, 726–​28 Islamic State in the Sinai, 573
large financial corporations and, 720 Islami Jamhoori Ittehad (IJI), 123–​24, 133–​34
majalla (civil code from Ottoman Empire) Islamism
in, 723–​24, 726 in Algeria, 109–​11, 793–​94, 831
in Malaysia, 9, 719–​21, 724–​30 anti-​US attitudes and, 433
shariʿa law and, 9, 719, 722, 724, 726–​29 Arab Spring and, 22, 597–​613
sukuk (Islamic bonds) and, 727–​29 audio cassettes as means of distributing
Islamic Call Society (Ad-​Dawah), 217 message of, 563–​64
Islamic Group (Egypt), 110 authoritarian governments’ suppression of,
Islamic Interbank Money Market, 726 101–​11, 214, 221
Islamic Jihad Group, 422 definition of, 7–​8
Islamic Jihad Union, 414, 425 donor networks and, 808
Islamic law. See also shariʿa law economic platforms among groups
consolidation among religious authorities promoting, 225–​27
in eighth century and, 629–​30 in Egypt, 16, 22, 101–​11, 213, 215–​22, 225–​31,
economic development and, 622–​23, 630–​31 256, 563–​77, 597–​98, 601–​3, 608–​10, 613,
inheritance laws and, 631 790–​93, 797, 799, 827, 833
religious legitimacy claims by rulers and, 626 in Indonesia, 15, 81–​85, 260, 263–​64, 790,
Sunni schools of legal interpretation and, 41 792, 794–​95, 798, 800
 Index   879

in Iraq, 104, 109–​11 Jaysh al-​Mahdi, 421

Islamic mandate effect and, 23, 833–​36, 840, Jews, 637, 640–​41, 643–​45, 647, 650
842–​43 Jinnah, Mohammad Ali, 120, 171, 275–​76,
LGBTQ rights and, 78 285–​86, 434
in Libya, 16, 213–​14, 217–​21, 225–​31 Joko Widodo, 84
in Malaysia, 15, 87, 89, 257, 264, 795–​96 Jordan
in Morocco, 22–​23, 104, 110–​11, 483–​92, 495, Bedouin population in, 755–​56
597–​98, 601, 604–​6, 609–​10, 828–​43 British mandate during interwar period in,
non-​Islamist audiences’ attitudes regarding, 650, 755–​56
105–​11 defense spending in, 780
organizational strength of groups democracy and, 256
promoting, 215, 222–​24 Desert Medical Unit in, 756
in Pakistan, 15, 22, 120, 125, 175–​76, 288, 434, East Bank in, 748, 755–​56
441–​43, 797–​98 education in, 748–​50, 755–​57, 773–​74
political mobilization and, 103–​4 elite coalitions in, 756, 771–​72
political parties and, 16, 22, 213–​32, 251–​66, food and energy subsidies in, 776
482, 789–​92, 796–​800, 833–​42 Hashemite Dynasty in, 756–​57
religious reasons for supporting groups healthcare in, 748–​51, 755–​57
promoting, 227–​30 Islamism in, 109–​11, 793, 845
satellite television as means of distributing Palestinian population in, 755–​57, 774
message of, 564 postcolonial era cabinets in, 745–​46
social media and, 561–​62, 565–​74, 577 religiosity in, 438–​39
social service provision and, 104, 215, 222, shariʿa law and, 440
226–​27, 231, 313, 325, 789–​801, 807–​8, 816, Six-​Day War (1967) and, 757
819, 833–​34 United Nations Relief and Works Agency
in Syria, 104, 109–​11, 533–​34 hospitals in, 750, 755–​57
in Tunisia, 16, 22, 109–​11, 213–​22, 225–​31, welfare service provision in, 740, 743–​44,
583, 585, 597–​98, 601, 606–​10 747–​51, 755–​57, 759, 771–​72, 780
in Turkey, 53–​55, 67–​70, 146–​50, 215–​16, 265, Jos (Nigeria), 540, 542–​43, 554
338–​40, 827 Junejo, Muhammad Khan, 123, 133
women’s experiences with, 22–​23, 78, Justice and Development Party (AKP)
827–​46 consolidation of power by, 143–​45, 159, 340,
Ismail Pasha Seddiq, 664 345
Israel, 119, 157, 254–​55, 757 democracy in Turkey and, 13–​14, 54, 336,
Izala Brotherhood, 551 346
economic performance under, 145, 148, 151,
J 154
Jabhat al-​Nusra, 532 electoral fraud benefiting, 148
al-​Jamaʿa al-​Islamiyya Party, 321 Erdoğan’s consolidation of power over, 150,
Jamaʿat-​i Islami (JI, Party of Islam), 120, 159
123–​25, 133–​35, 272, 287, 792, 797–​98 European Union and, 144, 155–​56, 215–​16,
Jamaʿat Nusrat al-​Islam wa-​l-​Muslimin 339
(JNIM), 816 Gülen movement and, 156–​58, 160
Jamaʿat Ulema-​i-​Islam (JUI), 120, 123–​25, 133, Islam’s role in Turkish society and, 13, 20, 54,
135, 272, 287 70, 145, 148–​49, 159, 265, 340, 350
Jan Sangh Party, 281 the judiciary and, 157–​58, 160, 340
Japan, 179 Kurdish movement and, 150–​51
880   Index

Justice and Development Party Kurdistan Workers’ Party (PKK), 151, 157, 338
(AKP) (cont.) Kurtulmuş, Numan, 149
media coverage of, 153–​54 Kuru, Ahmet, 77, 79t, 623, 633
origins of, 146–​47, 339 Kutlah al-​Amal al-​Watani, 486
pro-​government businesses and, 152–​54, Kuwait, 130, 769–​70
159, 265 Kyrgyzstan
religiosity among supporters of, 336 anti-​Bakiev protests (2010) in, 421
shariʿa law and, 344–​45 democracy and, 116
social service provision by, 154–​55, 791 hijab wearing in, 370, 374–​75, 377–​79, 385,
387–​88, 390–​91
K
illicit economy in, 420–​21
Kaduna (Nigeria) impact of violence on ethnic identities
Barnawa neighborhood in, 549–​50 in, 541
Christian-​Muslim mistrust in areas of, Islamic revival in, 375
551–​55 Osh region in, 379, 421
Christian-​Muslim segregation in, 548–​50, political mobilization as response to religious
555 regulations in, 370, 373–​80, 385–​93
Christian-​Muslim violence in, 539, 541–​42, Qurʿan study groups in, 370, 374–​75, 377–​79,
545–​48, 555 385–​89, 392–​93
Kakuri neighborhood in, 549–​50 religiosity in, 376, 379, 389
post-​traumatic stress disorder levels in, Soviet era in, 375
540–​41, 551, 554 terrorism in, 411, 414, 421, 425
religiosity levels in, 550–​51 Uzbek population in, 379
shari’a law and, 547–​48
State Assembly in, 548 L
Karamat, Jahangir, 124 Lahore Declaration (1999), 124
Karbala (Iraq), 9, 355, 358 Lashkar-​e-​Jhangvi, 434
Karimov, Islam, 422 Layene brotherhood, 296
Kashmir, 124, 271, 276, 283, 289, 434 Lebanon
Kazakhstan, 411, 414 clientelism in, 320–​24
Kefaya movement (Egypt), 603 democracy and, 256
Keïta, Ibrahim Boubacar, 810–​12, 819 education in, 774
Keïta, Modibo, 809 elections and political competition in, 15
Kerala state (India), 276–​77 France and, 650
Khartoum (Sudan), 541 healthcare in, 775
al-​Khatib, Abdelkrim, 831 Hizbullah and, 321, 860
Khedive Abbas, 663 Islamism in, 321
Khedive Ismail, 664–​65, 667 non-​Muslim minorities in, 639, 650
Khedive Sa’id, 663 religiosity in, 438–​39
Khedive Tawfiq, 667 religious power-​sharing arrangements in
Khyber Pakhtunkhwa (KP), 123, 132, 273, 288 political system of, 308, 772
Konaré, Alpha Oumar, 810 shari’a law and, 440
Kosovo, 541 welfare service provision in, 743, 771–​72, 784
Kotku, Mehmet Zahid, 68, 146 Leghari, Farooq, 124
Kouffa, Amadou, 817 Lembaga Tabung Haji (LTH), 725–​26
Kuran, Timur, 8–​9, 622, 631, 633, 643–​44, 711, Lewis, Bernard, 637, 643
721 LGBTQ rights, 78, 85
 Index   881

Liberal Union, 59 shariʿa law, 85–​86, 88–​89, 440, 725–​28

Liberia, 700, 702–​3, 706–​8 Singapore and, 725
Libya ulama in, 86
Arab Spring uprising (2011) and, 198 Malaysia International Halal Showcase
authoritarianism in, 214 (MIHAS), 728
democracy in, 221–​22 Malaysian Islamic Party (PAS), 87, 261–​62,
healthcare system in, 774 264, 726, 795–​96
Islamic State in, 450, 466 Malaysian Muslim Youth Movement (ABIM), 87
Islamism in, 16, 213–​14, 217–​21, 225–​31 Malaysian People’s Movement (Gerakan), 261
women in, 832 Malaysian United Indigenous Party (Bersatu),
Lieven, Anatol, 678, 686 261–​62
Lipset, Seymour Martin, 624 Mali
long divergence hypothesis, 622, 629–​33, 721 Algiers Accord (2015) and, 819
Lyautey, Hubert, 757 authoritarianism in, 809
clerics’ role in politics and government in,
M
807–​12, 819
madrasas, 40, 42, 45, 62–​63, 288, 630, 633, coup (2012) in, 812
814–​15 coup (2020) in, 811
Mahasabha Party, 275 democracy in, 808, 810
Mahathir Mohamad, 87–​88 education in, 813–​15, 818
Makarfi, Ahmed, 548 Family Code in, 811
Makdisi, George, 40–​42, 44–​45, 47n12 French colonial period in, 813
Makram, Umar, 661–​62 French intervention (2013) in, 812
Malawi, 700, 702–​3, 706–​8, 714 Islamism and jihadist groups in, 807, 810,
Malayan Chinese Association (MCA), 261 815–​20
Malayan Indian Congress (MIC), 261 Islamism in, 810
Malaysia judiciary in, 812–​13, 817
Asian Financial Crisis (1998) and, 262, 727 laicism in, 808, 818
British colonial era in, 85–​86, 724 separatist rebellion in northern regions of,
Chinese population in, 796 812, 816, 819
civil rights and religious freedoms in, 80–​81, shariʿa law and, 817
88–​89, 93 Sufism in, 808, 810
Council of Religious Affairs in, 86 trust in traditional authorities in, 809
democracy in, 80, 88–​89, 93, 245, 256 ulama’s role in social service provision in, 12
economic development in, 725 Maliwad, Kalubhai, 282
Employees Provident Fund in, 728 mamluks (military slavery)
household debt levels in, 727–​28 Abassid Dynasty and, 37
Islam as state religion in, 75, 78, 80, 85–​89, caliphate political legitimacy and, 36, 39, 43
93, 261 consolidation of political power via, 18
Islamic financial institutions in, 9, 719–​21, economic development and, 622, 625, 633
724–​30 iqta taxation system and, 40
Islamism in, 15, 87, 89, 257, 264, 795–​96 landed elites’ powers challenged by system
party competition in, 252, 257–​58, 260–​65 of, 39–​40
pre-​Islamic societal structures in, 43 political legacy of, 18
Public Sector Retirement Fund in, 728 “religionization” of civil society and, 33–​34,
racial riots (1969) in, 87, 93, 261 40, 45
religiosity in, 438–​39 tribal social structures and, 37
882   Index

The Management of Savagery (Naji), 858, Moroccan Association of Human Rights

860–​61, 864 (AMDH), 605
Mao Zedong, 853–​63, 865 Moroccan National Students Union, 488
marabouts (Sufi religious leaders), 296–​309, Morocco
485 Alternance government in, 489–​90
Mardin, Şerif, 337 Amazigh population in, 485, 491, 829,
Maronites, 640 830f
Masjumi Party, 259–​60 anti-​colonial movement in, 483, 485–​87
Mauritania, 119, 700, 702, 808 Arab Spring uprising and, 22, 491–​92,
Mbacké, Modou Kara, 298, 305 597–​98, 600–​1, 604–​6, 609–​12, 831
Mbacké, Serigne Falilou, 298, 303 Berber Dahir crisis (1930) in, 485–​86
Mecca, pilgrimages to, 303, 657, 725 constituency services in, 314, 325–​29
Megawati Sukarnoputri, 84, 180–​81 education in, 748–​50, 757–​58, 773–​74, 841
Melayu Islam Beraja (MIB), 91–​92 electoral authoritarianism in, 325, 482–​84,
Messadi, Mahmoud, 773 495–​96
Mian, Atif, 288 February 20th movement in, 491, 493, 601,
Mihna Era (Abassid Dynasty, 829–​849), 39 604–​6, 609
military as political actor food and energy subsidies in, 776
Arab Spring uprisings and, 193, 197–​98, 205 French colonial era in, 485–​87, 757
authoritarianism and, 21, 193–​96, 205 gender inequality in, 829–​30
democratic transitions and, 204–​6 Green March (1975) in, 493
foreign military aid and, 193, 201–​3, 205 healthcare in, 748–​51, 758, 775
in Indonesia, 82, 180–​81, 205 Islamism in, 22–​23, 104, 110–​11, 483–​92, 495,
Islamist groups suppressed by, 200, 273, 286 597–​98, 601, 604–​6, 609–​10, 828–​43
organizational cultures of militaries and, leftist opposition in, 488–​89, 491–​92
193, 203–​5 monarchical regime in, 325, 483, 486, 489–​94,
in Pakistan, 21, 115–​19, 126, 129, 132, 134–​35, 597, 605, 831
173, 176–​77, 204, 273, 286, 289–​90, 435, 681 Moudawana (law governing the family) in,
political parties aligned with, 129, 287, 435 489–​90, 845
protest suppression and, 21, 193, 195–​98, 205 nationalism in, 486–​87
Senegal and, 204–​5 parliamentary gender quota in, 829
threat perception and, 193, 198–​99, 205 postcolonial era cabinets in, 745–​46
Tunisia and, 204–​5 rural-​urban inequality in, 757–​58
in Turkey, 53, 151, 156, 338, 340 secularism in, 483–​84, 486–​88, 491
military slavery. See mamluks (military shariʿa law and, 610–​11
slavery) Sufism and political authority in, 693
Milli Nizam Partisi, 146 welfare service provision in, 743–​44, 747–​51,
Milliyetçi Hareket Partisi (MHP), 147 757–​59, 771–​72
Mills, C. W., 688 Western Sahara conflict and, 493
Mirzoev, Gaffor, 424 Morsi, Mohamed, 103, 106–​7, 221, 567–​7 1, 573
Miskawayh, 40 Mouride brotherhood (Senegal), 295–​307
Moatassim, Jama’a, 606 Mouvement Alternatif pour les Libertés
Modi, Narendra, 186, 271, 274, 281–​84, 289 Individuelles (MALI), 491
Mohajir Qaumi Movement (MQM), 126, 287 Mouvement de la Tendance Islamique (MTI),
Mohammed V (king of Morocco), 488 110, 220, 607
Mohammed VI (king of Morocco), 490–​91, Mouvement Populaire Démocratique et
604, 831 Constitutionnel (MPDC), 604
 Index   883

Movement for the Restoration of Democracy government suppression of, 101–​11, 220, 335,
(MRD), 126 568–​69, 577, 602–​3
Movement for Unity and Reform (MUR, At-​ Guidance Bureau of, 602–​4
Tawhid wa al-​Islah), 604, 831, 834 out-​group supporters of, 216
Movement of Islamic Tendency, 583. See also political mobilization by, 265–​66
Ennahda shar’ia law and, 601
Mubarak, Hosni social media and, 561–​62, 565–​73, 575–​76
Arab Spring uprising and, 597–​98, 603 social service provision by, 215, 222, 226–​27,
healthcare system and, 775, 782 604, 789–​90, 792–​93, 799, 833
Muslim Brotherhood suppressed by, 103–​4, websites as means of distributing message
106–​7, 256, 602–​3 of, 565
ouster (2011) of, 101 Muslim Brotherhood (Jordan), 109, 335, 798
state of emergency (1987-​2011) under, 201 Muslim Brotherhood (Libya), 220
Mufti Mahmud, 123 Muslim Brotherhood (Syria), 109, 522–​23
Mughal sultanate, 682 Muslim brotherhoods (Senegal), 11, 295–​307
al-​Muhajiroun, 436 Muslim League (Pakistan), 124, 170–​7 1, 272–​
Muhammad (Prophet Muhammad), 34, 36–​37, 73, 275–​76, 285–​86, 684–​85
627 Muslim League (PML-​N), 124–​25, 133–​34, 287
Muhammad Ali Muslim League (Q), 124–​25, 129, 287
Council of Consultation and, 663 Muttahida Majlis-​e-​Amal (MMA), 123, 125
Mamluk power reduced by, 660, 668
public schools established by, 645 N
public welfare system under, 753 Naguib, Mohamed, 752, 780
religious elites and, 661–​62, 668 Nahdlatul Ulama (NU), 82–​83, 181, 259–​60,
state centralization and government 794–​95
finances under, 19, 656, 660–​63, 667–​68 Naji, Abu Bakr, 858–​63
Muhammadiyah, 82–​83, 181, 260, 792, 794–​95, Nakba (Palestinian refugees from 1948 Israel
798 war), 756–​57
Murad Bey, 659 Namangani, Juma, 424–​25
Musharraf, Pervez Napoleon Bonaparte, 659
authoritarianism under, 126, 130, 134 Naqshbandi Order, 146
coup (1999) by, 124 Narayan, Jayprakash, 185
economic performance under, 128 Nasakom ideology, 180
elections under, 118, 123 Nasser, Gamal Abdel, 105, 602, 748, 752, 780
judiciary and, 129–​30 National Awakening Party (PKB), 259–​60, 263
Muslim League (Q) and, 129 National Forces Alliance (NFA), 224–​25,
ouster (2008) of, 125, 130, 132, 134 229–​30
religious political parties and, 435 National Mandate Party (PAN), 259–​60, 263
shrine families in Pakistan and, 684, 688 National Popular Movement, 831
Muslim Brotherhood (Egypt) National Rally of Independents (RNI), 838–​39
Arab Spring uprising and, 22, 597, 601–​3, National Trust Party (Amanah), 261–​62
606, 608, 611–​12 National Union of Popular Forces (UNFP),
coup (2013) against government led 488–​90
by, 106–​8, 110, 568–​69 Nawab Ismail Khan, 171
discipline and training within, 103, 217, 603 Nawaz Sharif, Mohammad, 124–​25, 129,
Freedom and Justice Party and, 213, 217, 222, 133–​34, 176
225, 229, 565, 570, 602 Nazimuddin, Khwaja, 172–​73
884   Index

Ndao, El Hadj Ibrahima, 303–​4 Ottoman-​Mamluk War (1516-​1517) and, 657

Nehru, Jawaharlal, 184, 186, 272, 275, 277 property rights and, 631
Neo-​Destour Party, 751–​52, 771 Provincial Law of 1864 and, 57
The Netherlands, 254 religious legitimacy claims and, 626, 628, 632
Niasse, Moustapha, 306, 307 secular-​religious divisions in, 337
Niasse, Serigne Mamoune, 298, 305 state financing in, 722–​24
Niger, 700, 702–​3, 706, 808, 816 tax registers from, 642, 650
Nigeria ulema in, 58–​61, 659
Biafra War (1967-​1970) in, 545 Öz, Zekariya, 157
British colonial rule in, 545
Christian missionaries and, 710 P
Christian-​Muslim mistrust in areas Pahlavi Dynasty (Iran), 772–​73
of, 542–​43, 551–​55 Pakatan Harapan, 262
Christian-​Muslim segregation in areas of, Pakistan
543–​44, 548–​50, 555 Ahmadis in, 120–​22, 135, 172–​73, 176, 273,
Christian-​Muslim violence in areas of, 539, 286–​87
541–​43, 545–​48, 555 anti-​colonial movement and nationalism
Islamic rule in northern regions of, 710–​11 in, 168–​7 1, 187
Kafanchan riot (1987) in, 545–​46 authoritarianism in, 117, 126–​34, 170
Middle Belt region in, 540–​41, 545–​46 Balochistan in, 121, 123–​24, 132, 273, 288, 433
Plateau State in, 546–​47 Bangladesh’s independence from, 133,
post-​traumatic stress disorder levels in 174–​75
conflict zones in, 540–​41, 544–​45, 551, 554 blasphemy laws in, 120, 122, 135, 175–​76, 288,
religiosity in, 438–​39 430, 442–​43
shariʿa law, 440, 547–​48 British colonial period in, 170, 677–​80
Zamfara state in, 547 civil rights and religious freedoms in, 115–​
Nour Party, 213, 223–​25, 229–​30 16, 120–​22, 134–​36, 285–​86
Nubar Pasha, 665 constitutions in, 119, 121–​22, 134–​35
coup (1958) in, 115, 172–​73
O coup (1977) in, 12, 115, 175, 273, 287, 681–​82,
Öcalan, Abdullah, 157 691–​92
O’Donnell, Guillermo, 53, 121, 127–​28 coup (1999) in, 115, 124
Oman, 772 democracy and, 13–​14, 115–​19, 122, 125–​26,
Omar Ali Saifuddien III (sultan of Brunei), 90 131–​36, 167, 172–​73, 177, 430, 435–​36, 440–​41
al-​Omari Mosque (Daraa, Syria), 527 economic performance in, 127–​28
Operation Cast Lead (Gaza), 157 education in, 681
Ottoman Empire. See also Turkey Hudood Ordinances (1979) in, 175
Basic Law of 1876 in, 58 Inter-​Services Intelligence Directorate (ISI)
decline of, 56, 632–​33 in, 124, 287
economic elite in, 632–​33 Islam and national identity in, 11, 14, 22,
Egyptian viceroys (pashas) of, 657–​58, 660 119–​22, 127, 135–​36, 167, 169–​76, 187, 272,
First Constitutional Period in, 58 275, 434–​35
majalla (civil code) in, 57, 723–​24, 726 Islamic political parties in, 271–​73, 285–​88,
The Meşīḫat in, 60–​61, 63 290, 431
Ministry of Justice in, 60 Islamism in, 15, 22, 120, 125, 175–​76, 288, 434,
modern secular state schools in, 56–​57 441–​43, 797–​98
Niẓāmiye court system, 57–​58 judiciary in, 129–​30, 134
 Index   885

Lahore Declaration (1999) and, 124 Party of Authenticity and Modernity, 491
local landlords in, 286 Party of Justice and Development (PJD)
marital law (1973) in, 172–​73 Alternance government and, 489–​90
military as political actor in, 21, 115–​19, 126, Arab Spring uprisings and, 22, 491–​92, 597,
129, 132, 134–​35, 173, 176–​77, 204, 273, 286, 601, 604–​6, 609–​12
289–​90, 435, 681 cabinet posts held by members of, 110
Northwest Frontier Province (NWFP) in, economic platform of, 495
121, 123 origins of, 606
Objectives Resolution in, 119, 172, 175 parliamentary elections of 1997 and, 489
One Unit representative system in, 121, 126 parliamentary elections of 2011 and, 491–​92
parliamentary elections (1985) in, 681 Parti de l’Authenticité et Modernité and, 605
partition from India of, 184, 272–​73, 276, shariʿa law and, 601
430, 436, 683–​84 women members of parliament from, 828–​
Political Parties Act of 1962 in, 122, 681 29, 831, 834–​36, 838–​40, 846
public opinion regarding India in, 430, 436, 441 Party of Progress and Socialism (PPS), 839
religiosity in, 430, 433–​43 party systems in Muslim-​majority societies
religious violence in, 430 colonial legacies and, 239, 245–​46
Retribution and Compensation Ordinances definition of party systems and, 239–​40
in, 176 institutional explanations for divergences
shariʿa law, 134, 288, 433–​34, 437–​40, 442 in, 242–​46
Shiʿi Muslims in, 176, 287, 434, 436, 442–​43 Islamic culture and, 245
shrine families in, 12, 676–​93 nature of competition within, 240–​42
terrorism and, 414 socioeconomic explanations for
ulama in, 171, 174–​75, 177 divergences in, 243–​44
United States and, 132, 201, 273 Patel, Sardar, 186
urban protests in, 130–​32, 134–​35, 272–​73 People’s Conscience Party, 259–​60
Pakistan Movement for Justice (PTI), 125, 134, People’s Rally Party (RP), 298
287 Peul ethnic group, 296
Pakistan Muslim League. See Muslim League The Philippines, 90
(Pakistan) Popular Movement Party (MP), 831, 838
Pakistan People’s Party (PPP), 124, 133, 286–​87, Prabowo Subianto, 260
681 Prophet Muhammad. See Muhammad
Palestinians (Prophet Muhammad)
Hamas and, 797 Prosperous Justice Party (PKS), 259–​60, 263,
in Jordan, 755–​57, 774 790, 795, 798, 800
Nakba (1948 Israel war) and, 756–​57 Protestants, 10, 632, 643
Operation Cast Lead (Gaza) and, 157 Punjab (Pakistan)
public opinion in Muslim-​majority Auqaf Department in, 679, 689
countries regarding, 490, 495 electoral politics in, 689
religiosity among, 438–​39 Islam and national identity in, 171–​73
shariʿa law and, 440 land inequality in, 683
Pancasila ideology (Indonesia), 82–​84, 179–​81, 187 Muslim League and, 170, 275
Pan-​Malayan Islamic Party (PAS), 87, 261–​62, partition of Indian subcontinent (1947) and,
264, 726, 795–​96 276, 683–​84
parallel cousin marriage, 33, 38, 43–​44 Sufi shrines in, 676–​80, 682–​83, 689, 693
Parti de l’Authenticité et Modernité (PAM), 605 zamindari contracts (landlord contracts)
Party for Truth and Development (PVD), 298 in, 683
886   Index

Q political participation levels and, 402–​5

Qaddafi, Muammar, 214, 217 populism and, 346–​51
Qaddafi, Saif, 220 shariʿa law and, 433–​35, 444
al-​Qardawi, Yusef, 564 voting behavior and, 253, 266
Qasr Al-​Aini Medical School, 753 Republican People’s Party (CHP), 337–​38
Qatar, 467, 769 Responsible Finance Initiative, 730
al-​Qazdaghl, Ibrahim Kahya, 658 Revolutionary Command Council (RCC), 780
al-​Qaʿida, 422, 441, 532, 858, 860 Riyad Pasha, 666
al-​Qaʿida in the Islamic Maghreb (AQIM), 816 Roman Empire, 622, 628
Quadriyya brotherhood, 296 Roy, Émile, 814
al-​Qurashi, Abu Ubayd, 858 Roy, Olivier, 4
Quraysh tribe, 37 Russia, 412, 417, 855, 857
Rwanda, 700–​3, 706–​8
R
Rabaa Massacre (Egypt, 2013), 107–​9, 569, S
571–​72 Ṣabāḥaddīn (prince of Ottoman Empire), 59
Rahman, Fazlur, 116, 124, 126, 132, 174 Sachar Committee, 278
Rahmon, Emomali, 419, 423 el-​Sadat, Anwar, 602, 793
Rashidun Caliphate, 626–​28 Ṣafvet Efendi, Sheikh, 65
Rashtriya Swayamsevak Sangh (RSS), 185, Said, Khaled, 101, 566–​67
275–​76, 281 Salafism
Rassd News Network, 566, 569 Ansar al-​Sharia and, 23, 581–​82, 586–​89,
Razak, Najib, 262 591–​93
Refah Partisi (Welfare Party, RP), 146, 151, Brunei and, 92
338–​39, 827 social media and, 573
religion in Syria, 532–​33
definition of, 4–​8 women and, 501, 503, 505, 509
doctrine and, 4–​8, 24 Salimov, Yaqub, 424
institutional aspects of, 8–​10 Samajwadi Party, 277–​78, 285
multidimensional nature of, 6–​8 Sanginov, Habib, 424
as social identity, 7, 12–​17, 24 Sarekat Islam (Islamic Union), 178
as social organization, 8–​12, 24 Saudi Arabia, 201, 434
religiosity Scheduled Castes, 277–​79
democracy and, 13–​14, 167, 336, 345–​46, 430, Schmitter, Philippe, 121, 128
434–​36, 439–​41, 444 Schulz, Dorothea, 811
devotion and, 14, 430, 433, 436–​37 Seck, Idrissa, 305
dogmatism and, 14, 430, 433, 436–​37 Sefarty, Abraham, 488, 490
Indonesia and, 438–​39 Selamet Hareketi (Salvation Movement), 146
Islamism and, 433, 441–​44 Seljuk Dynasty, 40–​42
jihad and, 436 Senegal
Justice and Development Party and, 335, Catholics in, 306–​7
341–​51 cross-​sectarian political partnerships in,
measures of, 341 295–​96, 298–​99, 306–​8
Pakistan and, 430, 433–​43 democracy in, 11, 204
piety and, 14, 429–​30, 433, 435–​37 France and, 297, 299, 304
political mobilization and, 371, 376, 379, 389, Grand Magal pilgrimage in, 300
432 Groundnut Basin in, 297, 301
 Index   887

laicism in, 808 British colonial government and, 677–​78

military as political actor in, 204–​5 central government’s relations with, 677–​78,
religiosity in, 437–​39 684
religious composition of population in, coup of 1977 and, 681, 691–​92
296–​97 economic development and, 677–​78, 693
religious political parties in, 295 education and literacy suppressed by, 12,
shariʿa law and, 440 677–​78, 681–​82, 691–​93
Sufism and Muslim brotherhoods in, 11, electoral politics and, 679–​80, 683–​90
295–​307 hereditary succession and, 678, 691
trust in traditional authorities in, 809 marriage alliances and, 690
Senegalese Democratic Party (PDS), 300, 307 Muslim League and, 684–​85
Şener, Abdüllatif, 147, 150 Shrine of Imam Hussein (Karbala, Iraq), 9,
Senghor, Léopold Sédar, 295, 299, 303–​4, 306 355, 358
Serer ethnic group, 296 shura (advisory committee), 245
Shaaban, Bouthaina, 529 Sierra Leone, 700, 702–​3, 706–​8
al-​Shabiba al-​Islamiyya (Islamic Youth), 110t, Sikhs, 274, 276
489, 493–​94 Singapore, 131, 725
Shah, Amit, 283–​84, 289 Singh Ballia, Surendra, 283
shariʿa law. See also Islamic law Sinha, Jayant, 283
Brunei and, 92 al-​Sisi, Abdel Fattah, 101–​2, 108, 575–​76, 782
Egypt and, 610–​11 Six-​Day War (1967), 757
Indonesia, 440 Snow, Edward, 856
Islamic banks and finance and, 9, 719, 722, Socialist Party of Senegal (PS), 300, 307
724, 726–​29 Socialist Union of Popular Forces (USFP),
Jordan and, 440 484, 488–​91, 495, 831, 838–​39
Malaysia and, 85–​86, 88–​89, 440, 725–​28 social media
Morocco and, 610–​11 authoritarian regimes’ use of, 575–​77
Nigeria and, 440, 547–​48 Egypt uprising (2011) and, 101, 561, 566–​67
Niẓāmiye court system in Ottoman Empire Facebook and, 101, 561, 566–​73, 575–​76, 608
and, 57–​58 Islamists’ use of, 561–​62, 565–​74, 577
Pakistan and, 134, 288, 433–​34, 437–​40, 442 polarization and extremism facilitated by,
al-​Qa’ida, 859 573–​75, 577
religiosity and attitudes toward, 433–​35, 444 resource mobilization theory and, 561
Turkey and, 63–​64, 66, 344–​45, 440, 610 Salafism and, 573
Shas Party, 255 Twitter and, 567–​7 1, 574
Shastri, Lal Bahadur, 184 YouTube and, 567, 569, 571–​72
Shiʿi Muslims Society of Muhammadan, 59
Alawites and, 522–​26, 528–​32 Soroush, Abdolkarim, 116
in Iran, 355–​66 Stepan, Alfred, 77–​78
in Iraq, 9, 355–​66 Sub-​Saharan Africa. See also specific countries
Islamic State and, 455, 469 child mortality in, 697, 701–​2, 704, 706,
in Pakistan, 176, 287, 434, 436, 442–​43 708–​9, 711
pilgrimage to Imam Hussein shrine by, 9, economic development in, 697–​7 15
355, 358 education in, 698, 701–​7, 710–​11, 713–​15
religious practices among, 355–​66 European colonialism and Christian
in Sub-​Saharan Africa, 701 missionaries in, 700, 710, 715
shrine elites (Pakistan) fertility rates in, 698, 701
888   Index

Sub-​Saharan Africa (cont.) Salafism in, 532–​33

gender inequality in, 698, 705–​6, 712 sectarian political appeals in, 21, 521–​34
Islam’s arrival in, 699–​700 shabiha (armed civilian groups) in, 530, 533
Muslim population growth in, 697, Sunnis in, 529–​32
699–​700, 715 war (2011-​) in, 21, 521–​22, 531–​34
Muslims’ socioeconomic status in, 697–​99, welfare service provision in, 771–​73
715
religiosity levels in, 701 T
religious conflict in, 701 Tablighi Jamaat, 288
Shiʿi Muslims in, 701 Tahrir Square protests (Egypt, 2011), 561, 602
Sudan, 541, 710, 730f, 732f, 733f, 769 Tajikistan
Suez Canal, 650, 663 civil war in, 423, 425
Suez War, 780 illicit economy in, 419–​21, 423–​25
Sufism state regulation of religious practices in, 375
economic development and, 675 terrorism in, 411, 414, 420, 423–​24
marabouts (religious leaders) in, 296–​309, The Taliban, 79t, 425, 434, 441
485 Tanzania, 700–​3, 706–​8
mysticism and, 643 Taseer, Salman, 176, 288
Naqshbandi Order and, 146 Tehreek-​e-​Taliban Pakistan (TTP), 434, 441
in Senegal, 11, 295–​307 Tehrik-​Nifaz-​i-​Fiqh-​i-​Jaafria (TNFJ), 176
shrines in Pakistan and, 676–​77 terrorism
Tijaniyya order and, 810 Central Asia and, 21, 411–​16, 420–​25
in Turkey, 66 civil wars and, 200
Suharto, 81–​83, 180–​81, 205, 260, 795 definition of, 412
Sukarno, 81–​83, 178–​80, 259–​60 in democratic regimes, 199
sukuk (Islamic bonds), 727–​29 illicit economies and, 412, 416–​25
Suleiman (sultan of Ottoman Empire), 722 individual-​level motivations for
Sultan of Johor, 724 committing, 413
al-​Suri, Abu Musab, 858 Islamic State and, 861
Sy, Ababacar, 304 Islamism and, 199–​200
Syria meso-​level motivations for committing,
Alawites in, 522–​26, 528–​32 413–​14
Arab-​Israeli wars and, 773 securitization and, 411
Arab Spring uprising (2011) in, 10, 197, socioeconomic factors and, 414–​16
524–​31, 534 topographic factors and, 414–​16
authoritarianism in, 522–​24 Tijan brotherhood, 296, 299, 306–​7
Daraa protest (2011) in, 527 Tijaniyya order, 810
education in, 773 Timbuktu (Mali), 815–​16
France and, 433–​35, 444 Touba (Senegal), 302
Hama uprising and massacre (1982) Traoré, Moussa, 809–​10
in, 109–​10, 112, 522–​23 Trinamool Congress Party, 284–​85
healthcare in, 774–​75 Tunisia
Iran and, 524, 530 Arab Spring uprising in, 22, 197, 204, 391,
Islamic State in, 450, 466, 470, 532, 853, 861 581, 585–​89, 597–​98, 600–​1, 603, 606–​12
Islamism in, 104, 109–​11, 533–​34 authoritarianism in, 214, 494, 583–​84
pension system in, 776 coup (1987) in, 583
religious demography in, 522 democracy in, 204, 221–​22, 245, 581, 589
 Index   889

education in, 748–​52, 773–​74 military as political actor in, 53, 151, 156, 338,
elections (1989) in, 607, 612 340
elections (2011) in, 612–​13 Ministry of Islamic Affairs and Religious
French colonial era in, 56, 650, 751–​52 Endowments in, 64–​66
healthcare in, 748–​51, 774–​75 Ministry of Justice in, 63–​64
hijab restrictions in, 335 National Intelligence Agency in, 157
Islamic State in, 470 neoliberal economics in, 151
Islamism in, 16, 22, 109–​11, 213–​22, 225–​31, Niẓāmiye court system in, 63–​64
583, 585, 597–​98, 601, 606–​10 populism in, 346–​51
laicism in, 214, 217, 231 Presidency of Religious Affairs (Diyanet)
military as political actor, 204–​5 in, 61, 64–​69
pension system in, 776 pro-​government businesses in, 152–​54, 159,
postcolonial era cabinets in, 745–​46 265
secularism in, 589, 613 religiosity in, 437–​39
Sousse terrorist attack (2015) in, 470 Savings Deposit Insurance Fund (TMSF)
US Consulate protests (2012) in, 591 in, 152–​53
welfare service provision in, 740, 743–​44, schools in, 61–​63
748–​52, 758, 769–​7 1, 783 Second Constitutional Period in, 62–​63
women in, 827–​28, 843, 846 secularism and, 53–​55, 61–​63, 67, 69, 82, 337
Tunisian General Labour Union (UGTT), 221, shari’a law and, 63–​64, 66, 344–​45, 440, 610
584–​85, 606, 608, 612, 771 Sheikh Saʿīd Rebellion (1925) in, 66
Turkey. See also Ottoman Empire state formation in, 54–​56, 61–​69
attempted coup (2016) in, 69–​70, 156, Sufism and, 66
158–​60 ulama in, 61–​68, 78
authoritarianism in, 53–​54, 130, Unification of Education Law (1924) in,
143–​60 62–​63
Balyoz trial in, 156–​58, 160 women in, 827
coup (1980) in, 69, 338 Turkmenistan, 411
democracy and, 13–​14, 53–​54, 69–​70, 156, Twitter, 567–​7 1, 574
159, 256, 346–​51
economic crisis (2001) in, 147 U
education in, 827 Uganda, 540, 700, 702–​3, 706–​8
Ergenekon trial in, 69, 155–​58 ulama (clergy)
European Union and, 144, 155–​56, 215–​16, in Brunei, 89, 92
339 economic development in Muslim-​majority
Gezi Park protests (2013) in, 399–​401, 406 societies and, 11–​12
Grand National Assembly in, 65 fatwa (certified opinion) and, 79
hijab wearing in, 335, 340 in Indonesia, 78, 81, 85
Housing Development Administration in Malaysia, 86
(TOKİ) in, 152–​54 in Mali, 807–​12, 819
Imam Hatip schools in, 63, 339–​40 in the Ottoman Empire, 58–​61, 659
Islamism in, 53–​55, 67–​70, 146–​50, 215–​16, Pakistan and, 171, 174–​75, 177
265, 338–​40, 827 social status of, 10
Israel and, 157 in Syria, 523, 528–​29
judiciary in, 144, 147, 156–​58, 160, 339–​40 in Turkey, 61–​68, 78
Kurdish population in, 150–​51, 157, 338 Umayyad Caliphate, 37, 626–​28
Land Office Law in, 153 Unified Socialist Party (GSU), 838–​39
890   Index

Unionist Party, 275 colonial legacies and, 739, 742–​48, 751–​59

United Arab Emirates, 773 colonial regimes and, 740–​42
United Development Party (PPP), 180, defense spending and, 773, 777–​83
259–​60, 263 in Egypt, 710, 743–​44, 747–​55, 758–​59, 771,
United Malays National Organization 773–​75, 780–​82, 784
(UMNO), 86–​87, 261–​62, 264, 726, elite coalitions and, 741, 744–​48, 754–​56,
795–​96 758–​59, 771–​72
United Nations Relief and Works Agency in Iran, 769, 772, 784
(UNRWA), 750, 755–​57, 774 in Jordan, 740, 743–​44, 747–​51, 755–​57, 759,
United States 771–​72, 780
Egypt and, 201 in Morocco, 743–​44, 747–​51, 757–​59, 771–​72
foreign military aid to authoritarian oil states and, 772
regimes from, 201–​3 postcolonial expansions of, 740, 747–​48
Indonesia and, 180, 201 private versus public institutions in, 740,
Iran and, 201 742–​44, 747, 751, 759
Pakistan and, 132, 201, 273 in Tunisia, 740, 743–​44, 748–​52, 758, 769–​7 1,
Saudi Arabia and, 201 783
United Tajik Opposition (UTO), 423–​25 Wiranto, 205, 260
UPS Party (Senegal), 299, 302, 304 Wolof ethnic group, 296–​97, 301
Uttar Pradesh (India), 275, 277, 280, 282–​83, women
285, 693 in Algeria, 832
Uzbekistan as authorities in Islam, 501–​4, 517
Andijan uprising (2005) in, 423 in Egypt, 827
illicit economy in, 421–​23 Islamic State and, 864
security apparatus in, 422 Islamist political regimes and, 22–​23, 78,
state regulation of religious practices in, 827–​46
375–​76 in Jordan, 846
terrorism in, 411, 421–​23, 425 in Libya, 832
as members of parliament, 827–​32, 834–​46
V in Morocco, 828–​32, 834–​46
Vaḥīdeddīn (Ottoman sultan), 65 online preaching by, 12, 501–​16
Vajpayee, Atal Bihari, 124, 281 as political candidates, 827
Venezuela, 130 political participation levels among, 397–​98,
Verma, Rahul, 280, 282t 401
Vietnam, 856–​57 Salafism and, 501, 503, 505, 509
Virtue Party (FP), 339 in Tunisia, 827–​28, 843, 846
in Turkey, 827
W
Wade, Abdoulaye, 300–​2, 304–​6, 307 Y
Al Wafd Party, 222–​26, 229–​30 Yahya Khan, 126, 128, 131–​33
Wahhabism, 92, 434 Yakış, Yaşar, 150
Wahid, Abdurrahman, 84, 774 Yarbay, Ersönmez, 149–​50
Weber, Max, 623–​24, 643, 713 Yassine, Abdesssalam, 110t, 484, 490–​91
Welfare Party (RP), 146, 151, 338–​39, 827 Yemen
welfare service provision. See also education; Arab Spring uprising (2011) and, 198
healthcare clientelism in, 320–​24
“cheap social policies” and, 781–​84 elections and political competition in, 15
 Index   891

Islamic State in, 450, 466 Zayd movement (Syria), 794

Islamism in, 321 Zia-​ul-​Haq, Mohammad
Yiğit, Korkmaz, 153 authoritarianism under, 126
Yoruba ethnic group, 545, 550 cabinets of, 123
Young Turk movement, 58, 63–​64 coup (1977) by, 12, 175, 273, 287, 677, 681, 691
al-​Youssoufi, Abderrahman, 488–​89 death of, 123, 132–​33, 176
YouTube, 567, 569, 571–​72 economic performance under, 128
Yudhoyono, Susilo Bambang, 84, 260 elections under, 118, 122–​23, 681, 683–​86
Islamism and, 175–​76, 272–​73, 287, 434, 684,
Z 688
Zafarullah Khan, 286 judiciary and, 129, 273
Zardari, Asif Ali, 176 shrine families in Pakistan and, 684–​89, 692
al-​Zarqawi, Abu Musab, 861 Ziyoev, Mirzo, 424–​25