JOBNAME: John PAGE: 1 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

How Far to Nudge?

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 1 / Date: 14/12
JOBNAME: John PAGE: 2 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

NEW HORIZONS IN PUBLIC POLICY

Series Editor: Wayne Parsons, Professor of Public Policy, Wales Governance Centre,
Cardiff University, UK
This series aims to explore the major issues facing academics and practitioners working
in the field of public policy at the dawn of a new millennium. It seeks to reflect on
where public policy has been, in both theoretical and practical terms, and to prompt
debate on where it is going. The series emphasizes the need to understand public policy
in the context of international developments and global change. New Horizons in
Public Policy publishes the latest research on the study of the policymaking process
and public management, and presents original and critical thinking on the policy issues
and problems facing modern and post-modern societies.
Titles in the series include:
Modernizing Civil Services
Edited by Tony Butcher and Andrew Massey
Public Policy and the New European Agendas
Edited by Fergus Carr and Andrew Massey
The Dynamics of Public Policy
Theory and Evidence
Adrian Kay
Ethics and Integrity of Governance
Perspectives Across Frontiers
Edited by Leo W.J.C. Huberts, Jeroen Maesschalck and Carole L. Jurkiewicz
Public Management in the Postmodern Era
Challenges and Prospects
Edited by John Fenwick and Janice McMillan
The Tools of Policy Formulation
Actors, Capacities, Venues and Effects
Edited by Andrew J. Jordan and John R. Turnpenny
Analysis and Public Policy
Successes, Failures and Directions for Reform
Stuart Shapiro
Public Policy Transfer
Micro-Dynamics and Macro-Effects
Edited by Magdaléna Hadjiisky, Leslie A. Pal and Christopher Walker
Policy Experiments, Failures and Innovations
Beyond Accession in Central and Eastern Europe
Edited by Agnes Batory, Andrew Cartwright and Diane Stone
How Far to Nudge?
Assessing Behavioural Public Policy
Peter John

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 2 / Date: 14/12
JOBNAME: John PAGE: 3 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

How Far to Nudge?

Assessing Behavioural Public Policy

Peter John
Department of Political Economy, King’s College London, UK

NEW HORIZONS IN PUBLIC POLICY

Cheltenham, UK + Northampton, MA, USA

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 1 / Date: 14/12
JOBNAME: John PAGE: 4 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

© Peter John 2018

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, electronic,
mechanical or photocopying, recording, or otherwise without the prior
permission of the publisher.

Published by
Edward Elgar Publishing Limited
The Lypiatts
15 Lansdown Road
Cheltenham
Glos GL50 2JA
UK

Edward Elgar Publishing, Inc.

William Pratt House
9 Dewey Court
Northampton
Massachusetts 01060
USA

A catalogue record for this book

is available from the British Library

Library of Congress Control Number: 2017953165

This book is available electronically in the

Social and Political Science subject collection
DOI 10.4337/9781786430557

ISBN 978 1 78643 054 0 (cased)

ISBN 978 1 78643 056 4 (paperback)
ISBN 978 1 78643 055 7 (eBook)

Typeset by Columns Design XML Ltd, Reading

02

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 2 / Date: 14/12
JOBNAME: John PAGE: 5 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

For Mike

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 3 / Date: 14/12
JOBNAME: John PAGE: 6 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 4 / Date: 14/12
JOBNAME: John PAGE: 7 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Contents
Preface viii
Acknowledgements x

1 Introduction 1
2 Behavioural public problems 21
3 The behavioural revolution in the social sciences 38
4 Nudge: All tools are informational now 55
5 Translating nudge into practice: Routes to innovation 69
6 Is nudge all it’s cracked up to be? Limitations and criticisms 88
7 The ethics of nudge 108
8 Nudge plus and how to get there 122
9 Assessing behavioural public policy 142

References 147
Index 167

vii

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 1 / Date: 14/12
JOBNAME: John PAGE: 8 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

Preface
Like many people, I drifted into nudge. In the 2000s, I was working on
measures to encourage citizens to participate in politics and civic life. As
someone interested in how to improve public policy outcomes, I had
become intrigued about how to encourage citizens to do a wider range of
acts for collective benefit. I wanted to motivate them with messages and
acts of persuasion to overcome constraints on collective action (see John
et al. 2011). In short, I discovered myself doing behavioural public policy
at the same time as Thaler and Sunstein’s (2008) book on nudge came
out. I found the ideas and language of the behavioural sciences very
helpful as I developed a research programme, especially as I was using
randomised controlled trials, which is the method of choice for testing
behavioural interventions.
I had of course long been aware of the strides of behavioural
economics. I recall going to hear Danny Kahneman give a keynote
lecture to the American Political Science Association in 1999. I had also
worked with David Halpern on a project on social capital at the end of
the 1990s. Anybody who talks to David cannot fail to be aware of his
wide interest in the interactions between government policy and policy
outcomes, which first manifested itself with his advocacy of the social
capital agenda and then of behavioural insights. I kept in touch with
David as he developed the behaviour change agenda in central govern-
ment, and he has inspired me.
I also discovered that people working in public agencies found the
skills I had developed in designing trials to be very useful in redesigning
their policies. Some of these interventions were carried out with the
Behavioural Insights Team (BIT) in its early days; latterly I worked with
local authorities and other public agencies on their own behavioural
insights. I became curious about the growing official interest in testing
behavioural ideas with experiments, and I wondered how a culture of
experimentation and behavioural redesign could be integrated into the
standard operating procedures of public bureaucracies. I also became
aware that I was one of the very few political scientists working on
behavioural public policy. I had all the skills of the nudger and
experimenter, but I saw the interventions much more in their political

viii

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 1 / Date: 14/12
JOBNAME: John PAGE: 9 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

Preface ix

context. As a result, I wanted to understand how behavioural public

policy could be useful to those who practised it. I was curious to find out
why politicians are attracted to nudge and what bureaucrats value in the
research and policy agenda. In short, I found that I had developed an
interest in the politics of behavioural public policy and the factors that
affected its diffusion and implementation. I believe that behaviour change
reforms can only work effectively when considered as part of a wider
system of institutions and interests, which involves understanding the
mechanisms of accountability, citizen reactions, and private advantage.
Nudging needs to operate within the current set of political constraints
and opportunities, and these choices are best understood within a
framework familiar to those who study political science and public
policy. I developed many of these ideas while designing trials for
behavioural public policy interventions and in thinking through their
implications as I dealt day-to-day with bureaucrats and politicians. Doing
behavioural public policy has inspired this book.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 2 / Date: 14/12
JOBNAME: John PAGE: 10 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

Acknowledgements
I have many people to thank. The first is Alex Pettifer at Edward Elgar.
Like a good publisher, he knows the value of the long game. Alex and I
had been meeting many times, long before the idea for the book was
hatched. Once I had the project in mind, he helped me develop it, and I
am very pleased this book is coming out under Edward Elgar’s imprint. I
am also very grateful to others at the press for helping me through the
preparation of the manuscript, in particular Helen Moss and Kaitlin Gray.
I am in debt to many academic colleagues with whom I have
developed behavioural interventions, in particular my co-researchers on
the Rediscovering the Civic project – Sarah Cotterill, Hanhua Liu, Alice
Moseley, Hisako Nomara, Liz Richardson, Graham Smith, Gerry Stoker,
and Corinne Wales – for which we developed tests of both nudge and
think (John et al. 2011). I also thank my friend Helen Margetts, with
whom I have worked on nudge-like interventions in the online sphere. In
the midst of writing our book on political turbulence (Margetts et al.
2016), I pitched her the idea of How Far to Nudge?, and she encouraged
me to do it.
I am particularly grateful to members of the Behavioural Insights Team
who welcomed me as a friend and advisor, especially in the early days.
My special thanks go to David Halpern, Laura Haynes, Michael Sanders,
and Owain Service. I am also grateful to my sister Ros who, as a
professor of developmental epigenetics, advised me on the discussion of
the role of epigenetics in affecting human behaviour in Chapter 2.
I appreciate the countless people with whom I have talked about
behavioural interventions. Because of the trendiness of the topic, I have
been invited to give many presentations to governments, third sector
bodies, and academic gatherings. Of course, talking about one’s ideas is
the same as developing them and helping them take shape. These
audience members listened patiently to my talks and asked very good
questions at the end. I also presented ideas about behavioural public
policy to the students who took my Making Policy Work course module,
based on the eponymous book (John 2011), which I taught when I was in
the School of Public Policy, University College London. Anyone reading
the book will find out that I move inexorably from the tools of

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 1 / Date: 14/12
JOBNAME: John PAGE: 11 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Acknowledgements xi

government to behaviour change, so the students knew instantly what

direction I was taking them in. I claimed that I was planting favourable
ideas about nudge and randomised controlled trials into the minds of the
students, but in truth they influenced me just as much.
I also thank the School of Governance at the University of Utrecht,
which invited me to spend a week there in November 2016, where I had
quiet space to work, and I gave a talk based on the book’s ideas. I thank
Thomas Schillemans, who invited me, as well as Lars Tummers and
Stephan Grimmelikhuijsen, who all made me feel very welcome. In the
middle of the week, I discussed the ideas in the book over a delightful
dinner with Joel Anderson, and he inspired me to develop them further.
I am very grateful to Oliver James, who reviewed the manuscript for
the publisher critically and sympathetically. I thank the research students
who work with me on behaviour change policies: Manu Savani, Anna-
belle Wittels, Eliza Kozman, and Patrick Taylor. We convened a mini-
seminar on my book on 20 March 2017, which was great fun as well as
offering me excellent feedback. Joel helped me once again with the
chapter on ethics (Chapter 7). Sarah Birch, Oliver Hauser, Sebastian
Jilke, and Gerry Stoker also made very useful comments on the manu-
script after its first draft. Needless to say, I have probably made many
mistakes by not listening enough to the great advice I have received.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 2 / Date: 9/1
JOBNAME: John PAGE: 12 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Prelims /Pg. Position: 3 / Date: 14/12
JOBNAME: John PAGE: 1 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

1. Introduction
We live in an age of large public policy problems that governments find
hard to solve. These include obesity, climate change, terrorism, race
discrimination, corruption, and youth unemployment, just to name a few
current policy challenges. The list could go on to include a large number
of other topics that range across all aspects of human experience. These
problems share their origin in human behavioural traits or habits that
persist over time. Behaviours create and sustain poor outcomes, both for
individual citizens and communities, mainly because effective policy
outcomes usually depend on some degree of citizen action and respon-
siveness. When citizens are engaged, motivated, and willing to change
their behaviours, it is much easier for governments to achieve their policy
objectives, partly because all citizens are joint participants in acts that
have collective benefits. When citizens are switched off, antagonistic to
governments, and focused on their short-term interests, public policy gets
much harder to implement and poor outcomes are the result.
Whether helping or worsening policy outcomes, behaviours are often
transmitted through peer groups and in close social networks. Behaviours
get embedded, as they are reinforced through habit and mutual depend-
ence. Even government policy has a role in sustaining behaviours that
might be negative for the individual and society in spite of good
intentions and many beneficial measures. In health, for example, the
policy of responsive medicine and fixing problems immediately can
undermine a more preventative approach based on promoting more
healthy behaviours, as people expect to attend hospitals. As a result, the
effort and resources of the public health services get concentrated on
interventionist medicine. Many structural inequalities in income and
capacity also influence human behaviour, which in turn reinforce existing
inequalities. As many policies sustain these disparities, poor behaviours
may continue over time, encouraged by governments as well as market
forces. The implication of these factors is a negative equilibrium trap
whereby collectively reinforced behaviours lead to poor outcomes. The
trap occurs because it is rarely in the short-term interests of people,
groups, economic enterprises, and even governments to change their
behaviours, partly because of the potential cost each participant faces

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 1 / Date: 9/1
JOBNAME: John PAGE: 2 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

2 How far to nudge?

from moving from the current state of affairs to one where all benefit.
The question that arises for policy-makers and social scientists is what
measures and approaches could be used to create a step-shift in policy
outcomes by reversing or changing embedded behaviours?
Governments or other public authorities are elected with a remit to try
to solve public problems. They have a number of tools and considerable
resources at their disposal to do so, whether taxes, laws, the capacity of
their bureaucracies, or by being at the centre of networks of experts and
sources of knowledge (Hood and Margetts 2007; John 2011). The
challenge is to be able to use these tools effectively and efficiently
change behaviours. The potential benefit of such official efforts is a
sustained improvement in outcomes for society and the advancement of
the common good. The deployment of human ingenuity and putting in
place the right combination of resources might shift long-entrenched
behaviours and lead to an upward movement in outcomes. Such a
transformation often operates in a virtuous circle of self-reinforcing
activities, such as a movement from a neighbourhood with high crime,
low employment, low trust, and high degrees of social dysfunction to one
where these attributes lessen over time and where positive actions
emulate each other to sustain local economies, increase access to
employment, and reduce anti-social activities.
Such entrenched behaviours, and corresponding opportunities, are
often thought to characterise only poor communities, which do not have
good opportunities for employment and are dependent on welfare. But
richer communities can fail too, even those that have economic advan-
tages and other opportunities but where not enough people contribute
enough to social outcomes or innovate in the local economy, or where
people become unwell from lack of exercise and unchecked affluent
lifestyles.
Governments have tended to focus on solutions that come from the
power of the state to change things, such as provide more units of public
activities to address public problems directly, through increased public
expenditure and employing greater numbers of public servants, such as
police officers, nurses, and soldiers. As societies got wealthier and more
settled they have been able to allocate resources and make regulations to
ameliorate public problems. These solutions often depended on influenc-
ing individual actions, such as patients attending hospital clinics or the
unemployed being ready for work; but changes to individual behaviour
were not core to these programmes, which were about using the
authoritative tools of the state to effect change.
In spite of the considerable successes of these initiatives, policy-
makers and experts have come to realise that the use of such powers and

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 2 / Date: 13/12
JOBNAME: John PAGE: 3 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Introduction 3

resources often does not fully address the problems under consideration,
mainly because they often cannot change the behavioural impulses and
habits that sustain the problems over time. In part, many of the harms
come from prosperity as people adopt more affluent lifestyles. In more
complex societies, there are more interdependencies, which means that
solving public problems entails tackling a range of human actions that
interlock. Greater social science knowledge about the causes of public
problems shows them to reflect these complexities and to depend on a
range of reinforcing behaviours and actions.
To an extent, governments have always been aware of the need to
change behaviour, whether it was asking people to volunteer, for example
to serve in the armed forces, or advising them to take precautions against
the spread of diseases, or ensuring they respond to new tax breaks or
increases. People or organisations receive incentives from government,
but to communicate these policies governments have often relied on
simple pleas, as in posters and media campaigns, and the more gradual
processes of education and information provision, as well as providing
clear benefits and costs for desirable and non-desirable public actions
(Becker 1968). In societies with high degrees of social deference and
respect of authority, it is possible for a fully paternalist approach to work
well. By communicating injunctions to change behaviour, backed up by
laws and incentives, governments can create the changes in outcomes that
are needed. It is possible to think of encouragements to eat healthily or to
volunteer or to report crime as obvious tasks of government, but they
depend on a respect for authority and expertise to work well. In today’s
society, such deference to government and experts has weakened; it has
been replaced by a belief in consumer sovereignty and the right to have
one’s own opinions as to the good life, which is the natural end point of
the liberal ideal. Expertise is often questioned and not trusted, which may
be a result of increasing polarisation of issues and attitudes. Commonly-
used measures based on incentives, command, or education may not
change behaviour in ways they once did or appeared to do. People even
resist commands because of the way in which they are put, as they do not
like being told what to do or how to behave. Then there is information
overload from multiple media sources, including social media.
If conventional methods of achieving change are not so efficacious, it
is not surprising that governments tend to act on the consequences of the
behaviours, putting right the effects rather than the causes of the
problems. They adopt policies that help them get elected, such as
providing more expenditure on hospitals or roads. It might be the case
that providing more policy outputs is the most feasible course of action
and the one that is politically the most possible. But many policy-makers

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 3 / Date: 13/12
JOBNAME: John PAGE: 4 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

4 How far to nudge?

want to make more of a difference. They know about limits to resources

and they wish to find whether new ideas and knowledge can work. There
is a common image of policy-makers operating according to a crude
version of the rational actor model, in seeking votes or expanding their
bureaus; but political action is more complex, and incorporates a public
service motivation to do good for society (Ritz et al. 2016), which is one
of the reasons people go into politics or public service in the first place.
With an awareness of the limitations of the tools of government and a
willingness to try new things out, the question becomes whether govern-
ments can address behaviour change head on with theories and
approaches that speak to the behaviours themselves and change them.

BEHAVIOUR CHANGE AND NUDGE

The point about measures to address behaviour change is that they need
to be based on theories about why the desirable behaviours are not
occurring and then interventions that address those behaviours and
change them. Better understandings of the drivers of behaviours and what
causes changes in individual actions have the potential to transform
public policies to ensure interventions and public decisions directly
address fundamental problems of society. A powerful and growing
movement in the social sciences has gathered pace in recent years with
its radical agenda and scientific backing. Over the last three decades or
so, the behavioural sciences and behavioural economics have incorpor-
ated insights from psychology about what motivates people to change
their behaviour, ideas that have taken root in more established research
programmes in economics and the policy sciences and have diffused to
subject disciplines, such as studies of health and education. There has
emerged an extensive research programme, which rigorously tests its
claims, increasingly done through field experiments. It complements
older traditions of research that use psychology to inform public policy,
such as in health psychology, with the difference that the behavioural
sciences are genuinely interdisciplinary and have a high level of public
visibility and acceptability, so that they can reach into all aspects of
public policy.
Out of studies of behavioural economics and behavioural sciences has
come the greater interest by public agencies in using these ideas. This
policy field is often called ‘nudge’ (Thaler and Sunstein 2008), which is
often thought to be the idea that low-cost small changes, attuned to
human psychology, can make large improvements to public policies,
largely from encouraging citizens to do things that they would upon

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 4 / Date: 9/1
JOBNAME: John PAGE: 5 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Introduction 5

reasonable reflection agree with. Policy-makers of all different types have

used theories and evidence on behaviour change in developing interven-
tions, often helped by nudge units, such as the UK’s Behavioural Insights
Team (BIT) and the White House’s Social and Behavioral Sciences
Team. As a proof of concept, they deploy randomised controlled trials
(Haynes et al. 2012), which randomise the message or nudge to an
intervention or a control or non-intervention group, comparing outcomes,
such as payment rates, between the groups. This procedure reaches a
high standard of evidence in attributing a causal relationship between the
nudge and the intervention (Gerber and Green 2012). It also creates
headline results that are easy to interpret as percentage point differences,
which can translate into benefits to the agency, such as increased revenue.
This combination between respectability in methods, easy-to-understand
headline results, and information for cost and benefit decisions has
helped assist the dissemination of behavioural insights, within and across
governments.
Armed with evidence from trials, agencies have been redesigning the
messages that they send, such as telling taxpayers who are behind paying
that nine out of ten others have already paid their bills, so conveying the
social norm (Hallsworth et al. 2017), or making text messages more
personal so people settle their court fines on time (Haynes et al. 2013).
Influenced by the high level of official interest, academics of all types
working in behavioural science now consider more policy-relevant appli-
cations of their research (Oliver 2013a; Shafir 2013), which also feeds
back into the design of policy interventions. A large research agenda has
been created on what is sometimes called behavioural public policy
(Oliver 2017).
The nudge agenda has been assisted by skilful advocacy by academics
and entrepreneurs, so its terms and understandings have entered the
mainstream of debate and public policy. Such acts of dissemination have
created the world of nudge and applied behavioural insights. Advocates,
such as academics and heads of nudge units, have presented behavioural
insights in a way that is pragmatic, apparently based on common sense,
closely linked to the concerns of bureaucracies and responsive to their
demands, in particular by addressing cost pressures. Partly as a result of
this translation and diffusion, and the essential pragmatism of the
enterprise, there has been an extraordinary success story in the use of
behavioural insights. The result is that governments and other public
agencies from a large number of jurisdictions have adopted a set of
techniques that allow them to address key problems in public policy and
improve the efficiency of public administration.
There are two dimensions of policies or procedures that need to be
adjusted by the policy-maker: one is the communication flow to the

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 5 / Date: 9/1
JOBNAME: John PAGE: 6 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

6 How far to nudge?

citizen, which is based on a plausible psychological mechanism to

change behaviour, such as peer pressure, norms, or availability; the other,
in tandem with the first dimension, is an institutional change, an
alteration of standard operating procedures, what Thaler and Sunstein
(2008) call ‘choice architectures’, which are the rules and routines of
public agencies that intersect and structure the choices of citizens and
other actors, such as by offering defaults, opportunities, or new ways of
doing business. The first dimension is what most people associate with
nudge, for example the messages sent by BIT mentioned above. But it is
the changes to institutional design that are core to behavioural public
policy and are the sources of its agenda for a change in the delivery of
public services. Otherwise nudge just becomes another version of what is
called ‘social marketing’ (see discussion by Hallsworth and Sanders
2016, and in Chapter 3). In other words, to be more than the science of
messaging, nudge involves a reform of political institutions and bureau-
cracies. As a consequence, nudge advocates propose changes in how
public organisations operate internally as well as about how well they
communicate with citizens. Partly as a result, the agenda of nudge
integrates closely with core questions in political science and the study of
public administration, in that it seeks to find out in what ways insti-
tutional design affects communication flows between citizens and gov-
ernment, so that citizens cooperate with each other more effectively and
achieve collective action, that is so they see it as in their interests to
contribute to public goods (Ostrom 1990).
Public goods are the positive externalities associated with pro-social
behaviour change, that is benefits that have collective advantage, but
which are not in the short-term interests of people to contribute to, partly
because they fear bearing the cost, or they think there is no point doing
small actions that are unlikely on their own to make a difference. Policies
designed with behavioural insights might be seen to cumulate to facilitate
collective action, and to promote greater trust in others and in govern-
ment. Such policies may convey signals that ask people to contribute to
public goods, either based on people’s understanding the wider social
purpose or with the policies operating subliminally without recipients
being aware of the social benefits. In this way, behavioural public
policies operate within the moral economy of private actions, whereby
people can be encouraged to act in ways that are pro-social and overcome
their sense of self-interest (Bowles 2016). Individuals are not entirely the
rational non-co-operators described in basic game theory, such as the
players in a prisoner’s dilemma. But there still remains the problem of
conveying an effective signal in a way that generates collective action,

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 6 / Date: 9/1
JOBNAME: John PAGE: 7 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Introduction 7

and encourages individuals to perceive societal and their own incentives

as working together, which is not easy to achieve.

THE TOOLS OF GOVERNMENT

Nudge can address more fundamental aspects of behaviour by working
with the existing tools of government. By reforming how the tools of
government work, such as with taxes or regulations, behavioural public
policy can improve their traction so that citizens willingly change their
behaviour for the public good. In that sense, the ambit of behavioural
public policy is much wider than the nudge stereotype indicates and
includes applying behavioural insights to all actions of government rather
than just to information flows between citizens and the state. For
example, a tax change, which appears to be as far from a nudge as can be
imagined, can change behaviour with no reference to behavioural
insights, but it will be all the more effective if designed with them in
mind (e.g. Chetty et al. 2009; Chetty and Saez 2013). Legal reforms need
to be introduced with awareness of the informational environment if they
are to be effective, even if the use of the behavioural sciences involves
some careful thinking about the purpose of law and its relation to
democratic processes (Alemanno and Sibony 2015). In these ways, the
conventional tools of government can be refashioned using the ideas of
behavioural science. The claim is that all tools of government are
informational at root and current research can refine and improve their
effectiveness (John 2013a).
The wider applicability of nudge to the tools of government has
prompted some to recommend that government should use behavioural
science to force citizens to move away from their existing behaviours
rather than nudge them (Marteau et al. 2011). Behavioural ideas can
more effectively compel people to change their behaviours than nudge,
which is seen to be more of an encomium. It involves using the power of
the state to sanction changes when they are needed rather than rely on
situations where citizens can opt out or refuse if they want. In spite of
wanting to go beyond the more limited range of messaging policies,
compulsion is not the approach of this book. It would be counter-
productive to develop a super-science of public administration so the
tools of government become directive in character, even though some of
these efforts can be useful in the right context. In the end, top-down
behavioural policies, even when refined by testing, will fail because of
the entrenched ideas and embedded behaviours of citizens. The result

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 7 / Date: 9/1
JOBNAME: John PAGE: 8 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

8 How far to nudge?

may be some short-term changes in behaviour, but not the uplift that
policy-makers and voters are looking for. Citizens increasingly resist
such commands and find ways to avoid regulations, which is the result of
the loss of authority of government and other expert voices, and the
spread of democratic ideas of self-ownership. Command and control
often just do not work on their own as strategies. It is better to work
within sets of human expectations and relationships, along the grain of
habits and values of both policy-makers and citizens, so that both connect
to each other in better ways, which involves respecting the autonomy of
each in the process. In that sense, no one wants nudge policies to be
effective in getting the police to arrest suspects more quickly, or in
inculcating fear among taxpayers; instead, the nudges of Thaler and
Sunstein appeal more to people’s good nature to get them to pay taxes on
time or report crimes and so on. It can be finely balanced between using
nudges to get better tool design and using psychological insights for
commands and controls. A middle approach is advocated by Oliver
(2013b), who argues for budges, which indicate a wider range of policy
options than nudges. Overall, the desired balance of behavioural insights
is to have as many responsive nudges as possible and choose ones that
build on democratically agreed policies and procedures rather than rely
on top-down controls. With these considerations in mind, policy-makers
are in a good position to seize the opportunity to develop more
decentralised and reflective nudges.
This focus on certain kinds of intervention has the advantage of not
extending the definition of nudge to every action of government that uses
a psychological insight in some way, which could be construed to be
anything that government does. It avoids concept stretching, which
Sartori (1970) complained about in studies in political science. Nudges
remain as light-touch human-centric policies and prompts. But opportun-
ities to find them exist right across the spectrum of public activities.

NUDGE AND THE IMPLEMENTATION OF PUBLIC

POLICIES
As well as operating across the range of tools and instruments, nudging
can operate within the implementation chain of policies. A successful
policy relies on a chain of interactions within bureaucracies as well as
feedback and interdependence between citizens and public organisations.
A behavioural approach can examine each link in the chain, find out
about the behavioural cues needed to improve the transmission of
commands and responsiveness to context, and then improve the delivery

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 8 / Date: 9/1
JOBNAME: John PAGE: 9 SESS: 6 OUTPUT: Fri Jan 26 10:51:11 2018

Introduction 9

of the policy. It is hard for a behavioural intervention on one element of

the implementation process to work if other parts of the delivery chain
are not fully functioning. Reforming more than one link in the chain,
making the improvements in delivery, can fan into the whole adminis-
trative process. Positive feedback can improve the working of all the
sequences in the implementation process and generate better policy
outcomes, as all actors benefit from the resulting synergy. This strategy
also requires repeated interventions to tweak the implementation process.
This opportunity for using behavioural public policy addresses con-
cerns long expressed by theorists of implementation about the likelihood
of policy failure and the importance of improving the causal links in the
implementation process by better theories about what works (Pressman
and Wildavsky 1973). A key challenge is how a sponsor agency can
ensure a delivery agency implements a policy faithfully. Behavioural
insights inform work on implementation by moving beyond seeing policy
outputs just as a summation of desired changes. Instead, as implied, for
example, by cybernetic approaches to implementation (Dunsire 1993),
improvement happens more effectively because of system-wide commu-
nications and positive feedback, which do not need to be guided
step-by-step. Other recent studies of implementation have focused on
learning, stimulated by better leadership and close attention to detail
(Coelho and Ratnoo 2015). Behavioural insights can be the route to
forging these better links, by careful appraisal of the causal steps and
encouraging positive feedback up and down the delivery chain.

ENTREPRENEURIAL PUBLIC ADMINISTRATION

Behavioural public policy is usually practised by bureaucrats who are
located in agencies. They have power and legitimacy to act from the
authorisation of politicians. In such a context, doing behavioural policy
might seem to be straightforward. It requires decision-makers to embrace
the ideas of behavioural science and improve their evidence collection
procedures. They then can put into place interventions that make best use
of these ideas: hence nudge units and behavioural policy initiatives tested
with randomised controlled trials (or other forms of rigorous evaluation)
as discussed above. Such a move can help reinvigorate bureaucracies by
making them more evidence based, more imaginative in their thinking,
and increasingly willing to challenge existing ideas and standard operat-
ing procedures. Behavioural public policy can give power to the experts
and those with track records of innovation, which can then spread across
agencies in a pattern of diffusion. In this way, a more entrepreneurial

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 9 / Date: 22/1
JOBNAME: John PAGE: 10 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

10 How far to nudge?

public administration might come about (John 2014). This expansion of

interest can widen the purview of behaviour change policies; the diffu-
sion of new practices might help bureaucrats develop policies without the
frequent need for experts or nudge units; and it might lead to a
try-and-test approach to developing and introducing policies.
Experimental or entrepreneurial public administration does of course
have to be matched against all the forces in bureaucracies that keep them
operating in path-dependent ways. They adapt in response to new ideas,
but they are probably more interested in long-term survival by effectively
gaming the adoption of new practices. As a result of this strategic
behaviour, organisations survive, ready for the next set of new ideas
promoted by think tanks and academic champions. In fact, to ensure
longevity, the successful adoption of nudge requires skilful advocacy on
the part of intellectuals and bureaucrats, who are sensitive to the needs of
agencies, keen to deliver improvements to services and to save money,
careful not to risk politically sensitive interventions, and are responsive to
the pragmatic instincts of heads of large departments and delivery units
(John 2017b).
With the success of nudge units in assisting government policy, it may
be the case that the very adaptability of behavioural insights reveals
nudge to be a conservative phenomenon, or more conservative than it
should be, largely because of its ability to be easily translated, that is to
be in tune with the preferences of administrators and politicians, whose
interests might limit the potential range of nudge and disguise its
potential radicalism. Such a concern about the underlying agenda of
behavioural change and its potential range is the reason for the question
in this book’s title: How far to nudge? Nudgers have taken pragmatic and
understandable routes to promote the successful use of behavioural
insights; the question is whether there is a potential for a wider and
deeper use of behavioural ideas so increasing the scope of nudge. In an
initial answer to how far to nudge, this book’s argument suggests that
reformers and academics could apply behavioural insights much more
extensively than they do now. The early innovations in behavioural public
policies are just the beginning rather than an end point for the research
and policy programme. It is a good time to bank the gains and move on.

THE PROMISE OF THINK: NUDGE PLUS

The secret of more radical nudges is to consider public policy in terms of
how citizens perceive their own actions and those of public officials,
especially as they unfold over the long term. Policy-makers need to

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 10 / Date: 9/1
JOBNAME: John PAGE: 11 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Introduction 11

respond to citizens and act consistently with their preferences. To an

extent, citizens need to be involved in and debate the policy changes that
might be directed to their own behaviours. This is the agenda of think,
which can be seen as an alternative to nudge (John et al. 2011), and is
close to ideas argued for by those who believe in deliberative public
policy. Debate, self-ownership, and collective decision-making are the
key ways of achieving behaviour change. People are given information to
debate the issues; they have the space to deliberate; and then policies
may be changed as a result for collective benefit (Fung 2006). A range of
democratic innovations across the world can harness this willingness of
people to get involved in different ways (Smith 2009). The main problem
with think is that it is hard to get these initiatives off the ground, ensure
policy-makers use these procedures, and implement the recommendations
of deliberative exercises. People self-select into these activities, making
for an unrepresentative sample of the population involved, and nowhere
near large enough to scale up to generate the shift in policy outcomes
needed for mass behaviour change. Citizen participation exercises on
their own often are characterised by low participation and minority
representation. Such is also the fate of citizen juries and other delibera-
tive forums. These interventions can also be expensive to implement.
They seem to be better when addressing thorny public problems, such as
the siting of a nuclear power plant, rather than the day-to-day issues of
behaviour change, such as weight loss. In contrast, nudge initiatives are
easy to implement, cheap to do, get the buy-in of policy-makers, do not
need much effort or even conscious awareness on the part of citizens, and
have an immediate impact on behaviours and in due course policy
outcomes.
But nudge is closer to think than might be imagined, especially in how
it works in practice. The claim of this book is that nudging is much more
than giving messages or creating defaults; it requires some thought on the
part of citizens to respond to signals from public organisations and their
political principals. In other words, for nudge to work in better ways
there should be a consideration about how citizens understand the wider
choices in public policy implied by nudges. Citizens may need to
understand the wider policy issues so as to be influenced by behavioural
cues. Consider, for example, a nudge seeking to get people to turn up to
a doctor’s appointment on time (Hallsworth et al. 2015), which points out
the costs to the publicly-funded health service of missed appointments. In
one sense, the nudge could be seen as just a prompt, activating a
subliminal belief that turning up on time is a good idea, and it is just the
busyness of everyday life that has caused the patient not to arrive on
time. But, in another, this nudge relies on the respondent understanding

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 11 / Date: 13/12
JOBNAME: John PAGE: 12 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

12 How far to nudge?

that resources are tight, that it is bad to waste resources and that turning
up on time has an effect on the quality of the service overall, and having
a belief that the money saved will be beneficial. In other words, the
nudge depends on people having thought about the complex public policy
issues, which is the reason that behavioural cues and prompts are being
considered by decision-makers. They need to know that it is a public
policy problem. It is more than just being reminded of their civic
obligations. Then also a degree of trust is needed on the part of citizens.
Citizens need to know that people in charge of the system overall are
striving to make it work better.
It is important for people to rein in their short-term desires and to act
collectively. It also implies that citizens are not acting from short-term
self-interest, where it may make sense not to turn up because of other
things happening that are more pressing, where citizens think that the
doctor will be free at another time and the appointment can be rear-
ranged. A message asking someone to turn up on time implies that
citizens think about their duties and obligations. It implies they know
what it means to be an effective citizen. The same is true of asking
people to pay their taxes on time. If taxpayers only thought about it
selfishly they would weigh up the costs of non-compliance against the
benefit in paying late, which yields the prediction that the nudge would
not work, because it is hard for the exchequer to chase up everyone
effectively, or people may behave strategically by paying at the last
minute or incurring the cost of a penalty that is small compared to the
benefits on cash flow and interest payments of not paying early. The
nudge works better if it is part of a long-term conversation between the
state and citizens about what is the role of citizens in public services, and
about the relationship between individuals and the state. This kind of
thinking might not need to be explicit, but it is implied in the ways in
which nudges actually work in practice, even if citizens are not always
thinking about the reasons and assumptions that are being conveyed.
The challenge for policy-makers is to ensure that some degree of
reflection or thinking is integrated into the design and delivery of
behavioural interventions so as to encourage citizen responsiveness. The
opportunity for nudge is to deliver interventions in such a way that
citizens can be involved and reflect upon their actions, an inclusive form
of nudge that is transparent with the public. There is a wider debate about
the use of ideas about behaviour change, but without the assumptions
about citizen capacity made by deliberative democrats.
There is then a radical agenda at the heart of behaviour change
policies, which can address core questions of political organisation and
even representation, as it is based on re-establishing good relationships

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 12 / Date: 12/1
JOBNAME: John PAGE: 13 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Introduction 13

between citizens and governments. It is only when the citizens are

involved with the decisions that affect them that long-term solutions to
public policy problems are possible. In this way, citizens can automate
the changes brought about by nudge and integrate them into their
routines and habits. They are nudged to think and then they habituate the
acquired behaviours. This claim is at the heart of crafting behavioural
public policy, that is policy founded on understanding the behaviours of
citizens, but also embraces their capacity to reflect and participate in joint
ventures by taking them seriously.
Such actions may be prompted by the nudge directly, but nudge can
aim to get citizens to address the issue themselves with a little help.
Citizens might lack the capacity to think through the issues or have the
right level of information to make an informed choice. Nudges become
what Hertwig (2017) calls boosts (see also Grüne-Yanoff and Hertwig
2016). Hertwig writes that ‘boosts explicitly seek to foster existing
decision-making competences and to develop new ones, thus enabling
individuals to translate their intentions (preferences) into behaviour – that
is, to exercise personal agency’. The effect is indirect, but moves away
from the technocratic approach so common with nudges. Boosts are a bit
like the thinks that John et al. (2011) tested, but are closer to nudges and
more deliverable. In Hertwig’s (2017) argument, people sometimes need
help in understanding complex choices, such as which medical procedure
to choose, which they are increasingly asked to do in an age of consumer
choice and sovereignty. They may not understand the statistics of
probability, but a few tips and guides might help them make better
decisions. These interventions are sometimes called ‘educative nudges’,
which are about transmitting skills and ideas, rather than automating
desired behaviours.
In the argument of John and Stoker (2017), these kinds of strategies
build on the original nudges, but they take them further, perhaps not as
far as think, but a little way toward that goal consistent with what social
scientists know about people’s limited capacity for reflection, constraints
on time, and their need for immediate effects, even if the approach is also
about building a long-term partnership between citizens and the state.
They call this approach ‘nudge plus’, which is to emphasise how nudges
can be broadened outwards, so as to incorporate boosts and citizen
thoughts about public policies. This stimulation of thinking is desirable
so as to move beyond the more technocratic considerations of the nudge
agenda as currently practised. Nudge plus amplifies nudges and makes
them more democratic.
There is a greater interest in the behavioural sciences in securing
long-term change rather than the short-lived uplifts that characterise the

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 13 / Date: 9/1
JOBNAME: John PAGE: 14 SESS: 6 OUTPUT: Fri Jan 26 10:51:11 2018

14 How far to nudge?

classic early nudge interventions, such as tax changes and prompts to

encourage charitable giving. Scientists and policy-makers know that the
effects of interventions can last a long time, but these are usually
maintained because of habit or discovering a new form of behaviour,
which might be unconscious. There is now much more of an interest in
developing self-sustaining behaviours that might operate across a range
of activities that complement each other, so that people do more exercise,
eat better, drink less, and so on, and this might translate into other
behaviours and activities, so citizens get a new perspective on their lives
as a result of reading across a range of their behaviours and seeing the
links between them. This is what psychologists call a ‘passion for
long-term goals’, or ‘grit’ (Duckworth et al. 2007). This commitment is
not related to IQ, and moderately to education: the implication is that
individuals can cultivate grit, in a long-term programme of behaviour
change. Some of the nudges that are seeking to change people’s goals,
such as commitment devices that are keeping people on track, are about
trying to get individuals on a path of change, and this requires people to
consider their situations and then put into place measures that will
change their behaviours, even if every step on the subsequent pathway is
not consciously reflected upon.
So far, the kinds of nudges described earlier in this chapter have not
fully addressed either thinks or the sustaining of interventions, partly
because of the pragmatic agenda followed by nudge units and other
advocates, designed to encourage agencies to use behavioural science and
to bed down behavioural insights into the normal business of making and
implementing policies. Given the success of nudge, now is the time to
broaden it out and to link it more closely to democratic practice.

SYSTEM-WIDE CHANGES
The aim with nudge, boost, or nudge plus is not to use a behavioural
insight to create a one-off change, but to promote a state or environment
where these interventions help establish a new equilibrium of self-
reinforcing and beneficial behaviours, whereby all benefit, and there is
not a huge daily effort to keep the new behaviours in place. The end
result is something like the Highway Code, which in the UK is a set of
codified rules that are not legally enforceable and that govern how people
drive their cars. They assist good driving to the benefit of all, but nobody
enforces this; it is a self-regulating system of rules. The behaviours
follow from a common understanding and internalised norms. People
don’t need to think about the code and may have forgotten it long after

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 14 / Date: 22/1
JOBNAME: John PAGE: 15 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Introduction 15

passing their driving test, having mugged up on it in order to get through.

But it still influences their behaviour and on reflection most people
consider it a good thing to follow. Of course, a highway code is not a
perfect system of nudging. People do not always obey it, especially in
metropolitan centres like London. A pedestrian might find it hard to rely
on cars following the code, for example acting on road signs. Many
motorists might behave selfishly and only follow the code strategically to
avoid hitting other cars. But such imperfections are bound to be the case
in any informal scheme, which is why there is a need for policy-makers
to try to shape these understandings by nudges. The key question is how
to get to a more self-sustaining implementation system and whether some
reflection as well as nudging is needed to improve such codes and
introduce them for every kind of behaviour. The nudge creates the
reflection that causes people to adopt the code; then automatic and
self-monitoring processes sustain the behaviours over time, as well as
daily interactions with others who also observe the practice. The code
creates a set of norms that are maintained over time by habit and fear of
not abiding by them. The nudge becomes the initial shift that sets off an
upward cycle of desired behaviours, one action that is part of a system of
reinforcing actions, with occasional repeated nudges to keep everyone on
track.
In fact, behavioural interventions need to be seen as part of a complex
system with moving parts (Spotswood 2016). The designer needs to think
about how to move from one equilibrium point to another, an insight
which reflects recent progress in the economic analysis of experiments
(Banerjee and Duflo 2014). The interest in using experiments is not just
to show that there is a one-off change, which might in fact put a system
out of equilibrium; rather they should show how the balance of interests
and interactions can shift to a new pattern of cooperation and regulation
that is both beneficial and self-sustaining.
Another example is smoking bans, such as the one introduced in
England that prohibited cigarette and other smoking in workplaces,
particularly in bars, cafes, and clubs where staff work but where people
used to like to smoke. These laws have the appearance of central
government compelling people not to smoke indoors. In reality, the laws
follow public opinion and considerable debate about what needs to
happen. There need to be considerable changes to the political and policy
agenda beforehand. Once the ground has been laid for a reform, the
legislation can create a new equilibrium point, which ends up with people
willingly not smoking in public places. Smokers internalise the norm of
not smoking inside, agreeing with the principle behind the legislation.
They nudge each other. Light-touch reminders from public authorities are

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 15 / Date: 9/1
JOBNAME: John PAGE: 16 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

16 How far to nudge?

all that is needed to sustain the practice over time. No enforcement needs
to occur, as those who smoke consent to the rules, which on the whole
they believe to be fair. They do not even need to think about obeying the
rules, which they come to do automatically. The question becomes what
interventions might start off the process of accepting regulatory changes
and the establishment of a stable system of self-reinforcing behaviours,
and whether people can be nudged in ways that support a system change.
This attention to the underlying support for institutional and behav-
ioural outcomes is a more radical version of nudge and takes the agenda
to a wide-ranging set of actions and behaviours. The agenda of this book
is to radicalise nudge and to extend its insights, as a way of answering
the question in the book’s title. With talk of the system of interactions, it
becomes vital to include the policy-makers as part of the conversation,
and these people should be expected to change their behaviours too,
which provides a further extension of nudge, as the next section
elaborates.

TURNING THE TABLES ON THE POLICY-MAKERS

There is a paradox: the use of behavioural policies requires the upholding
of the ideal of the classic rational politician or bureaucrat who carefully
weighs up the evidence and then follows the best course of action. But
this is just what behavioural science says does not occur: decision-
makers should be just as responsive to cues and limited in their capacity
to weigh evidence as the citizens they hope to influence. This bias occurs
in spite of official support and the backing of experts who advise
policy-makers. Often heuristics take precedence because of the speed of
decision-making. Decision-makers prefer to rely on their own instincts
and act out of habit (Loewen et al. 2014; Lodge and Wegrich 2016). This
argument appeared in the classic works of public administration, as set
out by its founding scholars in the 1950s, in particular by Simon (1947,
1955), who formulated the concept of bounded rationality. These ideas
reappeared in Allison’s famous book Essence of Decision (1971). Such
concepts are consistent with formal models of politics that emphasise
search costs.
The limited cognitive range of politicians and bureaucrats may mean
that nudge becomes the preserve of the enthusiast or the follower of
fashion. Politicians are locked into existing routines and sources of
information, with new ways of working involving risks and potentially
negative outcomes. Indeed, prospect theory says they would avoid
making decisions that have potentially negative or perceived negative

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 16 / Date: 9/1
JOBNAME: John PAGE: 17 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Introduction 17

outcomes. In practice, some politicians take risks and are not fully aware
of the negative outcomes. Rather than considering the costs of system
redesign and being aware of the risks of adopting new policies (testing
them through trials), they are more interested in being able to claim
credit for acting in the short term. The bureaucrats who adopt behavioural
insights may be those who benefit from being innovative or being
different from the mainstream, or risk-takers. There is the danger that the
behavioural interventions will be the wrong ones, and done from advo-
cacy, fashion, and the norms of behavioural science. It is a case of the
blind leading the blind.
In the end the limited human cognitions of elected politicians and
bureaucrats offer an opportunity for citizens and citizen groups to use the
self-same techniques of behavioural science to influence politicians in
ways that are more direct than elections and public deliberation. The
spread of the use of behavioural ideas beyond government gives an
opportunity for citizens to turn the tables on their governors. They can
nudge politicians and bureaucrats. In this way, there is an opportunity to
move away from the paternalistic and technocratic assumptions of the
existing version of behavioural policy and to open up government using a
variety of nudges and cues, which might be an achievable conversation to
have, and one that might be feasible given the resistance by citizens to
deliberative exercises that need high levels of time and commitment. The
result is that politicians, bureaucrats, and citizens are engaged in a shared
project, both to use behavioural techniques and to work out what they are
for. All can learn as part of this process, which involves interaction.
Citizens are happier to influence politicians who have more limited
formal rationality. All are human and need external stimuli to get to the
best course of action.

CONCLUSION
In this book, the meaning and scope of behavioural public policy are set
out. It is seen as an ideal to which bureaucrats, citizens, and politicians
can aspire. The claim is that thinking in terms of behavioural solutions
can help policy-makers address public problems, but that part of the
solution is the engagement of citizens in problem-solving activities. The
result is bureaucracies that are not only adept in using behavioural ideas
and testing them experimentally, but also are responsive. The solution to
current policy problems is about joint problem-solving between citizens
and bureaucrats, where each recognises the limits of the other in terms of
time and capacity in making decisions.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 17 / Date: 9/1
JOBNAME: John PAGE: 18 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

18 How far to nudge?

This book goes beyond many publications in the field of nudge, while
paying due homage to the important books and papers that have come
before. It seeks to apply the perspective of a political scientist and public
administration scholar: that is, to see the application of behavioural
policy as a feature of politics that needs to be viewed as a consequence of
conscious decisions to implement and to introduce behavioural policies,
and that has to be argued for within bureaucracies and implemented using
standard frameworks for delivery. The book aims to understand the
impact of behavioural policies within bureaucracies and how they oper-
ate. It seeks to set out the wider context of nudge in terms of citizen
responsiveness to the measures proposed, which implies some degree of
conscious consent about the proposals and an element of debate, even if
done internally in people’s minds as they make decisions. The book sets
out the ideal of a more behavioural policy system, where policy-makers
consider the interests of citizens when they formulate public policies and
where citizens reflect upon and automate the behavioural cues that are
produced by decision-makers. Citizens become part of the solution by
using the behavioural sciences to get responses from policy-makers, and
policy-makers stimulate other policy-makers in searching for a respon-
sive and experimental public administration. The aim is to create a
system of behaviours that positively reinforce each other.
There is a massive opportunity to be grasped. Too often large policy
programmes fail because they do not address the behavioural determi-
nants of policy outcomes; in particular, policy-makers often do not
understand the very human reasons why measures and interventions fail.
It is possible to use an understanding of human behaviour to solve these
problems and change policy outcomes for billions of people across the
world. It is desirable to use theories of behaviour change to improve the
capacity of the government so it works much better than it does currently.
Given the desirability of the basic objectives, the question is how far
should policy-makers and reformers go? Can they go much further than
they do currently? This book asks these questions and aims to provide
answers, with the implication that reformers and policy-makers can
nudge more than they do currently, and governments, when they are
formulating and implementing nudge policies, can act in ways that are
consistent with liberal ideals and uphold a responsive and rights-
respecting system of public authority that citizens in a democracy rightly
expect. By keeping a bottom-up perspective, working from the perspec-
tive of citizens and treating decision-makers as human actors, the
potential of nudge can be seized in ways that enhance democratic ideals
and are consistent with ethical goals.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 18 / Date: 9/1
JOBNAME: John PAGE: 19 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Introduction 19

PLAN OF THE BOOK

The approach of the book is bottom up: to explore the public problems
that give rise to the challenges and then to see how writers and thinkers
stumbled on the solutions. The starting place, in Chapter 2, is the nature
of the problems that governments face. This chapter sets out the key
policy problems facing government and why they are sustained by
entrenched behaviours. It contains a more detailed review of some policy
problems, such as obesity. It assesses why conventional policy instru-
ments often fail.
Once the problems have been reviewed, the intellectual history starts in
Chapter 3, as knowledge from the academy is very important in
explaining what happened later on when policies were being formulated.
The chapter reviews the extensive literature on behaviour change, and in
particular is interested in the debates that occurred in the 1960s and
1970s that emerged from the research of Kahneman and Tversky. Even
though the concepts of bias and judgement are not entirely original to
these authors, the pair developed these ideas into a research programme
that had a profound influence on economics and on social science
thinking in general. Even though there have been debates in sub-
disciplines, such as the psychology of health or social marketing, they
have not had the profound influence of the founders of behavioural
economics. Readers may be familiar with the review of behavioural
policy problems in Chapter 2, and they may know the story of behav-
ioural economics and other related studies in Chapter 3. If so, they may
skip or briefly skim these chapters without risking not understanding the
key arguments and claims that come later on.
The modern story starts in Chapter 4. The chapter is about the way in
which ideas in the academy got into the mainstream, through the natural
processes of influence and the role of advocates and entrepreneurs. There
is an account of how nudge has been enhanced by its popularisation and
has come to be seen to be common sense in its application, but also
limited in its radicalism in that certain routes have been taken but not
others. This argument, which is about path dependence, continues in
Chapter 5, which considers how nudge has been used by policy-makers,
which been propagated by nudge units and other agencies keen on
applying behavioural insights. The chapter contains a review of the
classic nudges, such as on letter redesigns and organ donation. This
chapter also tells the reader about the policy innovations brought about
by research in the behavioural sciences, but with an eye to what choices
have been made and whether other opportunities could have been seized.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 19 / Date: 9/1
JOBNAME: John PAGE: 20 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

20 How far to nudge?

Chapter 6 starts to build up the core argument of the book by looking

at some common criticisms of nudge, particularly over its range and the
power of its findings. Even though many of these criticisms of nudge are
overdone, the process of explaining them prompts the thought about what
kinds of intervention could be stronger and more self-sustaining. Chapter
7 is about the ethical dimension to behavioural interventions, which again
has attracted a lot of critical opinion, in particular about the consistency
of the arguments Thaler and Sunstein have put forward, and the sense of
unease people have felt about the way in which nudge appears to be
manipulative and to reduce people’s autonomy. Thinking about this
problem leads to the argument for a more agent-centred version of nudge,
which is the basis of the radical version argued for in this book. Chapter
8, called ‘Nudge plus and how to get there’, is a presentation of the
central proposals of the book, which are about the potential for a more
open, reflective, and decentralised form of nudging, with examples and a
discussion of the potential for their use. Chapter 8 also contains a
discussion of elite nudges and how these may be extended to a wider
range of phenomena as part of bottom-up nudging. Chapter 9 is intended
to pull the book together to answer the question ‘How far to nudge?’,
taking into account practical limitations and the salience of politics. It
comes to an overall assessment about the future potential of behavioural
public policy.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 01-Chapter1PJED /Pg. Position: 20 / Date: 9/1
JOBNAME: John PAGE: 1 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

2. Behavioural public problems

This chapter is devoted to explaining why human behaviour is so
important for achieving desired public policy outcomes. It is important to
establish this claim, so it becomes more than a truism. The truism would
take the following form: because all public policies are implemented
through changes or assumptions about human behaviour, behaviour is
important in public policy. But this statement can be seen to be
tautological in that any form of human behaviour has to be considered
part of public policy, as most policies involve humans as subjects in their
design and implementation. The behavioural argument has to be much
stronger. It should take the causal form that certain behavioural traits and
biases detrimentally affect policy outcomes in ways that are systematic,
knowable, and fixable. The whole point of behavioural public policy is to
correct for certain kinds of behaviour, in particular those traits that
reinforce poor policy outcomes. People do not make good choices for
themselves, which is the justification for the nudges of the Thaler and
Sunstein sort, which can be seen in many fields, such as health, saving
for retirement, and seeking better attainment in education.
One important point to take account of when considering behavioural
causes of public problems is that the nature of the behaviour varies
according to the outcome, which means that there are different behav-
ioural problems at play and varied solutions that may be adopted. With
some problems, it is question of individual choices, which are not greatly
dependent on each other, even though they may have an impact on the
costs of public services, such as in health. Other behavioural problems
require citizens to cooperate with each other and have a reasonable
expectation that others are going to act as well or else their own
contribution would be too small to make any difference and they would
bear all the costs. This feature of some social relationships is what social
scientists call the ‘collective action problem’ – the idea that people might
want to change their behaviour but because they are uncertain about the
choices of others they go for a safe strategy, which often means not
cooperating or doing anything (see Ostrom 1990). Some kinds of
behaviours, such as pro-environmental behaviours, suffer from this
problem and thus require different kinds of intervention from government

21

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 1 / Date: 13/12
JOBNAME: John PAGE: 2 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

22 How far to nudge?

or other agencies in response. The implication is that the tools of

behaviour change need to be adapted to deal with the nature of the
problems being addressed.
These behavioural factors need to be considered alongside other causes
of policy problems, whether they have their origins in technological or
scientific changes, economic and social factors or even the actions of
policy-makers themselves, which often reinforce behavioural public
problems. Different sectors of activity vary in the extent to which
individual behaviour is the cause. A judgement needs to be made about
the balance and nature of the interrelationships, and where individual
action depends on other policy levers being in place, rather than
individuals having to bear the burden of achieving behaviour changes and
delivering favourable policy outcomes on their own even if nudged by
government. The tendency to assume that individual behaviour is the sole
source of a public problem needs to be avoided when formulating
behavioural public policy.
To begin, this chapter explores the sources of behavioural problems by
examining the case of health, which provides a relatively straightforward
example of behavioural problems, before looking at more complex
examples, such as the environment.

THE NATURE OF THE PROBLEM: AN EXAMPLE

FROM FOOD AND DIET
Consider a basic activity such as the consumption of food. In past
generations, people had limited supplies, in particular of protein, and had
to make do with a mixed diet of some meat, but mainly vegetables and
fruits. This diet was particularly observed in countries bordering the
Mediterranean, where there were few sources of meat but a lot of natural
produce, such as olive oil, and where vegetables were easy to grow. But
it also was followed in countries in the north of Europe, where meat was
only a bit more plentiful, at least for the large majority of the population.
Most people were engaged in manual work. While these populations
could be hit by disease and while even minor cuts and wounds could not
be effectively treated before the arrival of penicillin, most people were
healthy and not overweight, because they had to exercise and the foods
available happened to be healthy. Move forward a couple of hundred
years to the 2010s, especially in the urban centres, and this diet has
largely vanished as the norm, with people consuming more meat, which
is high in fat. Fewer fresh vegetables are eaten. New products that are
cheap to buy and convenient to eat contain a lot of sugar, fat and salt, all

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 2 / Date: 13/12
JOBNAME: John PAGE: 3 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Behavioural public problems 23

of which are bad for health. People do not exercise very much, as they
drive cars or can move about without physical effort on public transport.
Many people have become sedentary. Some are overweight or obese.
They are likely to get strokes, heart attacks, and cancer as they get older,
in spite of advances in life expectancy from better health care, rising
incomes and changing occupations. In some ways, it was not the fault of
the individuals that the healthier (if riskier) environment that imposed
exercise and a better diet went away. Growing prosperity and technology
are to blame. Human beings like to eat until they are full, often eating
what is on the plate or just the available food to hand. People would
rather sit on the sofa than exercise. Sweet foods are generally nicer –
evolution told people to like them as sources of energy when food
supplies were meagre. Watching television or playing computer games is
much more fun than running on a machine in the gym. But the result is
that the behaviours that come naturally are very harmful and impose high
costs for the individual and for society too in terms of health care costs
and reduced labour productivity. An obesity epidemic is in train (Branca
et al. 2007). Individual human behaviour is the main cause, which is
influenced by the availability of foods, the activities of private com-
panies, the role of the media and advertising, and even government
policy.

BENEFICIAL BEHAVIOURS IN HEALTH

A change in eating habits and some modest exercise, such as walking,
can have large effects. To achieve these changes, governments have tried
to educate through public service adverts and advice by doctors. Healthy
lifestyles have been promoted by journalists and television programme
makers. The middle classes have become aware of these advantages, so
healthy eating and exercise have become fashionable without much
prompting by government. The hope is that these behaviours can diffuse
more generally across society, but in practice a lot of people do not want
to act on the information or wish to be influenced by the middle classes.
Alternatively, they may accept the argument and understand what the
science is saying, but not be motivated enough to change their behaviour.
Governments can try to get people to change their behaviour through
education and encouragement, but it is hard to shift behaviours. Consider
the need for regular exercise to probe this argument further. Exercise is a
classic form of behaviour that has strong beneficial effects. Research on
1 million adults found that exercise of about one hour a day can help
them avoid premature death, making a 60 per cent difference in outcomes

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 3 / Date: 13/12
JOBNAME: John PAGE: 4 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

24 How far to nudge?

(Ekelund et al. 2016). The researchers did a systematic review of six

databases to locate cohort studies (the same people observed over time),
which had information on daily sitting, television-viewing time and
physical activity, to look for the impact on mortality, heart disease, and
forms of cancer, and were able to account for other causes of ill health,
such as obesity and smoking. The clear finding is that lack of exercise
causes negative health outcomes. What is interesting from this study is
that it found that just one hour of exercise reduces the negative health
effect of sedentary behaviour, even of television-watching, though this is
still harmful. But it is important to be careful in forming conclusions
from a longitudinal study in which exercise co-varies with health, as
confounding factors that might cause people both to be sedentary and to
have ill health are not fully considered. This argument is stronger for the
television link, as people might retreat indoors if they have less inclin-
ation to go outside because they are isolated and depressed, which causes
them to become unhealthy. Yet overall it is true that lack of exercise is
correlated strongly with poor health outcomes. The scale of the problem
is massive, with only 25 per cent of adults in the US doing the very
minimum amount of exercise of half an hour a day five days a week
(U.S. Department of Health and Human Services 2010).

RESISTANCE TO BEHAVING WELL

Small changes in behaviour can change outcomes quite strongly. But why
do people keep behaving in harmful ways? One cause is a lack of
options: at work the job task and the design of the office make sitting
almost inevitable. There may be long tasks to complete, such as serving
customers or completing routines on the computer. People might not take
breaks, wanting to keep in front of their screens to check e-mails. There
may be nobody else who goes out of the office at lunchtime to encourage
others. It may be a noisy polluted street, so people feel it is not worth
going out, let alone exercising. Even motivated individuals might not
want to do some exercise during break times. In this way, change in the
external world (e.g. changes in work patterns) links to other incentives
and social norms that can embed low exercise as a repeated activity.
Behaviour comes about from mental processes, in which behavioural
intentions are important. Such intentions come from an underlying set of
attitudes based on beliefs and people’s evaluation of likely outcomes.
Affecting behaviour is people’s subjective norm, linked to their normative
beliefs and motivation to comply; important too is their perceived sense
of control, which is affected by their efficacy, or the belief that they can

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 4 / Date: 13/12
JOBNAME: John PAGE: 5 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Behavioural public problems 25

change matters for themselves. This is the theory of planned behaviour

(Fishbein and Ajzen 1975; Ajzen 1985). People may have intentions to
change their behaviour but in fact do not, so that the status quo and the
behaviours that sustain weak outcomes are maintained over time. Norms
and the social environment in which people live have an influence in
limiting the willingness to engage in behaviours that may improve
outcomes. People may want to engage in a particular behaviour, but
social factors and norms may prevent it from happening.
Behaviours shaped by peers and families are often enjoyable, even
addictive, so hard to change once adopted. The effects of these behav-
iours are delayed for many decades, so people cannot see the long-term
consequences of their actions, and may think that nothing will happen to
them. Because not everyone develops negative health outcomes – at first
at least – people might think themselves immune or lucky, which the lack
of immediate consequences might encourage them to think. Their calcu-
lation of risk is very poor, but hard to correct. Bald statistics are often
meaningless to many people who do not understand them. Such statistics
can be countered by a sense of being lucky, a feeling that you deserve
one luxury and a belief that you are cutting down (no matter how
unjustified this is). People deceive themselves and rationalise their
behaviours, screening out messages they do not like to hear. The question
becomes what measures and actions can be taken by government to break
through such arguments that people make to themselves.
Health is a field where there are clearly identifiable behaviours
individuals can change to make their own outcomes a lot better but that
also have large collective benefits in terms of the saving of public
resources, which can be reinvested in other areas where behaviour is less
of a cause and there are therefore other forms of treatment. Lung cancer,
for instance, is mainly caused by smoking tobacco. A behaviour change
involving giving up smoking tobacco will lead to large reductions in
health expenditure from treating cancer and other diseases caused by
smoking, such as emphysema and hardening of the arteries, but in
addition to the benefits to society the individual will still be massively
better off. A lower consumption of alcohol can also lead to reduced
public expenditure (or reduced insurance contributions for all), for
example on treatment for cirrhosis and other health disorders.

HABITS
The other dimension to the resistance of people to messages is the
prevalence of unconscious processes that cause them to continue the

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 5 / Date: 13/12
JOBNAME: John PAGE: 6 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

26 How far to nudge?

harmful behaviours. People automate the behaviours into their routines,

so they become habitual, adopted because people have behaved similarly
in the past, which makes their repetition reassuring. Over time habits
may strengthen (Denford et al. 2016: 56). Habits once set are hard to
break. Even after behaviour change, old habits can reappear, with
individuals trending back to their old behaviours. Habits can be useful in
the other direction. Once a new behaviour has been stimulated, it can
create a new habit, which beds down and is followed with just such a
lack of reflection and self-awareness, with the difference being the
benefit the individual gets from the action.
Behaviours are reinforced by the actions of outside bodies, which can
use a variety of behavioural interventions to keep individuals consuming
in the way they do, for organisational advantage. Private sector organ-
isations spend a considerable amount of money on advertising to
maintain expenditure on products that might not be healthy, especially
smoking and alcohol, but also processed foods. Retailers engage in subtle
interventions to change behaviours, such as in the design of super-
markets. As Marteau et al. (2011: 263) summarise,
Shaping environments to cue certain behaviours is extremely effective,
unfortunately often to the detriment of our health. The ready availability of
foods that are packaged, presented, and engineered to stimulate our automatic,
affective system has led us to consume more than is needed – consumption
that is further primed by advertising. The doubling in alcohol consumption by
young people over the past fifty years is attributed in part to its marketing and
ready availability, and the design of many neighbourhoods supports car
driving over walking or cycling.

Citizens are nudged then to act against their self-interest.

WHY CURRENT POLICIES DO NOT FULLY ADDRESS

BEHAVIOUR CHANGE
In terms of thinking about how to deal with the problem of no exercise,
governments seem to think that a large number of mainly educational
initiatives, and demonstration projects to scale up initiatives, work well
(see Reis et al. 2016), but very few are evaluated in a robust way, such as
with randomised controlled trials. In fact, a behavioural perspective
would be critical of such a blunderbuss approach that does not address
the reasons why people do not behave as public professionals want them
to. The public health approach would just see the lack of exercise as an
information deficit needing encouragement and education. But if the

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 6 / Date: 13/12
JOBNAME: John PAGE: 7 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Behavioural public problems 27

endogeneity issue is seen as fundamental then there may be causes of the

lack of willingness to exercise which are correlated with the outcome,
such as mental health factors that impact on well-being and health and
cause someone not to exercise. When considering many interventions
around exercise, it is common to find that exercise is addressed at the
same time as other behaviours, such as smoking, diet, and alcohol
consumption (see Kahn et al. 2002: 76), making it hard to assess how the
intervention actually addresses health outcomes.
Many information-based interventions have weak effects, for example
the placing of signs next to lifts to encourage use of the stairs. Mass
media campaigns also have moderate to no effects (Kahn et al. 2002: 78).
Classroom interventions that have had a behavioural component (the
learning of skills) have variable effects. Significant for this book, a more
behavioural approach has better effects, such as systems of social support
or the introduction of buddying. Effective too are environmental changes,
such as changing the availability of exercise in terms of facilities. The
conclusion to draw is that the information-informing approach is useful,
but does not yield much, and a more behaviourally informed approach
might work much better. It is possible to come to similar kinds of
conclusions about other interventions designed to change people’s behav-
iour, such as improved diet in low- and middle-income countries (World
Health Organization 2009). Light interventions based on human agency
do not appear to work, despite being favoured by many governments,
such as the UK government’s Change4Life programme, a campaign
launched in 2009, aimed at children, to change eating habits (also see
MangerBouger in France and Let’s Move in the United States).
These campaigns appear to be tackling human behaviour, but they are
essentially non-behavioural in that they are encouraging behaviour
change; they do not use the full range of psychological traits, and perhaps
focus on mechanisms that individuals do not find that easy to use:
‘Exerting agency requires individuals to rally their cognitive, psycho-
logical, time, and material resources’ (Adams et al. 2016). It requires too
much of an individual’s mental capacity to make sense of all that
information and then act. It means that a lot of resources are expended on
these campaigns, but they overload the individual and do not take
advantage of more powerful mechanisms that require less effort.
The measures that appear to work for diet are targeted and directive
(see Adams et al. 2016). The reason why such strong strategies can be the
only ones that have a chance of working is the embedded nature of the
behaviours that underpin the desired activities. It can be debated whether
strong interventions, such as banning smoking in public places, are
genuinely behavioural. The argument is that top-down interventions have

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 7 / Date: 13/12
JOBNAME: John PAGE: 8 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

28 How far to nudge?

to be behavioural or else they will not work. People can easily defy
smoking bans; the only way to achieve compliance is by managing
consent, which involves a lot of nudges before the activity is banned or
restricted. These interventions then are closer to the nudge approach
argued for in Chapter 1 where both information and institutional con-
straints are altered at the same time. In other words, is there a middle
way between hard and soft interventions that economises on cognitive
capacity but keeps individual freedom and agency?

RATIONAL ACTION AND BEHAVIOUR CHANGE

Often people do not act when it is clearly in their interests to do so. In
economic theory behaviour should be the consequence of a balance of
costs and benefits, so that actions that are harmful should be considered
and then individuals do something about it. With some of the examples,
such as health, it is possible to see clearly the advantage of a change in
behaviour, so behaviour change is the logical consequence of a govern-
ment providing more information. But, even with health, these kinds of
campaigns are limited. Another way of looking at the problem is through
an assessment of risk whereby individuals tend to overweight current
benefits and down-weigh the costs. This can be rational in the sense that
people might like smoking and drinking so much that they decide they
would rather have the benefit now than later. The role of government is to
inform people so they can find out whether they have exactly balanced
out the costs and benefits. It is often the case that smokers know the costs
of what they do but find it easy to down-weigh the actual costs or forget
about them or justify the costs by statements along the lines of ‘Smoking
is my only pleasure.’ This is at the heart of behaviour change interven-
tions: whether the secret is to try to discover a rational core self that
really wants to respond, which is the nudge approach, or whether the
policy-maker just assumes irrationality and it is fine just to change the
behaviour using this assumption. The key point is whether there are
automatic processes that govern human behaviour that do not have their
origins in cost–benefit thinking but derive from personal psychology.
This means that human behaviour can be shaped by a set of external
stimuli and prompts that are not about conveying the costs and benefits of
decisions, but may relate to mood, emotions, intuitions, and senses of
well-being in ways that are immediately responsive and not reflected
upon. This may be a reaction against the perception of loss, and wanting
to conform to the behaviour of other people.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 8 / Date: 13/12
JOBNAME: John PAGE: 9 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Behavioural public problems 29

There will be more on these points later in the book, but they show the
importance of embeddedness and suggest that simple interventions based
on altering the costs and benefits are usually limited in impact. That said,
there is evidence that some cost–benefit interventions work, such as
increases in the costs of alcohol per unit and paying for smoking
cessation, which do not crowd out motivation.

REACTANCE
People often resist messages that come from an authority. This can be
called ‘reactance’, which is a resistance to the activity when it appears
that choices are being limited (Brehm 1966; Brehm and Brehm 1981).
This phenomenon sounds surprising, because much of traditional public
policy rests on the idea of experts finding out new pieces of information
about what is beneficial or harmful for individuals. Governments then
design policies to embody the information through advice, persuasion
and regulation. Then citizens take note of the expert information, and
adjust their behaviour accordingly. That model works where people trust
those in authority and respect the role of experts; they put their faith in
the politicians who make decisions for them. There is also a gradual
diffusion of behaviours from early adopters to others, in processes of
emulation that are based on respect and deference more generally. This
model does not work in such a straightforward way today, partly because
of the declining trust in experts and in government. Across a large
number of indicators from surveys, standard messages of political trust
show a decline, which is related to general social factors as well as to the
performance of government (Dalton 2005). Dalton points out that the
rates of decline are highest among the most highly educated groups in
society. These are precisely the groups that are more receptive to
messages about changing their behaviours, so this is damaging for the
conventional tools of behaviour change. It is not just that citizens resist
the idea of better behaviour change, but that the means of achieving it is
different to a standard diffusion and deference model. Other steps are
needed, ones that are attuned to the reference groups involved. For many
interventions, there are different kinds of sub-groups that vary in their
receptiveness to an intervention.
There are more precise reasons why a message to change behaviour
might be resisted. People sometimes do not like to be told to do
something. This can work in a simple way, with people resenting being
told off, as a child does not like what might appear to be harsh words.
People like approval and do not want disapproval, which might cause

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 9 / Date: 10/1
JOBNAME: John PAGE: 10 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

30 How far to nudge?

them to comply with a request, but it can cause them to resent it and
maybe do the opposite to what is asked for. Sometimes the resistance can
work in a subtle way. People comply with the request, but their secret
self does not really want to do the activity: they might prefer to eat
chocolates and sit in front of the television. Even though they comply
with the intervention, after a while they return to the old activity, with the
added enjoyment of it being a guilty pleasure; they might then do the
condemned activity all the more because they like it, what is sometimes
called a ‘boomerang effect’, similar to reactance. It can also be described
as a ‘backfire effect’: when people receive corrective information their
willingness to believe incorrect information increases (Nyhan and Reifler
2010). It appears to be linked to the strength with which the fact is
associated with a person’s identity. Reactance can happen when there is a
lot of pressure to undertake the activity. Even strong eye contact with
someone can provoke reactance (Chen et al. 2013). People do not like to
be persuaded directly and will resist such messages, but this is what
information campaigns often do.

MULTI-CAUSAL AND EMBEDDED BEHAVIOURS

In examining behaviours, it is tempting to focus on one kind and a single
cause, and to see the solution as a focus on that behaviour, and then one
lever or tool of government might solve the problem. Some of the health
examples seem to point to this problem, so that smoking during preg-
nancy can be countered with a single intervention to change, such as an
incentive. But in practice there are linked behaviours that lead to poor
outcomes, which themselves have multiple causes, and which can then be
reinforced over time by social norms and the influence of peers. A
problem like alcoholism might be caused by personal factors, but in
addition by the urban environment, low education, the influence of
families, and cultural factors among social groups and their lifestyles
(Annis et al. 1990: 17). The key idea is that these problems are embedded
because behaviours are reinforced by the social environment and the
influence of peer groups.
Behavioural problems are not just caused by individual psychology;
they relate to how an individual is linked to other individuals and their
social behaviours, which can often reinforce each other. Take the example
of crime, which comes from people seeking a private benefit from illegal
behaviours, or in acting out of impulse or anger in ways that harm
people. There is a large literature about what causes crime, how it is
embedded in social networks, and how people rarely make cost–benefit

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 10 / Date: 13/12
JOBNAME: John PAGE: 11 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Behavioural public problems 31

decisions about committing a crime. They may understate its costs and
overestimate the benefits from the action, such as theft of property. It is
the social networks that influence this kind of behaviour, and affect how
information about criminal behaviour is transmitted. Once a person starts
behaving in a criminal way, it is hard for that person to break out of the
pattern, as a result of habit and the influence of peer groups. These
behaviours may be reinforced by the criminal justice system and prison,
which means that people’s life chances are reduced and that they cannot
get access to the job market, so making a life of crime apparently
inevitable and making it likely that the pattern of recidivism will
continue.
Problems themselves can be multiple and correlated with each other,
so children who have behavioural problems in schools might also have
problems at home; for example, through substance abuse they might
develop psychological problems. As Gardner and Shaw (2008: 883) write
on the problems of young children, ‘Based on the relative instability of
problem behavior during early childhood, it should not be surprising to
learn that many children demonstrate multiple types of problem behavior,
including co-occurring disruptive and emotional problems (e.g., opposi-
tionality coupled with depression or attention deficit/hyperactivity dis-
order.’ It follows that policies addressing such problems need to take
account of these multiple causes.
In terms of societal problems, these examples may appear to be
extreme, and also concentrated in sub-populations who are not numerous
and already have a large amount of attention paid to them by public
agencies with research programmes to back them up. But they form a
subset of a large world of reinforced behaviours, which are just as likely
to be the case in prosperous communities, but which have received less
attention and are less visible. The point to make is that it can make it
quite hard for an intervention to succeed, because of the force of these
social factors driving behaviour and reinforcing each other. The nature of
these behaviours can explain why informational or educational pro-
grammes often do not work, and it can then provide a justification for
programmes of behaviour change. It can also give rise to pessimism as to
whether these behaviours can be addressed in one-off interventions even
of the nudge sort.

COLLECTIVE BEHAVIOURS
The attraction of the diet and exercise examples, and even crime, is that
public interventions can be individually based without needing collective

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 11 / Date: 10/1
JOBNAME: John PAGE: 12 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

32 How far to nudge?

action on the part of the population. The intervention can be targeted to

the individual and tailored accordingly. Even then, it is hard to shift
behaviours, because they are tied up with people’s lifestyles, structured
by the work environment, supported by technology that everyone needs
to use such as cars, and then influenced by peers. In the case of the
environment, all these factors apply, but there is an additional problem
that means that behaviours are even harder to shift. The issue is that
individuals need to act jointly to ensure that environmental measures
work. Consider a very simple environmental issue, such as littering. Here
picking up an individual piece of litter does not make much of an impact
on an urban street and also will not improve the welfare of the individual,
who has to find a bin or pocket the piece of litter and take it home.
Whereas the individual act of exercise has a direct beneficial impact for
the individual (even feeling good after the exercise), the problem with not
littering is that there is a cost, and the only reward is a feeling of having
done some good. The problem is that interventions to improve the
environment have to appeal to people’s sense of the collective interest,
but it is hard to appeal to these beliefs and values, making it difficult to
shift behaviours in the long term. Also, if the intervention only shifts
behaviour to a small degree, no one will notice the effect. Several fewer
pieces of litter in the public park will hardly be spotted by anyone, which
may de-incentivise the few people who responded to the anti-littering
intervention. It is hard to question the claim that the behaviours will
continue over time in the absence of strong enforcement. The book’s
discussion later on seeks to deal with the problem of inducing collective
action without authoritarian and unfeasible interventions.
The same argument applies to many environmental interventions, as
these are seeking to affect public goods, such as pollution control. With
climate change, changes in behaviour are needed on the part of individual
citizens to reduce carbon, by using cars less, changing levels of consump-
tion of central heating or air conditioning, and so on. Even if many
changes are not directly targeted at citizens, such as on packaging and
materials for insulating homes, they need at base some degree of citizen
consent or else they are not going to be popular and government will fear
introducing them. There is a collective action problem with individual
behaviours, such as people in Western countries thinking there is no point
driving less when so many new cars are being driven in China (and vice
versa).
Behaviours are hard to shift, because citizens might think it is the
responsibility of those other than themselves, such as companies, elites,
and governments of other countries, to take action. In addition, people
are encouraged to consume more from society and to regard constraints

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 12 / Date: 13/12
JOBNAME: John PAGE: 13 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Behavioural public problems 33

on their behaviour as a loss of their rights, particularly if they can

observe other classes or people in other countries as already consuming
at a high level. It is also hard to address climate issues because of the
slow rate of change and because people base their views on current
weather, which may not be unfavourable. If people are not inclined to
change their behaviour, confirmation bias may cause them not to select
information that is relevant or provided from official sources. They read
dire warnings from climate change experts, but still go on a pleasant day
trip to the seaside or take a long-haul flight to an island that is about to
be engulfed by rising sea levels caused by global warming.

THE IMPACT OF POVERTY

A big set of behavioural questions concern poverty, which is the conse-
quence of a range of economic processes (World Bank 2015) and lack of
resources. Just coping with poverty poses a large cognitive burden on
individuals, which impinges on their ability to search for jobs and invest in
long-run activities to help their families (Shafir and Mullainathan 2013).
There is a further reason why poverty and poverty-reinforcing behaviours
can have long-term effects. A variety of adverse environmental exposures
during pregnancy and in early childhood can disrupt normal brain develop-
ment and influence learning and acquiring behaviours. Such exposures
include poor diet in pregnancy and recreational drugs (nicotine and
alcohol), as well as maternal stress and anxiety, and are often linked with
poverty. As the World Development Report 2015 authors write, ‘experi-
encing excessive stress and anxiety in infancy impairs the early develop-
ment of learning abilities and non-cognitive skills, with cascading negative
consequences for later achievements’ (World Bank 2015: 101).
One mechanism thought to link early-life adversity to later-life out-
comes concerns the phenomenon of epigenetics. Epigenetic marks such
as DNA methylation and histone modification lie on top of the gene
sequence (the inherited instructions for the organism) and play an
important role in switching genes on or off, required for normal
development and lifelong health. These marks are passed from cell to cell
and sometimes from an individual to the individual’s children – termed
‘epigenetic inheritance’. While the genetic instructions passed on by
parents to their children are not directly affected by the early environ-
ment, the epigenetic marks that control them can be (for an introduction
to this subject, see Carey 2011). Epigenetic marks are particularly
responsive to environmental factors during foetal development and early
in childhood when the brain is developing most rapidly. This means that

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 13 / Date: 10/1
JOBNAME: John PAGE: 14 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

34 How far to nudge?

genes may not be correctly marked, and this abnormal marking may
cause ill health later in life, with increased prevalence of mental health
issues such as depression, anxiety and schizophrenia as well as metabolic
disorders (see Kundakovic and Jaric 2017). Some studies suggest that
incorrect marks can also be passed on down the generations, with even
grandchildren affected by the environment of their grandparents. This
means that poverty experienced in one generation can get locked into
families, even when future generations might not experience so much
poverty. Factors such as depression may affect the individual’s ability to
find and hold a job, to have sustained family relationships and to
appropriately parent children, adding to adversity for the next generation.
This process raises the question of whether intervention that improves the
poverty of young mothers might prevent or even reverse incorrect
epigenetic marks. Behaviour change interventions could create beneficial
behaviours in young mothers during pregnancy and early motherhood
aimed at reducing smoking or increasing healthy eating, which would
benefit their own health and which could potentially have much longer-
term beneficial consequences for their children by reducing the acqui-
sition of incorrect epigenetic marks. It does suggest a large problem for
policy-makers seeking to address the behaviours that reinforce poverty, as
they face another hurdle, that of epigenetic causes of the behaviour,
which come from the negative outcomes of that behaviour, so adding to
self-reinforcing processes.

BLAME ATTRIBUTION
One area of great care when making the argument about the behavioural
sources of policy outcomes is the attribution of responsibility to citizens
when they do not bear the full responsibility for their actions or the
consequences. At one level this is about actual causes, or the extent to
which citizen action or inaction has caused the behaviour as opposed to
other actors such as private firms or governments, and it is inevitably the
case that all outcomes are multi-causal. There may be some responsibil-
ity, but other actors are also to blame. If the omission perspective is taken
into account, then government and other actors could have been doing
other things to put the situation right, so they bear responsibility for not
following it through. The other aspect of this problem is that the actors
interact in that each can shift blame on to the other as a reason for not
acting. With the blame-shifting perspective there can be other reasons for
inactivity, in that an opportunity for a cooperative relationship has been
lost and outcomes become worse than they would have been if steps

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 14 / Date: 13/12
JOBNAME: John PAGE: 15 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Behavioural public problems 35

toward collective action had been taken. The solution is a partnership

between government and citizens, which is often lost in a technocratic
approach that assumes a tools or instrumental perspective to getting
things done, or a social psychological perspective that regards behaviours
as trapped within social structures and unchangeable. In fact, the capacity
for citizens and bureaucrats to think differently can be an opportunity for
change, but lack of capacity is itself a reason why behaviours are the
source of negative policy outcomes and hard to break out of.
These arguments lead to a better understanding of the complex reasons
for policy problems. These are to do with the system properties of both
policy-making and human activity. Individuals behave according to their
own inclinations, but also as social actors, so they respond to the signals
and behaviours of friends and family and also to messages and incentives
from organisations such as governments. This is the role of government
as a regulator of the private sectors, and across a whole range of activities
including the sector under consideration. In terms of health people may
respond not just to their doctors and medical practitioners, but also to
signals from those in education and from people in the private sector.
There is a tendency for such social systems to tend to equilibrium, so all
the actors feel no need to alter their behaviour, because the actions of the
other players have caused them not to consider any changes. This can
create poor outcomes that are reinforced by these supporting factors.
It is important to set out these causes of public policy problems, and
not to assume that the behavioural aspect to public policy is purely a
consequence of individual deficiencies. There are social causes for poor
behaviours that get reinforced over time, and public agencies may
compound these behaviours. This does not mean that nudgers should
despair, but they need to realise the causes of the behaviours they wish to
change and that individual behaviours are connected to many social
processes and other government policies. Behaviours can be changed, but
sustained behaviour change needs to be linked to other reforms and
adjustments to standard operating procedures. Moreover, behaviours vary
in their degree of embeddedness, with some behaviours relatively easy to
shift, while others are more difficult. Policy-makers have seized on the
easier-to-shift behaviours as a way to start the research agenda and to
show the benefits of the nudge model. The greater challenge is to find out
how to shift these more embedded behaviours.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 15 / Date: 13/12
JOBNAME: John PAGE: 16 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

36 How far to nudge?

CONCLUSION
Behaviour change is crucial for achieving positive outcomes for society,
the economy, and politics, largely because behaviour contributes to
human welfare, aggregating the actions of individuals for societal benefit.
Change behaviour and individuals benefit, usually in the long term. As a
result, there are collective gains in the form of reducing costs to the
public purse, increasing the efficiency of the economy, and promoting a
more agreeable and safe society for citizens to be part of. Other
behaviours can enhance collective welfare without a great deal of benefit
to the individual, which can be a harder problem to solve. The challenge
is to change individual behaviours in combination with the decisions of
others, so society may overcome collective action problems and indi-
viduals are incentivised to act for the public good. Overall, there are few
public problems that could not be improved by changing individual
behaviours, but public policy needs to be sensitive to the mechanisms
causing those behaviours, and how they vary across domains.
In social science, there is a long and venerable tradition of understand-
ing policy problems as a consequence of social and economic structures,
such as power and inequality. Negative behaviours, such as in health,
may often be seen as a consequence of structural factors, such as income
inequality. The very reason behaviours are embedded, as discussed in this
chapter, is often the result of long-term social and economic factors and
even biology. These inequalities may be influenced by epigenetics, being
shaped by the environment and then passed down the generations. There
is a complex chain of causes and effects, which reformers should be
aware of, and that includes behavioural public policy-makers. But those
working on behaviour change do not ignore the important structural
determinants of public policies. Behaviour change reformers wish to
address what can be done in spite of the structural causes, by examining
what choices individuals could make in certain situations, so as to get an
advantage for themselves, or to understand a social or policy goal. These
small changes may operate at the margin, maybe for only a few
individuals, or just change outcomes in a small positive direction; but
when summed over time and supported by repeated interventions, which
may interact with other policy levers, they can lead to significant social
improvements. Policy-makers can adopt a form of radical incrementalism
to effect behaviour change over the long term (Halpern and Mason 2015).
That said, there is no doubt about the scale of challenge for policy-
makers when addressing behaviour change, given the structural deter-
minants of the problems, that so many aspects of behaviours are

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 16 / Date: 13/12
JOBNAME: John PAGE: 17 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Behavioural public problems 37

embedded into social networks, and the resistance people commonly

display to any incentive or message that might cause them to change
long-held habits. The question is whether an intervention has enough
power or salience to effect behaviour change, whether it is meaningful
for the individual and whether it can realistically be achieved. For that,
reformers need a theory of behaviour change that is attuned to the
psychological make-up of individuals so that government policy is more
salient and can overcome internal barriers and hurdles. Such interven-
tions need to address collective action problems, as individual behaviour
links to other behaviours. Policy-makers therefore need a powerful theory
to shape their interventions. The next chapter shows how behavioural
economists have provided such a theory.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 02-Chapter2PJED /Pg. Position: 17 / Date: 13/12
JOBNAME: John PAGE: 1 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

3. The behavioural revolution in the

social sciences
It is important to start this chapter with the acknowledgement that the
study of human behaviour is at the heart of social enquiry, a topic that
appears in the first published thoughts about society, politics and the
economy. As soon as intellectuals started thinking about society, they
considered human behaviour as an important object of study. Social
scientists in particular have sought to understand what drives human
behaviour, right from the founders of sociology and economics, such as
Emile Durkheim, who wanted to explain the prevalence of suicide in
modern societies. Almost all forms of social inquiry involve a consider-
ation of behaviour in some form, and even studies on what causes social
and political attitudes entail a consideration of how those attitudes are
affected by behaviour, which in turn affects attitudes. Yet there is
something about the contemporary period that has heightened the aca-
demic interest; perhaps the prominence of the behavioural public prob-
lems highlighted in the last chapter has stimulated academics. Most
importantly, researchers in the social sciences have managed to produce a
more tractable set of theories and models that are of direct use outside
the academy, which started from some esoteric research questions and
models, of interest only to a handful of intellectuals, but which over time
gradually developed into a larger body of knowledge.
This academic agenda is the field of behavioural economics. Over a
short period of time, its ideas became influential and entered the
mainstream of academic thinking, so laying the foundations for its later
influence in public policy. In understanding the influence of nudge and
behavioural public policy, it is important to find out how the ideas of
behavioural economics emerged and why they took the course they did.
This was not just the neutral process of scientific endeavour, but one
where ideas came to prominence because they had found their right time
and the appropriate context in which to emerge, and took a specific form
because of the path dependency that affects the production of academic
knowledge. The course the research took was influenced by an important
debate within the discipline of economics about the veracity of rational
actor models, which happened in spite of its two main protagonists,

38

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 1 / Date: 14/12
JOBNAME: John PAGE: 2 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The behavioural revolution in the social sciences 39

Kahneman and Tversky, being psychologists. By being shaped within

economics, the ideas emerged with a particular character and identity,
which influenced how the whole subsequent research agenda in the
behavioural sciences has developed, including the impact on public
policy.

SOCIAL SCIENCE AND BEHAVIOUR CHANGE

Theories of individual behaviour are core to many disciplines and fields
of enquiry, and occur in long-running research programmes that pre-date
behavioural economics. Psychology as a discipline involves the study of
human behaviour, often measured in the laboratory. A strong movement
in psychology called ‘behaviourism’ stressed that human activity should
be assessed by observing human behaviour (Skinner 1938), sometimes
called ‘radical behaviourism’. In political science, the term ‘behavioural-
ism’ became a framework within which to carry out research on voting,
participation, and interest group politics, addressing the family of behav-
iours that make up citizen and group involvement in politics, and was
associated with the collection of systematic data, such as from surveys
(see Sanders 2010).
The other social science activities that pre-date the behavioural public
policy agenda were done in specialist fields linked to policy and
professional practice, partly because of the employment by government
of social scientists. Work in the field of health psychology looks at the
utilisation of treatments and medicines and shows how individual cogni-
tion affects their success. It links closely with behavioural medicine,
which seeks to alter the underlying behaviours leading to ill health. This
discipline can trace its origins back to the start of the twentieth century,
though in practice owes more to developments in the 1970s, such as the
founding of the Society of Behavioral Medicine and the Academy of
Behavioral Medicine Research.
Researchers in the social sciences have been interested in collecting
data about behaviours without affecting the behaviour itself, such as
through unobtrusive measures (Webb et al. 1966). Another long-running
programme is research into design, which has influenced thinking about
the design of buildings and engineering projects (Norman 1988). There
is also a research field that examines how psychological factors affect
the behaviour of road users, with a focus on what interventions might
change that behaviour (e.g. Rothengatter and Groeger 1998). More
generally, psychologists have long been interested in setting out models
of human behaviour and in seeing how psychological theory can be

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 2 / Date: 10/1
JOBNAME: John PAGE: 3 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

40 How far to nudge?

used to effect changes in public policy (Segall 1976). Another pro-

gramme whose researchers are interested in the influence of messages on
human action is social marketing, which originates with consumer
research in the 1950s (Wiebe 1951–1952). It is a systemic attempt to
understand how to influence behaviours that have a wider social benefit
(Kotler and Zaltman 1971). This research area has expanded in recent
years, and its outputs can be found in specialist academic journals, such
as the Journal of Social Marketing. There are also a number of
prominent research institutes, such as the National Center for Health
Marketing at the Centers for Disease Control and Prevention (CDC), the
Social Marketing Institute (SMI) at Georgetown University and the
Institute for Social Marketing at the University of Stirling. Social
marketing is important for designing public information and promotion
campaigns to change behaviour (which was part of what nudge policies
were reacting against). As was discussed in Chapter 1, behavioural
economics differs through its focus on choice architectures and its more
direct attention to observed behaviour, but it is possible to observe
overlaps with research topics in social marketing, such as message
design (Tapp and Rundle-Thiele 2016). It should also be noted that
researchers in cognitive psychology have advanced knowledge of the
human brain, which did not affect the early work in behavioural
economics, but has now started to influence it and has become part of
the general intellectual flow of ideas. One example is psychological
work on cross-modal space, which is about the way in which sense
mechanisms interpret the world and give signals (see Spencer and Driver
2004). There are different parts of the brain that receive such messages,
which can be measured using scans. By understanding the neuro-
pathways, it may be possible to produce designs that appeal to particular
senses, such as designs of plates for suppressing hunger, as an aid to
diet. There is a wider set of applications, such as interventions that can
reduce smoking, which need careful designs to find out whether they are
effective (Varazzani 2017).
Health psychology, social marketing and brain sciences are all long-
running research programmes, which came about independently of the
current behavioural revolution, and continue in any case, inspiring the
agenda of behavioural public policy, and probably will carry on existing
irrespective of what happens to nudge and nudge units. In fact, research-
ers in these fields often think that those who advocate the new
behavioural agenda neglect existing research programmes and do not
give due credit to what has gone before. Nonetheless, what is interesting
about the current behavioural agenda is that it is more high-profile. It has

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 3 / Date: 10/1
JOBNAME: John PAGE: 4 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The behavioural revolution in the social sciences 41

the ability to operate across disciplines, incorporate many sub-disciplines

and be accepted by policy-makers and pundits. Nudge and behavioural
public policy’s prominence can increase the visibility of other research
programmes, such as social marketing, and ensure the work of psychol-
ogists, who are employed in government or as part of official evalu-
ations, is taken much more seriously. But the argument for the import-
ance of behavioural economics is not just that it is a good label for a
more psychological approach to understanding public policy, but that the
approach blends psychology and economics in a novel way, which
generates powerful insights for policy-makers.

THE RATIONAL MODEL

The reason why behavioural approaches are controversial owes much to
debates within economics, which has a claim to be the dominant social
science paradigm because of its link to and legitimation of wealth
generation. Economists use mathematical models to theorise about eco-
nomic relationships; these have a high degree of sophistication and their
implications are tested using advanced statistics, creating traction in a
world where scientific models hold sway in the generation of knowledge
in the academy and in government.
The fundamental building block in economics is the concept of
individual utility, which individuals seek to maximise through their
choices. To understand the efficient selection of such choices, economists
typically adopt a version of the rational model, where individuals have a
clear objective to maximise, which is stable over time. Individuals can
calculate the implications for realising their objective in terms of costs
and benefits, and then can select the option that best meets it. The
attraction of the model is that it offers micro-foundations for understand-
ing human action in that individuals are seeking to realise preferences
determined outside the economic model. There are system-level advan-
tages in that individual welfare can be advanced by these choices, and
can be aggregated for society as a whole through the expression of
preferences and in market exchanges.
If individuals are assessing costs and benefits in making their deci-
sions, then government policy can be adjusted to promote the collective
benefit. Becker (1968) developed the economic approach to public policy
when considering the costs and benefits of crime. Becker argued that, if
people considering committing a crime rationally weighed up the costs
and benefits, policy-makers could design a set of laws that could alter

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 4 / Date: 10/1
JOBNAME: John PAGE: 5 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

42 How far to nudge?

these costs and benefits. By changing penalties, certain kinds of crime

could be reduced, for example. As a result of these guided choices,
individual citizens would make decisions that in aggregate are closer to
following the public interest. The model provides a clear justification for
a paternalist approach to designing public policy, but it needs rational
individual choices to be in play.
The power of the rational model is that it offers clear guidance to
policy-makers about how to design a range of policies, not just laws and
for crime, but across the whole gamut of tools of government and for any
field. The model also tells decision-makers how to allocate public
policies and finances efficiently. It reaches back to the utilitarian ideal
that government is there to promote the greatest amount of happiness and
welfare in society. By using its resources to alter costs and benefits, the
politicians or bureaucrats can guide policy choices to improve welfare
overall. Individuals reveal their preferences, and government can use
these choices to produce beneficial outcomes. Given the complexity of
public policy, it has an attractive simplicity that helps policy-makers
solve problems.
A practical example helps understand how the economic model works.
Imagine that the manager of a public housing estate has a small budget
to prevent crime and has to make a choice about what to do with it,
such as whether to invest in crime-prevention technology to stop
break-ins, provide better street lighting to ensure people feel safe, or
commission a public information campaign to increase the public’s
vigilance, or fund more street patrol to deter criminals. Using the
rational model, the manager can decide to invest the budget on the basis
of known calculations on the part of the residents of the estate and the
likely activities of criminals. With a certain amount of expenditure on
better window locks, some criminals may decide it is not worth the
bother breaking in and decide to stay at home (or carry out the crime
elsewhere). The manager can find out how much expenditure leads to a
given reduction in crime; similarly, a certain amount of spending on
lighting can change behaviour, and similarly with all the other
measures, even expensive ones such as street patrols. The manager
can then work out where to spend the fixed pot of money to the best
effect, optimising the use of resources. The manager is being rational
in allocating resources based on criminals being rational in responding
to the incentives. And, to be clear, it is a behavioural model too,
but one of a certain kind that the current generation of behavioural

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 5 / Date: 10/1
JOBNAME: John PAGE: 6 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The behavioural revolution in the social sciences 43

economists would wish to question or at least modify. However, it might

be possible to keep a version of the rational model but incorporate biases
on the part of citizens.

CRITIQUES OF THE RATIONAL MODEL

Recent studies of behaviour change modify the assumptions of the
rational model, as they take into account human biases and heuristics
when individuals assess information, although this insight is not particu-
larly new. It goes back to the founders of economics. The older field of
political economy operated with a complex account of individual behav-
iour (Oliver 2013a: 1), not least the insights of the classic theorists in
political economy, such as Adam Smith, as well as earlier pioneers, such
as Vernon Smith (Cartwright 2011: 5–9). In fact, the rational model was
not fully present at the foundation of the study of economics, and was
not developed until the neoclassical revolution of the late nineteenth
century.
The focus on limits to rationality remains a core idea in psychology, in
particular that human cognition has biases that prevent the rational
weighing of options (see Sutherland 1992). Research on judgement that
continues today in mainstream psychology (Newell et al. 2007). The
critique of rational action has been very influential in other disciplines. It
is possible to go back to the work of the influential administrative
theorist Simon (1947) to find an attack on the rational model based on
limited cognitive capacity. Simon offered a new model that shows how
individuals follow informal decision-rules. They satisfice (that is take a
course of action to get to a goal based on minimal cognitive effort) based
on following past behaviours and economising on the costs of infor-
mation searching. One can see the modern twist in the title of Simon’s
(1955) paper ‘A behavioral model of rational choice’. Simon was very
interested in using the research of psychologists to help him understand
human behaviour, which linked to the research programme in the
Carnegie School. Such ideas of the limited cognitions and bandwidth of
human beings were central to the study of public administration, such as
in the work of writers who saw an incremental pattern to human
decision-making (Lindblom 1959), which influenced a generation of
students in public policy and turns up in the language of today’s
behavioural policy-makers (Halpern and Mason 2015). Such a model
underlies the influential book Essence of Decision (Allison 1971), about

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 6 / Date: 10/1
JOBNAME: John PAGE: 7 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

44 How far to nudge?

the Cuban missile crisis of 1962, in its critique of the rational actor and
preference for an organisational process model. Even critics of incremen-
talism, such as Jones (1995), rely on a model that assumes that people
vary in their attention to issues not in proportion to their actual threat, but
in relation to the framing of the issue by the media and the feedback of
many other influences, creating large shifts in policy, what are called
‘policy punctuations’.
In spite of the influence of these models of human behaviour, they
stayed largely within the academic study of public policy and public
administration rather than in economics and other applied social sciences.
The world of economics as it was put forward by the Chicago economists
moved ahead from the 1950s with more sophisticated accounts of
rational action. Rational choice models arguably became more visible in
the 1980s as the rational foundations built into increasingly complex
mathematical models. The world of evaluation and policy advice also
became dominated by versions of the rational model, partly from the
dominance of economics as a paradigm, and also from the elegant and
practical way in which economic models helped policy-makers achieve
their ends, such as through cost–benefit analysis.
It is important not to overstate the impact of a simple version of the
rational model. As ever, the intellectual history is complex, with even a
Chicago economist, Milton Friedman, questioning some assumptions of
the rational model, as well as other paradoxes being discussed in the
1950s and 1960s (Oliver 2017: 16–30). Economists and others have
usually been careful to claim that the rational model only offers a stylised
version of human behaviour. It is a simplified model, and is not intended
to be real. It should be used to clarify assumptions, to produce fresh
accounts of social and economic processes, and to generate claims and
predictions that can be tested in the real world in ways that lead to the
formulation of more complex formal models (see Dowding and King
1995).
In spite of these careful claims, the rational model has been questioned
or modified, as it is possible to question individuals’ basic skills at
solving problems or acting rationally, which concern the level of cogni-
tive capacity involved. When supported by a framework of pre-existing
choices, people might do fine in assessing them, but when faced with
something new they do not perform so well. Relatively basic experiments
show this, such as Schwartz’s experiments on choice, whereby increasing
the number of choices available, such as the number of brands of jam,
makes individuals hardly able to make effective decisions (Schwartz

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 7 / Date: 10/1
JOBNAME: John PAGE: 8 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The behavioural revolution in the social sciences 45

2005). Individuals may be thought of as ‘scaffolded’ by existing proced-

ures and routines, which don’t work at all well in new situations (Heath
and Anderson 2010: 233). The rational model might do fine to explain
marginal changes in stable situations in which individuals find them-
selves, but if the rules of the game change then individuals find it hard to
cope. Across a range of contexts, empirical tests of the basic propositions
of rational actor theories do not perform well. In political science, the
weight of the evidence for core political behaviours, such as voter
turnout, do not confirm the predictions from a simple rational choice
model (Green and Shapiro 1994). Other explanations of human behaviour
need to come into play to account for voter choice, even though it may
involve elements of rational calculation. Rational explanations and formal
models do not necessarily fall by the wayside; nor do behavioural
findings hinder the application of formal models to understand human
behaviour, as they can incorporate human biases and mental short cuts.
However, what is produced is a more nuanced and context-driven account
of human behaviour, where rational action plays a part but needs to be
integrated with a behavioural approach.

THE FOUNDING OF BEHAVIOURAL ECONOMICS

The development and current prominence of the psychological approach
in economics and public policy owe much to the work of Kahneman and
his collaborator Tversky (see Kahneman 1973, 2011; Kahneman and
Tversky 1979; Kahneman et al. 1982), whose work over several decades
influenced a large amount of research in economics and public policy
(see Lewis 2016). The scale of their influence is partly because they
worked for a long period of time in this field doing a series of
experiments that gradually added to knowledge. Importantly, they pub-
lished some of their work in top economics journals, and also worked
closely with or communicated with leading economists, so influenced
thinking from within the world of rational choice models rather than
from outside. Their work was never seen as a threat to the old orthodoxy,
more as adding stimulation and challenge. Partly for this reason, Kahne-
man won the Nobel Prize for Economics in 2002, which both reflected
and consolidated his reputation. Jones et al. (2013: 10), citing Sent
(2004), argue that the reason that Kahneman and colleagues got such
recognition was because they spoke the language of rational thinking,
which was their starting point, rather than working with a completely
different intellectual framework. Oliver’s (2017) history of behavioural

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 8 / Date: 10/1
JOBNAME: John PAGE: 9 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

46 How far to nudge?

ideas in economics shows that economic theorists in the 1950s had

already made a number of tweaks and modifications of rational models
by exploring paradoxes, which made economists receptive to ideas from
psychology.
As psychologists, Kahneman and Tversky worked on the field called
‘judgement and decision-making’. They developed the influential pros-
pect model whereby individuals overweight losses and underweight gains
because of the psychological process at work (Kahneman and Tversky
1979). Significantly their paper was published in the prominent economic
journal Econometrica and was set out in terms accepted by economists,
in particular in equation form. Their approach links strongly to the
endowment effect, which seeks to explain why people tend to trade much
less than economic theory predicts (Thaler 1980; Kahneman et al. 1990).
People value things just because they own them and do not want to lose
them.
Tversky and Kahneman (1974) also developed the important term in
behavioural economics of ‘anchoring’, which shows how the way in
which information is provided, in particular the content of the first piece
of information, influences current choices. They carried out a series of
experiments that show that even people schooled in statistics were not
able to solve basic arithmetic tasks. Another common term used in this
field is ‘mental accounting’, which captures the limits people face when
trying to calculate decisions which are quite complex in their heads and
where biases creep in through the use of short cuts. People tend to think
in relative terms rather than absolutely when considering these decisions
(Thaler 1999). They will treat money in terms of its use or its origin,
making decisions based on where it came from or where it is going rather
than in absolute terms. Money is not seen as fungible, so people might
treat a small win as cash to be spent, but a large win will be saved in
total. Even experienced statisticians tend to focus on results which could
not be supported by the data, believing in findings that were generated
from small numbers that could not be generalised (Tversky and Kahne-
man 1971).
What is striking about Kahneman and colleagues’ work is the care in
the way it has been advanced, over a long period of time, which may
have masked the underlying radicalism of the ideas, and maybe they even
hid the radicalism from themselves. It was in the course of a long
personal friendship, as Lewis’s (2016) history recounts, that their ideas
developed gradually. They questioned some of the axioms of economics,
those that form into the rational actor model. These ideas and heuristics
disrupt the framework of economic policy and undermine the assump-
tions behind claims of welfare efficiency. Economics assumes stable

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 9 / Date: 10/1
JOBNAME: John PAGE: 10 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The behavioural revolution in the social sciences 47

preferences, but in their work preferences may be shaped by short-run

considerations and how issues are framed. As a result, many of the
predictions of economic models are disrupted. However, the way the
behavioural agenda unfolded did not highlight these radical implications,
partly because academic work in economics had already anticipated
them.
Moreover, behavioural economics does not need to be inconsistent
with self-organised efficiency, as envisaged in the neoclassical model.
Sugden (2004) shows that it is still possible to have efficiency even if
contracts take place between non-rational consumers. Unstable or in-
coherent preferences can still yield efficient outcomes. Maybe because
relaxing basic assumptions is a familiar procedure in economic analysis,
behavioural economics can slip into the mainstream without much
controversy. Ironically, the influence of Kahneman and Tversky in
psychology has been more muted, with psychologists questioning the
findings of some studies through the failure to replicate them, such as
some of the priming studies (Engber 2016), though in fact the studies that
are hard to replicate are mainly those reported in Thinking, Fast and Slow
(Kahneman 2011) rather than the original experiments Kahneman carried
out with Tversky.

BEHAVIOURAL ECONOMICS ENTERS THE

MAINSTREAM
In the 1980s and 1990s, influenced by the work of Kahneman and
Tversky, and other founders, such as Thaler, a new generation of
behavioural economists started work on discovering how biases compli-
cate economic models. These ideas gained momentum in the 1980s and
1990s. The movement started out in the work of a small circle of
scholars, who benefited from their location in elite US universities, and
the support of influential institutions, such as the Russell Sage Foun-
dation. These scholars brought in younger students, who became the new
vanguard. It was a quiet revolution, where the links between psychology
and economics were gradually formed, as Thaler describes in his account
of the making of behavioural economics (Thaler 2015). In small gather-
ings and through exchanges of papers, Kahneman was able to challenge
directly some of the assumptions of economics, such as in his Asian
disease test (Tversky and Kahneman 1981), presented in front of influ-
ential economists (Thaler 2015: 159–160). Thaler was invited to contrib-
ute a regular column, ‘Anomalies’, to the Journal of Economic
Perspectives (later collected in his 1994 book The Winner’s Curse).

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 10 / Date: 10/1
JOBNAME: John PAGE: 11 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

48 How far to nudge?

Rabin’s (1998) paper on psychology and economics makes the case

for behavioural economics most strongly and points out the challenges,
in weak and strong form, to economic theory in general (see also
DellaVigna 2009). The objective was to introduce psychological realism
behind economic models as a way that improves them rather than to
overthrow the project and methods of the discipline of economics.
Economics remains characterised by methodological individualism, for-
mal or mathematical models that build from assumptions, and then
careful tests of the implications of these models using advanced statistics,
all of which remain in place in behavioural economics.
As time went by, models of individual behaviour became more
sophisticated. Theories of inter-temporal choice, which are about the
making of a decision and understanding the consequences of it over time,
have been important (Ainslie 1975; see the review by Berns et al. 2007).
Such choices are subject to bias, and hence what is called ‘hyperbolic
discounting’, which can be specified mathematically, and these papers
again appear in top economics journals (e.g. Ainslie 1991). More
informally, experiments can show that people make choices based in
heuristics about whether to accept a sum of money now or a different
sum later, which generally show a preference for a more immediate but
smaller sum now. Such lack of control is largely seen as the reason why
savings do not take place as much as conceived in classical economic
theory (see Harris and Laibson 2002).
Another concept that was developed at this time was ‘preference
reversal’, which was formulated by Tversky and Thaler (1990), though
was already set out in the work of Grether and Plott (1979). The concept
tries to account for why people change behaviours, revealing different
preferences during the course of making choices, which often follow
from the procedural set-up of those choices, influenced by the scales and
units people face. As with other concepts in behavioural economics, a
research programme has been set off, with the idea being developed by
academics such as Loewenstein (1987) and List (2002).
These studies formed part of a more general movement in economics,
whose members started to use field experiments to test real-world
interventions, moving away from laboratory experiments (see Levitt and
List 2007; Banerjee and Duflo 2014). This shift from using the more
psychological methods of the laboratory to field experiments has import-
ant consequences, as the following chapters indicate. Behavioural econo-
mists had started to engage policy problems through testing interventions
which were still interesting theoretically but also tried out policy initia-
tives in the real world.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 11 / Date: 10/1
JOBNAME: John PAGE: 12 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The behavioural revolution in the social sciences 49

A key idea that also had implications beyond the academy was of
status quo bias and inertia. This is the idea that, if people want to
conserve cognitive energy, they might simply choose the status quo
option, which may mean no change in behaviour even when it is clearly
in their interests to change, and of course links to the heuristics approach
discussed earlier. It also follows from Kahneman’s idea of loss aversion
as a reason why it might be better to stick with the status quo, as it does
not entail risk, or at least the perception of risk. Thaler was a notable
exponent of this concept, as he was the co-author of a famous article by
Kahneman et al. (1991). Even the original article, published in an
eminent journal, was to an extent homespun, with its starting example of
a friend’s wine-buying decision. The default option is introduced, again
in easy-to-understand examples of offers made by an electric power
company and US state laws on car insurance. The authors argue for the
importance of these findings by saying they violate the assumptions of
stable preferences, yet the style is not aggressive and rests on reason-
ableness and plausibility, as well as logical argument and evidence. As
well as conservatism, people suffer from overconfidence or optimism
bias, so they overstate their skills and make rash decisions (Sunstein
2014a: 48). The tendency to come up with a range of propositions that
appear to contradict each other is one of the characteristics of behav-
ioural sciences. Rather than elaborating a general theory, research seeks
to find the exact context in which a psychological trait will have an effect
on decision-making.

CHANGES IN MAINSTREAM ECONOMICS

At the same time as these behavioural concepts were being developed,
scholars in other branches of economics started to look at the impact of
human-centred motivations that do not strictly follow the rational model.
In tests of the propositions of game theory, it had started to be accepted
that players do not choose options that benefit themselves, but are
prepared to share the proceeds (e.g. Camerer 2003). Economists started
to accept that individuals display altruism in their dealings with others
and that they like to reciprocate, which forced theorists to modify the
rational model. These findings can be complemented by the way in
which economists acknowledge the idea that individuals may trust each
other in ways that do not follow the predictions of the tragedy of the
commons or prisoner’s dilemma games. This body of evidence from
laboratory experiments on game theory is reviewed in Ostrom’s (1997)
address to the American Political Science Association, which essentially

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 12 / Date: 10/1
JOBNAME: John PAGE: 13 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

50 How far to nudge?

says that a behavioural approach is needed to understand how collective

action works rather than the predictions of the prisoner’s dilemma game.
As Cartwright writes, ‘game theory was instrumental in the birth of
behavioural economics’ (2011: 9). This influence was because revisions
to the standard economic model were first explored in the laboratory,
then economists started to test such propositions in the field, for example
Gneezy and List’s (2006) investigation of gift exchange in labour
markets. That Ostrom, a political scientist, won the Nobel Prize for
economics is also evidence about the change in direction within the
discipline, which of course does not mean that fundamental ways of
understanding the world and core methods have been overthrown in
economics, but that the discipline had broadened out to take on board
non-economic motivations, which then form part of its intellectual
project. Some of these themes were already buried in arcane debates and
papers, but they were ready to be rediscovered.
Behavioural economics fits well with this non-radical reorientation of
the discipline, which has the consequence of protecting economics from
criticism and preparing it for a more challenging period, following the
certainties of the 1980s and 1990s, after the financial crash of 2008. In
this way, economics broadened out so as to retain its distinctiveness and
ensure its intellectual superiority through formal modelling, and the use
of advanced statistical methods to test the implications from the models.

THE SOPHISTICATION OF BEHAVIOURAL

ECONOMICS
Consistent with the mainstreaming of behavioural economics is the
greater use of formal models and the increased sophistication of the tests
applied. For example, Park and Sabourian (2011) use complex math-
ematical models to understand herding in financial markets, which can
help explain their instability. More recent developments show the power
of behaviourally inspired economics to offer a more integrated account of
how the economy works, which is shown by a recent move to behav-
ioural macroeconomics, which harks back to Keynes’s model, but is fully
updated and enhanced with the insights of behavioural economics (see
Akerlof 2002; Gabaix 2016). The macroeconomic model includes behav-
ioural findings on savings and the behaviour of asset markets. It uses
concepts from behavioural economics, such as loss aversion, which can
then feed into macroeconomic models.
Another route to advancing behavioural economics is through greater
understanding of individual psychology, between aspects of individual

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 13 / Date: 10/1
JOBNAME: John PAGE: 14 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The behavioural revolution in the social sciences 51

behaviour that focus on planning and other inclinations to be a doer or to

act out of habit. This again is a neat way to retain aspects of rational
actor models, but to suggest that behaviour is contingent and may depend
on prompts. So why do people procrastinate when they know that acting
immediately will benefit them and save the costs of procrastination
(Anderson 2016)? Then there is Thaler’s planner and doer framework,
whereby individuals might not be able to exercise self-control (Thaler
and Shefrin 1981), so they may need a device or a commitment to help
them achieve their goals. A similar idea appears in Kahneman’s (2011)
more recent work of synthesis, that of two systems of thinking: system 1
is fast, instinctive and done with gut feelings; system 2 is more
considered, deliberative and slower. Understanding these two systems
provides a route to understanding the general operation of individual
psychology in economics and is a way to understand generally how bias
operates at times to modify the assumptions of an earlier generation of
economic theorists. But it is also fair to say that, even though these
approaches are highly attractive intuitively and one can imagine indi-
viduals moving between the kinds of thinking – planner and doer – of the
two systems, there is no meta-theory that tells us when to expect
transitions from one approach to another, so there is no predictive
integrative model that can guide the whole system. This means that
behavioural economics broadens out to make predictions that are contin-
gent rather than universal. Given that behavioural economics tends to use
ideas from psychology rather than innovate itself, it tends to replicate
psychology’s tendency to subdivide processes into many different causal
mechanisms and heuristics, which can become very numerous as differ-
ent traits are identified and then given their respective labels. In this way,
which is not stated, behavioural economics is slightly different to the
classical variety in spite of sharing the basic approach, in that classical
economics built up its understanding of economic relationships from
micro-foundations. The elegance of economic theory gets complicated in
its behavioural variant.

ACADEMIC APPROACHES TO POLICY PROBLEMS

As well as pursuing theoretical developments, behavioural economics
started to address policy problems in concrete areas, which is entirely
consistent with the grounded way in which behavioural economists think
about the world and the general programme of behavioural economics,
which is to find contexts where specific mechanisms explain human

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 14 / Date: 10/1
JOBNAME: John PAGE: 15 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

52 How far to nudge?

behaviour. Even in advanced economics papers, there are often policy

prescriptions.
Thaler and Benartzi (2004) have pioneered this practical focus by
examining pensions choice, testing whether defaults encourage employ-
ees to sign into a pension scheme. The authors carried out interventions
with private sector companies to see if the default option and giving the
default at the same time as a pay rise would encourage people to make
savings for a pension, as a way of addressing people’s bias in not
wanting to forgo their current earnings. The promising results are useful
tests of status quo bias, but also have a practical application. Another
example is taxation, whether making local sales taxes more salient affects
consumer behaviour, which has been tested experimentally (Chetty and
Saez 2013). Health is another big area, such as in the work of Loewen-
stein et al. (2013) on the lack of public understanding of health
insurance. Formal work has looked at the question of moral hazard in the
purchase of health insurance, and that there is a human tendency to
undervalue health as a result. There are policy implications of adjusting
co-payment and of adjusting the nudges to respond to perceptions of
behavioural hazard (Baicker et al. 2015).

THE DIFFUSION OF BEHAVIOURAL ECONOMICS

The final set of developments has been the expansion of the ideas of
behavioural economics into other disciplines, which makes the research
programme genuinely interdisciplinary. One key area is law, whose
scholars took an early interest (Sunstein 2000), and which has developed
into a subfield of its own (Zamir and Teichman 2014; Alemanno and
Sibony 2015), which is about the reach and effectiveness of legal
instruments, and also concerns legal argument about giving reasons and
justifying autonomy (see Chapter 7). Other developments are the study of
behavioural public policy, which is the investigation of how behavioural
instruments can have an influence on policy outcomes (Oliver 2013a).
There is behavioural public administration, which is based on making a
closer link between psychology and public administration, and is
designed to help scholars better understand bureaucratic behaviour (Tum-
mers et al. 2016). Development economists have embraced behaviour
change theories and evidence, as shown by the World Bank (2015) report
cited earlier.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 15 / Date: 10/1
JOBNAME: John PAGE: 16 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

The behavioural revolution in the social sciences 53

CONCLUSION
This review of behavioural economics, as well as studies of human
behaviour in other parts of the social sciences, shows a remarkable set of
intellectual developments. What has happened is that the hegemony of
rational actor models in economics, and in other disciplines, has less-
ened, and alternative accounts of motivations for human behaviour have
become more commonly articulated and accepted. The argument in this
chapter has sought to acknowledge the pedigree of psychological
approaches to understanding human behaviour and the long-standing
critique of rational actor models, which have been conventional wisdom
in the study of public policy, in work going back to the 1940s. But the
key feature of the current period of debate, which has been more in
evidence since the late 1970s, has been its occurrence inside or proximate
to the discipline of economics.
It has been a quiet revolution, happening over many decades. It has
been easier than might have been expected, because economists have
always discussed different kinds of theoretical models and have been
open to puzzles and contradictions. Economists usually stress that their
models depend on simplifying assumptions so as to facilitate mathemati-
cal modelling, rather than believing the assumptions are necessarily true,
as they are often designed to be relaxed at a later stage. Partly for these
reasons, key papers in behavioural economics appeared in the key
academic journals in the discipline, such as the American Economic
Review, and were written by faculty in high-ranking departments, such
as from the universities of Princeton and Chicago. The economics
mainstream matured gradually in this period, with economists taking
more of an interest in motivations and findings that arise outside the
assumptions of the rational model, producing theories and research on
new topics such as trust, reciprocity, and altruism, and discussing the
mental processes that behavioural economists had uncovered. Economics
has broadened out to incorporate this line of thinking in a way that runs
alongside or is in dialogue with research using the rational actor model,
reflecting the hidden depths of the discipline as a whole. In the process,
economics retains its distinctive contribution by formalising the propos-
itions that behavioural economists have put forward and testing them
using advanced statistical methods. This integration of behavioural
models into the mainstream is an important development, given the
importance of economics in the hierarchy of forms of knowledge, which
is accepted in the academy and reinforced in the public sphere. Given
the legitimacy of economics, this prominence helped researchers outside

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 16 / Date: 22/1
JOBNAME: John PAGE: 17 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

54 How far to nudge?

economics take the concept of biased behaviours more seriously. Some

of those outside the discipline who adhered to a version of the rational
model could see how it could be adapted in line with the findings of
behavioural economics; others who were more critical could say that
they had been proved right all along. But it was not the case that
economics lessened in its dominance in the academy and in the public
sphere; in fact, the opening up of economics ensures its continuing
dominance and legitimacy, reinforced rather than undermined by these
new ideas.
The redefinition of economics was to have a large effect on the public
realm. As it became more realistic, while remaining rigorous, the
possibility was for a wide-ranging influence on how researchers and
academics sought to improve public policy outcomes through public
action, developments that the next chapter teases out.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 03-Chapter3PJED /Pg. Position: 17 / Date: 10/1
JOBNAME: John PAGE: 1 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

4. Nudge: All tools are informational

now
Behavioural sciences and behavioural economics have entered the main-
stream, at first within academic disciplines, but more profoundly within
the world of media debate and commentary, and thence into the ambit of
policy-makers. As will be documented in this chapter, this translation is
an important development, a case study of how an academic programme
can directly influence public policy. This phenomenon is not automatic.
Indeed, the conventional wisdom is that academics find it hard to make a
direct impact (e.g. Caplan 1975, 1979), for example through official
evaluations. It is only when there are special conditions in place that
academics can position themselves to affect policy choices (see John
2013b). Academics need to understand that it is mainly through the
propagation of ideas that they can have a profound influence (Weiss
1977), which is gradual and long-lasting, and this is the realm in which
advocates of behavioural economics so successfully operate.
The success of behavioural public policy derives in part from the
practical agenda of behavioural economics itself, discussed in the last
chapter, which is because of the realism that the introduction of psychol-
ogy entailed. It helped the consideration of policy questions within
academic research projects. The salience and impact of the new sets of
ideas came from the synthesis of economics and psychology that
modified rational actor models. As a result of the research programme
that emerged, academics soon realised that they had discovered a
powerful set of tools, which were capable of shifting behaviours with
relatively few inputs, so addressing the behavioural concerns discussed in
Chapter 2, but without implying large contributions from government
resources. The other reason for success is that the advocates of behav-
ioural economics were good translators of the science into claims and
propositions that could be easily understood. This chapter is about this
translation, and shows skilful advocacy by the proponents of behaviour
change, who simplified the propositions, created eye-catching summaries,
and made the behavioural cues seem just like common sense. In fact,
they used all the skills of nudge to sell nudge itself.

55

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 1 / Date: 14/12
JOBNAME: John PAGE: 2 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

56 How far to nudge?

THE SIMPLIFICATION OF BEHAVIOURAL

ECONOMICS
Many behavioural economists are talented advocates, not really needing
other translators. They are persuasive characters, used to the public world
of US elite universities, with their easy access to the op-eds of the New
York Times and the Washington Post. The findings they convey are
appealing because they are often based on easy-to-understand tests or
experiments. Taking pot shots at a simple version of the rational model,
replacing it with smart pieces of research that reveal a more human side,
especially for everyday activities and routines, is a good sell for the
quality newspaper market. For example, Johnson and Goldstein’s (2003)
seminal study on organ donations, which showed the importance of a
default, was published in a general science journal, which helped get the
results noticed. Thaler was instrumental in this process of dissemination,
which may have come from his relaxed and convivial personality. He is
known for making colourful asides and wry jokes, commenting on
human frailties, and peppering his talks and prose with funny but
apposite anecdotes. It is no surprise that the pensions default example
should have become so well known, and been communicated so effect-
ively to policy-makers (see Thaler 2015: 314), for example in a contri-
bution to the policy forum in Science (Benartzi and Thaler 2013).
Benartzi is a persistent advocate, who did his own publicity-seeking tours
and gave the obligatory TED talks. He published an advocacy book
(Benartzi 2012) on pensions choice, as well as having a website and a
consultancy. Of course, Thaler made great play of pensions defaults in
his co-authored nudge book (Thaler and Sunstein 2008).

POLICY-MAKERS AND THE BEHAVIOURAL AGENDA

Before discussing the nudge book (Thaler and Sunstein 2008), it is
important to emphasise that the findings of behavioural economics were
reaching outside the academy during the 1990s, as findings started to
emerge and discussion was taking place in the circles that Thaler (2015)
discusses so well in his book Misbehaving. The pensions example has
already been mentioned. The New York Times celebrated the arrival of
behavioural economics as early as 2001 (Uchitelle 2001), mentioning the
work of David Laibson and hyperbolic discounting. The winning of the
Nobel Prize by Kahneman in 2002 generated much commentary (see
Altman 2002).

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 2 / Date: 14/12
JOBNAME: John PAGE: 3 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge: All tools are informational now 57

Academics were not the only ones doing the translating. The classic
intermediators were those working for think tanks and policy institutes
who like to sift through the vast output of academic journals, find out
what research is hot, and then do literature reviews. In the UK, the New
Economics Foundation commissioned such a study. Shah and Dawney
(2005) produced their seven principles, which the authors had divined
from reading the literature reviewed in the previous chapter. They
distilled the findings into a list: other people’s behaviour matters; habits
are important; people are motivated to do the right thing; people’s
self-expectations influence how they behave; people are loss-averse and
hang on to what they consider theirs; people are bad at computation
when making decisions; and people need to feel involved and effective to
make a change. Lists have been important for the diffusion of behav-
ioural economics, and comes from the way in which they draw attention
to particular cognitive processes, and reflect the relative weakness (or
advantage, depending how it is seen) in not having an integrative theory.
Weakness in the pantheon of knowledge can be a strength in the world of
diffusing ideas, because lists are less threatening, and it is possible for the
reader to focus on one or two claims that are appealing rather than feel a
need to accept the programme wholesale. In particular, the subtitle of
Shah and Dawney’s (2005) publication, Seven Principles for Policy-
Makers, makes it clear the audience is people who make decisions in the
public realm. Each section of the publication has a subsection on
relevance to policy-makers. At the same time, and this is another
characteristic of this new applied field of enquiry, the publication refers
directly to academic studies, giving the title of papers and directly
referencing them, for example Frey et al.’s (2004) review of procedural
utility, the idea that a sense of procedural fairness encourages people to
comply, published in what for policy-makers must seem a fairly obscure
academic journal. The publication is replete with summaries of such
academic reviews, with academics doing the summaries, which are in
turn summarised by people who work in the policy world, as this needs
different skills of translation. This simplification of academic work
makes it more easily digestible, and it needs a particular aptitude, as
communicating and summarising the results of academic research require
different abilities to those needed to publish an academic article (see
Flinders 2013). The main question to ask is whether the simplification
loses something of the potency of the ideas, as these documents tend to
concentrate on the less threatening proposals.
The Strategy Unit in the Cabinet Office, under prime minister Tony
Blair, commissioned its own report on behavioural economics. Blair had
hired Halpern (see Halpern 2015) to work in the unit. He had become

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 3 / Date: 10/1
JOBNAME: John PAGE: 4 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

58 How far to nudge?

known for his work on social capital, and he had absorbed insights on
behaviour change. He has a background in social psychology, but also
worked directly with policy-makers in the Options for Britain project
(Halpern et al. 1996), which reviewed policy options shortly before the
1997 election, and in Nexus, the online think tank. Halpern appears later
as the director of the Behavioural Insights Team (BIT). Working for
Blair, Halpern and colleagues produced a report looking at the policy
options from research on human behaviour (Cabinet Office – Prime
Minister’s Strategy Unit 2004). The report begins with the importance of
behaviour for policy (similar to the argument articulated in Chapter 2).
Central to the report is a section on theories of behaviour change, which
contains reviews of the rational actor model. Tversky and Kahneman are
cited alongside other key authors. Then appear key concepts in behav-
ioural economics, such as scarcity, availability, and prospect theory, with
the terms ‘loss’ and ‘gain’, referring to the classic paper by Kahneman et
al. (1990). For a government report, there is much use of academic terms,
such as ‘fundamental attribution error’ and ‘social cognitive theory’. As
Halpern discusses (2015: 30), the report met with controversy; but it
illustrates how close the relationship between the academy and govern-
ment had become, with Kahneman visiting the Cabinet Office in this
period (Halpern 2015: 32).
At the same time, other parts of UK central government produced
reviews of behaviour change research. Government Social Research
(GSR), which coordinated research across government, commissioned a
report, published in 2008, which looked at the literature on behavioural
economics (Darton 2008), as well as producing a practical guide. Defra,
the former environment and farming ministry, produced a series of
research reports, which were essentially literature reviews of the behav-
ioural science literature (Pike et al. 2010).

THE NUDGE BOOK

The book by Thaler and Sunstein, Nudge: Improving Decisions about
Health, Wealth, and Happiness (2008), is the ultimate exercise in
translation. The argument is that policy-makers can be choice architects:
they can influence the choices people make by redesigning procedures
and institutions in ways that privilege socially beneficial outcomes.
Choice architects can take advantage of the argument that people tend to
choose the status quo or default option, which of course comes out of
status quo bias claims in the behavioural economics literature. People
tend to be cognitive misers, maybe even lazy, so they tend to go for the

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 4 / Date: 14/12
JOBNAME: John PAGE: 5 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge: All tools are informational now 59

easiest option or one that does not entail any change. In many circum-
stances, the no-change option can be the least beneficial. People end up
defaulting to choices that benefit them the least, such as no pension
scheme or a poor insurance renewal policy, and/or one from which
society does not benefit (e.g. not donating organs). But what if govern-
ment could organise procedures and rules so that the default option was
the one that had the most personal and societal benefit, such as the best
pension or insurance scheme, or where donating organs was the default
activity for anyone renewing a car licence?
The nudge aspect is the idea that such interventions are light-touch;
that is, they guide individuals to where they want to go or would want to
go if they gave the option enough thought, but does not force them,
enabling people to keep their liberty intact. The interventions still act
paternalistically, which is captured by the apparently oxymoronic term
‘libertarian paternalism’, which will be discussed in Chapter 7. Keeping
an element of freedom is key to Thaler and Sunstein’s definition of
nudge:
any aspect of the choice architecture that alters people’s behavior in a
predictable way without forbidding any options or significantly changing their
economic incentives. To count as a mere nudge, the intervention must be easy
and cheap to avoid. Nudges are not mandates. Putting fruit at eye level counts
as a nudge. Banning junk food does not. (Thaler and Sunstein 2008: 6)

The focus is on choice and how policy-makers respect the basic values
and long-term preferences of those who are nudged.
The nudge book (Thaler and Sunstein 2008) walks the reader through
other terms in behavioural economics, such as ‘anchoring’, ‘availability’,
‘overconfidence’, ‘gains’, and ‘losses’, yet again showing the love of lists
in behavioural public policy. As the book unfolds, examples and case
studies appear from the world of investment, organ donations, energy
saving, and many others, ending up with a recommendation for a dozen
nudges, so directly speaking to policy-makers. But what characterises the
book the most is its lightness of tone, the humorous quips, the hint that it
is Thaler who corresponds to the lazy person needing nudging, and then
the pretend jostling between the authors, all of which sugars the
behavioural public policy pill quite considerably. It was no surprise that
the book was a hit, and the funnier of the two authors became a media
star, especially in the UK. A phenomenon had been born, and a thousand
nudge jokes set forth, the word entering the vocabulary of commentators
and becoming something that policy-makers should think about doing.
The book, with its eye-catching yellow elephant on the cover, appeared in

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 5 / Date: 10/1
JOBNAME: John PAGE: 6 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

60 How far to nudge?

newsagents and airport booksellers, the ultimate trade publication. The

book and its positive reception then fed back into academia, which turned
nudge into a veritable industry of new investigations, moving beyond the
formulations of behavioural economics, more interested in how the
rational model could be modified and formalised, into more direct tests
of policy interventions. The question had become less about elaborating a
psychological model of economic behaviour and more about finding out
which psychological process would be best attuned to motivating human
beings to act. The translation had happened.

THE WORLD OF PRACTITIONER GUIDES

In the wake of Nudge, a range of publications emerged, partly influenced
by it, partly doing their own research. One was the popular MINDSPACE
report (Hallsworth et al. 2010). As with the earlier Cabinet Office report
on behaviour and policy, this piece of desk research emerged from the
heart of the UK government machine, commissioned by the Cabinet
Office, and this was still when Labour was in office, which shows that
nudge in the UK pre-dated the Conservative–Liberal Democrat Coalition
government. It partly reflects the enthusiasm of an unconventional and
highly talented Cabinet secretary, Gus O’Donnell, Britain’s leading civil
servant, who was an academic economist and really liked the practical
application of economics. Again, Halpern was involved, this time from
outside government, based in the Institute for Government, as well as
other researchers. It was a thorough review of the academic literature on
behavioural economics, with copious references to research pieces in the
academic journals, and had footnotes galore. But the key to the publica-
tion was the acronym MINDSPACE, which captures the importance of
thinking about mental processes and represents different nudges that can
be chosen depending on the situation policy-makers are faced with (again
replicating the tendency of guides in behavioural public policy to produce
lists of nudges). The acronym covers much of the ground of behavioural
economics: messenger, incentives, norms, defaults, salience, priming,
affect, commitments, and ego. The only one that appears to jar is
incentives, as behavioural economists try to move away from a simple
understanding of incentives (this is where the authors summon in losses,
but presumably they needed a vowel to make the acronym work). The
report, with its many examples, is attractive to policy-makers, but also
has academic credibility (it ended up as a paper in an academic journal –
see Dolan et al. 2012). MINDSPACE quickly became a hit, not as an
airport book, but as a download (the Institute for Government’s most

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 6 / Date: 14/12
JOBNAME: John PAGE: 7 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge: All tools are informational now 61

accessed). The word, like nudge, has become common currency and, as
will be shown in Chapter 5, is still used as a guide by which policy-
makers could apply behavioural insights, in spite of attempts to displace
it with other, shorter acronyms.

NUDGE GOES BACK TO THE ACADEMY

As was briefly discussed in Chapter 3, one impact of the greater policy
interest has been the reorientation of the academy itself to focus on more
concrete policy issues. Academics have always taken an interest in
behavioural policies, but these applications tended to appear in specialist
disciplines or field areas, such as the environment and health. What
distinguishes the new field of interest is that it works across many fields
and seeks to draw behavioural insights in a number of contexts, with the
driving interest being the identification of the psychological cue. A new
descriptor emerged, ‘behavioural public policy’, which denotes a more
serious tone than the enterprise of nudge, and by implication a wider
range of applications. In many ways academics are practising behavioural
public policy in carrying out applied interventions that are of use to the
policy world and in the academy. Advocates of this approach seek to
ensure that insights from psychology predominate in understanding how
policy interventions and bureaucratic routines work in practice.
Behavioural public policy became an option on master’s courses (e.g.
at the London School of Economics). It has been written about in the
quality press. It is represented as part of the ambit of the Behavioral
Science and Policy Association, which acts to translate findings from the
behavioural sciences into policy. It is the title of a new journal published
by Cambridge University Press (see https://www.cambridge.org/core/
journals/behavioural-public-policy) and of an edited book (Oliver 2013a),
the latter with contributions on environmental policy, health, and finan-
cial behaviour as well as more general applications. The Behavioral
Foundations of Public Policy (Shafir 2013) is a more ambitious volume
in the same vein, seeking to provide coverage across many theories and
venues to show the potential for a structured and wide-ranging debate
about the causes and remedies of behavioural public problems.
Some of these academic treatments were a conscious reaction to the
agenda carved out by nudge, partly in praise of it but also keen to ensure
that the terms of the debate remained open to a wide range of contexts
whereby behavioural interventions could apply. Nudge implies that the
space for behavioural interventions is where individuals have choice or
autonomy. Individuals really wanted to do an activity, but through lack of

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 7 / Date: 10/1
JOBNAME: John PAGE: 8 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

62 How far to nudge?

attention or status quo bias did not do it. Several writers have questioned
this limitation. John (2013a) argues that nudges apply across the tools of
government and can be used to improve the informational context within
which the tools of government are exercised. A tool of government, such
as a tax change or a new law, might at first appear authoritative and
capable of implementation only by virtue of the power of the state. But,
as every student of implementation knows, an insight that goes back to
Pressman and Wildavsky’s (1973) famous study, the various levers and
controls that a central authority can deploy are never fully going to work
(see John 2011). There are too many complex chains of command and
autonomous organisations that need to work together. The way to achieve
better implementation is to improve the information environment when
using the tools, which can help their effectiveness, such as the clear or
socially-informed communication of a new law or tax change.
What the nudge agenda adds is a close understanding of how a law or
tool is going to apply in behavioural terms. It can be redesigned so as to
work better, but in keeping with the choices and wishes of the person or
organisation involved. This is because a key part of responding to
government policy is the understanding and appreciation of the salience
of signals, which can vary in the extent to which they offer an impetus to
change behaviour. In this way, ‘All tools are informational now’ (John
2013a). Oliver (2013b) makes a similar argument, but he is interested in
the extent to which a policy change involves regulation or not. Non-
regulation is the world of nudge; but the world of regulation, which can
still use behavioural insights, is called ‘budge’ so as to distinguish it from
non-obtrusive measures, such as the defaults in the Thaler and Sunstein
(2008) formulation. In this way, the debate about behavioural public
policy, behaviour change, and nudge is an expanding canvas designed to
be the academic study of policy effects, that is decisions taken by
policy-makers to change outcomes using procedures and tools of govern-
ment to achieve their ends. Essentially, this is a literature about evalu-
ation, about the possibility of achieving changes in policy outcomes,
albeit grounded in behavioural sciences and theory. The tools approach
shows that nudge can deal with the core problems in public policy rather
than just be about improving communications. The chains of implemen-
tation can be examined using behavioural concepts, identifying the
precise mechanisms that can help make a policy work. Making those
chains work in better ways can then foster a spirit of innovation that
infuses the whole delivery process to make it more efficient and
responsive. This wide-ranging ambition not to leave any stone unturned
becomes part of the radical agenda for nudge that this book takes up in
later chapters.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 8 / Date: 10/1
JOBNAME: John PAGE: 9 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge: All tools are informational now 63

In tandem with nudge, work that uses ideas from psychology, particu-
larly in health, had also started to be simplified, creating guides for
policy-makers and others, maintaining a bridge between academia and
the policy world. Influential has been the behaviour change wheel, which
is a heuristic device invented by Michie and colleagues (Michie et al.
2011), who desk-reviewed a series of frameworks on behaviour change
and synthesised them into a framework, which was first promoted in
academic articles, then with the website and book, with a focus on
influencing practitioners. It uses the COM-B (‘capability’, ‘opportunity’,
‘motivation’ and ‘behaviour’) model to make a rounded assessment of
what tools to use and what potential they have. These tools are called
‘intervention functions’, and relate to the tools of government literature
discussed earlier. The different fields are represented as a wheel, with
processes and tools of government placed in successive circles so
observers can select which choice or set of choices is appropriate for the
situation they face. The wheel becomes an aid to decision-making, but
one rooted in academic knowledge. It probably appeals most to those
who work on health policy interventions, where there is a stronger
grounding of academic ideas and common use of ideas from academic
psychology, but the guide is more complex than many of the other
simplifications.

THE WIDER PHENOMENON OF TRANSLATION

The torrent of publications and outlets of behavioural public policy need
to be seen as part of the wider phenomenon of translation of findings
from the academy into practice. This transfer had always happened, and
right from the start of social science there have been those who have
written about the findings of the academy in an attractive style. The
eminent think tank Chatham House was founded in the 1920s. The social
affairs think tank Political and Economic Planning was founded in 1931,
and went on to become the research organisation Policy Studies Institute
(PSI). In the US, the Brookings Foundation, which has a wide public
policy remit, was founded in 1916. Think tanks provide this role of
translation by providing summaries and inviting academics to present
their findings to policy-makers, and have generally been regarded as
being influential in the policy process, in particular in helping shifts in
ideas (Stone 1996). And so it has been the case with behavioural
economics. In fact, it is possible to argue that the translation phenomenon
has become an industry in recent years, which has benefited the
behavioural sciences. Funding councils and foundations have helped too

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 9 / Date: 14/12
JOBNAME: John PAGE: 10 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

64 How far to nudge?

with their focus on ensuring impact as a condition of grant funding, and

with governments, such as that of the UK, enshrining policy relevance in
their evaluation of the outputs of higher education institutions, for
example the Higher Education Funding Council for England’s (HEFCE)
Research Excellence Framework (REF).
The role of academics simplifying their findings into a book that is
readable goes back to the paperback revolution of the 1960s, if not
before. In fact, Adam Smith’s An Inquiry into the Nature and Causes of
the Wealth of Nations ([1776] 1904) is sometimes regarded as a piece of
translation, although it is the age of cheap publishing and business books
that has seen the explosion. Of course, popular psychology has always
been important. Economics was a latecomer in spite of Keynes and
Galbraith’s early entries in this field. Since the mathematical strides of
the 1980s, there has been a swing back to the more practical concerns,
which has led to this spate of books. An example is Freakonomics (Levitt
and Dubner 2005), which sought to summarise examples of applications
of economics to solve policy problems, with state-of-the-art research
methods and designs (explained simply), often with counterintuitive
conclusions and recommendations. There have been sequels to this book,
a website with resources to read, and a regular podcast for listeners to
download. Of course, some of the insights in Freakonomics come from
behavioural economics broadly considered.
More directly behavioural is the book Scarcity: Why Having Too Little
Means So Much, which summarises research on the impact on behaviours
of scarcity, in particular the impact of poverty, which narrows the range
of choices families can make because of the pressure on cognitions
(Shafir and Mullainathan 2013). The book’s authors translate this work
through the consultancy Ideas42 (see http://www.ideas42.org), whose aim
is to ‘to use the power of behavioural science to design scalable solutions
to some of society’s most difficult problems’. Of particular prominence is
Ideas42, which reflects a general trend in the dissemination of behav-
ioural ideas, which is using behavioural insights to seek to solve
problems of implementing interventions in developing countries, for
example ensuring the behavioural insights apply to the uptake of finan-
cial services so that individuals understand the risks involved, or advo-
cating interventions such as pensions defaults. Reflecting this interest in
using behavioural insights, the World Bank (2015) commissioned one of
its research projects on ‘Mind, Society, and Behavior’. The report has
chapters on poverty and development, household finance and climate
change. It makes the case strongly for using a more psychologically
attuned account of human behaviour rather than the standard economic
model (see Demeritt and Hoff 2015). Behavioural economists have piled

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 10 / Date: 10/1
JOBNAME: John PAGE: 11 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge: All tools are informational now 65

in too, such as Gneezy and List (2013), whose The Why Axis: Hidden
Motives and the Undiscovered Economics of Everyday Life is an account
of how incentives operate in broad and social ways.
Such diffusion can be seen in other fields too. For the environment,
Jackson (2005) reviews sustainable lifestyles, using the tools of behav-
ioural science, though coming to some cautious conclusions about what
may be possible in the environmental field given the entrenched nature of
the behaviours to shift. Another big area for behavioural insights is
design. Architects can use these ideas to reshape public spaces, such as
walkways and public housing estates, which can be seen as part of a
more ethnographic approach to understanding organisations (Kimbell
2015). Similarly, academics who work in the field of social marketing
find their ideas being used as ways to promote behaviour change by
consultancies, such as the National Social Marketing Centre (see http://
www.thensmc.com/what-social-marketing).
Journalists have contributed to the popularisation of behavioural eco-
nomics, for example the Financial Times journalist Tim Harford. His
books on economic problems and solutions (e.g. Harford 2005), news-
paper column, and BBC Radio 4 programme More or Less, which he
presents, contain many examples from behavioural economics. The large
number of science policy writers have also contributed to the debate. In
these popular science books is the idea that the evidence base matters,
such as from the use of experiments. People who write about science
believe that scientific methods should have a wider application in public
policy (Henderson 2012). A similar approach is taken by the critic of the
use of trials in medicine, Ben Goldacre, a doctor who takes the
pharmaceutical companies to task for poor testing (Goldacre 2008),
believes that trials offer a way to test claims in public policy. He was a
co-author of the influential guide to trials published by the Cabinet Office
(Haynes et al. 2012). These books show the current importance of
rigorous methods in evaluation, and link well to the promotion of
behavioural insights which can also be tested by trials.
Psychologists have not been quiet in the rush to the bookshelves and
airwaves, not needing the intermediation of the economists. Cialdini has
worked on the role of norms in encouraging (or discouraging) pro-social
behaviours, for example work on littering, for many years, and his
insights have become the common wisdom in this field, such as the
experiments that showed that hotel guests were more likely to reuse their
towels if told that other guests also did it (Goldstein et al. 2008), and the
role of norms in discouraging or encouraging littering (Cialdini et al.
1990). Building on this work, he has written a general book on influence,
which again is a book in a trade imprint (Cialdini 2009) and is targeted to

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 11 / Date: 10/1
JOBNAME: John PAGE: 12 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

66 How far to nudge?

a general audience of people seeking to be more effective. Closer to the

behavioural agenda is The Small Big (Martin et al. 2014), which offers a
similar theme to the nudge idea that small things make a difference. This
book is linked to a consultancy offering services to the public and private
sectors, Influence at Work (see https://www.influenceatwork.com). Dan
Ariely has produced a flurry of books which address the behavioural
public policy agenda, such as Irrationally Yours: On Missing Socks,
Pickup Lines, and Other Existential Puzzles (2015).
The profusion of books offering psychological insights into public
problems and recommending solutions also gives an impression of the
uptake in this agenda. An example is the book Mindless Eating (Wansink
2006), which details laboratory research into why people eat indiscrimin-
ately and how they are often manipulated into making unwise food
choices. Another book, Reckoning with Risk: Learning to Live with
Uncertainty, examines risk in everyday life, such as how people respond
to figures and statistics, and takes forward Kahneman’s experiments
(Gigerenzer 2003). Another example is Chabris and Simons’s (2010)
book on intuition and perception, which shows how bias enters into
everyday decisions. In the UK, the Royal Society of Arts made a splash
on behavioural public policy with its Social Brain project, Transforming
Behaviour Change, which takes a broad approach to understanding how
cognitions affect behaviour change, deploying research from the neuro-
sciences, as well as more conventional sources in behavioural economics.
What is striking about this review of popular science and commentary
is the level of engagement and the sheer range of applications and
discussions. Even though these ideas had been developed over a long
period of time, there is a sense that from the mid-2000s there had been a
change in the intellectual weather that indicated that the time was right
for new policy initiatives. Maybe the economic recession that occurred
from 2008 caused people to doubt findings from conventional economics
and search out new sources of knowledge. As with much diffusion, it is
hard to trace the causal processes. It is easier to see connections and
detect a trend, but nonetheless it is a plausible line to argue, given the
timing and volume of these publications, that these books were respond-
ing to a period of intellectual uncertainty.

PUBLIC CRITICISM
Along with the praise comes the criticism, which is mainly reserved for
Chapter 6 on the limitations to nudge, and which comes from many
quarters. Much is theoretical, such as about the consistency of the term

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 12 / Date: 10/1
JOBNAME: John PAGE: 13 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge: All tools are informational now 67

‘libertarian paternalism’ or about nudge embodying a neoliberal version

of the state in austere times (Leggett 2014). There is also critique from a
Foucauldian perspective, such as Whitehead and colleagues, who argue
that the use of these techniques strongly impinges on the autonomy of
individuals, establishing what they call the ‘psychological state’ (Jones et
al. 2013). The point of raising these criticisms here is not to engage with
the attack on the use of behavioural sciences and nudge at this point in
the book, but to show how the critiques of nudge are part of the fanning
effect behind the phenomenon, reacting to the uptick in interest, in part
helping to diffuse it. Controversy can draw attention to a new phenom-
enon and highlight its importance.

CONCLUSION
The diffusion of ideas about behaviour change, which emerge out of the
academy, has been a considerable phenomenon in recent years, coming
into prominence over a short space of time. On the surface, this success
might seem to be a surprise, given the way in which academic ideas can
be difficult to grasp and that policy-makers and commentators find it hard
to be interested in theoretical problems. But the diffusion has been easier
in this case, partly because the academic advocates themselves were keen
to simplify the ideas. Even though there are quite a few technical issues
in the literature, at base there is a simple message that many audiences
outside economics were willing to hear, which is that the dictates of the
rational model need to be modified by a more plausible set of psycho-
logical assumptions. As stated in Chapter 3, the actual change in thinking
within economics has not been as strong or marked as might be thought.
Behavioural economics operates within the mainstream; but the challenge
to an orthodoxy is a good message to get across. Critique suggests
conflict, which makes for a good story to tell in the media. Advocates
write in ways that are easy to understand and consistent with the prior
expectations of not only other social scientists but more general readers,
who may have always thought the rational model was too demanding but
were afraid to challenge it. Thaler (2015), with his talk of ‘humans’ and
‘econs’, amplified the tale of rebellion.
The central ideas were synthesised and simplified and made ready for
policy-makers in easy-to-understand lists. These acts of translation and
simplification became essential in getting a new idea adopted, a con-
scious process of appealing to the world outside the academy (Flinders
2013). The notion of an idea that catches on is very familiar in studies of

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 13 / Date: 10/1
JOBNAME: John PAGE: 14 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

68 How far to nudge?

public policy. Critical in this literature is the presence of policy entre-

preneurs who can be a useful conduit for ideas (Kingdon 1984). With
talented advocates in the right place and at the right time, ideas for
reforming public policy can become normal and accepted, sometimes
quite quickly. It is a truism in studies of diffusion that innovation takes
the classic S-curve (Rogers 1983), whereby interest can be slow at first,
but then there is a rapid expansion, which can end up as policy. The
pattern appears to fit the expansion of nudge policies. In this process, it is
possible to observe that a certain kind of nudge was appearing, which
was more suited to the policy world. The question to ask is whether the
dynamics of diffusion and translation were encouraging policy-makers to
select certain kinds of nudges, those of a less threatening and uncontro-
versial kind, so they got more acceptability. These choices may have
structured how the research and policy agenda evolved into the nudge
policy agenda of today.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 04-Chapter4PJED /Pg. Position: 14 / Date: 10/1
JOBNAME: John PAGE: 1 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

5. Translating nudge into practice:

Routes to innovation
At some stage, for all the talk of behavioural insights and studies done in
the academy, there is a need to put these ideas into practice. With
behavioural public policy, the transfer usually involves an agency or
another public body, or sometimes a private company acting in a public
way, which can introduce a nudge or behavioural insight into its standard
operating procedures. By making a conscious choice to do behavioural
public policy, an agency can redesign a policy or procedure that may
have existed in the same form for many years, either from past legacies,
or from other ideas, or that may even have been adopted unthinkingly.
What the use of the behavioural insight involves is a different way of
working, a new kind of implementation that has the behavioural insight
incorporated as part of the communication or information flow from the
organisation. A lot needs to happen, which involves an internal person or
unit that wishes to oversee the insight, and then a delivery body, such as
an implementation agency, which wishes to carry out the intervention; it
requires some interest by a senior person in one of the organisations
involved; it then needs authorisation for the intervention taking place,
often at a political level; as a result of the authorisation some action has
to be decided upon and then followed; then ideally some change takes
place as a result of the intervention.
Just as in the world outside behavioural science, inertia and defaults
play a role in the implementation of behavioural insights, so the
challenge is to find out how to turn the initial enthusiasm for nudge into
implemented policies that embody these insights. The question then
becomes: what kind of organisational innovation needs to happen so that
insights can transfer? In this process, what role do the translators play in
disseminating the ideas? Then, within the organisation, are certain
officials or individuals important in convincing others to adopt behav-
ioural insights?
In this chapter, the process of organisational translation is investigated
to see how innovations happened and what are the mechanisms of
diffusion. The idea is to find out why nudge policies emerged when they
did and to explain the rapid dissemination and adoption of these ideas in

69

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 1 / Date: 10/1
JOBNAME: John PAGE: 2 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

70 How far to nudge?

agencies across the world. In addition, it is important to find out whether

the tendency toward simplification and the creation of a more practical
agenda show the limits to the transfer of nudge and illustrate the
pragmatic way in which it has emerged; translation may affect the
radicalism of the nudge agenda, which has become associated with
providing certain kinds of evidence and the use of robust forms of
evaluation that have the power to make findings legitimate and powerful.
The topic of experimentation is covered in the next section, before the
chapter maps out how the field of behavioural public policy has devel-
oped. It is important to understand how experimentation works within an
organisation, and how innovations using experimentally tested behav-
ioural public policies intersect with organisational interests and routines.
Understanding these processes allows for an assessment of the range of
nudge, to find out what it is possible to do in an organisation given
current levels of commitment, expertise, and engagement. The discussion
also indicates how the use of behavioural insights creates new opportun-
ities for public sector reform.

TESTING AND EXPERIMENTATION

Often involved in the translation into practice is testing – the initial
appraisal as to whether an intervention is going to work or not – which
becomes part of the initial selling of the idea to the organisation: test, see
what works, implement with the successful policy, and then tell others.
There are many standard ways of testing policies, such as a desk review
of the available evidence, doing a pilot, or carrying out a small study
with a few participants or areas. There are better methods of evaluation,
whereby what is being tested can be proven to have an effect, usually
when compared to a counterfactual or a comparison state of affairs. In
terms of translation, the better the test the more stakeholders can be
persuaded, as the test has more credibility, though this statement depends
on stakeholders believing and understanding the robustness of the
method. Sometimes more advanced methods are hard to understand or it
is more of a challenge to convey the results simply, whereas low-tech
stakeholder assessments and consultancy reports can at times be more
effective. The hope is that practitioners understand the robustness of the
methods on offer and select the best one for the problem at hand. Some
public managers understand the importance of knowing rather than
having policies simply legitimated by outside evaluations and consultant
reports.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 2 / Date: 14/12
JOBNAME: John PAGE: 3 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Translating nudge into practice 71

In the world of evaluating policy one method holds sway: this is the
randomised controlled trial (RCT) or field experiment (see John 2017a).
The RCT relies on randomising participants into an intervention or
comparison group and then comparing outcomes afterwards (and if
possible before). Randomisation ensures the groups are equivalent and
the only thing that distinguishes the groups is the intervention. Compar-
ing the outcomes between the treatment/intervention and control/
comparison groups yields the impact of the behavioural cue or other
intervention. Researchers and policy-makers can then make a causal
inference that the intervention caused the outcomes. RCTs address the
main weakness of most methods of evaluation that they cannot separate
out the intervention from other factors that affect outcomes (John 2017a).
In fact, RCTs are part of a family of methods, such as natural experi-
ments or even quasi-experiments, which make a causal inference by
ruling out the influence of observed and unobserved factors that are
correlated with the outcome. RCTs have a special place as a method
because they are associated with testing an intervention and can mimic
whether a policy is introduced or not. Their dominance in medical and
health-related evaluations gives them a lot of legitimacy with policy-
makers. Most people understand the basic rationale of a trial as having
one or more treatment arms and with results generated by comparing
outcomes with those in a control group, or comparing outcomes between
treatment groups. People know there is a need for a large number of
participants and that trials work well in particular situations, such as where
there is an opportunity to randomise and outcomes can be measured easily.
RCTs initially got a bad reputation because they were thought hard to
carry out in a policy context. This perceived failure was because the early
experiments tended to evaluate large-scale social interventions, such as
negative income tax, which turned out to be very complex to implement
(Ross 1970). These implementation issues meant policy-makers did not
get the benefit from trials that they expected, so the method tended to
appear in the textbooks and guides (e.g. HM Treasury 2011) but was not
often used in practice, except in the work and welfare domain, whose
researchers overcame the problems of the early social interventions
(Gueron and Rolston 2013). This relative fall from fashion in the 1970s
and 1980s has been countered by a recent interest from scholars across
social science disciplines, in particular economics (List 2011) and
political science (Green and Gerber 2003). Social scientists rediscovered
the usefulness of trials and discovered contexts where they could be
implemented effectively and yield cumulative advantages, such as Get
Out the Vote (GOTV) experiments, which can evaluate what kind of
intervention causes voters to turn out at the polls (Green and Gerber

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 3 / Date: 10/1
JOBNAME: John PAGE: 4 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

72 How far to nudge?

2015). It soon became normal for behavioural economists to use experi-

ments to test out behavioural theories (as highlighted in Chapter 3),
which of course links to the frequent use of trials in psychology.
Moreover, they can test out causal mechanisms, especially if there are
multiple treatment arms. Once the idea got raised that agencies can do
the experiments themselves, RCTs did not seem so radical, especially
when behavioural economists and others got involved in partnerships
with agencies to deliver these experiments.
Behavioural interventions work very well when tested with trials. The
key advantage is that prior research produces a number of recommenda-
tions that can be applied to relatively discrete stages of the delivery
process. Consider, for example, a favourite of behavioural interventions,
which is the redesign of a letter encouraging someone to settle up a
payment, which has been used by HM Revenue and Customs (HMRC),
the UK tax collection agency, to encourage people to pay their debts at
an earlier date (Hallsworth et al. 2017). The discretion of the policy-
maker is in the wording of these letters and whether different phrases can
be included alongside the normal and standard reminder text. The
wording can be changed without recourse to legal advice. It is possible to
give the instructions to send the different letters, then to examine the
payment accounts of those who have been written to, and to tie the
payment outcomes with the identifiers that say which treatment group
they are in. It is then possible to produce tables of payment rates and the
treatment allocation, and run standard tests of statistical significance
between the average of payments or proportions in each group. The key
is to design the different wordings so there is a comparison of feasible
alternatives. As discussed earlier, these trials do not provide a pure test of
a psychological mechanism that would always stand peer review in a
psychological science journal which would be better delivered in a
laboratory setting; they are simply tests that assess what it is practical for
government to do in particular situations. Often the trial arms are decided
in an interaction between the client and the researcher, with each
bouncing ideas back and forth. The idea is not to test a theory, but to find
something that works. At all times the agency is very much in charge,
and the science has to fit within political and organisational constraints.
The Behavioural Insights Team (BIT), discussed in more detail below,
has made great play with trials, though it took about 18 months to get
them fully established into its work programme. Trials were championed
by the publication Test, Learn, Adapt: Developing Public Policy with
Randomised Controlled Trials (Haynes et al. 2012). The taxes and fines
reminder studies were done as trials (Behavioural Insights Team 2012),
as was early work on charitable giving (Cabinet Office and Charitable

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 4 / Date: 10/1
JOBNAME: John PAGE: 5 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Translating nudge into practice 73

Aid Foundation 2013). Trials quickly became the signature of the unit,
and developed in scale and ambition over time, such as the trial of
welfare-to-work at Loughton, Essex, which was turned into a stepped-
wedge (staged) trial throughout the whole county (see Halpern 2015).
The use of trials continued when the team was spun out of government
and characterises its work on standards and attainment in education, for
example (see Behavioural Insights Team 2015). The team was keen to do
trials where they were relatively easy to do, that is when there were clear
and numerous units to randomise, as with tax reminders, and when the
agency was already collecting outcome data, so not placing huge burdens
on the agency, making the experience not too onerous but impactful in
terms of results that could be useful. Many of the early trials of the team
yielded large financial savings. With some extrapolation, which makes
use of assumptions, such as the effect of the trial being long lasting and
the results externally valid across a large UK or England population, it is
possible to come up with annual savings of many millions of pounds that
can be claimed by BIT as a benefit for the agency and a fulfilment of the
objectives of its political principals.
The demonstration work of the team probably encouraged a more
general use of trials throughout UK government, where they have
become a more normal practice. One example is a large-scale trial carried
out by the former Department for Business, Innovation and Skills, testing
whether getting online advice for business support and a growth voucher
would be good for businesses (Department for Business, Innovation and
Skills and Cabinet Office 2014), which was designed by BIT. Such use of
trials has been encouraged by their rapid adoption in the United States, in
particular for education, work mirrored in the interventions commis-
sioned in the UK by the Education Endowment Foundation and other
initiatives, such as the London Schools Excellence Fund, which has
worked with the evaluation champion Project Oracle and its partnership
organisation, The Social Innovation Partnership (TSIP). At the same time,
the government set up What Works research and evidence centres where
strong forms of evaluation, such as randomised controlled trials, are
privileged. It can be seen that a variety of influences going beyond
behavioural interventions have promoted the use of trials in government
and by other agencies. The level of experience of running them has been
growing strongly over time. It makes this kind of testing of programmes
more feasible and gives a more scientific underpinning to evaluation.
This is probably as significant a development in public policy as the use
of behavioural sciences, and again the expertise has come from eco-
nomics, while academic psychology does not feature very much, with
other disciplines, such as education and crime science, also contributing.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 5 / Date: 10/1
JOBNAME: John PAGE: 6 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

74 How far to nudge?

In fact, there is a huge diversity in the kind of work being carried out. It
is possible then to see that part of the reason for the fast diffusion of
behavioural insights is that they were tested with pragmatic and relatively
easy-to-do trials, which yielded savings, and that the details were
released to the media.

EARLY USES OF BEHAVIOURAL INSIGHTS IN

BRITAIN
As was discussed in Chapter 3, policy-makers in Britain took an interest
in the behavioural agenda right from the early days (see John and
Richardson 2012). In the UK, central government departments found out
about it by commissioning reports from their research staff. One of the
leading departments, before the reform of the structure of government in
2016, was the Department for Environment, Food and Rural Affairs
(Defra), which had been pioneering policies for climate change, such as
energy consumption and encouraging people to buy and use more
energy-efficient appliances. The department had a long history of
research into this area, and had been using behaviour change as a core
part of its mission, influencing the design of regulation as well as nudges.
It published a report, A Framework for Pro-environmental Behaviours, in
2008 (Defra 2008). The department had been increasingly aware of the
importance of seeking to change citizen behaviour in order to reach
policy goals, carrying out a series of internal reviews (e.g. Pike et al.
2010). HMRC also developed an early interest in behaviour modification
in relation to tax payments, and carried out pioneering experiments into
changing information about reminders for tax returns, consulting with
Steve Martin, one of Cialdini’s colleagues. The Department of Health has
a long tradition of research into health behaviours, so ironically is less
part of the wave of enthusiasm for nudge, having its own long-running
research and policy agenda. In a supplementary memorandum to the
House of Lords enquiry by the Department of Health (BC 151), the
department reported on the extensive amount of research it carries out
that has important elements on behaviour change. It helped set up a
Policy Research Unit on Behaviour and Health, located at the University
of Cambridge (see http://www.bhru.iph.cam.ac.uk/about-us/).
There were other routes for the influence of behavioural economics
and other research to enter into government policy, such as the role of the
Government Economics Service, which also produced guidance in 2008
about how behavioural economics could be used to inform policy (John
and Richardson 2012). The Government Social Research Service was

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 6 / Date: 10/1
JOBNAME: John PAGE: 7 SESS: 6 OUTPUT: Fri Jan 26 10:51:11 2018

Translating nudge into practice 75

another generic source of information (as was the Central Office of

Information); the Government Statistical Service and numerous advisory
bodies also play a role. Potential coordination was increased because
Government Economic and Social Research (GESR) was created from
merging the Government Economics Service and the Government Social
Research Service. There was increasing coordination across government,
as evidenced by the appointment of Rachel McCloy, a psychologist from
the University of Reading, in a Public Sector Placement Fellowship. With
these precursors, the Behavioural Insights Team built on existing good
practice in central government.
In a survey of the use of behavioural insights in 2012, a report
concluded that, ‘In spite of the large amount of social science advice that
government receives overall, it is hard to resist the conclusion that the
overall provision for behaviour change at the moment tends to be patchy,
dependent on departmental initiatives and ministerial interest’ (John and
Richardson 2012: 20). The Green Alliance in its submission to the House
of Lords’ behaviour change inquiry captures the criticism very well: ‘We
are starting to see movement in the right direction with behavioural units
being set up in DECC, CLG and DfT, and pockets of people in the
Cabinet Office. Yet these posts are not yet core to the policy creation
process, and the rational actor model is still largely prevailing’ (House of
Lords 2011). These views were iterated by the House of Lords Select
Committee on Behaviour Change, which thought that the focus on
non-compulsory or non-regulatory policies risked neglecting the more
powerful interventions that government knows work. Though there was a
lack of applied research on changing behaviour at a population level,
there is other available evidence that the government needs to use to
better effect. The House of Lords’ review of behaviour change policies
found that, although it received some examples of evidence-based
policies, such as policies on energy-efficient products and smoking
cessation services, there were many examples of policies that had not
taken account of available evidence, including policies on food labelling
and alcohol pricing (House of Lords 2011: 5). The question to ask is
whether, over six years after the publication of the report, the policy
agenda has moved on.
As with central government there has been a long tradition of seeking
to promote behaviour change in local government, but confined to
particular areas such as transport and the environment, often occurring in
the larger strategic councils, such as the county and metropolitan
authorities. The growth of interest in behavioural economics was con-
fined to a few pioneering councils, often because chief executives, other
officers or politicians got interested in the idea. The way the process

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 7 / Date: 22/1
JOBNAME: John PAGE: 8 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

76 How far to nudge?

works is through active local authorities examining the whole of their

policy machinery and finding out where to input behaviour change ideas.
For example, in December 2010 West Sussex had a challenge session
where it set out the principles of behaviour change, inviting representa-
tives from departments. Other councils have had initiatives more focused
on individual services, such as work by Coventry City Council and
iMPOWER, and also Croydon. Other examples are the London Borough
of Sutton, which sought to remodel its transport service, and the London
Borough of Barnet, with its policies on green champions. Kent has
extensive policies, such as employing a behaviour change manager.
Salford has been pioneering work on health. Many of these nudge-based
policies pre-dated the Coalition Government (2010–2015) and even the
publication of Nudge (Thaler and Sunstein 2008). Such examples of good
practice were diffused in existing networks of good practice sponsored by
the Local Government Association, such as its communities of practice
(CoPs).

THE BEHAVIOURAL INSIGHTS TEAM

The setting up of the Behavioural Insights Team, a unit in the Cabinet
Office working directly to the prime minister, in the summer of 2010 is a
story that has been well told (e.g. John 2014), in particular by its director,
David Halpern, in his book Inside the Nudge Unit (2015). It is a tale of
small beginnings and the way in which the unit built on the gradual
diffusion of ideas about behavioural sciences described above. It bene-
fited from the backing of the Conservative leader David Cameron and his
advisors, who were interested in behaviour change and social policy, and
what could be done to create a more pro-social citizenry, a project called
the Big Society. Gus O’Donnell was still in place as head of the civil
service and was ready to propose the idea to the incoming government,
and of course it had as its director the same David Halpern who had
played such a critical role in generating enthusiasm for behavioural ideas
when he was in the Strategy Unit in the previous Labour Government.
It started with just seven members, taking up just about half of a long,
shared desk in the cavernous offices at 1 Whitehall. Most were career civil
servants seconded from other units, with only Halpern as a psychologist,
later joined by Laura Haynes. Paul Dolan, a behavioural economist, was a
team member for a short period of time. The team could not be called a
psychological unit, as its approach was more practical, keen on building
relationships around Whitehall and its environs and getting a sense of
excitement going. Although it was plugged into the academic community,

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 8 / Date: 10/1
JOBNAME: John PAGE: 9 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Translating nudge into practice 77

with the Academic Advisory Panel set up in 2011,1 the approach was to
take some appealing ideas and apply them to practical problems of
government departments and other agencies. A long cast list of celebrity
academics, such as the psychologist Dan Ariely, came through Whitehall
to give talks and attend meetings, and they also provided proposals to
test. Team members gave presentations to government departments
around Whitehall and had many meetings with officials to discuss how
behavioural ideas could be applied to policies being developed. From
these interactions, some officials came forward with proposals for behav-
ioural interventions that were developed in partnership with members of
the team. The team was in effect providing a service to these depart-
ments, which was free at first. Team members also did desk research of
different options as studied in the literature to identify ones to test.
The approach was to convey behavioural ideas as clearly as possible.
This was symbolised by the simplification of the MINDSPACE acronym
(Hallsworth et al. 2010), with its basis in the economics and psychology
literatures, to the shorter and more straightforward EAST (easy, attract-
ive, social and, timely) (Service et al. 2014). In presentations and their
publications, the team stressed the straightforward nature of their inter-
ventions and how big problems often have simple solutions. For example,
over measures to encourage households to install energy-saving devices
in their homes, team members emphasised that the main reason for lack
of adoption of insulation was that people did not want to clear their lofts,
and uptake increased fourfold when households were offered free loft
clearance (Department of Energy and Climate Change 2013). When
presented with a slide showing a cluttered loft, audience members could
not help smiling as they saw their weak selves in the example. The result
is that officials in Whitehall and elsewhere felt they owned such
unthreatening findings and research, and wished to apply these new ideas
in their own departments or agencies.
The team was very public about its work, and there is no sense of it
being a secret cabal unleashing a programme of control of citizens
around Whitehall (see John 2017b). It was interested in trail-blazing the
proof of the concepts. This was very much true of the work on speeding
up debt recovery using behavioural insights, summarised in the document
Fraud, Error and Debt: Behavioural Insights Team Paper (Behavioural
Insights Team 2012). The document is a guide to the work that has been
done, containing studies by the team as well as summaries from desk
reviews. It very much spoke to the agenda of government departments,
seeking to add value on the top of policies that had already been agreed.
The changes it encouraged were not radical step-shifts in citizen compli-
ance with government ideas and policies, but modest changes based on

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 9 / Date: 10/1
JOBNAME: John PAGE: 10 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

78 How far to nudge?

working along the grain of citizen preferences. Many behavioural inter-

ventions, although having their origins in academic work, were easily
comprehensible, such as personalising SMS text messages to encourage
citizens to pay their court fines (Haynes et al. 2013). This use of nudge is
consistent with a more citizen-friendly form of governance whereby
innovations are adopted that take the viewpoint of hard-pressed citizens
seeking to manage their lives. Nudge needs bureaucrats who understand
how people think rather than those who roll out interventions designed in
a psychological lab (John et al. 2011).
Like any small unit operating in the relatively decentralised context of
British government, where the centre does not have absolute power and
itself resembles a feudal court rather than a commander and enforcer (see
Burch and Holliday 1996), the team needed to persuade other parts of
government to take part and then share in the glory. It worked in a
competitive environment of credit-claiming and searches for ministerial
attention, but it survived relatively unscathed in spite of the inevitable
spats and turf wars. It may have benefited from a shift in focus from the
centre of government toward a concern with delivery and implemen-
tation, which had happened under prime ministers Blair and Brown. The
1997–2010 Labour Government’s focus on delivery was epitomised by
the employment of the arch-moderniser Michael Barber. The implemen-
tation focus carried on under the Coalition Government (2010–2015),
especially as it was hard to agree major policy objectives between the
two political parties.
BIT achieved a positive impact in the media, getting a good press from
both left-of-centre publications (e.g. Benjamin 2013) and the right-wing
press (e.g. Bell 2013). What many of these write-ups on the unit show is
that journalists started out with a story about manipulation and the role of
psychologists using their expertise to trick the citizens. In the end, they
were won over by the common-sense approach and added value of such
techniques. There was a similar kind of reaction in the public commen-
tary and assessment of nudge policies. Initially there was some resistance
to the use of nudge policies and the role of the BIT, with the argument
that such an approach was designed to prevent the use of the strong tools
of government. Critics believed nudge was a covert way of justifying a
retreat of the state, and entrenched the role of the private sector in
government in fields such as food labelling (House of Lords 2011).
Academics have made the same kind of argument: nudge was not strong
enough to achieve effective behaviour change and a more robust
approach would be better (Marteau et al. 2011). However, few critics
directly criticised nudge-based initiatives. Behaviour change advocates
have instead criticised the idea that nudge implies that behavioural

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 10 / Date: 10/1
JOBNAME: John PAGE: 11 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Translating nudge into practice 79

interventions apply just to information campaigns and the manipulation

of defaults, whereas in fact behavioural science can be applied more
generally across the tools of government to inform how they work (see
Oliver 2013b).
In short, the team was a success and has been adroit in managing its
external reputation. Its good fortune depended on responding to the
political agenda, and working with government departments and agen-
cies. It mainly tweaked existing policies and implementation procedures
rather than imposed an integrated vision. Above all, it carried out its tasks
in a way that was public and transparent. Its record seems a long way
from the image of the psychological state (Jones et al. 2013), and more
about the way in which a relatively small team was able to craft useful
improvements to standard procedures and help public officials get better
traction on how policies are implemented (John 2017b). The team
publicised its policies to encourage more debate and take-up of behav-
ioural insights.

BEHAVIOURAL INSIGHTS, RCTS AND INNOVATION

Work in organisational theory examines the conditions that can help
organisations become more innovative. In the private sector, such innov-
ation is associated with new products and higher profitability (Argyris
1993; Nooteboom 2000), but in the public sector it is often about policy
changes and better approaches to public sector management (Kelman
2005; NAO 2009). Writers in this literature outline the forces for
conservatism within organisations, which are due to the power of
routines, psychological factors and standard operating procedures, which
tend to benefit those in power. Existing power-holders may resist new
policies because they are associated with younger post-holders whose
careers might benefit. Against this can be arrayed forces for innovation
that can overcome such resistance, whose advocates are often concen-
trated in small groups in the organisation’s bureaucracy. But these reform
groups need nurturing and must build a successful coalition to overcome
change. In addition, there is a long-held assumption that innovation is
hard to achieve in the public sector because of the lack of the profit
motive. Moreover, lines of accountability to political office-holders mean
that bureaucrats have limited discretion to innovate independently. None-
theless, in the right conditions innovation can occur (Borins 2002), and it
is possible to read across the private and public sectors, taking note of the
context of each when assessing what factors drive innovation.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 11 / Date: 10/1
JOBNAME: John PAGE: 12 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

80 How far to nudge?

BIT has all the hallmarks of an innovator: it promoted new ideas and
ways of working, and it has been successful in promoting innovation (see
John 2013b, 2014). The literature on organisational change provides a
good description of how the unit was set up and operated, and how senior
sponsorship allowed it to do new things and operate in part against the
grain. One assumes that the politicians and civil servants knew what they
were doing when they hired Halpern to direct the unit, as he had
performed a similar role with the Labour Government. The Cabinet
Secretary must have expected that the unit would attempt to foster
innovation, and this fitted with his outlook as an unconventional civil
servant. The retirement of O’Donnell did not affect the success and
legitimacy of the unit. Halpern and his unit were able to deliver to the
prime minister successes in an otherwise bleak environment for the
government. The costs of operating the unit were low, and there were few
risks, especially since journalists liked it too. One can imagine why the
politicians were happy to sponsor it and that other parts of government
were content to follow this central lead.
It shows what can be done with a modest level of investment by the
centre. Whether such units are time limited is beyond the scope of this
book (and it may be the case that BIT got out at the right time – see
below), and it is clear that bureaucratic routines and the demands of
governing take priority in the long term. But when there is the right
balance of environment, structures, and people it is possible to produce
more innovation at the centre of British government, and similar stories
can be told for other countries and locales. The key is relative openness
in a complex institutional structure and a willingness to work across
boundaries.

FROM SPECIAL UNIT TO INTERNATIONAL

NOT-FOR-PROFIT
In the end, the directors of BIT realised this particular model of
innovation as a unit within the bureaucracy could not be sustained
indefinitely. Unlike a company that needs a unit to ensure product
development over many decades, the tendency for units to follow fashion
is more the case in the febrile world of central government. In fact, the
surprise about BIT was that it had lasted so long, and that it had been
able to network and build political support rather than having automatic
protection from a few senior people, especially when the unit grew in
size over time. David Halpern and members of the team still work closely
with senior officials at the permanent secretary level, and also network

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 12 / Date: 10/1
JOBNAME: John PAGE: 13 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Translating nudge into practice 81

across government and agencies, such as by running seminars and the

training events for civil servants, the policy schools.
In government, BIT found itself constrained by rules and regulations
which made it hard to carry out behavioural interventions, even purchas-
ing the materials to carry out trials, and to engage other experts, to bid
for funds, or to massively expand staff at the same time as the civil
service wanted to reduce numbers of employees. In 2014, it ran a bidding
round for outsiders to own the organisation along with government so
that it could carry out its central government work on a framework
contract. On receiving bids, the government invited Nesta, the organ-
isation that has an endowment from the National Lottery and works on
projects for the public good, to be the partner, so that the current owners
of the unit are the government, Nesta, and the employees.
When BIT moved out of government, staff numbers expanded
dramatically, and it opened new offices in New York, Singapore, and
Manchester (UK). It had previously had a staff member working in
Sydney for the New South Wales Government, in the unit Behavioural
Insights, and it has expanded its operation in Australia. Numbers of staff
increased to nearly 90 in the London office alone, and the range of
projects increased to include international work on behavioural insights,
such as in development and corruption. This growth reflects the diffusion
of behavioural insights and BIT’s success in promoting this agenda.
Many of the insights were still quite conventional, but the range of
applications was increasing. BIT was careful to avoid being like an
international consultancy, keeping its academic contacts and research
agenda, and also getting funding from foundations rather than just from
private or governmental sources. The charge of following private interests
through what appears to be privatisation of a government asset can be
refuted by reference to the government, Nesta, and its own employees
having shares, with the first two firmly as public service organisations
and key to the governance of BIT. It is important to counter appearances.
Moreover, much of the dynamism remains intact even though the unit
progressed from using insights in an accessible and easy way. BIT has
played a role in the international diffusion of behavioural insights as used
by government and agencies across the world, through its extensive
network of international contacts and where other governments have seen
BIT as a model to emulate. Although the team employs social psycholo-
gists, the modal background is economics, and Michael Sanders, director
of research (chief scientific officer) at the team, is a behavioural
economist.
Meanwhile, the behavioural agenda continues in government with the
formation of a behavioural economics unit within HMRC, Behaviour

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 13 / Date: 10/1
JOBNAME: John PAGE: 14 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

82 How far to nudge?

Change Knowledge Network, with good links across government. Behav-

ioural ideas diffuse to other jurisdictions, such as to local government,
with local councils redesigning the considerable amounts of information
and regulations they convey to citizens, such as local tax reminders (see
Blume and John 2014), or encouraging a channel shift to the online
delivery of services (John and Blume 2017). The behavioural agenda
works well in a decentralised arena where agencies focus on delivery and
relationships with customers and citizens. Agencies can try out different
ideas and customise their messages and regulations in different contexts.

TRANSLATION TO OTHER CONTEXTS

From their origins in the UK and USA, behavioural insights have been
translated into practice in a large number of jurisdictions, where they
reflect the developments highlighted in previous chapters, and are based
on a diffusion of the success of behavioural insights as pioneered by BIT.
One survey found 51 countries that have central state-led policy initia-
tives (Whitehead et al. 2014: 7). The number has probably gone up
considerably since 2014. The survey found that non-government organ-
isations were important in disseminating these ideas, which relate to the
policy diffusion literature that focuses on intermediators and translators
(Graham et al. 2013). In addition, 136 states out of 196 had initiatives
that were influenced by the behavioural sciences. International organ-
isations have been very influential, as the World Bank’s (2015) World
Development Report indicates, but it appears that behavioural insights
have been thought to be essential in testing out aid and development
policies, especially as trials are a very important tool of evaluation in this
field. Overall, the view of many commentators is that nudge has gone
beyond fashion and faddishness and has become embedded as a standard
practice in public policy. In 2017, the Organisation for Economic
Co-operation and Development (OECD) produced an enumeration of
case studies of behavioural public policy, which came from across the
world; its foreword states: ‘The report finds that the use of behavioural
insights has moved beyond a trend’ (OECD 2017: 3).
Nudge units have become important elsewhere. In the USA, where it is
to be expected that behavioural ideas would be influential, the political
context can also be critical of social sciences. The long-running anti-
intellectualism of US life and suspicion of the elites of the East coast go
back to the populist movements of the 1890s, anti-Communism in the
1950s, George Wallace’s presidential campaign, the Reagan revolution,
and extend to the election of Trump to the presidency in 2016. The vast

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 14 / Date: 10/1
JOBNAME: John PAGE: 15 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Translating nudge into practice 83

investment in social and behavioural science is countered by this negative

political culture and its effect on policy-makers. Nevertheless, behav-
ioural initiatives found favour under President Obama, first with Cass
Sunstein’s appointment as head of the Office of Information and Regula-
tory Affairs, though much of this role was about simplification of
administrative procedures rather than the full application of behavioural
insights (see Sunstein 2014b). A more recent initiative was the Social and
Behavioral Sciences Team (CBST), which was organised under the remit
of the National Science and Technology Council, receiving support from
the Office of Evaluation Sciences and the General Services Adminis-
tration. It implemented Executive Order 13707, ‘Using behavioral sci-
ence insights to better serve the American people’. As with BIT, there
were message-based interventions, such as a newly designed letter that
increased farmers’ access to credit, messages to low-income families, and
a single email that went to service members that doubled the rate of
involvement for a savings plan, followed by face-to-face meetings to fill
out a form requiring a forced choice (Dubner 2016). The unit had
projects that ranged from increasing the efficiency of the bureaucracy to
dealing with responses to climate change. Predictably, President Trump
did not support the unit, and its staff were dispersed to other agencies on
21 January 2017.
These changes in organisational priorities in the UK, USA and
elsewhere highlight the importance of the political sanction of nudge, as
it depends on political principals who are sympathetic, such as Cameron
and his Cabinet ministers in the UK case, and Obama in the USA. A
change of leader can alter all that, such as in the UK in 2016, where the
new prime minister, Theresa May, was less interested in nudge. In part,
that is because the UK government was facing the bigger problem of
delivering Brexit. It also reflected May’s brand of Conservativism, which
meant she was less interested in following modern trends than Cameron
(though she was part of the reform of the Conservative Party, and her
phrase ‘the nasty party’ was an injunction to modernise). Under her
leadership, the Home Office was more resistant to using behavioural
insights than other government departments. Such changes show the
limits to the diffusion of nudge as well as the opportunities. Prime
ministers come and go. At the time of writing, May’s future is less
secure, since the Conservative Party lost seats and became a minority
government after the 2017 General Election.
The contingent nature of the diffusion in the US and UK is replicated
elsewhere, and the adoption of nudge reflects the local circumstances. In
the Netherlands, there has been considerable interest, but the adoption of
behaviour, insights into policy reflects departmental strengths and there is

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 15 / Date: 10/1
JOBNAME: John PAGE: 16 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

84 How far to nudge?

no central nudge unit as a result. The use of behavioural science in the

Netherlands – as elsewhere – is dispersed, which reflects different
understandings of the research programme as well as institutional frag-
mentation (Feitsma and Schillemans 2016). There is a unit operating in
the German federal government. In some places nudge units do not find
favour, such as with the Australian federal government: proposals were
put to the incoming Liberal administration in 2013 (John 2013c), but the
government did not at first set up a nudge unit or the equivalent. Then
there was a change of heart and the Behavioural Economics Team of the
Australian Government (BETA) was set up, starting its work in February
2016. In France, the employment of Olivier Oullier in the Centre for
Strategic Analysis of the Prime Minister seemed to indicate the value of
a cognitive behavioural scientist at the centre of government. But nudge
ideas were ridiculed in the French press, and the subsequent adminis-
tration under President Hollande did not go down the behavioural policy
route. French intellectuals and commentators were more critical of what
are thought to be American or Anglo-Saxon imports that smack of
neoliberalism (even though there is no bias toward competition and the
small state in nudge). Consistent with the story told in Chapter 4 and
above, there is a varied response to behavioural insights from organ-
isations, with resistance to the new ideas as well as enthusiasm for them.
The OECD report notes ‘a concern among practitioners that knowledge
transfer will not happen. Networks of practitioners have experienced a
certain resistance from some of their members in sharing the results of
their work’ (OECD 2017: 19).
Diffusion has gone upwards and downwards. It has gone upwards to
international organisations, such as the World Bank, where behavioural
insights have started to be applied to international aid programmes to
improve their effectiveness. The OECD published a report on behavioural
economics (Lunn 2014). It set up a network to promote its ideas, The
European Nudging Network (TEN), managed by the Initiative for Sci-
ence, Society and Policy (ISSP) in collaboration with the OECD and
HEC Paris; it was created in 2014 to disseminate applied behavioural
insights in Europe and elsewhere. It held a seminar, Behavioural Insights
and New Approaches to Policy Design, on 23 January 2015, with people
attending from central and local governments, regulators, staff of inter-
national organisations, and academics. Another recipient of behavioural
ideas is the European Commission, which also produced a report (Sousa
et al. 2016) and where the Joint Research Centre (JRC) offers support to
other Commission services by providing behavioural insights. The OECD
produced its own report (OECD 2017) with the idea of disseminating
good practice. Private consultancies have also assisted the diffusion, as

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 16 / Date: 10/1
JOBNAME: John PAGE: 17 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Translating nudge into practice 85

part of their work is enthusing and communicating as well as being

commissioned to do evaluations. An example is Ideas42 (discussed in
Chapter 4). Output from the journals of the professional associations,
such as the Behavioral Science and Policy Association (BSPA), is
intended to be read by practitioners, diffusing ideas in blogs and tweets
linked to research articles and shorter pieces.
Local authorities have been in the vanguard though have been less
celebrated. Given the multiplicity of local authorities, there are bound to
be some innovators even if the rest are laggards. Policy often diffuses
through the local level by imitation and emulation (see Ward and John
2013). As already noted for the UK, much of the earlier innovation
happened at the local level. It is often regional and local governments
that are very interested in behavioural innovation.
Behavioural policy has diffused in ways that depend on the context and
willingness to deploy these ideas. Nudge needs to be seen as feasible and
acceptable by the elected politicians and elites in a society or else it is
not going to work or be adopted. It needs champions to push it forward
as a project that has to have publicity and good reception to behavioural
initiatives. In this way, the progress of behavioural ideas is political in the
wider sense of that term, in that ideas need to work alongside political
interests, and nudge advocates need to respond to the agenda of existing
power-holders. Nudge then has not been used to transform the state, such
as by being a new set of ideas that can be used by authoritarian
politicians seeking greater control and compliance over the citizens
(though it can be used in this way, such as for combating terrorism or
discouraging migration); rather it usually works within the fragmented
and complex structure of government and agencies, and where advocates
are seeking to advance new ideas. It is important to get the attention of
senior policy-makers and those responsible for the delivery of policies.
Nudge has been successful because it has worked within the existing
agenda of public policies and according to the standard operating
procedures of the bureaucracies. It appeals to the prior aims of bureau-
crats and to those wanting more efficiency and traction from public
policies rather than system-changers or people seeking uniform controls
over citizens.

CONCLUSION
How far to nudge is a product of what is possible politically and
organisationally. As nudges need sanction from real-world institutions,
such ideas are constrained and structured by who uses them as well as

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 17 / Date: 10/1
JOBNAME: John PAGE: 18 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

86 How far to nudge?

who promotes them. A research agenda needs successes; it has to work

for politicians and bureaucrats in terms of responding to and solving
current policy problems. In previous chapters, it was shown how the
behavioural agenda expanded outward from some narrow theoretical
debates to more practical concerns. In this chapter, the next step has been
taken, and it is shown how an agenda is helped by advocacy within
public organisations, and the impact of implementing such policies that
appeal to politicians and bureaucrats located in agencies. It is no surprise
that nudges based on simplification work well. It is a relatively small
number of nudges that work most often, such as norms (though this may
be a product of the nudges that have been tried rather than a full record
of what is possible from the population of possible nudges). There is a
focus on high-volume financial transactions, where it is relatively easy to
modify communication flows. Cases where the architecture shifts, such
as with Driver and Vehicle Licensing Agency (DVLA) experiments on
organ donation, are rare, and the vast bulk of behavioural insights build
on various kinds of enhanced messaging.
It is no surprise that such specialist units work in this practical way
and seek to adapt to existing policies and routines. For example, feminist
think tanks of the 1980s in the end had to operate in the mainstream, but
this compromise may have diluted their radicalism (Eisenstein 1991).
Even with this constraint, a special unit in the bureaucracy can promote
innovation, which happened with BIT, especially in the early days when
it made advances across Whitehall. The unit worked collaboratively
across government, translated ideas, and ensured that they were regarded
as pragmatic. The work of the unit as an internal propagator of ideas was
enhanced by its role in translating ideas internationally, making the later
expansion of the unit an extension of this role.
These feedback effects continue to the time of writing in 2017, often
from the actions of advocates and nudge units promoting behavioural
insights that work. They attract attention, which legitimates the ideas,
gets other agencies interested, and so sustains the agenda. The question
arises as to whether this attention to behavioural insights is going to
continue and whether the agenda can deliver a more radical set of
reforms. To what extent are only positive findings reported, limiting the
extent of knowledge acquisition? That nudge policies are contingent on
the context makes them depend on pre-existing alliances and interests.
When those interests change, so does the research and policy agenda. To
anticipate the potential for a backlash, the alleged limitations of nudge
are considered in the next chapter.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 18 / Date: 10/1
JOBNAME: John PAGE: 19 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Translating nudge into practice 87

NOTE
1. The author is a member of the panel.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 05-Chapter5PJED /Pg. Position: 19 / Date: 10/1
JOBNAME: John PAGE: 1 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

6. Is nudge all it’s cracked up to be?

Limitations and criticisms
In assessing how far to nudge, it is important to acknowledge possible
criticisms of this kind of intervention, in particular of the power and
range of its effects. As has been pointed out in the earlier chapters, there
is a considerable momentum behind the agenda of behaviour change, and
a large number of positive findings have been produced from evaluations
and academic studies, which give an impression that everything nudge-
like works and that public services are being transformed as a result. To
move forward, is it simply a case of expanding the remit of behavioural
insights to include more agencies and more contexts, harvesting the
gains, and rolling out the reformed procedures and policies? Naturally, no
policy and research agenda is going to be transformative in quite such a
way, no matter what benefits there are. Thus, this chapter is the place for
talking more cautiously about the agenda of nudge, not to knock it down,
but to understand some of its limitations and come to a more balanced
assessment of how far to nudge. To this end, some of the conventional
criticisms of nudge are reviewed, and the chapter also assesses the more
damning attacks that have been mounted.

LIMITED RANGE
The first criticism of nudge is that the range of nudges is limited because
they tend to focus on a limited subset of activities for which messages
can be delivered. Nudges work well for messages, texts, and other forms
of communication, which can be directed to the people who get them and
are easy to receive and read. They work well when there is a clear
interest for people in complying with the prompt or responding to the
cue. In other cases, they may be well disposed to do the act if it is in
accordance with their values, such as when deciding to donate their
organs for use after their death. They may do something when prompted
to do so or, in the Thaler and Sunstein (2008) version of behaviour
change, respond to a default option that corresponds to their underlying
preferences. But in spite of the massive expansion of these kinds of

88

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 1 / Date: 14/12
JOBNAME: John PAGE: 2 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Is nudge all it’s cracked up to be? 89

nudges in many contexts, ranging from local to international, it is easy to

see that there may be a limited range for these kinds of interventions.
Although there are a lot of communications from government, there are
limits to the numbers of letters, emails, and texts that can be sent.
Though this limit to the capacity to use nudge may vary across
governments, with HMRC, for example, sending many millions of letters
each year, many agencies use fewer communications. For example, a
defence agency might be interested in recruitment, so might want to send
messages to get a better representation in its employees, and there might
be many internal messages that need to be conveyed. But in practice
defence is about maintaining security: conducting operations and getting
the logistical back-up in place for them. If managing these operational
matters is what a defence facility is all about, how far can behavioural
insights be applied in this sector? Behavioural insights might be out of
range for some government activities.
The other issue is whether there are enough defaults to generate
pro-social outcomes by status quo bias. A default relies on there being a
status quo option of complying with a pro-social or beneficial action, but
there are many areas where none exists, or where default is the
non-beneficial do-nothing decision, such as not applying to university. It
is not possible to design a default of attending university, say through an
automatic application: the student has to choose actively this option. It is
of course possible to design websites with defaults in them, and pension
schemes can offer this, but it is easy to see that such a range of these
kinds of offers can be limited and that default options cannot be
standardised across all government services. In the end, defaults are
going to be useful but they are hardly a panacea to improve public policy
outcomes.
In other cases, individuals might not be willing to be well disposed to
doing the action, who may be hard to influence, making the nudge
difficult to apply for these people. With actions that involve entrenched
behaviours, such as in health and crime, where people like doing the
behaviours and also get benefits from them, light-touch nudges are not
going to work well. There are a number of eminent academics and
practitioners who make the argument that these limitations are critical for
nudge. Marteau et al. (2011) are prominent in this debate, as they suggest
that nudge does not address fundamental issues in behaviour change.
In answer to these questions, it is possible to return to the core claims
set out at the start of this book, which are about the link between the
tools of government and the behavioural public policy reform agenda.
Behavioural public policy can reshape the information environment of
public policy, which requires the simultaneous implementation of steps in

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 2 / Date: 14/12
JOBNAME: John PAGE: 3 SESS: 6 OUTPUT: Fri Jan 26 10:51:11 2018

90 How far to nudge?

a causal chain, where there are a variety of instruments to be deployed,

which range across laws, commands, incentives, organisational levels,
and networks. The tools can operate in a customary manner, with the key
decision being about their authorisation. But with behavioural insights, it
is possible to redesign them so they function much better, whether
operating outside the organisation on citizens or other organisations, or
inside the organisation for its operating procedures and internal commu-
nications. The approach to bias and the discovery of more powerful
mechanisms apply to the implementation chain just as much as to the
classic soft forms of communication like letters. The task then becomes
to identify the links, to find out the key joins and implementation gaps,
and then to devise behavioural measures to improve the working of each
part. By identifying and meeting these gaps, positive feedback may be
unleashed in the organisation, so increasing overall performance, often by
more than the sum of the impacts of each link in the chain.
A good example of this approach to applying behavioural insights is
the Behavioural Insights Team (BIT) study with the UK Department for
Work and Pensions of the employment service at Loughton (see Behav-
ioural Insights Team 2015). The team was charged with reforming
searching for work in the organisation and sought to look at all the
activities that claimants were obliged to do and how effective were the
organisational procedures, such as the number of signatures that were
needed to complete the applications. Through the use of behavioural
knowledge, it was possible to reform the procedures in the light of the
experience of the individual and the cognitive burdens imposed. It was
also possible to insert other behavioural measures, such as asking
claimants to set out a plan using behavioural science to increase
motivations, using the work of Gollwitzer (1999) on ‘implementation
intentions’. As a result of these innovations, the whole organisation can
work better, as was the case in Loughton, for example with a 5 per cent
increase in people moving off benefits, which was replicated (with lower
effects) in a stepped-wedge trial, where the implementation was done
across the county in random (or near-random) order of job centres. In this
context, it is also possible to show that behavioural interventions work
with groups beyond the normal citizens well disposed toward carrying
out tasks and apply to a wider range of actions, and within the complex
implementation chains of a bureaucracy. Thinking about the defence
example just mentioned, it is possible for responsiveness and efficiency
to be improved from the use of behavioural insights through thinking
through links in an implementation chain, which might even end up on
the battlefield!

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 3 / Date: 22/1
JOBNAME: John PAGE: 4 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Is nudge all it’s cracked up to be? 91

New generations of experimenters are increasing the range of interven-

tions and their traction, for example by looking at persuasion effects,
which deploy stronger techniques to shift attitudes and behaviours. The
reason these experiments are interesting is that for many years it has been
assumed they need to go with the grain of people’s expectations, which
explains the apparently limited range of experiments. But this research
shows it is possible to customise interventions that move people from
their settled preferences. Broockman and Kalla (2016) find that using
transgender canvassers reduces prejudice, even as the result of a ten-
minute conversation. This intervention is not a classic behaviour change
cue, but it might be expected that it could lead to changes in behaviours
in due course, such as voting, even where people do not agree in the first
place with the proposition.

WEAK EFFECTS
The second main criticism of nudge is that its interventions do not deliver
strong effects or have such small effects they are not worth doing when
compared to other options. Overall, they do not offer much of a panacea
in spite of the trumpeted claims for benefits. Although a lot of attention is
given to nudge policies, when they are examined they tend to have effect
sizes that are small percentage point differences, which can appear large
when presented in attractive bar charts compared to a control arm, but in
fact do not deliver as much uplift as many common interventions already
do. The criticism is that this is precisely because they are nudges, that
they do not have much power to change behaviour and are really gentle
interventions. They are not able to address fundamental social problems
that have embedded causes. Moreover, behaviour change theories focus
on individual-level rather than societal change (Jones et al. 2013: 51).
It is true that many behavioural insights give percentage uplifts of less
than 10 per cent; however, in the light of their cheapness, and because
many other factors determine the outcomes concerned, these changes are
significant, especially given the light-touch nature of the interventions. It
is a question of perception. If the interventions are seen as part of a
general movement to activate citizens, then these effects may cumulate
over time. When done in an organisational context it is possible for these
influences to sum together and feed off each other.
The other issue is to question whether they are weak effects in any
case. While many interventions have small effect sizes, others are large.
Even Marteau et al. (2011: 264), critics of nudge, report that placing fruit
by the cash register increases the amount of fruit bought by school

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 4 / Date: 14/12
JOBNAME: John PAGE: 5 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

92 How far to nudge?

children at lunchtime by 70 per cent (Just and Wansink 2009). The

example shows how very simple interventions can lead to large changes
in behaviour across a whole range of activities, which is a simple
example of changing choice architectures. The weak treatment argument
also misses the fundamental point about nudge policies, which is that
they achieve strength from the precision of their psychological effects.
The argument is that ‘small is big’ when interventions are attuned to
people’s perceptions and cognitions. Small changes in communication
can have beneficial effects, especially when it is considered how modest
and low-cost these interventions are.
Many scholars argue that the weak effects are a result of poor theory.
Better theory would lead to stronger effects because it makes the
proposed change more plausible and strengthens the link between the
nature of the intervention and the psychological mechanism that is
going to ensure the change in individual behaviour (Davis et al. 2014).
The argument is that the more the intervention is based on good theory
the stronger is the effect, though studies of the relationship between
theory and effectiveness show mixed results, partly because theory is
itself a varied entity, with some theories not being very explicit. Davis
and colleagues’ review of studies concluded that the problem was not
lack of effectiveness, but that the range of theories that are tested is a
small subset of a larger number of theories. They suggested that social
science and by implication policy-makers are missing out on these
potential effects. Nudgers are usually quite cautious and want to deliver
a successful result for their clients, often by adding in many mech-
anisms in a treatment design, making it hard to work out which one is
doing the work. Taking the approach of crowding out, too many
interventions can reduce the impact of a clear intervention based on one
theory. Adding treatments also has the same problems, and may in fact
reduce the impact of an intervention. For example, Fieldhouse et al.
(2013) tested for different kinds of nudges in their voter turnout
experiment, with some households receiving two letters and two tele-
phone calls over successive elections. The treatment arm with the largest
number of treatments showed reduced impact compared to treatment
arms with fewer interventions.
Moreover, the stronger the insight, the stronger is the effect. It is the
case that norms and softer insights have weak effects, but these are never
intended to be strong interventions. It is not surprising that effect sizes
might be in the 5 per cent range (which yields millions of pounds in
revenue in any case). But there are stronger psychological techniques that
might be applied, which reach more powerfully into cognition by
shaming, making people visible to others, or even making them agitated,

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 5 / Date: 21/12
JOBNAME: John PAGE: 6 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Is nudge all it’s cracked up to be? 93

all of which are behavioural interventions. In the history of Get-Out-the-

Vote studies, there are a number of light-touch strategies that are more
about contact than persuasion. They have effect sizes of between 2 and 7
percentage points, with the more personalised interventions yielding the
strongest effect, which is consistent with a light-touch behavioural
approach (Green and Gerber 2015). If the salience of the psychological
device is increased, effect sizes duly improve. In an experiment to test for
the impact of exposing a voter to the voting record of neighbours and the
record of neighbours back to the voter, Gerber et al. (2008) found that
this intervention increased voter turnout by 13 percentage points, a
doubling of the treatment effect from more light-touch interventions. By
promising to tell respondents the voting history of their neighbours, and
the voting history of respondents to the neighbours, researchers can
increase turnout from conveying visibility and applying social pressure.
The argument is that the reason nudges have been seen to produce what
are thought to be relatively low effect sizes is because they have been
chosen to be non-controversial. If the nudge unit had chosen stronger
interventions the effects would have been greater. It is merely the effect
of the choice of instrument, which of course reflects what is acceptable
from the sponsoring organisation’s point of view and is politically
feasible. The nudges that exist are ones that are allowed by policy-
makers, not that nudges are necessarily weak treatments. As the rhetoric
of behavioural insights is that it is small things that count, it is hardly
surprising that that is just what is offered to nudge units to test in
interventions. Given that so many resources from government depart-
ments are tied up with delivering existing interventions, it is also not
surprising that not a huge amount of attention and resources are diverted
to delivering behavioural interventions.

TEMPORARY EFFECTS
The third criticism is that nudge policies only have temporary rather than
long-lived effects. It is possible for an agency to get an uplift in outcomes
but for the change not to be sustained over time. People get excited about
changing their behaviours, which lead to an uplift in outcomes, even in
the short-term, which makes it look as though a nudge is what was
needed; but over time they return to a baseline of habitual behaviours,
which is where they want to be in terms of long-term preferences or what
suits their lifestyles and that of their peers. This argument comes from
the embeddedness claim that people’s behaviours are entrenched though
their own habits, reinforced by peer-group pressure and social networks,

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 6 / Date: 10/1
JOBNAME: John PAGE: 7 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

94 How far to nudge?

and influenced by the media and the advertising from commercial

organisations. This is the argument of Chapter 2. It is easy to see how a
small nudge might have only a temporary effect on human behaviour
because of these long-term factors. The warm glow given by a nudge can
easily fade. People might not even remember the nudge six months after
it was given.
One answer is that nudges may be repeated over time, which reflects
their low-cost nature. Yet behavioural researchers know that such
repeated treatments can fade in effectiveness over time as respondents get
used to the nudge. Sometimes nudges depend on novelty, and once they
are delivered a few times the effect wears off. Other nudges rely on
recipients not knowing they are being nudged. This is because the
response needs to be automatic, that is come from unconscious non-
reflective processes. If people find out they are being nudged, they will
think about it, and they might not like being manipulated. If nudges are
repeated over time, they are more likely to be detected. But nudges can
be varied, and this is normal practice in the private sector, with
companies like Google and Amazon altering their nudges on the basis of
feedback and A/B testing. These companies survive and prosper with
such strategies.
A more sophisticated idea, which reaches beyond the habit and
repetition of nudges, is the claim that nudges can be part of a journey to
a new set of outcomes, a way of moving from one equilibrium with
negative outcomes for individuals, and the communities of which they
are part, to a more positive equilibrium position. At the low equilibrium
point behaviours reinforce each other largely as a consequence of the
failure of collective action, where it is in nobody’s interest to change
behaviours, whereas at a more positive equilibrium point, once collective
action has been achieved, these positive behaviours reinforce each other,
with no one wanting to risk precipitating collective action failure. The
nudge then can help citizens solve the collective action problem by
conveying signals that encourage others to act. People indicate to one
another that they are prepared to do the action. In the words of one
report, it is about ‘nudging the S-curve’, that is affecting the pattern of
diffusion and rate of change of adoption (Brook Lyndhurst 2006). This
phenomenon links to work on the critical mass whereby there are points
of breakthrough in changes in behaviours, and where the nudge might be
critical in moving from a slowly rising level of change to one where there
is rapid acceleration and norms spread so that everyone adheres to the
new desired set of behaviours. As Brook Lyndhurst (2006: 8) reports,

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 7 / Date: 10/1
JOBNAME: John PAGE: 8 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Is nudge all it’s cracked up to be? 95

A range of fundamental features of the social system mean that a model of

policy intervention predicated on the steady refinement of interventions
towards a set of policies that ‘work’ may be ill-founded. Rather, given the
complexities of ‘behaviour change’, a model of ceaseless innovation, within
broad parameters of focus and in a network setting, offers a potentially
valuable conceptualisation of how to move forward.

WEAK EXTERNAL VALIDITY

One key criticism of nudge, which relates to other limitations, is the
extent to which its findings are transferable to other locations from the
places where it has been tested, and that the science of behaviour change
is based on a selective reading of the evidence which gives the impres-
sion of wide-ranging effects but in fact indicates that the range of
subjects upon which it works is limited, partly because of the availability
of people to participate in studies. This potential weakness concerns
external validity, which is about the ability to generalise. This constraint
comes from the fact that most populations for studies are selected to
receive the intervention, particularly for more personalised interventions.
Because of the need to recruit participants and to get their consent, they
are not representative. Marteau et al. (2011: 342) comment on this point:
‘To date, few nudging interventions have been evaluated for their
effectiveness in changing behaviour in general populations.’ Of course, it
is possible to sample whole units like local councils, or carry out national
evaluations once the pilot has taken place. Fieldhouse et al. (2013)
carried out a national-level Get-Out-the-Vote experiment after doing a
local intervention (John and Brannan 2008), which allowed them to show
the external validity of interventions to all England (Fieldhouse et al.
2014). As nudges become increasingly tested, so more evidence about the
range of application of nudges becomes available.

NUDGES DON’T ALWAYS WORK

Linked to the limits of external validity and range is the claim that
nudges do not always work in all contexts as expected. Alternatively, the
context might be varied just slightly and the nudge yields results that are
puzzling or cannot be considered to be the direct result of the research
design. A null result from a study does not often give much evidence of
why nudges don’t work, often just the result, as there are little data to
analyse. In the early days of nudge, it seemed that the interventions
always worked, such as norms on tax compliance. In fact, examining at

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 8 / Date: 14/12
JOBNAME: John PAGE: 9 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

96 How far to nudge?

the reports of the interventions more closely, such as BIT’s work on error
and debt (Behavioural Insights Team 2012), shows that not all the
interventions worked or worked in the way the behaviourists expected.
For example, a BIT trial seeking to impress on doctors that there is a
social norm of paying their taxes on time did not work. Another trial
sought to replicate Ariely and colleagues’ experiment, which asked
people to sign an insurance claim at the top rather than at the bottom to
prompt the moral nature of the communication, rather than to confirm
what had already been written on the form when signed at the bottom
(Shu et al. 2012). When trying to replicate the results on prompting
people to be honest in declaring whether they had claimed a single-
person discount for the local council tax in England, the researchers
found the prompt worked best for signing at the bottom rather than the
top (Behavioural Insights Team 2012). Norms do not always work when
trying to encourage repayment. Silva and John (2017) found that social
norms do not encourage students to settle their fees. In another example,
Van de Vyver and John (2017) tested for the impact of social norms in
seeking to improve the implementation of a government policy that asked
local councils and parish councils to register a local community asset,
randomly providing half of the councils with social norm information,
which did not work either.
In a recent paper, ‘Nudges that fail’, Sunstein (2017b) draws attention
to these negative or null findings, such as information disclosure about
the number of calories in an item of food. There are studies showing that
placement of food near a checkout can affect choices, but sometimes it
does not. Sunstein (2017b) argues that nudges can face the strong
counter-preferences of users, which can override defaults. It also seems
plausible to suggest that reactance, discussed in Chapter 2, still occurs in
reaction to nudges, especially when they are known about by the
individuals affected by them. Those who are opposed to the nudges, such
as private sector interests, might counter-mobilise, for example banks
wanting to keep profitable arrangements when faced with regulations
trying to stop them. It is also likely that there are many other nudges that
do not work but are not reported. Null findings are hard to publish, and
researchers and agencies usually do not want to publicise them. It is
impossible to know how many failed nudges there are, in spite of efforts
in the major disciplines to encourage the publication of null findings.
Sunstein (2017b) argues for the greater development of stronger
nudges and for nudgers not to give up. In particular, better regulation is
the answer. Nudges operate better when the grain of public policy
changes is in the favoured direction. A more general consideration is that
a mature research programme is not dependent on continual successes

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 9 / Date: 10/1
JOBNAME: John PAGE: 10 SESS: 6 OUTPUT: Fri Jan 26 10:51:11 2018

Is nudge all it’s cracked up to be? 97

but needs to learn from failure as part of a genuine desire for reform. If
all nudges worked, then there would not be much point in researching
them; instead policy-makers should not expect all of them to work, but
through failures and successes they can design better nudges and reform
wider choice architectures.

WEAK KNOWLEDGE BASE

The irony may be, as Marteau et al. (2011) comment, that nudges have
not been evaluated properly in that, although nudges are thought to be
cheap, some interventions, such as those on the environment, may be
expensive, so are not cost-effective. In spite of the expansion of field
trials in this area (see Chapter 5), their number is still relatively low for
the purposes of evaluation and cost-benefit analysis. In particular, it is
still not possible to do many meta-analyses, which depend on a large
number of studies with an identical or similar design. The House of
Lords’ (2011) report authors commented that their review of the evidence
indicated gaps in knowledge in
aspects of the automatic system, particularly in relation to how emotional
processes regulate everyday behaviour; a lack of comparative research into the
limits to the transferability of behaviour change interventions across cultural
differences; uncertainty about how genes interact with environmental and
social factors to cause behaviour; and, a lack of understanding about the effect
of social dynamics on behaviour. Other witnesses commented on the chal-
lenges involved in integrating the numerous theories of behaviour which were
emerging from across the range of sciences of human behaviour. (House of
Lords 2011: 17)

Since then, more evidence has appeared on all these aspects of behaviour,
but there is not the level of knowledge about the impact of behaviour
change that appears in medical evaluations, for example, or in the
Campbell Collaboration, which is based on a large number of studies to
create meta-analyses.
The claim of lack of evidence, however, is too easily made, as there
are always likely to be gaps in knowledge in such a vast field,
particularly as behaviour change is a diverse set of activities, with some
elements that are relatively new, such as savings and taxation. In all
these fields, a large number of studies are emerging, for example studies
of norms in public administration settings (see John et al. 2014). The
Campbell Collaboration itself reports a large number of behaviour
change studies; this increase in reporting reflects the gradual growth of

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 10 / Date: 22/1
JOBNAME: John PAGE: 11 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

98 How far to nudge?

this area in many disciplines, which partly pre-dates the expansion of

interest in behavioural economics and the nudge agenda. The interested
reader can find reviews of the evidence in such studies as ‘Behaviour
change programs to prevent HIV among women living in low and
middle income countries’ (https://www.campbellcollaboration.org/library/
women-hiv-behaviour-change-programmes-low-middle-income-countries.
html) or ‘The effects of school-based social information processing
interventions on aggressive behavior: Part I: Universal programs: A
systematic review’ (https://www.campbellcollaboration.org/website-search
?limitstart=20&searchword=behaviour+change), just to name two of the
many studies listed on the Campbell Collaboration website. The Behav-
ioural Insights Team website, www.behaviouralinsights.co.uk/, contains
summaries of the many studies that have been done by the team. New
journals in the field, such as Behavioural Public Policy and Behavioral
Science and Policy, as well as the long-running journals, such as
Behaviour Change, publish recently-completed studies, which are more
than matched by the large research output in journals in established
specialist fields, such as Health Psychology.

POTENTIALLY HARMFUL
Some critics, such as Marteau et al. (2011), believe that nudges can be
harmful. The argument is that conveying light-touch nudges might
convince people that some foods are healthy, for example through
labelling, and then encourage them to eat more carbohydrates than they
should, for example overeating muesli health bars full of honey. Marteau
et al. (2011) cite a study of food labels that shows that labels can falsely
convey reassurance (Wansink and Chandon 2006). The additional argu-
ment that makes this position more decisive is that nudges may not be
enough for high-risk individuals or may be artificially reassuring, as what
such individuals need is more targeted interventions. This limitation leads
to the argument that nudges more generally may not address core
behaviours, may prevent individuals from considering the full range of
options, and allow them to be reassured by messages that go with the
grain of their intuitions. The argument is that individuals might not wish
to be confronted with information that they do not like and do not feel
comfortable with, or even feel angry about. Getting a more upfront
message might cause individuals to reflect upon the problems they face,
for only when they have thought about these issues sufficiently can they
act. Gentle nudge messages might not be noticed by subpopulations with
low education and resources, those people who are focusing on the

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 11 / Date: 14/12
JOBNAME: John PAGE: 12 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Is nudge all it’s cracked up to be? 99

necessities of life. In short, the argument is that the state should be trying
to shove individuals to change their behaviour rather than nudge them.
These arguments on their own do not do enough to knock down nudge.
The argument for more targeted interventions for more needy people
does not mean that nudges do not work or cannot play a role. The
harmful argument does not work here. A more general argument has
more merit. This is the idea that encouraging an unreflective process
might cause too much dependence and not enough reflection, so nudge is
not the solution for long-term problems and might even disguise them.
Even if the argument that disguising larger problems might have some
force, it does not show that nudges are harmful. A lot depends on the
kind of nudge, and the nudge could conceivably be the first step to wider
reflections by individuals, with the nudge being the precipitator rather
than the obstacle. It is a hard argument to sustain that the nudge could
harm the individual’s ability to receive a welfare-enhancing message,
having a kind of soporific effect.

DIVERTING PUBLIC ATTENTION

Even mainstream authors in behavioural science (e.g. Marteau et al.
2011) as well as critics (Goodwin 2012) have come to believe that the
use of behavioural sciences in government and the rise of behavioural
policies has de-privileged more interventionist policies on the part of
government. Stressing light-touch, non-regulatory, and fiscally cheap
policies, diverts public attention from the tools and measures that are
more effective. This was the view of a critical government report from
the UK House of Lords (2011). This criticism is partly a function of a
restrictive focus on nudge: ‘interventions which may be described as
“nudging” are not synonymous with, but rather are a subset of, non-
regulatory interventions’ (House of Lords 2011: 12). As was argued in
Chapters 3 and 4, this is an overly restrictive view of nudging which
means that many of the customisations of tougher policy instruments
would be ruled out. While nudges are cheap to do, nudge is not going to
rule out another strategy. In fact, nudging works well in combination with
standard approaches to effecting change, except that what behavioural
insights offer is the chance to make the policy work in a better way. The
policy-maker gets all the advantages of the existing tools of government
with their proven advantages but with all the added value of behavioural
insights.
It is not plausible to argue that the state is locked into a set of activities
that rule out either nudges or more stronger interventions. It would be

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 12 / Date: 10/1
JOBNAME: John PAGE: 13 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

100 How far to nudge?

different if the two strategies cost a lot of money and nudges were
marginally cheaper, encouraging government to go for them. Of course, it
is still possible that governments face limits to the use of resources, so it
could be the case that if the cost per outcome was significantly in nudge’s
favour it would lead to a redistribution of resources away from in-depth
interventions to more population-based ones. This does not pose a large
problem in the efficient allocation of government resources, because
society benefits. Government might want to weight the targeted resources
to ensure that needy individuals were protected, but even this would not
deny the argument that nudge and targeted strategies are complementary
and that a government might want to find the efficient balance between
the two.
The critical argument can, however, go further. The critic knows that
the government will not be rational in allocating its time, but in fact the
government might be under different kinds of pressure. One might be to
cut resources generally. The availability of a cheap measure would appear
to satisfy a government that it was meeting a need so could justify cuts
made to more targeted and expensive interventions elsewhere. A govern-
ment might have an ideological preference to cut back the state, and the
nudge strategy might give it the ability to do that, because it depends on
a light-touch intervention that does not affect the autonomy of the
individual. In this view, nudge plays into a right-wing agenda while
appearing neutral and pleasing to all. Appearing shiny and new, it can be
a cover for worse things happening.
This argument again rests on nudge taking a particular form, that is
favouring only light-touch techniques, but it can just as easily be used to
favour stronger interventions done in a nudge way. When seen in this
way, there is no political project behind nudge: it is neutral, really about
implementation, not making policy choices and does not rule out certain
kinds of choice. The fact that a government of the right chose nudge in a
period of fiscal austerity might show enthusiasm for low-cost approaches;
but, in fact, as Chapter 5 shows, it owed more to timing and context as
behavioural approaches had become fashionable, when policy entre-
preneurs were at work in and around government, and that the prime
minister, David Cameron, happened to have a social agenda which was
about promoting good behaviour and better collective outcomes, which
worked well with nudges. In other parts of the world, such as the
Netherlands and the USA, the agenda was taken up by non-right-wing
governments; and even right-wing governments, such as those led by
Theresa May or Donald Trump, are not guaranteed to like nudge. It does
not fit into a simple political box.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 13 / Date: 10/1
JOBNAME: John PAGE: 14 SESS: 7 OUTPUT: Fri Jan 26 10:51:11 2018

Is nudge all it’s cracked up to be? 101

LIMITS TO INCREMENTALISM
In essence, the behaviour change programme is not about a radical
overhaul of public policy, as it relies on a series of tweaks to the
policy-making process. It does not impose a rationalist vision upon the
policy-maker, and produces more responsive and considered policies as a
result. This is incrementalism, an approach to policy-making that the
Behavioural Insights Team celebrates (Halpern and Mason 2015), in
particular a version called ‘radical incrementalism’, the idea that through
many small steps radical transformation may come about.
The use of this term draws attention to a familiar debate in public
policy, studied first in the 1950s (to which starting students of public
policy are introduced in their first weeks of study). A number of
influential public policy scholars argued that the incremental pattern of
decision-making, that of making a series of experimental small steps and
then learning based upon experience, creates an adaptive kind of govern-
ance, which draws upon the knowledge of organisations and their
competing elements, and celebrates a decentralised approach to decision-
making. Even the term ‘radical incrementalism’ has a long pedigree,
going back to at least the 1960s (Wildavsky 1966). Lindblom argued
strongly for an incremental approach to policy-making (Lindblom 1959,
1965; Braybrooke and Lindblom 1963), which has been contested on
both empirical and normative grounds ever since. On empirical grounds,
it has been observed that such a stepped approach to decision-making is
not how modern democracies work, as they often have long periods when
not much happens, followed by rapid changes or policy punctuations
(Baumgartner and Jones 1993; Jones and Baumgartner 2005). If these
ideas about policy-making are taken as read, then the incremental
approach is not in the grain of how decision-makers operate in that
reformers need to be aware that they are at their most powerful when a
new regime starts to change, and the reforms and changes can start to
happen quickly, reinforced by media and political interest, which causes
a radical rethink of routines and existing policy choices. Simply burrow-
ing up from below will not transform public policy. Nudge is likely to
receive resistance from within bureaucracies, and also supporters in the
wider policy community, who will not support such changes without an
agenda shift. Radical incrementalism in this view does not work.
The other critical point of view, is best summarised in the critique of
Goodin (1982). He argues that policy-makers do not want incremental-
ism, as it does not provide a route map of where they want to go.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 14 / Date: 22/1
JOBNAME: John PAGE: 15 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

102 How far to nudge?

Incremental decision-making could lead to cliff edges without the

policy-makers knowing they are there, or have sleeper effects which are
negative but which are not realised in a series of incremental gains, each
of which looks plausible and beneficial but where all get worse off in the
long run.
Of course, there are easy ripostes to these points of view. For the
empirical critiques, even the policy agendas and punctuated equilibrium
perspectives do not rule out some form of incrementalism in periods of
slow change. Moreover, advocates of this perspective often acknowledge
the beginning of changes in an approach during the period of stability
whereby old ways of doing things do not work. As incremental changes
do not offer benefits in the same way as before, policy-makers start to
learn that, even though there may be strong upswings in attention, these
are often based on a lot of groundwork beforehand. In this way, the use
of behavioural insights can be the baseline and a form of preparation for
radical changes that can get unleashed. If bureaucracies do become more
experimental and willing to try out new measures, the expectation is that
for much of the period not much happens, but then this very experimen-
tation can contribute to the radical changes happening later on. In terms
of the normative critique, this criticism also depends on a characterisation
of incrementalist decision-making as a completely blind set of policies,
which is not a realistic position to adopt. Just as much intelligence and
guidance can apply to decision-making under incrementalism, which can
resemble rational action; it is just that the means of getting to decisions is
different. Incremental decision-makers can review their options, look
forwards and backwards, even if the overall strategy is adaptive. In fact,
blindness is more associated with rational decisions, trusting the policy
analysis or consultant’s report, rather than incrementalism, which is more
about responsiveness and gaming the environment. The critique of radical
incrementalism fails.

IDEOLOGICAL AND NEOLIBERAL

Another type of criticism comes from within social science, from
researchers interested in the politics of knowledge and its use. It concerns
the use of psychology in government, which is thought to be manipula-
tive and to exclude more political underpinnings of government interven-
tion (e.g. Jones et al. 2013). In part, this critique of nudge policies relies
on accepting a number of other claims, one of which is that modern
politics and policy are characterised by an ideology, that of neoliberal-
ism, which is a set of ideas that legitimate market institutions, and that

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 15 / Date: 14/12
JOBNAME: John PAGE: 16 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Is nudge all it’s cracked up to be? 103

reproduce over time in ways that extend individualist and market-

orientated thinking (see Jones et al. 2013: 6). The way in which nudge
emerges does not lead to the consideration of more radical solutions to
policy problems. The approach to public policy tends to focus on the
individual causes of policy problems rather than the wider structural
reasons linked to the operation of a market economy. Nudge posits a
public policy response based on influencing individual choices. This
chimes with the views of other critics, like Dobson (2011), who criticises
nudge for its de-politicisation and neglect of the wider context of political
debates, such as green nudges that do not consider the wider reasons for
environmental problems. This involves denying the ethical foundation of
these debates:
In thinking of sustainability as a matter of tweaking behaviour, nudgers
commit what philosophers call a ‘category mistake’. Ethics, norms and values
are not an optional extra in sustainability – they are constitutive of it. From
this point of view, it is as absurd to see sustainability as a matter of resizing
waste bins as it would have been to nudge slave owners towards ending
slavery by making their ships a little shorter and narrower. (Dobson 2011: 9)

Of course, there are problems with this argument in that nudge may be
seen to be partly a critique of those same market-driven policies, which
are based on rational calculations, whereas behavioural economists want
to move beyond this. Jones et al. (2013: 10) argue that Kahneman and
Tversky’s work ‘embodies a normative assumption of rationality’, which
seems to go too far and does not appear to do them justice, as many
behavioural economists, such as Sugden, do not operate with strict
rational assumptions. A neoliberal argument would regard behavioural
economics as critical of the economics of the new right and questioning
of the rational model. This blunderbuss approach is the main problem
with these functionalist arguments in that any government policy can be
fashioned to be seen to support an existing order, when in fact ideas and
policies may be independent and different from each other. For example,
it is possible to imagine a socialist state with no market mechanisms
suffering from a weakness of internal commands and needing a nudge-
based policy to improve them. Nudge might in fact be more important in
a centralised or market-constrained economy, as there is a need to
motivate citizens to act with a less clear perception of the reward. In
market societies, given how long nudge has been used by the private
sector, the use of it by the public sector may be regarded as a reclaiming
of these techniques for public purposes, again the antithesis of a
neoliberal approach. After all, the big criticism of nudge from libertarians
is that it is paternalist, not that it is libertarian (see Chapter 7 on ethics).

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 16 / Date: 10/1
JOBNAME: John PAGE: 17 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

104 How far to nudge?

An acceptance of the lack of stable preferences as implied in behavioural

economics may suggest that the state should not intervene (see Sugden
2008), but it can also mean that policy-makers need to pay more attention
to the general design of policies and the constitutional issues governing
their operation. As Hargreaves-Heap (2017: 10) writes, ‘behavioural
public policy should also be concerned with the character of constitu-
tional rules that constrain and enable action’. Once the unstable nature of
preferences is considered, it justifies a more egalitarian approach to state
intervention, such as for taxes and benefits, so that people’s interests are
protected.
That is not to say there is not conservatism within the behavioural
programme, as was recounted in Chapters 3 and 4, in that potentially
radical ideas became integrated within mainstream economics, and also
were fashioned in a way that could respond to the concerns of policy-
makers. Even though much customisation of academic ideas had
occurred, this is a different argument to saying that nudge is a manifes-
tation of neoliberalism. Within nudge there are radical ideas that do not
sit with the existing order, which can build on the work that has been
done, which in effect is what Jones et al. (2013) are arguing. The
difference with the position adopted here is that more agency should be
attributed to the ideas and people involved so the policies do not result
necessarily from an all-encompassing ideology. By introducing the
overarching set of ideas from Foucault, the analysis becomes rather
determinist. It seems a better course to say that ideas emerged where they
were consistent with the existing order, but that their radicalism remained
intact waiting to be rediscovered, which Jones et al. (2013) hint at in their
concluding chapter. Overall, the difference between Jones et al. (2013)
and the position here is that the Foucauldian approach does not give
enough emphasis to the agency of the actors involved in the policy
process. The argument in Chapter 8 points out that a technocratic and
top-down approach to behavioural policy is not a necessary consequence
of its adoption, but that a more decentralised and open approach can be
adopted whereby a range of actors can use these ideas. Agency is the way
forward.

NEGLECTING CITIZEN INTELLIGENCE AND

FEEDBACK
The final criticism links to these earlier points, but is more specific. This
is the idea that nudge policies do not take advantage of citizen involve-
ment and feedback, because of the nature of nudge policies, which are

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 17 / Date: 10/1
JOBNAME: John PAGE: 18 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Is nudge all it’s cracked up to be? 105

designed behind closed doors, particularly if the nudges only work if they
are kept secret from the citizens so that citizens do not get the chance to
be involved. This means that designers of behavioural public policies do
not take advantage of the correction of errors and feedback. As Farrell
and Shalizi (2011) write, ‘There is no reason to think technocrats know
better, especially since Thaler and Sunstein offer no means for ordinary
people to comment on, let alone correct, the technocrats’ prescriptions.
This leaves the technocrats with no systematic way of detecting their own
errors, correcting them, or learning from them.’
It is commonly assumed that nudge policies can only work if kept
secret, but this is not right. People often like being nudged, and this is the
whole point of the Thaler and Sunstein agenda: that people have
preferences to do good things or to help themselves in the long term.
They even want to be nudged. It is a way to get to their good selves. If
the large amount of public opinion evidence that Sunstein (2016) has
collected is accepted, citizens approve of and welcome nudge policies,
which means that if they realise they are being nudged this might even
improve the chances that policies might be accepted. Even though nudges
work on the non-reflective and automatic systems, it is possible to
consider a stage of reflection that might accompany or reinforce these
automatic system-based processes (see the argument in Chapter 8). There
might be a realisation that a nudge has taken place, which might help the
individual keep on track afterwards and put in place plans to keep the
good behaviour going. This might become an automatic and self-
regulating process later on when the individual does not need to think
about the new set of behaviours.
There is nothing to stop nudges involving citizen contact and feedback,
and incorporating more democratic processes, which is implied by the
public nature of nudging and that there are public good aspects to what is
being influenced. This view is closer to the world of Ostrom (1990) and
supports self-reinforcing forms of governance rather than the science of
manipulation. But there is some truth to the criticism based on obser-
vation of how nudge is currently practised. It is mainly technocratic,
carried out by administrators advised by behavioural scientists who use
randomised controlled trials to select the best option. Although nudges
are usually approved by political principals and they are not done in
secret, they are introduced without a lot of public involvement or wider
debate, even in the bureaucracy, as well as in other democratic forums.
Most nudges are not done in consultation with citizens, do not involve
feedback whereby citizen views can be incorporated.
The argument that nudges could be widened to take account of this
criticism was made in Nudge, Nudge, Think, Think (John et al. 2011).

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 18 / Date: 10/1
JOBNAME: John PAGE: 19 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

106 How far to nudge?

That book suggested there are two traditions of examining citizen roles in
public policy, the nudge route and the think route, the latter drawing on
ideas in deliberative democracy. These ideas are regarded as opposites, as
they draw from different intellectual traditions, so might not appear to be
comparable. The two traditions are closer than might be imagined, partly
because successful nudging will involve some reflection and understand-
ing of the wider policy issues. Deliberation, which has been limited in
the way in which it has been implemented in recent years because of
self-selection by and inequalities among citizens, needs to be made more
practical and egalitarian. The use of behavioural ideas can encourage
some reflection as part of a nudge-based intervention, and where all
citizens can be encouraged to think. More of this argument appears later
in the book, especially in Chapter 8.

CONCLUSION
Nudge has attracted criticism from all quarters, which is entirely to be
expected given it is an idea and set of policies popular with policy-
makers and experts, and that a lot is claimed from the approach.
Academics and critics like to knock new approaches down, so the
question becomes how much do these criticisms hold, and whether they
imply limitations to the range of nudges that can be used, which brings
the argument to the question and title of this book. If all these criticisms
are taken seriously, they would severely limit how far to nudge. Of
course, often it is a question of degree, because it is hard to see that there
are fundamental objections to nudge, in the sense that these policies must
have some use and that few people can disagree with measures that
improve the payment of taxes and achievement of other desirable
outcomes, especially as they are low-cost and unobtrusive. It would take
someone of a very critical turn of mind to see such measures as
conspiratorial and ideological, having poor long-term consequences for
political action, and increasing citizen passivity. Even critics such as
Marteau et al. (2011) do not go that far, nor do those writing from a
critical perspective (e.g. Jones et al. 2013) who are sceptical about the
current manifestation of nudge rather than believe it is necessarily bad.
The key question is about limits – the range of applications, weak
effect sizes, and longevity. These factors appear to relegate nudges to a
useful but limited range of applications which do not add very much to
the conventional tools of government. Nudges from this viewpoint do not
address the fundamental causes of behaviours and moreover cannot
address them. This is the argument considered in the first half of this

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 19 / Date: 10/1
JOBNAME: John PAGE: 20 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Is nudge all it’s cracked up to be? 107

chapter. Even though it is possible to see some truths to this criticism,

overall it is a misconceived position, partly because it is based on what is
easily observed in the current wave of changes and ignores the power of
behavioural interventions as they are developing currently. It also fails to
appreciate the key idea behind behavioural interventions, which is the
rooting of public action in an understanding of the cognitions of citizens
and helping them work through their constraints. The critique is based on
a false conception of nudge: that behavioural public policy is limited just
to light-touch interventions. Instead most public actions may be shaped
by behavioural science and improved by a better understanding of
individual cognitions. It so happens that thinking through these limita-
tions and cognitions does encourage a gentler, more decentralised, and
human-centric form of governance, but this is not something that
automatically follows from altering choice architectures.
The more conceptual critiques rest and fall on their own premises. If
the taste is for the psychological state then the modest interventions that
make up nudge are going to confirm this approach no matter how they
are introduced. But it is hard to support the idea of the psychological
state from the way in which these interventions have been introduced.
Nudge adoptions reflect the pluralism of central state organisations. They
tend to be patchily introduced, even though they can be transformative.
Behavioural insights and randomised controlled trials have largely been
introduced in an open and transparent way, which does not conform to
the image of psychologists secretly using their power to manipulate
citizens for government ends. In practice, the nudge agenda has emerged
with a lot of public attention, rather than the secret state taking control.
The same can be said for other critiques, which are about diverting
attention from the so-called serious tools of government, as there is no
evidence that this has taken place as a result of nudge. It is not possible
to prove the assertion that more nudge equals less regulation, even
though certain governments may have said that. Nudge works well for
both left and right, and for governments of any complexion. It is mainly
neutral, merely saying that the designers of any intervention need to
consider carefully how largely autonomous social and economic actors
are going to respond to the sanction or request. It can make governments
of the left, the centre, and the right more agile and effective. Nothing is
ruled in or out in behavioural public policy, at least only in terms of
practicability. Of course, the policy-maker may wish to rule out certain
activities on ethical grounds, which is the topic of the next chapter.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 06-Chapter6PJED /Pg. Position: 20 / Date: 10/1
JOBNAME: John PAGE: 1 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

7. The ethics of nudge

An important consideration for how far to nudge, which goes beyond
practicability and effectiveness, is moral and ethical constraints. Nudge
policies might work, but governments should not use them if they offend
moral values, are inconsistent with other values that are important to
uphold, or have negative effects on achieving other important objectives
that are based on following ethical standards. There is some overlap with
the criticisms of the psychological state, discussed in Chapter 6, in
arguments about loss of autonomy, but considerations about ethics move
the debate more toward the realm of philosophy and moral argument, so
they deserve a separate chapter.
Nudge has attracted a lot of argument in this vein, partly because its
advocates Thaler and Sunstein made great play of the term ‘libertarian
paternalism’, which implies that policies that are carried out in the
interests of the citizens without their consent could at the same time
uphold their freedom. As Thaler and Sunstein realise, this is a contro-
versial position to hold, and arguments for and against it are reviewed in
this chapter. In the end, this book’s argument is that it is better to accept
that nudge can at times be paternalist, but that paternalism is not
necessarily an ethically problematic position to hold. It might be accept-
able or ethically defensible to constrain freedom in certain circumstances,
provided various safeguards have been implemented, so that extreme or
controversial interventions have been ruled out or modified. There are
different ways of being paternalist, some of which uphold citizen consent
and legitimacy in a democratic state, and they can form into a fruitful
line of interventions in behavioural public policy.
It matters ethically whether the decision-making before the nudge was
carried out following democratic procedures and was fully transparent.
As nudges are directed to all the tools of government, thus affecting the
design of all or most public policies, they involve the same consider-
ations on human freedom as increasing taxes, introducing new laws, and
even going to war, in that democratically agreed policies that respect
human rights can be followed if there is enough provision for their
review and their effectiveness is demonstrable. In this sense, the argu-
ment in this chapter is that the debate about libertarian paternalism is a

108

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 1 / Date: 22/1
JOBNAME: John PAGE: 2 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

The ethics of nudge 109

red herring or diversion. That is not to say there are not moral arguments
to consider, or that respecting human choice cannot be beneficial in some
nudge policies, but getting beyond this debate is useful because it allows
academics and policy-makers to focus on issues that are of more
importance and can help guide citizens to make better choices in public
policy. Moreover, the debate about libertarian paternalism disguises other
ethical questions about the use of the tools of behavioural public policy.
Questions of the desirability of manipulation come to the fore, as does
the question of when deception and secrecy might be sanctioned, or when
some degree of anxiety or harm caused by an intervention can be
justified, or whether there are some nudges that are off limits to the
policy-maker and researcher. The debate about ethics and nudges
involves considering wider questions about the ethical constraints on
public action, and on the correct use of research to gain knowledge from
human subjects.

LIBERTARIAN PATERNALISM
The starting point for discussions of the ethics of nudge is libertarian
paternalism. Thaler and Sunstein (2008; see also Camerer et al. 2003)
argued for this position as a way to justify the extensive use of
behavioural interventions. The authors recognised that the term could be
seen as an oxymoron, as it appears contradictory: how could something
be paternalistic, which is about taking away autonomy, making a decision
in place of people acting purely on their own, but at the same time be
libertarian, where individuals are free from constraint to follow their
preferences and tastes no matter what the consequences are for them?
Although apparently contradictory, it is possible to defend this position.
The basic idea behind the libertarian paternalist argument is that
individuals are free not to follow a nudge and to take the opposite action
if they decide to do so. The whole set-up behind a default encourages
this. For example, the driver licensing website that prompts organ
donation, or even defaults to that choice if individuals take no action,
may be designed so people are free to reject the prompt or default and
with relatively little cost choose the option that they want so as to satisfy
their preferences. The paternalistic side is that the public authority is
hinting or arranging matters to influence the choices of individuals, so
that they end up choosing the one that is socially desirable and/or better
for them, perhaps unconsciously or by being prompted to do so. There is
a real-world influence of the state and public authority, as the whole
point of the evidence for nudge, as reviewed in the previous chapters, is

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 2 / Date: 14/12
JOBNAME: John PAGE: 3 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

110 How far to nudge?

that the message and its framing matter. People do change their behav-
iour as a consequence of nudges, so their choices are being affected by
what people in public office decide. But as choice is respected as well as
being influenced so there is libertarian paternalism: ‘choice architects can
preserve freedom of choice while also nudging people in directions that
will improve their lives’ (Thaler and Sunstein 2008: 252).
One key issue in the use of behavioural science – and one that
becomes important when reflecting on think later in the book – is the
extent to which people are consciously exercising their freedom when
responding to a nudge. For the libertarian, it matters that individuals are
aware of the choice being offered. If the nudge is designed to prevent
such awareness emerging, or working entirely automatically, it is hard to
see it as an exercise of free choice. Nudge advocates defend the use of
automatic processes because if individuals did reflect on their choices
and agreed with them, and might even have been trying to achieve the
outcome and failed, they would probably thank the public officials for
making the choice easy for them. Officials can say that rather than
overriding freedom the public authority is removing the framing biases
that limit making effective choices.
This argument becomes one of the selling points about nudge in that
people’s freedom is respected. Light-touch policies can be used by
politicians who do not want to intrude on people’s liberty, for example
avoiding a policy of banning some food or regulating its content. This
characteristic has the advantage of building support for nudge and
creating the conditions for successful policies. The light-touch nature of
nudge affects the ethical justification of nudge in that these policies may
have more legitimacy and acceptability.
It is possible to attack libertarian paternalism because even soft
paternalism requires some constraint on autonomy. If people are short of
time, they don’t have the opportunity to reflect on all the choices. If
government limits the information upon which choices are based and
individuals do not have enough resources to challenge the decision, the
effect is just as if the proposal had been mandatory. Had people been
offered a more balanced choice, leading to a more active decision on
their part, they could have made a freer selection of their course of
action.
There is also a defence of more libertarian arrangements even when
there are human biases. Some economists, such as Sugden (2008), claim
that markets can organise themselves even with incoherent individual
preferences and still generate welfare gains. Sugden takes the view that
the argument of libertarian paternalism makes an assumption that people
are rational underneath, merely that they have been diverted from their

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 3 / Date: 10/1
JOBNAME: John PAGE: 4 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The ethics of nudge 111

true interests by biases. But this reasoning is contrary to the findings of

behavioural economics, which does not always assume that rationality
operates fully in the first place. It is just not clear what the true
preferences of citizen actors are, so lack of certainty weakens the
justification for the paternalist claim for overriding them. The implication
is that there is no planner underneath the doer. From this perspective, it is
not coherent for the state to override preferences as if there were market
failure. Sugden’s position is that overriding those preferences limits
freedom as much as in classic paternalism.
Much turns on the term paternalism. Its use by Thaler and Sunstein is
thought to be ambiguous (Sugden 2016), wandering between what people
would choose for themselves and what is good for them. Moreover, as
Sugden writes,

self-acknowledged self-control problems are a lot less common than many

behavioural economists seem to think. Even if behavioural economists or
policy makers feel confident that people’s lifestyle choices are based on some
kind of error, they should not jump to the conclusion that the error is a
self-acknowledged failure of self-control, or that (as Thaler puts it) it is what
those people would themselves call an error. (Sugden 2016: 122, italics in
original)

These nudges are turning what affects a small number of people and their
actions into a more general constraint on human action.
It is hard to defend the libertarian aspect of libertarian paternalism
when libertarians say that the whole point of freedom is the latitude to
make bad choices if that is what people want to do. It is a hard argument
for nudge advocates to extract themselves out of. Thaler and Sunstein try
by referring to the duty to explain things to people, so if a government
knows that an activity is bad it must explain this. One way of thinking
about this problem is to consider a public information sign warning
people of a cliff edge, which might appear to be paternalistic (similar
examples appear in Mill, and it is also the argument made by Dworkin
(1988) that government can act in the case of insanity), but is placed in a
prominent position to nudge people not to go further; they are neverthe-
less free to take the risk. As Anderson (2010: 372) notes, ‘the appeal to
soft paternalism will work for Thaler and Sunstein only if all choice-
improving nudges are to be understood along the lines of misinformation
and temporary insanity’. But, of course, that is not the case, as most
people have some freedom and rationality, even if they do not exercise
them at all times. Many situations make less clear what is the legitimate
role of the state.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 4 / Date: 10/1
JOBNAME: John PAGE: 5 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

112 How far to nudge?

Thaler and Sunstein are also vulnerable to an attack that freedom has
been limited by the nudge itself. Citizens were not free to choose an
alternative option because the choice architecture guided them sub-
consciously to the government’s ends. When delivering behavioural
public policy, the decision-maker cannot realistically inform the citizen at
the point of decision, either in terms of the effectiveness of the nudge or
delivering a reflective choice, for example, at the point of say donating
organs (though, as this book argues, this can be improved). The whole
point is that the opportunity to nudge needs to be wrapped up with other
procedures, such as a standard letter or the design of a website, and
nudges needs to work quickly, easily, and automatically. This point
relates to a more general criticism of nudge, that it is manipulative and
autonomy limiting through using psychological techniques (e.g. Mols et
al. 2015).
To get out of the problem that the practice of nudge appears to limit
choices as they are being made, and that consent cannot be sought,
Thaler and Sunstein (2008: 244) argue for the publicity principle that
‘bans government from selecting a policy that it would not be willing to
defend publicly to its citizens’. They hold back from making all nudges
public at the point of carrying them out, saying that many nudges work
best when focused on automatic processes. Their proposal respects
citizens by policy-makers being willing to argue for the benefits subse-
quently. These are dangerous waters, as Thaler and Sunstein well know,
and amount to an after-the-event justification, like for secrecy. The moral
force of an argument does not depend on when it has happened: it should
apply both beforehand and afterwards. As Anderson writes, it puts them
in the company of planners who say about their decisions that no one
will really mind. He goes on to write: ‘the espousal of transparency and
publicity constraints comes across as an artificial and ad hoc declaration
of values that belies a lack of real interest in the importance of ensuring
that those subjected to these subtle forms of state power understand the
underlying rationale’ (Anderson 2010: 374). Even if people agree with
the decisions, they often like process to be followed and to be consulted
in any case. Part of the problem is that almost anything can conceivably
be justified after the event. Most writers and thinkers know that sophis-
ticated arguments can be put forward for almost any position, but this is
not the same as arguing face-to-face with citizens to justify their
restriction of choice.
Thaler and Sunstein (2008) also face a cliff-edge argument of their
own, as they acknowledge that there is nothing to stop any infringement
of the liberty of the citizen from being seen as a nudge: there is no
hard-and-fast distinction between hard and soft paternalism. Anticipating

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 5 / Date: 10/1
JOBNAME: John PAGE: 6 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The ethics of nudge 113

this argument, Thaler and Sunstein (2008) argue that difficult cases
emerge and can be dealt with by additional procedures for consideration.
It might be said in their defence that public policy is full of these
balancing decisions where no outcome perfectly respects one principle,
but where policy-makers need to make the best choice that maximises
their values and ensures the policy does not lead to perverse effects. There
is no slippery slope, providing good reasoning has been undertaken.
One strong argument for defending the claims about freedom is that
choice is anyway constrained by government policies and a range of
private sector interventions, as well as by social institutions. As Thaler
and Sunstein (2008: 252) write, ‘nudges are everywhere, even if we do
not see them’. Liberals have classically argued against libertarians, such
as Nozick (1974), that there is no first state of liberty from which an
intervention or law removes freedom. Whatever a public agency does,
whether to nudge or not to nudge, affects freedom. Choice architectures
have to be created whatever the public agency does. The question is
whether they are designed with the right choices embodied in them.
In the view of critics, in spite of all the gentleness of nudge, there is no
getting away from the fact that nudges deprive individuals of their
autonomy or do not respect individuals’ full autonomy. The objection, as
Anderson (2010: 374) writes, ‘concerns the marginal status of respect for
the autonomy of those targeted for nudges’. There is an elite whose
members make decisions that affect others. It might be more straight-
forward just to argue for a strong paternalist policy to justify this
ethically, as Conly (2013) does on the basis of a duty to prevent harm.
But can a version of the autonomy argument work when delivering
nudges? Given the reality of policy-making and modern society where
citizens are busy with their own lives, pay taxes, and delegate decisions
about welfare to politicians who run expert and professionalised bureau-
cracies, some notion of encouraging the full autonomy of citizens to own
the nudges and to debate them is not realistic, but the question is whether
the balance could be tipped a little bit in the direction of citizen control
and autonomy. There may be some halfway house between secretive
nudging and full-scale deliberation: ‘a choice architecture that makes it
easier to avoid regrettable decisions or, rather, various measures to
improve individuals’ decision-making capacities, say, through education,
“buddy” arrangements, decision-making heuristics, etc.’ (Anderson 2009:
375). It is instructive that one of the limiting conditions that Thaler and
Sunstein (2008) offer is the chance for citizens to be given an opportunity
to think and reflect before they make a bad choice. The authors get closer
to the deliberative nudges favoured in this book, and Chapter 8 suggests
a way forward.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 6 / Date: 10/1
JOBNAME: John PAGE: 7 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

114 How far to nudge?

FREEDOM AND THE PSYCHOLOGICAL STATE

One consideration is whether the behavioural sciences, in particular those
that rest on cognitive science, have re-orientated the role of the state to
take greater power over the citizen based on detailed knowledge about
individual behaviour and how to manipulate it. This argument appears, as
was discussed in Chapter 6, in the concept of the psychological state
(Jones et al. 2013), which implies that more harnessing of behavioural
sciences may reduce autonomy and freedom. There is a kind of tech-
nological argument at work here, which is that former sorts of authority
were not efficient and could not control individual behaviour effectively.
Citizens benefited from a degree of inefficacy in government bureau-
cracies that worked with imperfect knowledge. The argument about
behaviour change knowledge is that it can realise the Weberian ideal of
the rational bureaucracy that oppresses freedom, much as Weber feared.
This argument depends on a degree of homogeneity which it is hard for
any scientific programme to achieve, let alone the diverse field of
behavioural sciences, not least because it has not been fully adopted and
the same messy bureaucratic structures and involvement of politicians
and many other actors are all in play, as discussed in Chapter 5. In short,
the use of different forms of knowledge is an important development, but
it cannot be seen to be a step-change in the dominance of knowledge that
the argument about the psychological state implies. The argument is also
that certain categories of people are subject to behaviour change inter-
ventions, which may not be true; and, in any case, certain categories of
people, notably the poor, have always been on the receiving end of
government policies and research efforts, such as those who are on
welfare. Nudge is no different in that respect. And a lot of nudges are
applied to the whole population, such as tax reminders or organ donation
prompts. Of course, everyone is in receipt of nudges from the private
sector, such as marketing, advertising, social media content, and tele-
vision campaigns.
As Thaler and Sunstein (2008) argue, much depends on the neutrality
of the policy-makers and the proposals they advocate. If nudges have a
political angle or are ideological or are about imposing the preferences of
policy-makers for a particular way of life, then nudges are doomed. It
gets worse if the policy-maker advances what may be considered to be
bad objectives from nudges, such as reducing the influence of political
opponents, or getting re-elected. This is one of the key worries about
nudging; as Schmidt (2017: 406) argues, ‘the granting the state the power
to systematically nudge its citizens indeed puts some people in a position
to impose their will on others’.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 7 / Date: 10/1
JOBNAME: John PAGE: 8 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The ethics of nudge 115

It is possible to sustain this argument by referring to the arguments of

the previous chapters, in that nudges have appeared from scientific
investigations and so are evidenced by social scientists, based on research
from trials. Moreover, the objectives of nudge policies, at least in their
first wave, are ones that few will object to, such as paying taxes on time
or improving health outcomes. Moreover, the nudge argument implies
some faith in the political process in that citizens expect that politicians
are there for the public interest and bureaucrats want to improve
outcomes. The democratic state has as part of its function the improve-
ment of welfare, and that public officials are motivated by a public
service motivation, for which there is some evidence as witnessed by
surveys. When taking into account the power of the private sector, which
has no clear path of democratic accountability in spite of progress on
corporate governance, to use nudges which are not designed to be in the
public interest but to make profit, objections to democratically controlled
nudges weaken. The contrast between using nudges in the private sector,
where the consumer can usually exit from the provider using nudges, and
the situation of no exit from the monopoly state does not apply, because
the consumer might be genuinely unaware of the nudges in the private
sector, but it is possible to find out about public sector nudges (Schmidt
2017: 415).
On the other hand, politicians do authorise some policies to get
re-elected, bureaucrats act from self-interest as well, and private interests
sometimes capture bureaucrats, a point acknowledged by Thaler and
Sunstein (2008: 240). This lapsing from a generally good path is another
reason to argue for more accountable nudging, which is done in public so
that the actions can be properly scrutinised, not to establish after the
event whether people would have agreed with them. In part, the slightly
ambiguous text at the conclusion of Nudge (Thaler and Sunstein 2008)
supports such an open process.

SUNSTEIN’S LATER ARGUMENTS

In later books, Sunstein elaborates on his defence of libertarian paternal-
ism. In Why Nudge? (2014a), he argues strongly against the limitations of
Mill’s harm principle, which justifies limits on freedom if actions harm
others. He argues that Mill was writing when it seemed reasonable to
leave people to their own actions, based on rational self-assessment. With
social science research, it is now known that people make massive errors
when they make judgements. That provides the justification for inter-
vening. Just as there is a justification for the state intervening in market

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 8 / Date: 10/1
JOBNAME: John PAGE: 9 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

116 How far to nudge?

failure, so it is legitimate for the state to intervene in what he calls

‘behavioural market failure’. Mill in any case provided a justification for
overriding preferences in the case of children or so-called primitive
peoples. Sunstein holds back from saying these behavioural biases are a
good reason to be paternalistic on their own, as other reasoning needs to
be engaged in, but it is a good starting point. Sunstein still defends
nudges as superior, as they are liberty-enhancing, defending what he calls
‘soft paternalism’ against a harder kind. Of course, as Sunstein realises,
the distinction has its problems, not least the lack of a clear division
between the different kinds of paternalism and the possibility of slippage
between them. But he defends a welfare-enhancing policy in the form of
soft nudges, particularly if welfare is seen in a broad sense.
These ideas are picked up in The Ethics of Influence (Sunstein 2016),
which summarises his earlier position that ‘ethical issues largely turn on
whether nudges promote or instead undermine welfare, autonomy, dig-
nity, and self-government’ (Sunstein 2016: 11). A government might in
fact have an ethical obligation to nudge to prevent disease and to stop
people acting in bad ways. The arguments are similar to those for the use
of other tools of government, such as fines, which is close to the position
adopted in the summary below. At times nudging might undermine
people’s dignity, so it might be right not to do it, such as if people feel
tricked by the nudge. For example, the social pressure argument of
exposing voter turnout to neighbours might not be the right way to
mobilise citizens to turn out at the polls (Gerber et al. 2008). Overall,
public officials should aim to be trustworthy to build trust with the
citizens when executing nudges, as they are expected to be more
generally.
There is still the preference for choice-enhancing interventions, but the
defence is less strong on libertarian grounds than in Nudge (Thaler and
Sunstein 2008), which is entirely understandable given the attacks that
have been made against libertarian paternalism even though Sunstein still
keeps an element of this earlier defence. He maintains his view that there
are already many nudges in operation, which means that government is
simply replacing one nudge with another. He makes an argument for
deciding the moral content of nudges on a case-by-case basis. The
promotion of autonomy is an important goal. In fact, many nudges do
promote reflection and are essentially educative (Sunstein 2017a).
Sunstein (2016) still prefers defaults to more active choices, at least in
some circumstances, where people have very low knowledge about the
choice and thinking about policies is difficult. In these cases, defaults are
better. Manipulation is not to be approved of, largely because of the ends
to which manipulation is targeted, and that individuals would not on

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 9 / Date: 10/1
JOBNAME: John PAGE: 10 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The ethics of nudge 117

reflection agree with the position advocated. Most nudges pass through
the manipulation test, but not all do, and some are on the borderline.
Here there is a more complex debate to be had.

PUBLIC OPINION AND NUDGING

Sunstein’s (2017a) book has a chapter on public opinion and nudging,
which is about whether people like nudging, that in the main they do as
they approve of these policies (see also Hagman et al. 2015). Ultimately, as
Sunstein fully realises, ethical arguments do not turn on public opinion.
Just because the public approves of the policy does not make it right. But
the findings help the arguments that Sunstein wishes to make. If the public
really disapproved of nudges then it could be argued that they were
manipulated, as on reflection they would not agree with the interventions;
but approval indicates that they would not mind too much being nudged
and approve of the arguments for promoting behaviour change, giving
some support to Sunstein’s position. Of course, it is important to be
cautious about public opinion, in that such answers might not be the
consequence of much reflection, and that opinion poll responses can be
framed by the question wording. But this reservation can be assuaged by
the basic finding that people agree with nudges if they approve of the end
being promulgated. The main difference is that partisan affiliation affects
the approval rating, which indicates that ideology may play a role,
undermining in part the neutrality criterion in nudge.
Other distinctions in the willingness of the public to approve nudge
appear in Jung and Mellers’s (2016) study, which manipulates frames as
well as tests for support. The authors focus on whether the nudge is
stated up front and is overt, or whether it is covert and subconscious,
which replicates the distinction between system 1 and system 2 forms of
decision-making. Unsurprisingly the US public tend to favour the more
overt kind, which supports the think approach to nudging argued for in
Chapter 8 (see also Felsen et al. 2013). People also prefer the pro-social
nudges to the more pro-self ones. Those that are preferred are about
collective benefit, such as recycling and other collective goods. The less
paternalistic nudges get more approval from citizens. Overall the vast
majority of nudges were supported in this study. Political views also
condition the acceptability of nudges. The responsiveness to nudges
shows considerable heterogeneity, whereby some people are more sus-
ceptible to nudges than others, while some show reactance. As in
Sunstein’s approach, the acceptability of nudges is altered if they are
personalised, suggesting a future direction for the design of nudges.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 10 / Date: 10/1
JOBNAME: John PAGE: 11 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

118 How far to nudge?

The heterogeneity is not just limited to type of person but may vary
cross-nationally, which may affect the extent to which policy-makers can
justify nudges to their citizens. Petrescu et al. (2016) also manipulated
the information conveyed, finding that the acceptability of limiting
portion size was less acceptable in the USA than in the UK (any UK
person who has visited a US restaurant away from the metropolitan
centres won’t be surprised by this conclusion), though there was no
difference for other government interventions to reduce consumption of
sugar-sweetened beverages. In this study, education was preferred to
regulation, with support for nudges somewhere in between. Interestingly,
this study did not find a difference in support between conscious and
non-conscious processes.

MOVING BEYOND LIBERTARIAN PATERNALISM

As the opening paragraphs of this chapter indicate, the debate about
libertarian paternalism is somewhat fruitless. It has encouraged behav-
ioural scientists to defend positions they should not need to. The position
that can be adopted is that nudges should be preferred not because they
maintain liberty, but because they are good public policies. Policy-makers
may adopt them if they are authorised correctly, have evidence behind
them, are evaluated properly, and where any potential negative effects are
outweighed by the benefits, so long as individual rights are not violated.
In this way, the justification of public policy remains within the classic
utilitarian tradition of ensuring the happiness of the greatest number,
subject to constraints on moral and rights grounds. Policies are needed
because they improve welfare. In fact, in later writings Sunstein (2014a:
18) moves closer to this position rather than defending libertarian
paternalism. Nudges are defended as a form of paternalism, but the aim
is to make policies more accountable, democratic, and responsive at the
same time. They can be better defended the more policy is transparent
and democratically controlled. This is the essence of Schmidt’s (2017)
argument.
Such an argument is helped by the ‘All tools are informational now’
position of this book. If there were a clear choice between nudge policies
and tool-based measures for all policy choices, then it might be easier to
distinguish more clearly autonomy-enhancing nudges from authoritative
commands. But, in fact, nudges are tied up with every kind of policy
intervention, making it impossible to distinguish between a tax change
and the nudge for example, because the nudge is embodied in how the
message is communicated. There is no autonomy that is violated in a

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 11 / Date: 10/1
JOBNAME: John PAGE: 12 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The ethics of nudge 119

norm-based message, because it is impossible not to pay taxes. If the

person does not pay, there are fines, all of which information is contained
in both the nudge letter and a non-behavioural equivalent. What the
behavioural science is doing is to make the initial communication more
social and more effective as a message.
With public policy, there is a danger that the intervention will violate
ethical constraints, much as a research project would (see John 2017a),
so that policy-makers like academics have to check whether any harm has
been caused, such as distress from having the policy, and whether this
can be justified by the benefits generated or can be limited in some way
in the delivery of the policy. As with university ethical guidelines, the
moral argument is not that deception is necessarily wrong but that it may
or may not be justified in a case. If there is no other way of doing the
policy, then after the event measures to inform citizens should be in
place. As Sunstein (2016) discusses, these situations may not always be
needed, as it is possible to get results even when people are being told
about being nudged (Loewenstein et al. 2015), though this study was
with an online sample and needs more replication and extension. It still
remains possible that telling the respondent will create reactance, where
people resist messages. In this case, it might be ethically defensible not
to convey the message.
Deception is an important problem and should be avoided, especially
with nudges that work well when delivered without the knowledge of the
participant (see Bovens 2009). A more general concern comes from
manipulation, that people might be encouraged to do things they do not
really want to do, which arises from the way the information is
presented, even if just short of deception. After-the-event justification that
it is in the respondents’ interests is not a strong enough defence, and, as
with other kinds of more controversial nudges, certain strong kinds of
manipulation will be off limits as a nudge, or the nudge can be altered to
make it less manipulative (see Wilkinson 2013). There are also special
cases where people might be more vulnerable and make bad choices as a
result of nudges, for example if a mentally impaired person chooses a
default pension scheme when it is not in the person’s interests to do so.
Overall, it is important to bear in mind that nudges are different from
each other. Contexts and populations vary, and this adjusts the moral
arguments in play (Bovens 2009).
It could be possible to build in more autonomy to the nudge than a first
design might envisage. It might involve more publicity about the nudge
at the point of delivery or ensuring that all the facts conveyed in the
nudge will improve welfare, for example a statement of what current
income would be lost if a pension default were not reversed. The

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 12 / Date: 10/1
JOBNAME: John PAGE: 13 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

120 How far to nudge?

information could also say that some people prefer not to save for
pensions because they value current pleasure so much. In addition nudges
could go the extra mile in ensuring that their intentions are good for the
long-run benefits of those who are nudged. In the view of many critics of
public policy (e.g. Dobson 2011), nudges come across as depoliticised
and rather technocratic: by building in more thinking about the public
choices, and being aware of different stages and joins in the political
process, a broader and more legitimate form of intervention could
emerge.
The joining of ethical thinking about public policy to conventions in
research governance comes about because of the use of trials. In research,
trials are thought to have special ethical problems from getting people to
act in ways they would not otherwise choose to do, which need
justification. It is the same with public policies that involve changing
direct contact with the citizens, which is not altered from some ex-
ogenous factor but because of the science and knowledge of the
researcher and policy-maker. While policy-makers can assume that a
form of ethical approval comes from political authorisation, that politi-
cians are put there to implement public policies for which they are
judged at a subsequent election, and that their actions are accountable in
the public sphere, such as parliamentary committees, questions in the
chamber, or the actions of the courts, in practice democratic accountabil-
ity and scrutiny cannot extend to the bureaucrats and their decisions at
such a detailed level; hence there is a need for some kind of ethical
sign-off or the building of ethical thinking into public policy. In this way,
some of the more extreme nudges, which might involve deceiving people
or making them think they are doing better than they really are, say in an
employment search, can be ruled out or modified as a result of ethical
scrutiny.
It so happens that nudges are less likely rather than more likely to
breach an ethical constraint, because of their gentleness, at least in
general, and because they tend to go with the grain of human prefer-
ences. In choosing between compelling people to do something, punish-
ing them if they do not do it, and a nudge, where people are encouraged
to do something, the criterion is not just respect for freedom but
non-coercive approaches being more likely to work in today’s anti-
authoritarian climate. People resent being told what to do, and they resist
strong messages. They are more likely to be responsive to the state or
public agency that nudges, because the approach implies a conversation
with citizens rather than the state or agency stamping its authority over
them. The listening and sensitive state is what this book is seeking to
encourage.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 13 / Date: 10/1
JOBNAME: John PAGE: 14 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

The ethics of nudge 121

CONCLUSION
The core argument in this chapter is that although there are important
ethical issues to consider when carrying out behavioural public policies
or nudges, which may even result in their modification or abandonment,
there is no fundamental ethical objection that rules nudges out or
prevents a sizeable number of them from being implemented. This simple
claim is because nudges are public policies authorised by democratically
elected governments that are subject to review. Those who introduce
nudges are accountable just as any other policy-makers seeking to
improve welfare. Ethical acceptability depends on how open and effective
are the procedures for authorisation. Like public policies, nudges raise
ethical issues of harm to people and need to ensure the right balance of
trade-offs between objectives citizens care about, such as welfare on the
one hand and equality, justice, and rights on the other. Nudges are in the
mainstream. Because nudges are so closely linked to other tools of
government, such as laws, fines, incentives and so on, it does not make
sense to distinguish particular moral claims about using a nudge from
those deriving from using another instrument. Issues around proportion-
ality, fairness, responsiveness, and reasonableness appear just as much
with laws and fines as they do with nudges, in that policy is more
desirable if it accords with democratic values and is more likely to work
in terms of getting compliance if these values influence the design of the
policy instruments.
To the extent that there is a clear choice between a nudge and a more
authoritative instrument of government, there is good reason to choose
the nudge, because, as Sunstein (2017b) argues, it is less likely to cause
reactance; in other words, nudges might be more effective as a result of
being light-touch. They have been discovered as a policy partly as a
result of the failure of traditionally applied tools of government (see John
2011). Some nudges (but not all of them) might enhance autonomy and
can help guide citizens to make more autonomous choices in the longer
term. These are the types of nudges that it would be best to approve of
and to develop. In that sense, there could be a moral purpose behind
nudging, but only if additional aspects to the intervention are introduced,
in particular to the way individuals are guided. Nudges are not by
necessity autonomy enhancing, but they can be. Nudges could be further
developed to be more in line with the values of democracy. This project
is the task of the remainder of the book.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 07-Chapter7PJED /Pg. Position: 14 / Date: 10/1
JOBNAME: John PAGE: 1 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

8. Nudge plus and how to get there

Previous chapters of this book have provided an extensive review of the
development of behavioural public policy, largely with approval of the
valued added of such a programme. There has been an acknowledgement
of the ease with which ideas in behavioural economics have transferred
into concrete policy proposals that sit well with the preferences of
bureaucrats. These measures have been tested with randomised evalu-
ations and then introduced into the standard operating procedures of
agencies. The central parts of the book show the diffusion of these
policies across fields of activity and jurisdictions, and the general
welfare-enhancing benefits of such changes are in evidence. The criti-
cisms of nudges, at least in a simple form that they do not work and offer
no benefits, do not stand up. Moreover, there are relatively few special
ethical constraints on the use of nudges, even though ethical constraints
should not be waved away and are important considerations in the design
of these policies.
Having shown the development of nudge, its benefits, and the lack of
constraints on its use, the task now is to see if the policy agenda can be
extended and expanded; whether there is greater potential for nudging,
especially given the variation in the kinds that can be done, as there are a
considerable number of choices about how such nudges are carried out
and which ones are followed. The argument is that nudges can move
more in a direction that involves some kind of reflection on the part of
the individual, at least as an element of the nudge, or a process of
reflection before or after a nudge has been delivered, ensuring that a
nudge operates on the boundary of system 1 and system 2 thinking.
When more reflective processes are engaged, the idea is that nudge gets
closer to the ideal of democratic self-government, as a means for the state
to help its citizens take back some control over their lives in the long run,
which is part of the attractiveness of Sunstein’s soft paternalism.
As well as expanding the scope of nudging, it is possible to alter the
subject of the nudge, that is the person or organisation that gives the
nudge, so that nudging is more inclusive and can address inequalities of
political power. This inclusion involves more equality between the citizen
and the state. It can be the basis for a more open conversation about

122

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 1 / Date: 10/1
JOBNAME: John PAGE: 2 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge plus and how to get there 123

public policy, which is based more on think and deliberative mechanisms.

The argument is that a more extended and radical form of nudging could
come about that builds on the strength of what has been achieved, which
can challenge elites as well as be used by them.

THE THINK IN NUDGE

A core idea in behavioural sciences is system 1 or automatic processes:
that humans act from unconscious mechanisms in their brains, which
have biases or heuristics baked into them, so that individuals make many
decisions, sometimes important ones, based on these intuitions, ‘thinking
fast’ to use the language of Kahneman (2011). There may be many
advantages to working in this way, in that it reduces the need for
cognitive capacity, makes many choices unproblematic, and allows
human beings to concentrate on things they want to think about, so that
many actions, such as driving a car, doing the shopping, or even carrying
out routine tasks at home like settling bills and doing housework, are
routine and do not burden people with unnecessary choices. These
shortcuts can be useful in making larger decisions if they remove a long
set of calculations that would in any case have caused the individual to
come to the same decision. If an individual is loyal to a political party
that represents his or her interest and so long as the party does act in the
interests of the individual, then what is the point of assessing the choices
and reading all the party programmes if in the end the person is going to
make the choice to vote for that party? Early work in political science
recognised this, for example Downs (1957), who regarded heuristics as a
way to economise on information searching, called ‘rational ignorance’.
Big and important decisions may be made in the same way. But how
do individuals know whether to let the automatic processes take hold or
to stop to consider options, that is move to Kahneman’s system 2
processes of slow thinking and more careful consideration? For example,
house purchase, as any real estate agent knows, is highly automatic, with
many decisions taking place in a matter of minutes as the potential
purchaser walks round the property. This rapid decision-making can lead
to poor decisions. It also means that purchasers are subject to manipu-
lation by agents and disguising of the quality of the property by vendors.
But a lot of purchases are done in communities known to the buyers,
with say rows of terraces where the houses are quite similar, and the
buyers typically know these areas and properties well, as they tend to
move within the area in which they already live, so the final decision can
be based on matters of taste about decor and small variations in design.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 2 / Date: 14/12
JOBNAME: John PAGE: 3 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

124 How far to nudge?

The level of risk may be reduced by the choice of a standard property.

Perhaps individuals know about this risk, which explains the conservative
nature of many moving choices. Moreover, regulators can be used to
check up on details later on, and rules may allow a period of reflection
afterward. But, of course, many decisions, as has been found in health,
happen over a long period of time where the error is not detected or
corrected. Some hasty decision-making causes people to buy poor
financial products or household appliances. If people can avoid getting
feedback on decisions they fear or regret, they do. There is hence a
possibility of having the opportunity to rethink options but choosing not
to. With political parties, supporters may stay loyal to their brand long
after their party has changed its policies to please new types of voters.
Rather than pursuing rational ignorance, people need to keep half an eye
on information flows and update their views from time to time.
Rather than impose their own ideas or try to reason with people,
behavioural scientists find that it makes sense to take advantage of
automatic processes to design proposals that move people away from
their own automatic processes or those used by the private sector. They
combat foe with foe in effect. The nudges that work at an automatic level
can correct for errors without the citizen being really aware of them. As
the initiatives in this book have shown, there are a number of advantages
of taking this approach when there is discretion in any case over the
delivery of policies. These initiatives can have large effects. But there are
some limitations, as Chapter 6 indicates. One is that the changes may be
focused on the minority of people who are willing to change their
behaviour. The effects may diminish over time. The approach concen-
trates on small behaviour changes, whereas dealing with a large public
problem, such as the environment, needs sustained and greater behaviour
change over a range of issues. There is also the problem of the lack of
consent of the citizens to those changes, which can cause the problem
that if they do not agree with the objectives of the change they may react
negatively from reactance if they find out. However, it may be the case
that it is easy to apply this approach at the early stage of the behavioural
agenda when the choices are ones that no one would disagree with and
individuals would not mind. More complex behaviour changes might
involve making more fundamental changes where it might not seem
obvious to individuals and it could take time for them to realise it is in
their long-term interests to change their behaviour.
The way out of the limitation is to say that all the nudges that have
been carried out can be banked as good pieces of public policy, but also
there is a case for reviewing where the agenda is right now, and to say
that an opportunity exists to move it toward a different kind of nudging

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 3 / Date: 10/1
JOBNAME: John PAGE: 4 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge plus and how to get there 125

that is more inclusive and based more on self-control. But policy-makers

should be cautious in considering more deliberative ways of involving
citizens in public policy, which might sound a surprise given the way this
argument is proceeding. But there are lessons from the fate of such
initiatives in deliberative public policy.
The main argument for a more deliberative approach derives from the
idea that citizens can shift their preferences if they debate about a policy
initiative with more time, access to expert information, and appropriate
facilitation (Dryzek 2000). This idea can be built into the design of many
institutions as a form of democratic innovation (Smith 2009) beyond
citizens’ juries and deliberative polling (e.g. Fiskin 1995) favoured in
much of the pioneering work on deliberation. In particular, a more
deliberative approach can be incorporated into public policy, through
providing genuine citizen input into decision-making, before major
decisions are made rather than afterwards just to legitimate them or as
meaningless consultation. In the design and also the delivery and
implementation of public policies, citizen input can guide public deci-
sions, which feeds into more responsive and efficient policies. This
benefit has been illustrated in case studies of citizen initiatives, such as
participatory budgeting in Porto Allegre or policing in Chicago (Fung
2006). From the public policy perspective, it is not a case of policies
emerging top down, but being discussed and owned by the people
involved. As a result of ownership, policy outcomes improve, partly from
public bureaucracies working more efficiently and responsively, but also
because citizens are cooperating in acts of collective endeavour. It is not
hard to see this as a form of behaviour change policy whereby behaviours
are modified as a result of what happens in the policy process. Behaviour
change can happen individually as a form of response to public policies
and decisions taken in the public realm, and it can be a form of collective
action whereby people overcome the obstacles to behaving in more
beneficial ways for society as a whole.
When the deliberative project has a particular policy focus it can be
compared to the nudging as described in this book. This was the idea
behind Nudge, Nudge, Think, Think (John et al. 2011), a book based on
collaborative research that tested the efficacy of nudge as opposed to
think strategies, using trials as the main method of evaluation. What was
found was that most of the nudge policies worked in the sense of
delivering changes in behaviour, with reasonable effect sizes given the
mild nature of the interventions, but it was much harder to conclude that
many of the think policies were as effective. In one case the think
strategy was good at getting the citizens to consider behaviour change.
The agency responded to their complaints by asking them to consider

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 4 / Date: 10/1
JOBNAME: John PAGE: 5 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

126 How far to nudge?

volunteering, but the organisation found it hard to respond with activities

for the citizens to participate in. The citizens were mobilised to turn from
complainers into deliberators, but the infrastructure and support were not
there to allow them to do so.
Another experiment tested whether a nudge or a think would be more
effective in getting students to consider donating their organs in the event
of their death, comparing the nudge (nudging information), a think (a
discussion), and a placebo. The nudge worked the best and then the
placebo, but the think did much worse than the placebo. Thinking and
debating do not necessarily lead to the desired policy outcome, and
nudging might be the best choice in many circumstances. But maybe it
was good that human subjects did not conform to what was expected of
them – they were entitled to consider the policy area in their own way, so
they decided that donations were not for them, having thought about the
problem. How can any democrat argue against this kind of reasoning on
the part of citizens? Otherwise thinking would only be fine if it came up
with the right decision from the state’s point of view. The example also
illustrates some of the costs of a deliberative approach to public policy in
that it is not usually possible to have the space to discuss things fully, as
in citizens’ juries or even budget discussions or policy implementation.
Could such small thinks really be a recipe for general policy advocacy? It
is hard to get citizens to deliberate, as was found in the online scaling-up
exercise in Nudge, Nudge, Think, Think (John et al. 2011), where few
engaged in this way, preferring to listen (or lurk) rather than engage (see
also Smith et al. 2013). It was hardly surprising there were modest effects
in this experiment. This limitation goes back to the conventional criti-
cisms of non-representative democracy mechanisms that they place a
high burden on citizens in terms of their time and capacity to process
large amounts of information. Partly as a result, they tend to attract a
minority non-representative sample of the population, including those
who are available to be manipulated, and the subset of people who do
change their behaviour is very small. The online deliberation reported in
Smith et al. (2013) still only attracted a minority of the citizens, who
themselves were self-selected respondents in the polling company’s
online survey bank. These were further self-selected into those who
agreed to participate and did not drop out. But if deliberation does not
easily scale up, the question is whether nudges, which can reach
everyone, can include a deliberative and thinking element, ensuring that
all affected do this to a certain extent rather than self-select. Can nudges
do justice to the complexity of the policy choice under consideration?

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 5 / Date: 10/1
JOBNAME: John PAGE: 6 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge plus and how to get there 127

MORE OPEN NUDGES: NUDGE PLUS

In order to move beyond the dichotomy between nudging and thinking, it
is important to return again to the important distinction between system 1
and system 2 thinking. None of the proponents says there is a hard-and-
fast distinction. It is true that different parts of the brain are involved and
there is secure knowledge in neuroscience behind these findings. But it is
not possible to say that automatic processes are only engaged with
nudges and that other parts of the brain are not working to a small
degree. Even in the most impulsive moment most people have an
awareness of the choices they face, or hold back just for a second. Going
back to that house-buying decision, which is sometimes made at the
viewing, there is often a moment in the car or on the bus on the way back
home when the person is struck by the importance of the decision made
and may feel a sense of regret , which is again an automatic response, but
is based on an awareness of the choice and can lead to a reconsideration
of the decision down the line. The example is a reminder that the time
period for the operation of a nudge can be long, which might include a
rash decision, reflection, and then another change of course (which
justifies the creation of choice architectures that allow for changes in
decisions or cooling-off periods). Another example might be an unrealis-
tic diet that is not kept to, but the person thinks of a more reasonable way
to reach the goal of achieving weight loss in the long term. The thought
process is that someone can understand a breach of the diet for special
occasions or treats, but the idea is that the individual finds a sustainable
strategy over the long term which involves steering, responding to
situations, and knowing her or his limitations and strengths.
The other factor to bear in mind is that nudges often require prior
understanding, which implies a degree of cognitive engagement and the
ability to understand the causal links. In behavioural science, for example,
it has been shown that conveying the costs of attendance or carrying out an
action for the public purse has positive effects. Hallsworth et al. (2015)
randomised SMS messages to outpatients in the NHS with a treatment
message that indicated the costs of missing an appointment and which led
to fewer people missing their appointments. It could be that this is purely
automatic, activating a norm of attendance. But it is more complex than
this. What is happening is that patients are being asked to think about the
consequences of their decisions. It is not a simple nudge based on an
automatic response, at least not in full, but requires the respondent to
understand the argument that missed appointments cost money and that

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 6 / Date: 10/1
JOBNAME: John PAGE: 7 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

128 How far to nudge?

turning up helps the public service. This kind of act cannot be seen as
purely system 1 even if system 1 needs to be in play.
In a more sophisticated application of this kind of treatment, John and
Blume (2017) sought to nudge holders of blue badges, which allow
disabled people to park their cars in designated places so it is easier for
them to get to shops and public facilities, to renew online, which is much
cheaper for the local authority that issues the badges. The researchers
tested a nudge in a reminder letter that said that the public authority
would save money if people renewed online, which worked in generating
considerable savings for the authority. The implication of the successful
message is that recipients understood the argument that was being made
and then believed the council would spend the saved resources on public
services. For the intervention to work, there had to be trust in the local
elected public authority not to waste the money, which is not easily
assumed in this age of declining faith in politics and those in public life.
Even more standard nudges require some thought on the part of the
respondent, as they are often seeking to convey an action in a complex
public policy system. The humble nudge ‘Nine out of ten people have
already paid their taxes’ requires the respondent to understand what this
phrase means. As well as following the norm, which might be automatic,
the taxpayers might also think about the likelihood of being caught and
whether paying up rectifies this risk, which requires a conception about
how payment systems work.
Nudges to change health behaviour often require that the people in the
trial have gone through a thought process about how the desired actions
will affect them, as otherwise the interventions would not have a chance
of working. Consider commitment devices (Thaler and Shefrin 1981).
These are concrete and public commitments people make or are encour-
aged to make to do an action so as to commit themselves to it. Although
the nudge operates through the psychological sense of commitment and
not wanting to go back on something for fear of feeling guilty, in fact
entering into a commitment device requires some degree of thought and
understanding of what a commitment device is in the first place. In many
ways, communicating this set of understandings is the only way to ensure
a successful nudge. Imagine people being duped into accepting a
commitment device. It is unlikely that this device would have any worth,
because the people would not really understand it. Even if they did, they
might say they did not really want to do it. A lot of mental ground needs
to be covered before some nudges can work. A nudge also works much
better when the respondent thinks about it. There is quite a bit of
evidence that commitment devices work, such as in health (Savani 2017).
But the focus in the debate is how such devices work once they have

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 7 / Date: 14/12
JOBNAME: John PAGE: 8 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge plus and how to get there 129

been agreed to (the doer) rather than the decision to take on a commit-
ment device in the first place (the planner). The latter involves thought
and reflection on the part of the individual, which is part of the
programme of nudge plus that is the core argument of How Far to
Nudge?
Consider the placement of healthy options next to the tills in cafeterias.
The automatic nudge works like this: people are stimulated to buy
chocolate and sweets, as they are near the till. They have made their main
food choices and are waiting to pay. Their eyes focus on the products
conveniently placed at eye level. The love of chocolate and sweet things
plays a role in that the consumer almost subconsciously places the bar on
the tray. It gets paid for and eaten pleasurably later on. Now consider the
placement of fruit near the checkout. Unless someone has a craving for
fruit, then that person will see the fruit but will not be prompted in the
same way. It works rather differently. The person sees the fruit and then
has to think along the following lines: ‘Well, have I had enough fruit and
vegetables today, my five portions, so it is OK not to buy more fruit?’ Or
the person could say: ‘No, I had better have some fruit, as I only had
cereal this morning.’ The person feels a lot better from having fulfilled a
moral commitment, which is consciously acted upon, even though the
person was nudged when waiting to pay for the food. There is no benefit,
especially in the long term, in people half-accidentally putting fruit on
their trays. They will probably leave the fruit behind when they come to
stack the tray.
Another example of these thought-provoking nudges is the work on
aspirations to motivate people to make better choices, such as to go to
university. The choice to go to university often is not based on ability, but
on students feeling they need to go and leave their families and friends to
advance their careers and life chances. Experimental research shows that
people can be influenced by communications, especially from someone
they admire, to make the choice to attend university. One example is a
letter to the student from someone at the university (Sanders et al. 2017).
Silva et al. (2016) found that role models giving talks to students works
too. What is going on with these interventions? It is not the classic
prompt of the subconscious nudge. Students need to understand and think
about the message. They need to think through a set of linkages, which
involves the idea that someone like them might attend university or go on
to better employment.
One of the key ideas in Kahneman (2011) is that not only are system 1
and system 2 separate, but they closely relate to each other. The active
self guides more automatic actions and lays in place ways in which these
automatic actions can work as if the more thinking self were in charge.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 8 / Date: 14/12
JOBNAME: John PAGE: 9 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

130 How far to nudge?

Service and Gallagher (2017) explore the idea of scaffolding in their

book Think Small, the idea that supporting structures are needed to help
individuals achieve their goals, which may be several, and where nudgers
should be focused when trying to develop the goals in the first place.
These are what they call ‘self-nudges’, and include focusing on choosing
the appropriate goal, focusing on a single objective, and breaking it down
to manageable steps. Service and Gallagher (2017) use examples from
the practical application of behavioural insights, such as with welfare-to-
work, to make their case. The result should be self-sustaining. But what
they show is not the classic nudges where an automatic behaviour is
stimulated, but where individuals take control of their own nudges. The
key word in the book’s title is ‘think’, of course!
Another application is where nudges are made more personal or
personalised to appeal to the individual. An example is including the
person’s name as part of a request to settle court fines (Haynes et al.
2013). This again can be regarded cynically as a ploy to make citizens
think that someone is taking a personal interest but in fact could be a way
to stimulate people’s interest and engagement with the problem, in that
someone in officialdom is taking an interest in them: if someone takes an
interest, they take more of an interest in themselves, which may involve
some reflection and conscious thought.
A recent evaluation of crime re-education policies for poor youth in
Chicago (Heller et al. 2017), using evidence from three randomised
controlled trials, showed that the Becoming a Man (BAM) programme
developed by the Chicago non-profit Youth Guidance reduced total
arrests during the intervention period by 28–35 per cent, reduced
violent-crime arrests by 45–50 per cent, improved school engagement,
increased graduation rates by 12–19 per cent and reduced readmission
rates to a correctional facility by 21 per cent. Heller et al. (2017) tested
whether people in the programme show more slow thinking by playing a
game, which they do. They conclude that they have ‘suggestive support
for the hypothesis that the programs work by helping youth slow down
and reflect on whether their automatic thoughts and behaviours are well
suited to the situation they are in, or whether the situation could be
construed differently’ (Heller et al. 2017: 2). Recent work shows the
influence of therapy-based interventions on social outcomes. Blattman
et al. (2017) tested whether providing cognitive behavioural therapy
(CBT) would improve outcomes, crime, and violence for unemployed
youth in Liberia, which they found to have strong effects. There has been
a more general interest in using ideas in CBT as a tool to increase
awareness of people’s own behaviour changes, influencing initiatives
such as mindfulness, which can be taught and conveyed so as to achieve

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 9 / Date: 10/1
JOBNAME: John PAGE: 10 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge plus and how to get there 131

behaviour change, and can even be targeted to policy-makers (see Lilley

et al. 2014). What is interesting from the nudge plus perspective is the
extent to which the behaviour changes come from measures that stimu-
late reflection and awareness.
In recent years, there has been a lot of focus on what characteristics
might be associated with long-term behaviour change, or self- and
societal benefiting behaviours. Is it possible to engage in a programme of
achieving long-term goals? In the view of psychologists like Duckworth
et al. (2007), it requires the development of orientations akin to determin-
ation and playing for the long term, what they call ‘grit’. In studies of
health it requires having a particular mind-set to engage in change (Burd
2016). Though some of these characteristics might derive from genes or
family context, the message from advocates of grit is that individuals can
consciously work at improving them. This activity must involve some
thinking and reflection on the part of individuals, even if the later actions
might follow more automatically. That is because the main factor beyond
successful behaviour change is effective motivation (see Michie et al.
2011), and this underlying feature can only be tackled by some degree of
reflection and consideration on the part of individuals.
The grit idea sounds tautological (it describes the success rather than
the nudge, and grit describes people who are successful, not how they got
there). There is little evidence for it in experimental evaluations (e.g.
Heller et al. 2017), and it does not correlate with outcomes, for example
in education (Rimfeld et al. 2016). But it does make sense that capabil-
ities can be cultivated. To get there, they may need a ‘boost’, which is a
term that has recently appeared in work on nudging (Grüne-Yanoff and
Hertwig 2016).
The boost argument works in many medical situations where people
need capacity to make decisions. In the past, doctors used to dispense
their decisions from on high for grateful patients to receive. Now, in the
days of consumer sovereignty, patients are given choices between alter-
native courses of action, say different treatments. These choices require
some understanding of the statistics, and it is easy to make simple
mistakes. This problem was identified in Kahneman and Tversky’s
research (e.g. Kahneman et al. 1982), which affects elite decision-makers
like doctors too. From a boost perspective, the question is whether there
are things that can increase capacity, or nudges that encourage people to
exercise choices with better knowledge of the consequences. At the same
time, it is important to realise it is not possible to offer an introductory
lesson on probability. What people need is nudges to help them to make
better choices but still based on their own reasoning. An example
Hertwig (2017) gives is the information about the risk of different

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 10 / Date: 21/12
JOBNAME: John PAGE: 11 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

132 How far to nudge?

treatments expressed as natural frequencies rather than in conditional

probabilities, so avoiding a need to understand Bayesian statistics. This
training strategy has proved effective (Sedlmeier and Gigerenzer 2001).
Other examples include rules of thumb to interpret financial investment
decisions for retirement, or simple rules to follow for a diet. In Hertwig’s
(2017) words, ‘The goal of boosts is to make it easier for people to
exercise their own agency in making choices.’ The proposals help get
over the paternalist problem as well as make choices better because they
put the individual more in control rather than being pushed along by a
discrete nudge.
The benefit from thinking more broadly is that a greater number of
behaviours can be altered. In the above examples, the increase is over
time, so actions are repeated. There could be expansion across the range
of actions, what is called ‘behavioural spillover’, whereby a new behav-
iour in one sphere causes behaviour change in another (Dolan and
Galizzi 2015). Part of the reason for this phenomenon is that ‘we may
derive satisfaction not just from tangible behavioral outcomes, but also
from the accumulation of signals and beliefs about our own identities’
(Dolan and Galizzi 2015: 4). The chances of behavioural spillovers are
much greater if individuals have shifted their approach to life so that a
change in one area, for example improving diet, is complemented by a
change in another, such as taking exercise, drinking less, or recycling
more, taking on the attributes of good citizens. Nudges that target the
motivation of citizens or appeal to their identity have more potency. They
work better when consciously thought about, which can help reverse
some tendencies for negative spillover, such as coasting after carrying out
a good deed.
This kind of nudging works better if there is a long-term relationship
between the individual and the public agency. Consider the process of
giving feedback, for example, on the performance of someone doing the
action, such as giving people a smiley if residents on their street have
recycled their waste more than the average for the area (John et al. 2011).
Here there are stages to the action, not just a one-off nudge. The start is
the state before the nudge, then the nudge to change behaviour, then the
feedback and then the change in behaviour as a result of the feedback.
All of this may seem to be capable of being manipulated, but in fact the
stages show how the thoughts of the nudged and the nudger need to be
synchronised. Most of all, to respond to the feedback the citizen needs to
understand what is going on: that the objective is good, that it makes
sense to respond, that keeping up with others as part of the agenda for
change is desirable, and that being reminded is to be part of the
conversation. It would be even better if the citizens, having changed their

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 11 / Date: 14/12
JOBNAME: John PAGE: 12 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge plus and how to get there 133

behaviours, got a thank-you message from the authority. Indeed, research

on the impact of thanking someone for voting indicates it also has the
benefit of encouraging further participation at a subsequent election
(Panagopoulos 2011).
The goal of nudging must be some long-term change to which the
individual has given some thought. The commitment is a way of
grounding the new form of behaviour. There is a need to put into place a
self-equilibrating system so things happen automatically and are self-
sustaining. The goal is a rule-based system which is not enforced, rather
as with the Highway Code, which is really about social cooperation, with
the state only as a back-up when behaviours go badly awry.
In some ways, this argument is an extension of the original Thaler and
Sunstein (2008) claim that people would not mind being nudged if they
had the chance to think about it. It is a small extra step to believing that
the nudge operates well on the basis of previous thought, which is then
acted upon. There will be some reflection, which soon becomes embed-
ded in automatic processes. Nobody wants to be permanently thinking
about public policy as in the deliberative ideal, but people want to be
autonomous in the way they respond to nudges. The idea is that a
combination of nudging and thinking can help the creation of an
automated and self-regulating system whereby people get to their goals
and where there is a synergy between social and individual aims. Rather
than nudges only being acceptable and practicable if people think about
them when being influenced in their behaviour, people can consider and
reflect on them, even if they do this only very briefly, before or after the
nudge or both. The actual nudge can be automatically responded to, but
individuals benefit from prior or subsequent thought about the context
and justification of the interventions. This argument, in Mols et al.’s
(2015) view, is more psychologically plausible and in line with how
long-term behaviour change comes about: ‘enduring behaviour change
involves an identity-change process whereby people proactively choose
to engage in behaviour that is perceived as identity-consistent and
therefore seen as “the right thing to do”’ (Mols et al. 2015: 82–83). They
argue that nudges will be more effective if they key into people’s social
identity. Whether the end result is automatic and non-conscious behav-
iour matters less. It is more convenient and self-sustained if people do not
always think about the choices. It is merely that their current choices
have been considered and understood at some point.
People can be affected by the environment within which choices are
made and whereby people may seize effective opportunities. In this
sense, the automatic processes can affect how people think. It is possible
to redesign buildings so that people interact rather more, and stand rather

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 12 / Date: 14/12
JOBNAME: John PAGE: 13 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

134 How far to nudge?

than sit, which increases well-being and the ability to make decisions
(Engelen et al. 2016). A systematic review of workplace reorganisations
(Bambra et al. 2007) stresses the importance of control in the workplace
for well-being and argues for more participation.
In many ways, these nudges, which are about public goods and general
public policy outcomes, are different to the private sector nudges which
are about clicking on some part of a website, or ensuring customers are
‘hot’ when making a purchasing decision, or showing how many rooms
are left on a hotel booking website and how many people are currently
watching them. Often there is much more involved with the action than
individualised purchasing decisions. There are complex public policy
issues where some awareness of the issues and of the behaviour is needed
as the baseline for the nudge to work, as people know this is not a simple
request. Then there is the communication about the public nature of the
goods being provided in that the nudge is really one part of a larger set of
actions about public provision.
In this sense, no one should be surprised by the finding that most
nudges are not carried out in secret, but are authorised publicly and
appear on the website of the Behavioural Insights Team or another
nudging organisation. It is easy to find out what the state is doing to
individuals. There is even publicity about those nudges that find their
way into the press and other forms of media. This is not just some act of
organisational survival; it is actually an important part of nudging. It is
all part of the conversation, which need not happen all in one go, but is
iterative and a long-term project of better communication between the
government and citizens. The result is implied consent to behavioural
public policy, which is backed up by detailed survey evidence. Sunstein
(2016) shows that citizens approve of nudges when they learn about them
so long as these interventions are consistent with their long-term object-
ives. If the research by Loewenstein et al. (2015) is a full guide, and to an
extent it is preliminary, nudgers should be encouraged that telling people
about nudges does not undermine support for them. Moreover, Bruns et
al. (2016) find that transparency does not reduce the effectiveness of
nudges. People do not display reactance when they find out.
The argument is that nudges already contain a thinking component,
even if understated, which means that the next generation of nudges
could expand this aspect. This wider range for nudge falls short of the
think argued for in John et al. (2011), but is more feasible than, and an
improvement on, the classic nudges. As a result, the research programme
is an expansion or enhancement of the nudging programme, which John
and Stoker (2017) call ‘nudge plus’.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 13 / Date: 14/12
JOBNAME: John PAGE: 14 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge plus and how to get there 135

INCREASING THE RANGE OF SUBJECTS AND

OBJECTS WITH NUDGE
If the previous section was about focusing on and selecting nudges that
have a deliberative and reflective component, this section is about
increasing the range of nudges to different kinds of people or organ-
isations so they can be both nudgers and nudged. The current manifesta-
tion of nudge has been focused often on a central government unit that
has been set up in the centre of government with the authority of the
prime minister or leading figures. The nudges are as a result top down, as
is the operation of a nudge unit, which appears to be a strong centre
using the insights of behaviour change to effect change. In fact, as
Chapter 5 shows, this image is a false one, in that the central units did
not have much power and relied on persuading and enthusing others in
the bureaucracy to engage in behaviour change, generally in pilot or
demonstration projects. The agenda relied on diffusing ideas across the
complex and myriad agencies of the state. Nudges need to be owned by
an agency and do not usually work in a top-down way. They rely on
enthusiasm and embedding behaviour change into bureaucracies and for
agencies to adopt the ideas in a decentralised way. Entrepreneurs flourish
in this environment, and they do not work well with strong control
procedures. Pluralist nudges help avoid organisational mistakes and limit
the effects of hierarchy. They ensure a responsible bureaucracy with
multiple sources of information, and provide correction and control.
It is relatively easy to extend this model of diffusion to outside central
government, which reflects the pattern of adoption of behaviour change.
Of course, any agency can use behaviour change ideas, especially as such
ideas work well in implementation processes, which are myriad across
the complex web of government and agencies. Such a process was
identified by Thaler and Sunstein: ‘Workplaces, corporate boards,
universities, religious organizations, clubs, and even families might be
able to use, and to benefit from, small exercises in libertarian paternal-
ism’ (Thaler and Sunstein 2008: 252). It is possible to see the extension
of nudge to local government and its agencies. The voluntary sector is a
useful conduit for nudges, which are already being used to reach clients
and to market charitable acts. There is no reason not to include
community and citizen groups as part of the diffusion, with the main
constraint being the capacity of small organisations to introduce nudges
and to evaluate them with trials. Trials can diffuse to the private sector, of
course, and this can be part of the intellectual climate of change, but
raises ethical issues about using nudges for private means. The same can

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 14 / Date: 14/12
JOBNAME: John PAGE: 15 SESS: 6 OUTPUT: Fri Jan 26 10:51:11 2018

136 How far to nudge?

be said for the consultancies and private organisations that sell nudges to
the public and private sector, with the danger that the provision of nudges
becomes a commercial activity and reduces the chance that a more open
and feedback-orientated version of nudge can prevail. The changes
highlighted so far may challenge standard models of public adminis-
tration based on relationships between politicians and bureaucrats, and
between politicians/bureaucrats and citizens. This connects to a wider
movement and set of theorising about a more decentralised and citizen-
centred form of public management. Behavioural public policy resonates
with such changes, and the rest of this chapter explores the linkages,
possibilities for reform, and what model of bureaucracy and politics
follows from behavioural policy.

CONTROLLING ELITES WITH NUDGES

Even with the decentralisation to smaller organisations, the relationship
between governments and citizens or organisations or client groups
remains the same: leaders and bureaucrats designing nudges to effect
change on citizens. The final move in the argument is to consider
whether nudges can be expanded to a more general political tool, which
can be used by citizens as well as those in government. The starting point
of the argument is that nudging rests on a belief that rational policy-
makers use knowledge to design policies to correct the biases of citizens.
Of course, the policy proceeds from the setting of objectives and then
creating the means to achieve those objectives (using evidence), at least
in formal terms (see John 2013b), but deviation from rational decision-
making is common wisdom among policy scholars, as was reviewed in
Chapters 1 and 3. Even though there are procedures in place to ensure the
weighing of options, and options analysis to help decision-making, in the
end gut feeling, intuition and emotion play their role in the selection of
policies and even in monitoring implementation processes. Brest (2013)
reviews the policy-making literature and identifies a number of common
behavioural errors that policy-makers make: hindsight bias, anchoring,
confirmation bias, overweighing slender evidence sources, poor interpret-
ation of statistical evidence, availability bias, affect heuristic (using
emotion to make decisions), psychic numbing (from overexposure to
statistical evidence), overconfidence, stereotyping, choice overload, loss
aversion, myopia, use of social proofs, and classic groupthink, which are
quite resistant to what he calls ‘de-biasing’, removing those biases and
improving the evidence base of decision-makers. In short, the same
biases affect policy-makers as citizens.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 15 / Date: 22/1
JOBNAME: John PAGE: 16 SESS: 6 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge plus and how to get there 137

The large literature on policy failures and successes is a tribute to the

non-rational factors in the policy process, whereby policy-makers often
ignore evidence in favour of their own gut feelings or just their lust for
power and fame (Dunleavy 1995; McConnell 2010; King and Crewe
2013). The experiment by Loewen et al. (2014) is very instructive. The
researchers recruited politicians to do an experiment based on the famous
one carried out by Kahneman et al. (1982), the Asian disease test, to see
if their reasoning was different to that of non-politicians considering the
same problem. This is also a reminder that the original work of
behavioural economists, such as Tversky and Kahneman (1981), was
directed to elites and small groups (such as statisticians or pilots) rather
than to the population at large. The experiment asks people to respond to
the following scenario: ‘An exotic disease is coming, and it’ll kill 600
people. You have two options. Choose the first, and 400 people will die.
Choose the second, and you take a risk: There’s a two-thirds chance that
everyone will die.’ When you change the frame to a third chance of
saving everyone there is movement upwards, as people are not avoiding
the negative frame. When the Asian disease question was put to 154
Belgian, Canadian, and Israeli members of parliament, they behaved just
as Kahneman et al. (1982) said, with no difference from the citizens, who
may have done better than the politicians (who fear negative outcomes).
The question is whether these biases are permanent features of decision-
making, so creating persistent errors in the policy-making process, or
whether there is scope for policy-makers to correct their mistakes. In a
very pessimistic point of view, politicians appear to repeat errors, which
is the implication of King and Crewe’s (2013) book on policy fiascos.
But other parts of this literature are more optimistic, such as
McConnell (2010) and Bovens et al. (2001), who look for the conditions
for success in public policies. So, given biases, how can policy-makers
correct their errors? A key form of correction comes from communica-
tions from citizens, who can feed back to politicians to show how well
they are doing. These kinds of communication are in any case usually
part of an intelligence strategy in the field of policy implementation, or
act as a fire alarm to overcome the limits of the principal–agent
relationship. But the problem is that, if the information is being commu-
nicated in a way that assumes a rational agent is receiving it, there is a
danger that the information may overload, or the individual might not
receive the information or see it as important. The politician in this
respect is no different to the citizen who gets the outcome of a public
information campaign. The result might be a lack of concentration and
sense of overload, and no indication that the information is important.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 16 / Date: 22/1
JOBNAME: John PAGE: 17 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

138 How far to nudge?

There is an example of how this happens in practice. Richardson and

John (2012) conducted an experiment to see whether a better-argued
lobby would get more of a response than a worse-argued one. They
recruited lobby groups to make a case for a policy change, and got them
to send a councillor, on the basis of random allocation, either a
well-worded letter or one that used poorer and less coherent arguments.
The result of the experiment was that there was no difference in
responses to the letters: the intervention made no difference to the
response rate. The conclusion to draw from this experiment is that
communications from citizens or citizen groups need to use some of the
behavioural techniques to ensure politicians take notice and correct their
errors. There is not much sense in using rational techniques to correct
biases which are entrenched. Just as with nudging, it is better to confront
bias with bias. There might be normative and practical reasons why
citizens would not even want their leaders to be rational calculators, but
would prefer them to have the same human emotions as they do
themselves and to react in a human way to crises and problems. In the
television series Star Trek, Spock with his rational facilities and willing-
ness to use them was a useful and essential part of the crew, but many of
the episodes showed the limitations of this kind of thinking. In contrast,
the verve of the more impulsive Kirk won the day in the end. Aside from
media convention, much of this appeal is about leadership, which is
about politicians being human. If there is negativity bias among elites, it
might come from understanding of the negativity bias of citizens. The
biases of the two are equivalent. But communication between the two
could be a form of correction, ensuring that politicians are emotional
when they need to be and more rational when that is better. Or it could be
that communicating and acting on the biases where appropriate would be
a better form of decision-making, where both governing and governed
realise the importance of heuristics, a sort of conscious bias. Gigerenzer
and Brighton (2009) refer to ‘homo heuristicus’, who may be able to
handle decisions in a more efficient way than someone with the infor-
mation overload implied by the rational model: ‘a biased mind can
handle uncertainty more efficiently and robustly than an unbiased mind
relying on more resource-intensive and general-purpose processing strat-
egies’ (Gigerenzer and Brighton 2009: 109). This factor is why program-
mers insert heuristics into artificial intelligence systems to make them
smart. But rather than having elites working on their own, it could be
argued that heuristics might be more successful if done in communi-
cation between elites and non-elites. It can encourage both to reflect on
new information brought by the other rather than just being nudged.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 17 / Date: 10/1
JOBNAME: John PAGE: 18 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge plus and how to get there 139

However, work on judgement and decision-making indicates that heuris-

tics perform well but not significantly differently from other decision-
making aids (Chater et al. 2003).
There is a considerable amount of evidence from what are called ‘elite
experiments’ that politicians and bureaucrats do take notice of behav-
ioural cues. It is possible to randomise these communications, in what are
called ‘experiments on institutions’ (Grose 2014). For example, Grose
(2010) randomised the legislators of a state to receive a message saying
they were being observed by an outside group. This message led to
increased attendance at committees. Nyhan and Reifler (2015) carried out
an experiment which fact-checked legislators in nine US states.
It might reasonably be thought that the inequality of resources will
continue in the age of behavioural sciences. The advantages of nudge
accrue to those who have power and access to these resources. They can
use the techniques of behavioural sciences to control citizens rather than
vice versa, which would fit the pattern of the development of the science
so far. But this use of nudge does not reflect the development of modern
politics, which has been transformed by the expansion of information in
recent years, and in particular the sharing of information across citizens
and a wide range of groups. The expansion of the internet and the use of
social media mean that people are connected in real time and that they
have the ability to contribute content with maximum impact at next to
zero costs. This facility increases the chances of political participation
and creates large swings in movements and ideas that challenge political
elites (Margetts et al. 2016), which has put those in government on the
back foot in many cases, though of course they can adapt. The rise of a
more fast-moving and turbulent politics will necessitate the use of the
ideas of behavioural sciences, both by citizens seeking policy change and
by politicians changing their courses of action in response. Because of
much of internet politics being done so quickly and where communica-
tions between people can be responded to in real time, in particular
where the internet is able to communicate social information, such as
how many other people are carrying out an activity (social norms), then
behaviourally coded information works very well on the internet and can
be used by groups and citizens for any purpose. Social media politics
then provides an arena within which behavioural nudges can thrive and
be highly influential. Can this readiness to use social media be harnessed
in a more responsive form of politics, which leads to correction, or does
it imply that the already manifest tendency to bias that already exists will
become exaggerated and maybe there will even be dangers to security
and effective policy-making?

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 18 / Date: 10/1
JOBNAME: John PAGE: 19 SESS: 6 OUTPUT: Fri Jan 26 10:51:11 2018

140 How far to nudge?

The danger with the reliance on citizen nudges or nudges from

lower-level organisations is that there may be an inequality of power,
which is only partly rectified by the dispersion of power from social
media, as it may be the case that more powerful organisations are more
skilful in using nudges. The other main disadvantage is that there is the
possibility of evidence and expert correction in government nudges, no
matter how cognitively bounded the decision-makers are, but such
correction might not exist with citizen nudges. Lack of evidence, together
with emotions, might actually bias the policy process even more than it
does currently without them. Dependence on heuristics and short-term
biases might foster anti-politics just as much as technocratic and expert
policies do currently.

CONCLUSION
This chapter has sought to move the argument of the book beyond a
review of existing evidence and research about the development of
behavioural public policy to a consideration of where the research and
policy agenda might go. It started from the difference between nudges for
public policy concerns and nudges done for commercial advantage, as the
former need citizens to be aware, to a degree, of complex linkages and
citizen roles. It has been argued that public policy nudges are rarely
purely system 1, that is just automatic, though some are. They usually
involve some understanding of the objective of the nudge, which may
involve some reflection, to ensure the nudge can work. Even where a
nudge is unconsciously acted on, it is in fact based on a set of
understandings and arguments that have already been established by the
citizen. The effectiveness of intervention for diet and more exercise, for
example, relies on the respondents being aware of the evidence and
arguments about behaviour change, which ironically might have been
conveyed by prior information campaigns so derided by nudge experts.
Then the nudge might provoke some reflection after the event. The
chapter gives some examples of these kinds of nudges and suggests
nudge plus is an important area for expansion of behavioural public
policy so as to incorporate some reflective component or element. The
argument is that an enhanced nudge is much more desirable than full-on
deliberation, which takes up too much time, assumes too much capacity,
and only selects a minority of the population. As part of this process,
government may seek to increase people’s capacity to make choices, the
‘boost’ in Hertwig’s (2017) terminology.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 19 / Date: 22/1
JOBNAME: John PAGE: 20 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

Nudge plus and how to get there 141

The expansion of nudge needs to go beyond the activities of centrally

sponsored units, and should be used much more by decentralised
organisations. At the heart of nudge is a more responsive form of
decision-making, one that is more agile and attuned to where citizens are
at cognitively. It can be used by anyone to promote the aims of other
organisations or people, thereby helping rectify the inequalities of power
and information between groups. There is the possibility of expansion
downwards to lower-level organisations and to citizens themselves, to use
the science for the ends of controlling decision-makers as well. The use
of science that takes account of bias can be a more effective way of
communicating with decision-makers, because the decision-makers suffer
from bias too. They may not have enough time and mental capacity to
respond to more informed communications.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 08-Chapter8PJED /Pg. Position: 20 / Date: 10/1
JOBNAME: John PAGE: 1 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

9. Assessing behavioural public policy

The aim of this chapter is to come to an overall assessment about the
potential for nudges. What is the full range of their application and the
likelihood of their continuing success? The idea is to use the argument and
evidence presented so far in this book to come to an appreciation about the
current and future scope of nudging, which could be just using it in a
minimal way, mainly about improving communications and transactions,
or extending right up to the maximum possible use. The latter position is
close to the argument of Chapter 8 with its advocacy of a decentralised
nudge plus. If the intellectual device of a continuum of possible nudges is
accepted, ranging between minimal and maximal, then the question is
where to place preferences for actual policies and practices on it. Readers,
given the extent to which they agree or disagree with the arguments made
in previous chapters, may wish to place themselves on this scale and so
answer the question of the book – How far to nudge? – themselves.
Overall, there is a need to stand back and take on board the criticisms of
nudge, as well as fully respond to the ethical challenges, paying attention
to the debate about limitations and constraints introduced in Chapters 6
and 7. These cautionary considerations should not be lost in the enthusi-
asm for next-generation interventions.
In deciding how far to nudge, one option need not have serious
consideration, which is the case for not nudging at all, which regards the
use of behavioural economics as fundamentally negative when deployed
by policy-makers, voluntary groups, and citizens. To come to such a
pessimistic view, readers would need to have complete faith in the model
of rational decision-making and simply want more information for
citizens, with more effective applications of incentives and sanctions, in
just the way Becker (1968) argued for, which is consistent with
autonomy and transparency. Given common knowledge about human
cognitions, a simple version of this model cannot apply: policy-makers
need to take some account of citizen biases in order to achieve their
objectives. It is possible to use some of the tools of rational action,
modified in some way, to apply behavioural economics in ways that
optimise behaviours, but this is a different position from not using the
insights of behavioural economics.

142

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 09-Chapter9PJED /Pg. Position: 1 / Date: 10/1
JOBNAME: John PAGE: 2 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Assessing behavioural public policy 143

The other argument for not using the science of behaviour change,
which emerged in Chapter 6 on the criticisms of nudge, is that it
represents an unacceptable loss of autonomy because it embodies forces
in society and the economy that limit the freedom of individuals to make
their own choices. The science of behaviour change offers to policy-
makers a further means to reduce individual autonomy and gives them a
more comprehensive set of tools with which to exercise power over
citizens. It ensures the passivity of citizens in deciding matters of
importance to them. Moreover, it depoliticises much of public policy into
a set of technical questions rather than focusing on societal values and
power relations, which affect the formulation and implementation of
public policy. It further hands power to the experts and professionals. In
this view, it is no surprise then that citizens are currently disengaged
from politics and keen to follow a populist or radical agenda when taken
advantage of in this way.
The anti-technical argument, however, is hard to sustain given how
much is still up for choice by today’s citizens and the openness of much
behavioural public policy. To hold such a position, it needs to be assumed
that powerful forces in society always control government and politics,
limiting the autonomy of citizens. If this view is accepted, by definition
any policy that comes from government is bound to have particular
ideologies of control built into it. Democracy then becomes a sham. But
even critical work on behaviour change holds back from this position
whilst at the same time offering criticisms of the policies on offer.
If these two extreme positions – full rational action and ideological
public policy – are rejected or at least limited as critiques, as it is sensible
to do, then the question concerns the extent of the usage of behavioural
techniques and assessing their full potential, which is affected by how
long these innovations are going to last and whether they can bed down
as permanent tools of government, after the flurry of interest of the last
few years has passed. In this book, the agenda and practice of behav-
ioural public policy have been reviewed in a balanced way, taking
account of both strengths and weaknesses. In spite of taking on board the
antecedents of this intellectual programme in the work of an earlier
generation of public policy scholars and psychologists, for example the
work of psychologists in areas such as health interventions and transport,
and research on social marketing, the claim is that the strides made in
behavioural economics are genuinely new because they are based on an
empirical programme that modifies simple rational actor approaches.
Behavioural economics was founded on identifying precise psychological
mechanisms involved with human decisions, in a way that appealed to
formal reasoning, and testing them in experiments and by other advanced

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 09-Chapter9PJED /Pg. Position: 2 / Date: 14/12
JOBNAME: John PAGE: 3 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

144 How far to nudge?

empirical methods. It was the closeness of the founders of behavioural

economics to the concerns of economists about the nature of decision-
making that was the source of the salience of the intellectual programme.
It kept the link to the overall aim of building a system of decision-
making, which could (in time) be based on micro-foundations, and
helped the programme maintain its realism and cogency.
It is no surprise to discover that once these ideas were tested they
started to transform public policy. It is also consistent with the novelty of
the programme that the way these ideas proved to be so acceptable was
through the diffusion processes of translation, simplification, and advo-
cacy. Skilled advocates managed to diffuse these ideas through the
academy and into the think tank world, and then the ideas were picked up
by elite newspapers such as the New York Times and the Financial Times.
The promotion of behavioural public policy by units such as the
Behavioural Insights Team was helped by this developing public agenda.
With these developments, there has been proof that these innovations
work when applied by central government and its agencies. Policies have
diffused across jurisdictions and levels of government, from the European
Union to Australian state governments. The benefit is a more curious
form of public administration based in trying things out and the regular
use of rigorous evidence. Naturally, the pattern of adoption is patchy and
reliant on politics rather than reflecting a uniform adoption. Not every-
thing works either, but trials allow for easy testing and, if need be, the
rejection of ideas as well as their acceptance. The process has a high
degree of agency in terms of who adopts nudges and how they are
implemented, rather than policy-makers being driven by the experts. As
the argument in Chapter 7 shows, there are relatively few ethical
constraints on the use of current kinds of nudges, except where they
cause harm and deception cannot be justified. Nudges compare well with
other policy instruments in standards of openness, review, and democratic
justification. If nudges fail on ethical grounds, then most democratically-
agreed policies would probably not pass muster either.
One approach to the current wave of reforms is to accept normalisation
and good practice. In this sense, some versions of behavioural insights
might be adopted as part of all policies to a degree and form part of the
research support for government. This is the natural pattern of innovation.
The early days were of specialist units and waves of experts. These units
may no longer be needed so much if everyone practises behavioural
public policy. In answer to the question of the book, nudging could be
more embedded in organisational processes and be part of general
policy-making. Part of this normalisation entails that nudges are used less
directly and appear more generally as part of policy design. What is

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 09-Chapter9PJED /Pg. Position: 3 / Date: 14/12
JOBNAME: John PAGE: 4 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Assessing behavioural public policy 145

happening now shows a natural maturation point for nudge. The question
then becomes whether a more maximal version of nudge could be
adopted, which could increase the pace of innovation as well as yielding
the benefits of the gradual roll-out of behavioural public policy and its
greater use by public agencies.

A MORE RADICAL APPROACH

An alternative approach reflects the argument for nudge plus as set out in
Chapter 8. Undeveloped in current nudges is a more reflective com-
ponent, which relates to the complexities of the policy problems that
citizens are being asked to help solve. Making the information com-
ponent more available is one of the ways forward, in successive nudging
that sometimes asks for reflection. The danger here is that there is not
enough time for citizens to consider complex policy issues in the course
of receiving a message, and citizens will jump to the wrong conclusions.
The response to this argument is that coming to either right or wrong
conclusions is what being a citizen is all about, and it is hard anyway
saying what right or wrong conclusions actually are in the first place.
What is important is that some reasoning has taken place, which in any
case can be part of a longer journey. The wider constituency covered in
nudge interventions is well worth it when compared to the limited
participation of citizens in more deliberative exercises. The survey and
experimental evidence appears to show that citizens prefer nudges when
they are conveyed more overtly rather than covertly (Jung and Mellers
2016). Moreover, there is potential to make nudges more sensitive to
citizen context by paying greater attention to the heterogeneity of
impacts, such as from personalising the nudges according to known
personality types. New technology, which is rolled out as part of
e-government, could provide an easier way of achieving this objective by
linking up databases, though some may find this moves the state to a
more manipulative strategy.
The other main change is to expand the use and range of nudges being
tested beyond the remit of central nudge units or even behaviouralists
working in central government departments. There is a considerable local
potential from nudge in the range of organisations that work in the
locality and have to face the delivery issues in public policy that nudge
policies are so good at addressing. There is a real issue of external
validity with the use of behavioural insights in that it is not known for
sure whether tests of interventions in one place generalise to another to
the same degree. It is possible to get around such a limitation if national

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 09-Chapter9PJED /Pg. Position: 4 / Date: 14/12
JOBNAME: John PAGE: 5 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

146 How far to nudge?

studies cover populations and have large sample sizes so as to assess the
impacts of treatments in local areas, but most empirical studies are in fact
not that large. There is no substitute for studies that are commissioned by
organisations that have contextual knowledge about conditions in the
locality.
It is often the case that local organisations are part of a delivery
process involving many organisations, which involve understanding the
links in the chain. Behavioural insights have been mainly used for
one-off improvements to delivery, either of standard procedures or of new
programmes, but there have been few behavioural evaluations in the
round where each link in the implementation chain has been assessed,
and where behavioural insights, as opposed to command-and-control, or
incentives, have been found more likely to work. Most of all, the
potential is for citizens and community groups to use behavioural
insights for their own ends, whether lobbying or seeking to control their
elected representatives or bureaucrats. The diffusion of new technology
can make using such techniques all the easier. In these ways, behavioural
insights could be used much more than they have been so far. The first
wave of influence reflected the pragmatic way in which new ideas met
the concerns of policy-makers; the second-wave programme can build on
these achievements in a more citizen-controlled and reflective way.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 09-Chapter9PJED /Pg. Position: 5 / Date: 22/1
JOBNAME: John PAGE: 1 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

References
Adams, Jean, Oliver Mytton, Martin White and Pablo Monsivais (2016),
‘Why are some population interventions for diet and obesity more
equitable and effective than others? The role of individual agency’,
PLoS Med, 13 (4), e1001990.
Ainslie, George (1975), ‘Specious reward: A behavioral theory of impul-
siveness and impulsive control’, Psychological Bulletin, 82 (4), 463–
496.
Ainslie, George (1991), ‘Derivation of “rational” economic behavior
from hyperbolic discount curves’, American Economic Review, 81,
134–140.
Ajzen, Icek (1985), ‘From intentions to actions: A theory of planned
behaviour’, in Julius Kuhl and Jürgen Beckman (eds), Action-Control:
From Cognition to Behaviour, Heidelberg: Springer, pp. 11–39.
Akerlof, George (2002), ‘Behavioral macroeconomics and macro-
economic behavior’, American Economic Review, 92 (3), 411–433.
Alemanno, Alberto and Anne-Lise Sibony (eds) (2015), Nudge and the
Law, London: Bloomsbury.
Allison, Graham T. (1971), Essence of Decision: Explaining the Cuban
Missile Crisis, New York: HarperCollins.
Altman, Daniel (2002), ‘A Nobel that bridges economics and psychol-
ogy’, New York Times, 10 October.
Anderson, Joel H. (2010), ‘Review of Thaler and Sunstein, “Nudge:
Improving Decisions about Health, Wealth, and Happiness”’, Eco-
nomics and Philosophy, 26, 369–406.
Anderson, Joel H. (2016), ‘Structured nonprocrastination: Scaffolding
efforts to resist the temptation to reconstrue unwarranted delay,’ in
Fuschia M. Sirois and Timothy A. Pychyl (eds), Procrastination,
Health, and Well-Being, San Diego, CA: Academic Press, pp. 43–63.
Annis, Helen M., Howard D. Cappell, Frederick B. Glaser, Michael S.
Goodstadt and Lynn T. Kozlowski (1990), Research Advances in
Alcohol and Drug Problems, Vol. 10, Berlin: Springer.
Argyris, Chris (1993), Knowledge for Action: A Guide to Overcoming
Barriers to Organizational Change, joint publication in the Jossey-
Bass management series and the Jossey-Bass social and behavioral
science series, 1st edn, San Francisco, CA: Jossey-Bass.

147

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 1 / Date: 14/12
JOBNAME: John PAGE: 2 SESS: 5 OUTPUT: Fri Jan 26 10:51:11 2018

148 How far to nudge?

Ariely, Dan (2015), Irrationally Yours: On Missing Socks, Pickup Lines,

and Other Existential Puzzles, New York: HarperCollins.
Baicker, Katherine, Sendhil Mullainathan and Joshua Schwartzstein
(2015), ‘Behavioral hazard in health insurance’, Quarterly Journal of
Economics, 130 (4), 1623–1667.
Bambra, Clare, Matt Egan, Sian Thomas, Mark Petticrew and Margaret
Whitehead (2007), ‘The psychosocial and health effects of workplace
reorganisation. 2. A systematic review of task restructuring interven-
tions’, Journal of Epidemiology and Community Health, 61, 1028–
1037.
Banerjee, Abhijit V. and Esther Duflo (2014), ‘The experimental
approach to development economics’, in Dawn Langan Teele (ed.),
Field Experiments and Their Critics: The Uses and Abuses of Experi-
mentation in the Social Sciences, New Haven, CT: Yale University
Press, pp. 78–114.
Baumgartner, Frank R. and Bryan D. Jones (1993), Agendas and
Instability in American Politics, Chicago, IL: University of Chicago
Press.
Becker, Gary S. (1968), ‘Crime and punishment: An economic approach’,
Journal of Political Economy, 76, 169–217.
Behavioural Insights Team (2012), Fraud, Error and Debt: Behavioural
Insights Team Paper, London: Cabinet Office.
Behavioural Insights Team (2015), The Behavioural Insights Team
Update Report 2013–2015, London: Behavioural Insights Team.
Bell, Chris (2013), ‘Inside the Coalition’s controversial “Nudge Unit”’,
Telegraph, 11 February.
Benartzi, Shlomo (2012), Save More Tomorrow: Practical Behavioral
Finance Solutions to Improve 401(k) Plans, New York: Portfolio.
Benartzi, Shlomo and Richard H. Thaler (2013), ‘Behavioral economics
and the retirement savings crisis’, Science, 339, 1152–1153.
Benjamin, Alison (2013), ‘David Halpern: We try to avoid legislation and
ordering’, Guardian, 5 February.
Berns, Gregory S., David Laibson and George Loewenstein (2007),
‘Intertemporal choice – toward an integrative framework’, Trends in
Cognitive Sciences, 11, 482–488.
Blattman, Christopher, Julian C. Jamison and Margaret Sheridan (2017),
‘Reducing crime and violence: Experimental evidence on cognitive
behavioral therapy in Liberia’, American Economic Review, 107 (4),
1165–1206.
Blume, Toby and Peter John (2014), Using Nudges to Increase Council
Tax Collection: Testing the Effects through a Randomized Controlled
Trial, London: Lambeth London Borough Council.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 2 / Date: 10/1
JOBNAME: John PAGE: 3 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

References 149

Borins, Stanford (2002), ‘Leadership and innovation in the public sector’,

Leadership and Organization Development Journal, 23 (8), 467–476.
Bovens, Luc (2009), ‘The ethics of nudge’, in Till Grüne-Yanoff and
Sven Ove Hansson (eds), Preference Change: Approaches from Phil-
osophy, Economics and Psychology, Berlin: Springer Science and
Business Media.
Bovens, Mark, Paul ‘t Hart and B. Guy Peters (2001), Success and
Failure in Public Governance: A Comparative Analysis, New Horizons
in Public Policy, Cheltenham, UK and Northampton, MA, USA:
Edward Elgar Publishing.
Bowles, Samuel (2016), The Moral Economy: Why Good Incentives Are
No Substitutes for Good Citizens, New Haven, CT: Yale University
Press.
Branca, Francesco, Haik Nikogosian and Tim Lobstein (eds) (2007), The
Challenge of Obesity in the WHO European Region and the Strategies
for Response, Copenhagen: World Health Organization, Regional
Office for Europe.
Braybrooke, David and Charles E. Lindblom (1963), A Strategy of
Decision: Policy Evaluation as a Social Process, Glencoe, IL: Free
Press.
Brehm, Jack W. (1966), A Theory of Psychological Reactance, New York:
Academic Press.
Brehm, Sharon S. and Jack W. Brehm (1981), Psychological Reactance:
A Theory of Freedom and Control, New York: Academic Press.
Brest, Paul (2013), ‘Quis custodiet ipsos custodes? Debiasing the policy
makers themselves’, in Eldar Shafir (ed.), The Behavioral Foundations
of Public Policy, Princeton, NJ: Princeton University Press, pp. 481–
493.
Broockman, David and Joshua Kalla (2016), ‘Durably reducing trans-
phobia: A field experiment on door-to-door canvassing’, Science, 352
(6282), 220–224.
Brook Lyndhurst (2006), Triggering Widespread Adoption of Sustainable
Behaviour, Behaviour Change: A Series of Practical Guides for
Policy-Makers and Practitioners No. 4, London: Defra.
Bruns, Hendrik, Elena Kantorowicz-Reznichenko, Katharina Klement,
Marijane Luistro Jonsson and Bilel Rahali (2016), Can Nudges Be
Transparent and Yet Effective?, WiSo-HH Working Paper Series No.
33, https://ssrn.com/abstract=2816227 (accessed 12 June 2017).
Burch, Martin and Ian Holliday (1996), The British Cabinet System,
London: Prentice-Hall.
Burd, Hannah (2016), ‘Five factors for supporting people to take a more
active role in health and wellbeing’, Behavioural Insights Team blog,

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 3 / Date: 14/12
JOBNAME: John PAGE: 4 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

150 How far to nudge?

http://www.behaviouralinsights.co.uk/health/five-factors-for-supporting-
people-to-take-a-more-active-role-in-health-and-wellbeing/ (accessed
19 June 2017).
Cabinet Office and Charitable Aid Foundation (2013), Applying Behav-
ioural Insights to Charitable Giving, London: Cabinet Office, https://
www.gov.uk/government/publications/applying-behavioural-insights-to-
charitable-giving (accessed 12 June 2017).
Cabinet Office – Prime Minister’s Strategy Unit (2004), Personal
Responsibility and Changing Behaviour: The State of Knowledge and
Its Implications for Public Policy, London: Cabinet Office.
Camerer, Colin F. (2003), Behavioral Game Theory: Experiments in
Strategic Interaction, Roundtable Series in Behavioral Economics,
Princeton, NJ: Princeton University Press.
Camerer, Colin, Samuel Issacharoff, George Loewenstein and Ted
O’Donoghue (2003), ‘Regulation for conservatives: Behavioral eco-
nomics and the case for “asymmetric paternalism”’, University of
Pennsylvania Law Review, 151, 1211–1254.
Caplan, Nathan S. (1975), The Use of Social Science Knowledge in
Policy Decisions at the National Level: A Report to Respondents, Ann
Arbor: Institute for Social Research, University of Michigan.
Caplan, Nathan S. (1979), ‘The two-communities theory and knowledge
utilization’, American Behavioral Scientist, 22 (3), 459–470.
Carey, Nessa (2011), The Epigenetics Revolution: How Modern Biology
Is Rewriting Our Understanding of Genetics, London: Icon Books.
Cartwright, Edward (2011), Behavioral Economics, London: Routledge.
Chabris, Christopher and Daniel Simons (2010), The Invisible Gorilla:
And Other Ways Our Intuition Deceives Us, London: HarperCollins.
Chater, Nick, Mike Oaksford, Ramin Nakisa and Martin Redington
(2003), ‘Fast, frugal, and rational: How rational norms explain behav-
iour’, Organizational Behavior and Human Decision Processes, 90,
63–86.
Chen, Frances S., Julia A. Minson, Maren Schöne and Markus Heinrichs
(2013), ‘In the eye of the beholder: Eye contact increases resistance to
persuasion’, Psychological Science, 24, 2254–2261.
Chetty, Raj and Emmanuel Saez (2013), ‘Teaching the tax code: Earnings
responses to an experiment with EITC recipients’, American Economic
Journal: Applied Economics, 5 (1), 1–31.
Chetty, Raj, Adam Looney and Kory Kroft (2009), ‘Salience and
taxation: Theory and evidence’, American Economic Review, 99 (4),
1145–1177.
Cialdini, Robert B. (2009), Influence: The Psychology of Persuasion,
London: HarperCollins.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 4 / Date: 14/12
JOBNAME: John PAGE: 5 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

References 151

Cialdini, Robert B., Raymond R. Reno and Carl A. Kallgren (1990), ‘A

focus theory of normative conduct: Recycling the concept of norms to
reduce littering in public places’, Journal of Personality and Social
Psychology, 58 (6), 1015–1026.
Coelho, Miguel and Vigyan Ratnoo with Sebastian Dellepiane (2015),
Political Economy of Policy Failure and Institutional Reform, London:
Institute for Government.
Conly, Sarah (2013), Against Autonomy: Justifying Coercive Paternalism,
Cambridge: Cambridge University Press.
Dalton, Russell J. (2005), ‘The social transformation of trust in govern-
ment’, International Review of Sociology, 15, 133–154.
Darton, Andrew (2008), Reference Report: An Overview of Behaviour
Change Models and Their Uses, London: Government Social Research.
Davis, Rachel, Rona Campbell, Zoe Hildon, Lorna Hobbs and Susan
Michie (2014), ‘Theories of behaviour and behaviour change across
the social and behavioural sciences: A scoping review’, Health Psy-
chology Review, 9 (3), 323–344.
Defra (2008), A Framework for Pro-environmental Behaviours, London:
Department for Environment, Food and Rural Affairs.
DellaVigna, Stefano (2009), ‘Psychology and economics: Evidence from
the field’, Journal of Economic Literature, 47 (2), 315–372.
Demeritt, Allison and Karla Hoff (2015), ‘Small Miracles’ – Behavioral
Insights to Improve Development Policy: World Development Report
2015, Washington, DC: World Bank, http://documents.worldbank.org/
curated/en/505731468127152574/pdf/WPS7197.pdf.
Denford, Sarah, Charles Abraham, Jane Rebecca Smith, Sarah Morgan-
Trimmer, Jenny Lloyd and Katrina Wyatt (2016), ‘Intervention design
and evaluation: Behaviour change imperatives’, in Fiona Spotswood
(ed.), Beyond Behaviour Change, Bristol: Policy Press, pp. 49–70.
Department for Business, Innovation and Skills and Cabinet Office
(2014), Growth Vouchers Trial Protocol, London: Department for
Business and Skills and Cabinet Office, http://www.behavioural
insights.co.uk/wp-content/uploads/2015/07/bis-14-561-growth-vouchers-
programme-trial-protocol.pdf (accessed 12 June 2017).
Department of Energy and Climate Change (2013), Removing the Hassle
Factor Associated with Loft Insulation: Results of a Behavioural Trial,
London: Department of Energy and Climate Change, https://
www.gov.uk/government/uploads/system/uploads/attachment_data/file/
236858/DECC_loft_clearance_trial_report_final.pdf (accessed 12 June
2017).
Dobson, Andrew (2011), Sustainability Citizenship, Weymouth: Green
House.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 5 / Date: 21/12
JOBNAME: John PAGE: 6 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

152 How far to nudge?

Dolan, Paul and Matteo M. Galizzi (2015), ‘Like ripples on a pond:

Behavioral spillovers and their implications for research and policy’,
Journal of Economic Psychology, 47, 1–16.
Dolan, Paul, Michael Hallsworth, David Halpern, Dominic King, Robert
Metcalfe and Ivo Vlaev (2012), ‘Influencing behaviour: The MIND-
SPACE way’, Journal of Economic Psychology, 33, 264–277.
Dowding, Keith and Desmond King (eds) (1995), Preferences, Insti-
tutions, and Rational Choice, Oxford: Clarendon.
Downs, Anthony (1957), An Economic Theory of Democracy, New York:
Harper.
Dryzek, John S. (2000), Deliberative Democracy and Beyond: Liberals,
Critics, Contestations, Oxford: Oxford University Press.
Dubner, Stephen J. (2016), ‘The White House gets into the nudge
business’, Freakonomics Radio, http://freakonomics.com/podcast/white-
house-gets-nudge-business/.
Duckworth, Angela L., Christopher Peterson, Michael D. Matthews and
Dennis R. Kelly (2007), ‘Grit: Perseverance and passion for long-term
goals’, Journal of Personality and Social Psychology, 92 (6), 1087–
1110.
Dunleavy, Patrick (1995), ‘Policy disasters: Explaining the UK’s record’,
Public Policy and Administration, 10 (2), 52–70.
Dunsire, Andrew (1993), Manipulating Social Tensions: Collibration as
an Alternative Mode of Government Intervention, Discussion Paper
No. 93/7, Köln: Max-Planck-Institut für Gesellschaftsforschung.
Dworkin, Gerald (1988), ‘Paternalism: Some second thoughts’, in The
Theory and Practice of Autonomy, New York: Cambridge University
Press, pp. 121–129.
Eisenstein, Hester (1991), Gender Shock: Practising Feminism on Two
Continents, Sydney: Allen & Unwin.
Ekelund, Ulf, Jostein Steene-Johannessen, Wendy J. Brown, Morten
Wang Fagerland, Neville Owen, Kenneth E. Powell, Adrian Bauman
and I-Min Lee (2016), ‘Does physical activity attenuate, or even
eliminate, the detrimental association of sitting time with mortality? A
harmonised meta-analysis of data from more than 1 million men and
women’, Lancet, Physical Activity Series 2 Executive Committee,
Sedentary Behaviour Working Group, http://dx.doi.org/10.1016/S0140-
6736(16)30370-1.
Engber, Daniel (2016), ‘The irony effect: How the scientist who founded the
science of mistakes ended up mistaken’, Slate, 21 December, http://
www.slate.com/articles/health_and_science/science/2016/12/kahneman_
and_tversky_researched_the_science_of_error_and_still_made_errors.
html (accessed 29 December 2016).

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 6 / Date: 21/12
JOBNAME: John PAGE: 7 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

References 153

Engelen, Lina, Haryana M. Dhillon, Josephine Y. Chau, D. Hespe and

A.E. Bauman (2016), ‘Do active design buildings change health
behaviour and workplace perceptions?’, Occupational Medicine, 66
(5), 408–411.
Farrell, Henry and Cosma Shalizi (2011), ‘“Nudge” policies are another
name for coercion’, New Scientist, 2837, 2 November.
Feitsma, Joram and Thomas Schillemans (2016), ‘Behaviour experts in
Dutch government: From newcomers to professionals?’, unpublished
paper.
Felsen, Gidon, Noah Castelo and Peter B. Reiner (2013), ‘Decisional
enhancement and autonomy: Public attitudes towards overt and covert
nudges’, Judgment and Decision Making, 8 (3), 202–213.
Fieldhouse, Ed, Dave Cutts, Paul Widdop and Peter John (2013), ‘Do
impersonal mobilisation methods work? Evidence from a nationwide
Get-Out-the-Vote experiment in England’, Electoral Studies, 32, 113–
123.
Fieldhouse, Ed, Dave Cutts, Peter John and Paul Widdop (2014), ‘When
context matters: Assessing geographical heterogeneity of Get-Out-the-
Vote treatment effects using a population based field experiment’,
Political Behavior, 36, 77–97.
Fishbein, Martin and Icek Ajzen (1975), Belief, Attitude, Intention, and
Behaviour: An Introduction to Theory and Research, Reading, MA:
Addison-Wesley Publishing.
Fiskin, James (1995), The Voice of the People: Public Opinion and
Democracy, New Haven, CT: Yale University Press.
Flinders, Matthew (2013), ‘The tyranny of relevance and the art of
translation’, Political Studies Review, 11 (2), 149–167.
Frey, Bruno S., Matthias Benz and Alois Stutzer (2004), ‘Introducing
procedural utility: Not only what, but also how matters’, Journal of
Institutional and Theoretical Economics, 160, 377–401.
Fung, Archon (2006), Empowered Participation: Reinventing Urban
Democracy, Princeton, NJ: Princeton University Press.
Gabaix, Xavier (2016), ‘A behavioral new Keynesian model’, unpub-
lished paper, http://pages.stern.nyu.edu/~xgabaix/papers/brNK.pdf (ac-
cessed 14 January 2017).
Gardner, Frances and Daniel S. Shaw (2008), ‘Behavioral problems of
infancy and preschool children (0–5)’, in Michael Rutter, Dorothy
V.M. Bishop, Daniel S. Pine, Stephen Scott, Jim Stevenson, Eric
Taylor and Anita Thapar (eds), Rutter’s Child and Adolescent Psy-
chiatry, 5th edn, Oxford: Blackwell, pp. 882–893.
Gerber, Alan S. and Donald P. Green (2012), Field Experiments: Design,
Analysis, and Interpretation, New York: William Norton.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 7 / Date: 21/12
JOBNAME: John PAGE: 8 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

154 How far to nudge?

Gerber, Alan S., Donald P. Green and Christopher W. Larimer (2008),

‘Social pressure and voter turnout: Evidence from a large-scale field
experiment’, American Political Science Review, 102 (1), 33–48.
Gigerenzer, Gerd (2003), Reckoning with Risk: Learning to Live with
Uncertainty, London: Penguin.
Gigerenzer, Gerd and Henry Brighton (2009), ‘Homo heuristicus: Why
biased minds make better inferences’, Topics in Cognitive Science, 1,
107–143.
Gneezy, Urim and John A. List (2006), ‘Putting behavioral economics to
work: Testing for gift exchange in labor markets using field experi-
ments’, Econometrica, 74 (5), 1365–1384.
Gneezy, Urim and John A. List (2013), The Why Axis: Hidden Motives
and the Undiscovered Economics of Everyday Life, New York: Public
Affairs/HarperCollins.
Goldacre, Ben (2008), Bad Science, London: Fourth Estate.
Goldstein, Noah J., Robert B. Cialdini and Vladas Griskevicius (2008),
‘A room with a viewpoint: Using social norms to motivate environ-
mental conservation in hotels’, Journal of Consumer Research, 35,
472–482.
Gollwitzer, Peter M. (1999), ‘Implementation intentions: Strong effects
of simple plans’, American Psychologist, 54, 493–503.
Goodin, Robert E. (1982), Political Theory and Public Policy, Chicago,
IL: University of Chicago Press.
Goodwin, Tom (2012), ‘Why we should reject “nudge”’, Politics, 32,
85–92.
Graham, Erin R., Charles R. Shipan and Craig Volden (2013), ‘The
diffusion of policy diffusion research in political science’, British
Journal of Political Science, 43 (3), 673–701.
Green, Donald P. and Alan S. Gerber (2003), ‘The underprovision of
experiments in political science’, Annals of the American Academy of
Political and Social Science, 589, 94–112.
Green, Donald P. and Alan S. Gerber (2015), Get Out the Vote: How to
Increase Voter Turnout, Washington, DC: Brookings Institution Press.
Green, Donald P. and Ian Shapiro (1994), Pathologies of Rational Choice
Theory: A Critique of Applications in Political Science, New Haven,
CT: Yale University Press.
Grether, David M. and Charles R. Plott (1979), ‘Economic theory of
choice and the preference reversal phenomenon’, American Economic
Review, 69 (4), 623–638.
Grose, Christian (2010), ‘Priming rationality: A theory and field experi-
ment of participation in legislatures’, paper presented at New York
University–Cooperative Congressional Elective Study Experimental
Political Science Conference, 5–6 February, New York.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 8 / Date: 21/12
JOBNAME: John PAGE: 9 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

References 155

Grose, Christian (2014), ‘Field experimental work on political insti-

tutions’, Annual Review of Political Science, 17, 355–370.
Grüne-Yanoff, Till and Ralph Hertwig (2016), ‘Nudge versus boost: How
coherent are policy and theory?’, Minds and Machines, 26, 149–183.
Gueron, Judith M. and Howard Rolston (2013), Fighting for Reliable
Evidence, New York: Russell Sage Foundation.
Hagman, W., D. Andersson, D. Västfjäll and G. Tinghög (2015), ‘Public
views on policies involving nudges’, Review of Philosophy and Psy-
chology, 6 (3), 439–453.
Hallsworth, Michael and Michael Sanders (2016), ‘Nudge: Recent devel-
opments in behavioural science and public policy’, in Fiona Spotswood
(ed.), Beyond Behaviour Change, Bristol: Policy Press, pp. 113–134.
Hallsworth, Michael, David Halpern, Dominic King and Ivo Vlaev
(2010), MINDSPACE: Influencing Behaviour through Public Policy,
London: Cabinet Office and Institute for Government.
Hallsworth, Michael, Dan Berry, Michael Sanders, Anna Sallis, Dominic
King, Ivo Vlaev and Ara Darzi (2015), ‘Stating appointment costs in
SMS reminders reduces missed hospital appointments: Findings from
two randomized controlled trials’, PLoS ONE, 10 (9), e0141461.
Hallsworth, Michael, John A. List, Robert D. Metcalfe and Ivo Vlaev
(2017), ‘The behavioralist as tax collector: Using natural field experi-
ments to enhance tax compliance’, Journal of Public Economics, 148,
14–31.
Halpern, David (2015), Inside the Nudge Unit, London: W.H. Allen.
Halpern, David and Danielle Mason (2015), ‘Radical incrementalism’,
Evaluation, 21, 143–149.
Halpern, David, Stewart Wood and Gavin Cameron (1996), Options for
Britain: A Strategic Policy Review, Aldershot: Dartmouth Publishing.
Harford, Tim (2005), The Undercover Economist, Boston, MA: Little,
Brown.
Hargreaves-Heap, Shaun (2017), ‘Behavioural public policy: the consti-
tutional approach’, Behavioural Public Policy, 1 (2), 252–265.
Harris, Christopher J. and David Laibson (2002), ‘Hyperbolic discount-
ing and consumption’, in Mathias Dewatripont, Lars Peter Hansen and
Stephen J. Turnovsky (eds), Advances in Economics and Econo-
metrics: Theory and Applications, Vol. 1, Eighth World Congress,
Cambridge: Cambridge University Press, pp. 258–298.
Haynes, Laura, Ben Goldacre and David Torgerson (2012), Test, Learn,
Adapt: Developing Public Policy with Randomised Controlled Trials,
London: Cabinet Office.
Haynes, Laura, Donald P. Green, Rory Gallagher, Peter John and David
J. Torgerson (2013), ‘Collection of delinquent fines: An adaptive

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 9 / Date: 21/12
JOBNAME: John PAGE: 10 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

156 How far to nudge?

randomized trial to assess the effectiveness of alternative text mes-

sages’, Journal of Policy Analysis and Management, 32, 718–730.
Heath, Joseph and Joel H. Anderson (2010), ‘Procrastination and the
extended will’, in Chrisoula Andreou and Mark D. White (eds), The
Thief of Time: Philosophical Essays on Procrastination, Oxford:
Oxford University Press, pp. 233–252.
Heller, Sara B., Anuj K. Shah, Jonathan Guryan, Jens Ludwig, Sendhil
Mullainathan and Harold A. Pollack (2017), ‘Thinking, fast and slow?
Some field experiments to reduce crime and dropout in Chicago’,
Quarterly Journal of Economics, 132 (1), 1–54.
Henderson, Mark (2012), The Geek Manifesto: Why Science Matters,
London: Bantam Press.
Hertwig, Ralph (2017), ‘When to consider boosting: Some rules for
policymakers’, Behavioural Public Policy, 1 (2), 143–161.
HM Treasury (2011), The Magenta Book: Guidance for Evaluation,
London: HM Treasury, https://www.gov.uk/government/uploads/
system/uploads/attachment_data/file/220542/magenta_book_combined.
pdf.
Hood, Christopher and Helen Margetts (2007), The Tools of Government
in a Digital Age, Basingstoke: Macmillan.
House of Lords (2011), Behaviour Change, Science and Technology
Select Committee 2nd Report of Session 2010–12, London: HMSO.
Jackson, Tim (2005), Motivating Sustainable Consumption: A Review of
Evidence on Consumer Behaviour and Behavioural Change, report to
the Sustainable Development Research Network, Brighton: University
of Sussex, http://www.sustainablelifestyles.ac.uk/sites/default/files/
motivating_sc_final.pdf.
John, Peter (2011), Making Policy Work, London: Routledge.
John, Peter (2013a), ‘All tools are informational now: How information
and persuasion define the tools of government’, Policy and Politics, 41
(4), 605–620.
John, Peter (2013b), ‘Political science, impact and evidence’, Political
Studies Review, 11, 168–173.
John, Peter (2013c), ‘My nudge tour of Australia’, http://bi.dpc.nsw.
gov.au/blog/my-nudge-tour-of-australia-professor-peter-john/ (accessed
15 January 2017).
John, Peter (2014), ‘Policy entrepreneurship in British government: The
Behavioural Insights Team and the use of RCTs’, Public Policy and
Administration, 29 (3), 257–267.
John, Peter (2017a), Field Experiments in Political Science and Public
Policy, New York: Routledge.
John, Peter (2017b), ‘Behavioural science, randomised evaluations and
the transformation of public policy: The case of the UK government’,

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 10 / Date: 21/12
JOBNAME: John PAGE: 11 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

References 157

in Jessica Pykett, Rhys Jones and Mark Whitehead (eds), Psycho-

logical Governance and Public Policy: Governing the Mind, Brain and
Behaviour, London: Routledge, pp. 136–152.
John, Peter and Toby Blume (2017), ‘Nudges that promote channel shift:
A randomized evaluation of messages to encourage citizens to renew
benefits online’, Internet and Policy, early view, http://onlinelibrary.
wiley.com/wol1/doi/10.1002/poi3.148/full (accessed 13 June 2017).
John, Peter and Tessa Brannan (2008), ‘How different are telephoning
and canvassing? Results from a Get Out the Vote field experiment in
the British 2005 General Election’, British Journal of Political Science,
38, 565–574.
John, Peter and Liz Richardson (2012), Nudging Citizens towards
Localism?, London: British Academy.
John, Peter and Gerry Stoker (2017), ‘From nudge to nudge plus: The
future of behavioural policy’, unpublished paper.
John, Peter, Sarah Cotterill, Alice Moseley, Liz Richardson, Graham
Smith, Gerry Stoker and Corinne Wales (2011), Nudge, Nudge, Think,
Think: Experimenting with Ways to Change Civic Behaviour, London:
Bloomsbury Academic.
John, Peter, Michael Sanders and Jennifer Wang (2014), ‘The use of
descriptive norms in public administration: A panacea for improving
citizen behaviours?’, https://ssrn.com/abstract=2514536.
Johnson, Eric and Daniel Goldstein (2003), ‘Do defaults save lives?’,
Science, 302, 1338–1339.
Jones, Bryan (1995), Reconceiving Decision-Making in Democratic Poli-
tics, Chicago, IL: Chicago University Press.
Jones, Bryan and Frank Baumgartner (2005), The Politics of Attention,
Chicago, IL: University of Chicago Press.
Jones, Rhys, Jessica Pykett and Mark Whitehead (2013), Changing
Behaviours: On the Rise of the Psychological State, Cheltenham, UK
and Northampton, MA, USA: Edward Elgar Publishing.
Jung, Janice Y. and Barbara A. Mellers (2016), ‘American attitudes
toward nudges’, Judgment and Decision Making, 11 (1), 62–74.
Just, David R. and Brian Wansink (2009), ‘Smarter lunchrooms: Using
behavioral economics to improve meal selection’, Choices, 29 (3), 1–6,
www.choicesmagazine.org/magazine/article.php?article=87 (accessed
12 June 2017).
Kahn, Emily B., Leigh T. Ramsey, Ross C. Brownson, Gregory W.
Heath, Elizabeth H. Howze, Kenneth E. Powell, Elaine J. Stone,
Mummy W. Rajab, Phaedra Corso and the Task Force on Community
Preventive Services (2002), ‘The effectiveness of interventions to
increase physical activity: A systematic review’, American Journal of
Preventative Medicine, 22, 73–107.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 11 / Date: 21/12
JOBNAME: John PAGE: 12 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

158 How far to nudge?

Kahneman, Daniel (1973), Attention and Effort, Englewood Cliffs, NJ:

Prentice-Hall.
Kahneman, Daniel (2011), Thinking, Fast and Slow, London: Penguin.
Kahneman, Daniel and Amos Tversky (1979), ‘Prospect theory: An
analysis of decision under risk’, Econometrica, 47 (2), 263–291.
Kahneman, Daniel, Paul Slovic and Amos Tversky (1982), Judgment
under Uncertainty: Heuristics and Biases, Cambridge: Cambridge
University Press.
Kahneman, Daniel, Jack L. Knetsch and Richard H. Thaler (1990),
‘Experimental tests of the endowment effect and the Coase theorem’,
Journal of Political Economy, 98 (6), 1325–1348.
Kahneman, Daniel, Jack L. Knetsch and Richard H. Thaler (1991),
‘Anomalies: The endowment effect, loss aversion, and status quo bias’,
Journal of Economic Perspectives, 5 (1), 193–206.
Kelman, Steven (2005), Unleashing Change: A Study of Organizational
Renewal in Government, Washington, DC: Brookings Institution Press.
Kimbell, Lucy (2015), Applying Design Approaches to Policy Making:
Discovering the Policy Lab, Brighton: University of Brighton.
King, Anthony and Ivor Crewe (2013), The Blunders of Our Govern-
ments, London: Oneworld Publications.
Kingdon, John (1984), Agendas, Alternatives, and Public Policies, Bos-
ton, MA: Little, Brown.
Kotler, Philip and Gerald Zaltman (1971), ‘Social marketing: An
approach to planned social change’, Journal of Marketing, 35 (3),
3–12.
Kundakovic, Marija and Ivana Jaric (2017), ‘The epigenetic link between
prenatal adverse environments and neurodevelopmental disorders’,
Genes (Basel), 8 (3), 104.
Leggett, Will (2014), ‘The politics of behaviour change: Nudge, neo-
liberalism and the state’, Policy and Politics, 42, 3–19.
Levitt, Steven D. and Stephen J. Dubner (2005), Freakonomics: A Rogue
Economist Explores the Hidden Side of Everything, New York: William
Morrow.
Levitt, Steven D. and John A. List (2007), ‘What do laboratory experi-
ments measuring social preferences reveal about the real world?’,
Journal of Economic Perspectives, 21 (2), 153–174.
Lewis, Michael (2016), The Undoing Project: A Friendship That
Changed the World, London: Allen Lane.
Lilley, Rachel, Mark Whitehead, Rachel Howell, Rhys Jones and Jessica
Pykett (2014), Mindfulness Behaviour Change and Engagement in
Public Policy: An Evaluation, http://www.sps.ed.ac.uk/__data/assets/
pdf_file/0004/180787/MBCEPPreport.pdf (accessed 1 May 2017).

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 12 / Date: 21/12
JOBNAME: John PAGE: 13 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

References 159

Lindblom, Charles E. (1959), ‘The science of muddling through’, Public

Administration Review, 19, 79–88.
Lindblom, Charles E. (1965), The Intelligence of Democracy, Glencoe,
IL: Free Press.
List, John A. (2002), ‘Preference reversals of a different kind: The more
is less phenomenon’, American Economic Review, 92 (5), 1636–1643.
List, John A. (2011), ‘Why economists should conduct field experiments
and 14 tips for pulling one off’, Journal of Economic Perspectives, 25
(3), 3–16.
Lodge, Martin and Kai Wegrich (2016), ‘The rationality paradox of
nudge: Rational tools of government in a world of bounded rational-
ity’, Law and Policy, 38 (3), 250–267.
Loewen, Peter John, Lior Sheffer, Stuart Soroka, Stefaan Walgrave and
Tamir Shaefer (2014), ‘Expertise and efficacy in elite political decision
making’, unpublished paper, https://sites.duke.edu/2014bmp/files/2014/
10/Loewen_et_al.pdf (accessed 15 January 2017).
Loewenstein, George (1987), ‘Anticipation and the valuation of delayed
consumption’, Economic Journal, 97 (387), 666–684.
Loewenstein, George, Joelle Y. Friedman, Barbara McGill, Sarah Ahmad,
Suzanne Linck, Stacey Sinkula, John Beshears, James J. Choi, Jona-
than Kolstad, David Laibson, Brigitte C. Madrian, John A. List and
Kevin G. Volpp (2013), ‘Consumers’ misunderstanding of health
insurance’, Journal of Health Economics, 32, 850–862.
Loewenstein, George, Cindy Bryce, David Hagmann and Sachin Rajpal
(2015), ‘Warning: You are about to be nudged’, Behavioral Science
and Policy, 1 (1), 35–42.
Lunn, P. (2014), Regulatory Policy and Behavioural Economics, Paris:
OECD Publishing.
Margetts, Helen, Peter John, Scott Hale and Taha Yasseri (2016), Political
Turbulence: How Social Media Shape Collective Action, Princeton, NJ:
Princeton University Press.
Marteau, Theresa M., David Ogilvie, Martin Roland, Marc Suhrcke and
Michael P. Kelly (2011), ‘Judging nudging: Can nudging improve
population health?’, British Medical Journal, 342, 263–265.
Martin, Steve J., Noah J. Goldstein and Robert B. Cialdini (2014), The
Small Big: Small Changes That Spark Big Influence, London: Penguin.
McConnell, Allan (2010), Understanding Policy Success: Rethinking
Public Policy, Basingstoke: Macmillan.
Michie, Susan, Maartje M. van Stralen and Robert West (2011),
‘The behaviour change wheel: A new method for characterising and
designing behaviour change interventions’, Implementation Science, 6
(42), 2–11.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 13 / Date: 21/12
JOBNAME: John PAGE: 14 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

160 How far to nudge?

Mols, Frank, S. Alexander Haslam, Jolanda Jetten and Niklas K. Steffens

(2015), ‘Why a nudge is not enough: A social identity critique of
governance by stealth’, European Journal of Political Research, 54 (1),
81–98.
NAO (National Audit Office) (2009), Innovation across Central Govern-
ment, report by the Comptroller and Auditor General, HC 12 Session
2008–2009, London: HMSO.
Newell, Benjamin R., David A. Lagnado and David R. Shanks (2007),
Straight Choices: The Psychology of Decision Making, London: Psy-
chology Press.
Nooteboom, Bart (2000), Learning and Innovation in Organizations and
Economies, Oxford: Oxford University Press.
Norman, Donald (1988), The Design of Everyday Things, New York:
Basic Books.
Nozick, Robert (1974), Anarchy, State, and Utopia, New York: Basic
Books.
Nyhan, Brendan and Jason Reifler (2010), ‘When corrections fail: The
persistence of political misperceptions’, Political Behavior, 32 (2),
303–330.
Nyhan, Brendan and Jason Reifler (2015), ‘The effect of fact-checking
on elites: A field experiment on U.S. state legislators’, American
Journal of Political Science, 59, 628–640.
OECD (Organisation for Economic Co-operation and Development)
(2017), Behavioural Insights and Public Policy: Lessons from Around
the World, Paris: OECD.
Oliver, Adam (ed.) (2013a), Behavioural Public Policy, Cambridge:
Cambridge University Press.
Oliver, Adam (2013b), ‘From nudging to budging: Using behavioural
economics to inform public sector policy’, Journal of Social Policy, 42
(4), 685–700.
Oliver, Adam (2017), The Origins of Behavioural Public Policy, Cam-
bridge: Cambridge University Press.
Ostrom, Elinor (1990), Governing the Commons: The Evolution of
Institutions for Collective Action, Cambridge: Cambridge University
Press.
Ostrom, Elinor (1997), ‘Behavioral approach to the rational choice theory
of collective action: Presidential address, American Political Science
Association, 1997’, American Political Science Review, 92 (1), 1–22.
Panagopoulos, Costas (2011), ‘Thank you for voting: Gratitude expres-
sion and voter mobilization’, Journal of Politics, 73, 707–717.
Park, Andreas and Hamid Sabourian (2011), ‘Herding and contrarian
behavior in financial markets’, Econometrica, 79, 973–1026.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 14 / Date: 21/12
JOBNAME: John PAGE: 15 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

References 161

Petrescu, Dragos C., Gareth J. Hollands, Dominique-Laurent Couturier,

Yin-Lam Ng and Theresa M. Marteau (2016), ‘Public acceptability in
the UK and USA of nudging to reduce obesity: The example of
reducing sugar-sweetened beverages consumption’, PLoS ONE, 11 (6),
e0155995.
Pike, Tony, Andrea Collier, Andrew Cotterill, Tim Everett, Rachel
Muckle and Amy Vanstone (2010), Understanding Behaviours in a
Farming Context: Bringing Theoretical and Applied Evidence Together
from across Defra and Highlighting Policy Relevance and Implications
for Future Research, Defra Agricultural Change and Environment
Observatory Discussion Paper, London: Defra.
Pressman, Jeffery L. and Aaron Wildavsky (1973), Implementation: How
Great Expectations in Washington Are Dashed in Oakland, Berkeley:
University of California Press.
Rabin, Matthew (1998), ‘Psychology and economics’, Journal of Eco-
nomic Literature, 36, 11–46.
Reis, Rodrigo S., Deborah Salvo, David Ogilvie, Estelle V. Lambert,
Shifalika Goenka, Ross C. Brownson and Lancet Physical Activity
Series 2 Executive Committee (2016), ‘Scaling up physical activity
interventions worldwide: Stepping up to larger and smarter approaches
to get people moving’, Lancet, 388 (10051), 1337–1348.
Richardson, Liz and Peter John (2012), ‘Who listens to the grass roots? A
field experiment on informational lobbying in the UK’, British Journal
of Politics and International Relations, 14 (4), 595–612.
Rimfeld, Kaili, Yulia Kovas, Philip S. Dale and Robert Plomin (2016),
‘True grit and genetics: Predicting academic achievement from person-
ality’, Journal of Personality and Social Psychology, 111 (5), 780–789.
Ritz, Adrian, Gene A. Brewer and Oliver Neumann (2016), ‘Public
service motivation: A systematic literature review and outlook’, Public
Administration Review, 76, 414–426.
Rogers, Everett M. (1983), Diffusion of Innovations, 3rd edn, London:
Collier Macmillan.
Ross, Heather (1970), ‘An experimental study of the negative income
tax’, Ph.D. thesis, Massachusetts Institute of Technology, Department
of Economics, http://hdl.handle.net/1721.1/13874 (accessed 12 June
2017).
Rothengatter, J.A. and John A. Groeger (1998), ‘Traffic psychology and
behaviour’, Transportation Research Part F: Traffic Psychology and
Behaviour, 1, 1–9.
Sanders, David (2010), ‘Behavioural analysis’, in David Marsh and
Gerry Stoker (eds), Theory and Methods in Political Science, 3rd edn,
Basingstoke: Macmillan, pp. 23–41.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 15 / Date: 21/12
JOBNAME: John PAGE: 16 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

162 How far to nudge?

Sanders, Michael, Raj Chande and Eliza Selley (2017), Encouraging

People into University, London: Department for Education.
Sartori, Giovanni (1970), ‘Concept misformation in comparative politics’,
American Political Science Review, 64 (4), 1033–1053.
Savani, Manu (2017), ‘The effectiveness of commitment devices: Field
experiments on health behaviour change’, Ph.D. thesis, University
College London.
Schmidt, Andreas T. (2017), ‘The power to nudge’, American Political
Science Review, 111 (2), 404–417.
Schwartz, Barry (2005), The Paradox of Choice: Why More Is Less, New
York: Harper Perennial.
Sedlmeier, Peter and Gerd Gigerenzer (2001), ‘Teaching Bayesian rea-
soning in less than two hours’, Journal of Experimental Psychology:
General, 130, 380–400.
Segall, Marshall H. (1976), Human Behavior and Public Policy: A
Political Psychology, New York: Pergamon Press.
Sent, Esther-Mirjam (2004), ‘Behavioral economics: How psychology
made its (limited) way back into economics’, History of Political
Economy, 36 (4), 735–760.
Service, Owain and Rory Gallagher (2017), Think Small, London:
Michael O’Mara Books.
Service, Owain, Michael Hallsworth, David Halpern, Felicity Algate,
Rory Gallagher, Sam Nguyen, Simon Ruda and Michael Sanders with
Marcos Pelenur, Alex Gyani, Hugo Harper, Joanne Reinhard and
Elspeth Kirkman (2014), EAST: Four Simple Ways to Apply Behav-
ioural Insights, London: Behavioural Insights Team.
Shafir, Eldar (ed.) (2013), The Behavioral Foundations of Public Policy,
Princeton, NJ: Princeton University Press.
Shafir, Eldar and Sendhil Mullainathan (2013), Scarcity: Why Having Too
Little Means So Much, London: Allen Lane.
Shah, Hetan and Emma Dawney (2005), Behavioural Economics: Seven
Principles for Policy-Makers, London: New Economics Foundation.
Shu, Lisa, Nina Mazar, Francesca Gino, Dan Ariely and Max H.
Bazerman (2012), ‘Signing at the beginning makes ethics salient and
decreases dishonest self-reports in comparison to signing at the end’,
Proceedings of the National Academy of Sciences, 109 (38), 15197–
15200.
Silva, Antonio and Peter John (2017), ‘Social norms don’t always work:
An experiment to encourage more efficient fees collection for
students’, PLoS ONE, 2 (5), e0177354, https://doi.org/10.1371/journal.
pone.0177354.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 16 / Date: 21/12
JOBNAME: John PAGE: 17 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

References 163

Silva, Antonio, Michael Sanders and Aisling Ni Chonaire (2016), Does

the Heart Rule the Head? Economic and Emotional Incentives for
University Attendance, London: Behavioural Insights Team.
Simon, Herbert A. (1947), Administrative Behavior: A Study of Decision-
Making Processes in Administrative Organization, 1st edn, New York:
Macmillan.
Simon, Herbert A. (1955), ‘A behavioral model of rational choice’,
Quarterly Journal of Economics, 69, 99–118.
Skinner, B.F. (1938), The Behavior of Organisms: An Experimental
Analysis, Cambridge, MA: B.F. Skinner Foundation.
Smith, Adam ([1776] 1904), An Inquiry into the Nature and Causes of
the Wealth of Nations, London: Methuen, http://www.econlib.org/
library/Smith/smWN.html (accessed 10 June 2017).
Smith, Graham (2009), Democratic Innovations: Designing Institutions
for Citizen Participation, Cambridge: Cambridge University Press.
Smith, Graham, Peter John and Patrick Sturgis (2013), ‘Taking political
engagement online: An experimental analysis of asynchronous discus-
sion forums’, Political Studies, 61, 709–730.
Sousa, Joana Lourenço, Emanuele Ciriolo, Sara Rafael Almeida and
Xavier Troussard (2016), Behavioural Insights Applied to Policy:
European Report 2016, EUR 27726 EN, Luxembourg: European
Commission.
Spencer, Charles and Jon Driver (2004), Crossmodal Space and Cross-
modal Attention, Oxford: Oxford University Press.
Spotswood, Fiona (2016), ‘Introduction’, in Fiona Spotswood (ed.),
Beyond Behaviour Change, Bristol: Policy Press, pp. 1–8.
Stone, Diane (1996), Capturing the Political Imagination: Think Tanks
and the Policy Process, London: Frank Cass.
Sugden, Robert (2004), ‘The opportunity criterion: Consumer sover-
eignty without the assumption of coherent preferences’, American
Economic Review, 94, 1014–1033.
Sugden, Robert (2008), ‘Why incoherent preferences do not justify
paternalism’, Constitutional Political Economy, 19 (3), 226–248.
Sugden, Robert (2016), ‘Do people really want to be nudged towards
healthy lifestyles?’, International Review of Economics, 64 (2), 113–
123.
Sunstein, Cass R. (ed.) (2000), Behavioral Law and Economics, Cam-
bridge: Cambridge University Press.
Sunstein, Cass R. (2014a), Why Nudge?, New Haven, CT: Yale Univer-
sity Press.
Sunstein, Cass R. (2014b), Simpler: The Future of Government, New
York: Simon & Schuster.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 17 / Date: 21/12
JOBNAME: John PAGE: 18 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

164 How far to nudge?

Sunstein, Cass R. (2016), The Ethics of Influence: Government in the Age

of Behavioral Science, Cambridge: Cambridge University Press.
Sunstein, Cass R. (2017a), Human Agency and Behavioral Economics:
Nudging Fast and Slow, Cham: Palgrave Macmillan/Springer.
Sunstein, Cass R. (2017b), ‘Nudges that fail’, Behavioural Public Policy,
1 (1), 4–25.
Sutherland, Stuart (1992), Irrationality: Why We Don’t Think Straight!,
New Brunswick, NJ: Rutgers University Press.
Tapp, Alan and Sharyn Rundle-Thiele (2016), ‘Social marketing and
multidisciplinary behaviour change’, in Fiona Spotswood (ed.), Beyond
Behaviour Change, Bristol: Policy Press, pp. 135–156.
Thaler, Richard (1980), ‘Toward a positive theory of consumer choice’,
Journal of Economic Behaviour and Organization, 1, 39–60.
Thaler, Richard H. (1994), The Winner’s Curse, Princeton, NJ: Princeton
University Press.
Thaler, Richard H. (1999), ‘Mental accounting matters’, Journal of
Behavioral Decision Making, 12, 183–206.
Thaler, Richard H. (2015), Misbehaving: The Making of Behavioural
Economics, New York: Allen Lane.
Thaler, Richard H. and Shlomo Benartzi (2004), ‘Save More Tomorrow:
Using behavioral economics to increase employee saving’, Journal of
Political Economy, 112, 164–187.
Thaler, Richard H. and H.M. Shefrin (1981), ‘An economic theory of
self-control’, Journal of Political Economy, 89 (2), 392–406.
Thaler, Richard H. and Cass R. Sunstein (2008), Nudge: Improving
Decisions about Health, Wealth, and Happiness, New Haven, CT: Yale
University Press.
Tummers, Lars G., Asmus Leth Olsen, Sebastian Jilke and Stephan G.
Grimmelikhuijsen (2016), ‘Introduction to the virtual issue on behav-
ioral public administration’, Journal of Public Administration Research
and Theory, Virtual Issue 3, 1–3.
Tversky, Amos and Daniel Kahneman (1971), ‘Belief in the law of small
numbers’, Psychological Bulletin, 76 (2), 105–110.
Tversky, Amos and Daniel Kahneman (1974), ‘Judgment under uncer-
tainty: Heuristics and biases’, Science, 185 (4157), 1124–1131.
Tversky, Amos and Daniel Kahneman (1981), ‘The framing of decisions
and the psychology of choice’, Science, 211 (4481), 453–458.
Tversky, Amos and Richard H. Thaler (1990), ‘Preference reversals’,
Journal of Economic Perspectives, 4 (1), 193–205.
Uchitelle, Louis (2001), ‘Following the money, but also the mind’, New
York Times, 11 February.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 18 / Date: 21/12
JOBNAME: John PAGE: 19 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

References 165

U.S. Department of Health and Human Services (2010), Healthy People

2010, conference edn, Washington, DC: U.S. Department of Health
and Human Services.
Van de Vyver, Julie and Peter John (2017), ‘A field experiment: Testing
the potential of norms for achieving behavior change in English
parishes’, Journal of Applied Social Psychology, 47, 347–352.
Varazzani, Chiara (2017), ‘The risks of ignoring the brain’, Behavioral
Science, http://behavioralscientist.org/risks-ignoring-brain/ (accessed
28 June 2017).
Wansink, Brian (2006), Mindless Eating: Why We Eat More Than We
Think, London: Hayhouse.
Wansink, Brian and P. Chandon (2006), ‘Can “low-fat” nutrition labels
lead to obesity?’, Journal of Marketing Research, 43, 605–617.
Ward, Hugh and Peter John (2013), ‘Competitive learning in yardstick
competition: Testing models of policy diffusion with performance
data’, Political Science Research and Methods, 1 (1), 3–25.
Webb, Eugene J., Donald T. Campbell, Richard D. Schwartz and Lee
Sechrest (1966), Unobtrusive Measures: Nonreactive Research in the
Social Sciences, Chicago, IL: Rand McNally.
Weiss, Carol H. (1977), ‘Research for policy’s sake: The enlightenment
function of social research’, Policy Analysis, 3 (4), 531–545.
Whitehead, Mark, Rhys Jones, Rachel Howell, Rachel Lilley and Jessica
Pykett (2014), Nudging All Over the World, Swindon: Economic and
Social Research Council.
Wiebe, G.D. (1951–1952), ‘Merchandising commodities and citizenship
on television’, Public Opinion Quarterly, 15 (4), 679–691.
Wildavsky, Aaron (1966), ‘Toward a radical incrementalism: A proposal
to aid Congress in reform of the budgetary process’, in Congress: The
First Branch of Government, Washington, DC: American Enterprise
Institute, pp. 115–165.
Wilkinson, T. Martin (2013), ‘Nudging and manipulation’, Political
Studies, 61 (2), 341–355.
World Bank (2015), World Development Report 2015: Mind, Society, and
Behavior, Washington, DC: World Bank.
World Health Organization (2009), Interventions on Diet and Physical
Activity: What Works, summary report, Geneva: World Health Organ-
ization.
Zamir, Eyal and Doron Teichman (eds) (2014), The Oxford Handbook of
Behavioral Economics and the Law, Oxford: Oxford University Press.

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 19 / Date: 21/12
JOBNAME: John PAGE: 20 SESS: 4 OUTPUT: Fri Jan 26 10:51:11 2018

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: 10-ReferencesPJED /Pg. Position: 20 / Date: 21/12
JOBNAME: John PAGE: 1 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Index
academics 5, 55, 57, 61, 65, 78, Behavioural Economics Team of the
106 Australian Government (BETA)
in behavioural public policy 61–3 84
role of 64 behavioural errors 136
alcoholism 26–7, 30 behavioural ideas 7, 17, 64, 77, 82,
Anderson, Joel H. 111–13 85, 106
Ariely, Dan 66, 76, 96 behavioural insights 9, 10, 79–80,
89–90, 93, 107, 146
backfire effect 30 application of 130
Barber, Michael 78 in Britain 74–6
Becker, Gary S. 42, 142 in France 84
Becoming a Man (BAM) programme in Netherlands 83–4
130 translation 82–5
The Behavioral Foundations of Public in UK 82–3, 85
Policy (Shafir) 61 in USA 82–4
‘A behavioral model of rational Behavioural Insights Team (BIT) 5,
choice’ (Simon) 43–4 72–3, 76–9, 90
Behavioral Science and Policy achievement 78
Association (BSPA) 61, 85 behavioural ideas 81–2
behaviour change 24–5 critics 78
critiques of rational model 43–5 innovation 80
and nudge 4–7 rules and regulations 81
policies 12–13, 26–8, 75, 125 from special unit to international
rational action and 28–9 not-for-profit 80–82
reactance 29–30 work on error and debt 96
‘behavioural market failure’ 116
reformers 36
behavioural problems
social science and 39–40
cause of 30
behavioural economics crime and 30–31
‘anchoring’ 46 from food and diet 22–3
books for 64–6 behavioural public administration 52
changes in 49–50 behavioural public policy 5, 9–10,
diffusion of 52 34–5, 85, 136
mainstream 47–9 academics 61–3
‘mental accounting’ 46 assessing 142–6
in 1980s and 1990s 47 entrepreneurial public
policy problems 51–2 administration 9–10
simplification of 56 radical approach 145–6
sophistication of 50–51 success of 55

167

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Index /Pg. Position: 1 / Date: 22/1
JOBNAME: John PAGE: 2 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

168 How far to nudge?

behavioural revolution in social collective action failure 94

sciences collective behaviours 31–3
behavioural economics see COM-B (‘capability,’ ‘opportunity,’
behavioural economics ‘motivation’ and ‘behaviour’) 63
brain sciences 40 commitment 128–9, 133
interdisciplinarity 41 communication flows 5–6
rational model 41–3 Conly, Sarah 113
critiques of 43–5 costs and benefits 28–9, 41–2
behavioural sciences, public criticism Couturier, Dominique-Laurent 118
66–7 criticism of nudge
behavioural spillover 132 don’t always work 95–7
behaviourism 39 limited range 81–91
Benartzi, Shlomo 52, 88, 105 limits to incrementalism 101–2
beneficial behaviours neglecting citizen intelligence and
in health 23–4 feedback 104–6
during pregnancy 34 neoliberalism 102–5
Benz, Matthias 57 potentially harmful 98–9
Blair, Tony 57, 58 temporary effects 93–5
blame attribution 34–5 weak effects 91–3
Blattman, Christopher 130 weak external validity 95
Blume, Toby 128 weak knowledge base 97–8
boomerang effect 30 Cutts, Dave 92
bounded rationality 16
Bovens, Mark 137 Dalton, Russell J. 29
brain sciences 40 Dawney, Emma 57
Brest, Paul 136 deception 119
Brighton, Henry 138 decision-makers 16
Britain, behavioural insights in 74–6 decision-making 46, 108, 117, 136,
Broockman, David 91 138, 139, 141
Brookings Foundation 63 hasty 124
Bruns, Hendrik 134 incremental 101–2
Bryce, Cindy 134 rapid 123
bureaucratic innovation 79–80, 86 deliberation 125–6
Department for Business, Innovation
Cameron, David 76, 83, 100 and Skills 73
Campbell Collaboration 97 Department for Environment, Food
Cartwright, Edward 50 and Rural Affairs (Defra) 74
CBST see Social and Behavioral depression 34
Sciences Team (CBST) diet 22–3, 127, 132, 140
Chabris, Christopher 66 in pregnancy 33
Change4Life programme 27 see also exercise
‘choice architectures’ 6 diverting public attention 99–100
Cialdini, Robert B. 65 Dobson, Andrew 103
citizen intelligence and feedback, Dolan, Paul 76
neglecting 104–6 Downs, Anthony 123
cognitive behavioural therapy (CBT) Duckworth, Angela L. 131
130 Dworkin, Gerald 111

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Index /Pg. Position: 2 / Date: 10/1
JOBNAME: John PAGE: 3 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

Index 169

Education Endowment Foundation 73 Gigerenzer, Gerd 138

educative nudges 13 Gneezy, Urim 50
effective motivation 131 Goldstein, Daniel 56
‘elite experiments’ 139 Gollwitzer, Peter M. 90
embedded behaviours 30–31 Goodin, Robert E. 101
entrepreneurial public administration Government Economic and Social
9–10 Research (GESR) 74–5
epigenetic inheritance 33 Government Social Research (GSR)
epigenetic marks 33–4 58
The Ethics of Influence (Sunstein) 116 Green, Donald P. 93
ethics of nudge 144 Grether, David M. 48
freedom 114–15 grit idea 131
libertarian paternalism 108–13 Grose, Christian 139
moving beyond 118–20 Guryan, Jonathan 130
psychological state 114–15
public opinion 117–18
habits 25–6
The European Nudging Network
Hagmann, David 134
(TEN) 84
Hallsworth, Michael 127
Executive Order 13707 83
Halpern, David 57–8, 76, 80
exercise
Harford, Tim 65
for beneficial effects 22–4, 32
Hargreaves-Heap, Shaun 104
lack of 26–7
experimental public administration 10, harmful nudges 98–9
18 Haslam, S. Alexander 133
‘experiments on institutions’ 139 healthy eating 23
Heller, Sara B. 130
Farrell, Henry 105 Hertwig, Ralph 13, 131–2
Fieldhouse, Ed 92, 95 Highway Code 14–15 , 133
food 22–3 HM Revenue and Customs (HMRC)
Foucault, Michel 67, 104 72, 74, 81
France, behavioural insights 84 Hollands, Gareth J. 118
Fraud, Error and Debt: Behavioural ‘homo heuristicus’ 138
Insights Team Paper 77 House of Lords Select Committee on
Freakonomics 64 Behaviour Change 75
freedom 114–15 ‘hyperbolic discounting’ 48
freedom of choice, libertarian
paternalism 110, 113
Frey, Bruno S. 57 Ideas42 64, 85
Friedman, Milton 44 ideological public policy, criticism of
nudge 102–5, 143
Galbraith, John Kenneth 64 incremental decision-making 101–2
Gallagher, Rory 130 Initiative for Science, Society and
game theory 49–50 Policy (ISSP) 84
Gardner, Frances 31 innovation
Gerber, Alan S. 93 bureaucratic, 79–80, 86
Get Out the Vote (GOTV) translating nudge into practice
experiments 71–2 79–80

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Index /Pg. Position: 3 / Date: 10/1
JOBNAME: John PAGE: 4 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

170 How far to nudge?

An Inquiry into the Nature and Luistro Jonsson, Marijane 134

Causes of the Wealth of Nations
(Smith) 64 mainstream economics, changes in
interdisciplinarity 41 49–50
internet politics 139 Marteau, Theresa M. 26, 89, 91, 95,
Irrationally Yours: On Missing Socks, 97, 98, 106, 118
Pickup Lines, and Other Matthews, Michael D. 131
Existential Puzzles (Ariely) 66 May, Theresa 83, 100
McConnell, Allan 137
Jackson, Tim 65 Mellers, Barbara A. 117
Jamison, Julian C. 130 ‘mental accounting’ 46
Jetten, Jolanda 133 Michie, Susan, 63
John, Peter 13, 62, 92, 96, 128, 138 ‘Mind, Society, and Behavior’ report
Johnson, Eric 56 64
Jones, Bryan 44 Mindless Eating (Wansink) 66
Jones, Rhys 103, 104 MINDSPACE report 60
Jung, Janice Y. 117 Mols, Frank 133
justification 115–16 More or Less program 65
Mullainathan, Sendhil 130
Kahneman, Daniel 45–7, 49, 51, 58, multi-causal behaviours 30–31, 34
129, 131, 137
Kalla, Joshua 91 neglecting citizen intelligence and
Kallgren, Carl A. 65 feedback 104–6
Kantorowicz-Reznichenko, Elena 134 neoliberalism, criticism of nudge
Kelly, Dennis R. 131 102–5
Kelly, Michael P. 26, 89, 91, 95, 97, Netherlands, behavioural insights
98, 106 83–4
Keynes, John Maynard 64 New Economics Foundation 57
Klement, Katharina 134 Ng, Yin-Lam 118
Ni Chonaire, Aisling 129
Larimer, Christopher W. 93 nudge 117–18
Lewis, Michael 46 advantages of 139
libertarian paternalism 67, 108–13 behaviour change and 4–7
defence of 115, 118, 119 belief 136
freedom of choice 110, 113 to change health behaviour 128
moving beyond 118–20 commitment device 128–9
light-touch nudges, 59, 89, 93, 98, controlling elites with 136–40
110, 121 criticism and limitation
light-touch policies 110 don’t always work 95–7
limited range, criticism of nudge limited range 81–91
81–91 limits to incrementalism 101–2
linked behaviours 30 neglecting citizen intelligence and
List, John A. 48, 50 feedback 104–6
Loewenstein, George 48, 134 neoliberalism 102–5
London Schools Excellence Fund 73 potentially harmful 98–9
long-term behaviour change 131 temporary effects 93–5
Ludwig, Jens 130 weak effects 91–3

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Index /Pg. Position: 4 / Date: 10/1
JOBNAME: John PAGE: 5 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

Index 171

weak external validity 95 Oliver, Adam 8, 45–6

weak knowledge base 97–8 open nudges 127–34
dispersion of power 140 Organisation for Economic
ethics of 144 Co-operation and Development
deception 119 (OECD) report 82, 84–5
freedom 114–15 Ostrom, Elinor 49, 50, 105
libertarian paternalism 108–20
psychological state 114–15 Park, Andreas 50
public opinion 117–18 peer review 72
public policies 118–20 pension scheme 52
as form of decision-making 141 personal nudges 130
goal of 133 persuasion effects 91
implementation of public policies Peters, B. Guy 137
8–9 Peterson, Christopher 131
personal 130 Petrescu, Dragos C. 118
think 123–6 planned behaviour 25
thought-provoking 129 Plott, Charles R. 48
translating into practice policy failures and successes 137
behavioural insights 79–80 policy-makers 16–17
behavioural insights in Britain and behavioural agenda 56–8
74–6 challenge for 12
Behavioural Insights Team 76–9 policy-making 136–9
innovation 79–80 policy punctuations 44
randomised controlled trial (RCT) Pollack, Harold A. 130
79–80 poverty, impact of 33–4
from special unit to international poverty-reinforcing behaviours 33–4
not-for-profit 80–82 ‘preference reversal’ 48
testing and experimentation Pressman, Jeffery L. 62
70–74 psychological state 114–15
translation to other contexts 82–5 public administration 18, 44, 52, 97,
Nudge: Improving Decisions about 136, 144
Health, Wealth, and Happiness behavioural, 52
(Thaler and Sunstein) 58–60 entrepreneurial 9–10
Nudge, Nudge, Think, Think 105–6, public criticism, behavioural sciences
125, 126 66–7
nudge plus 10–14, 127–34 public goods 6, 32, 134
‘Nudges that fail’ (Sunstein) 96 public opinion and nudging 117–18
‘nudging the S-curve’ 94 public policy 118–20, 134
behavioural see behavioural public
Obama, Barack 83 policy
obesity 23, 24 implementation of 8–9
O’Donnell, Gus 60, 76, 80 issues 12
OECD report see Organisation for Pykett, Jessica 103, 104
Economic Co-operation and
Development (OECD) report Rabin, Matthew 48
Ogilvie, David 26, 89, 91, 95, 97, 98, radical behaviourism 39, 145–6
106 radical incrementalism 101–2

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Index /Pg. Position: 5 / Date: 22/1
JOBNAME: John PAGE: 6 SESS: 3 OUTPUT: Fri Jan 26 10:51:11 2018

172 How far to nudge?

radical nudges 10 banning 15, 27–8

radicalism 103–4 during pregnancy 30, 34
Rahali, Bilel 134 Social and Behavioral Sciences Team
Rajpal, Sachin 134 (CBST) 5, 83
randomised controlled trial (RCT) 5, Social Innovation Partnership (TSIP)
9, 26, 71–3, 79–80, 105, 107, 73
130 social marketing 40
rational action 143 social media politics 139
and behaviour changes 28–9 social science
rational bureaucracy 114 and behaviour change 39–40
rational ignorance 123 behavioural revolution in see
rational model 41–3, 60, 138 behavioural revolution in social
critiques of 43–5 sciences
reactance 29–30 soft paternalism 116
Reagan, Ronald 82 Soroka, Stuart 137
Reckoning with Risk: Learning to Live Steffens, Niklas K. 133
with Uncertainty 66 Stoker, Gerry 13
Reno, Raymond R. 65 Stutzer, Alois 57
resistance to behaving well 24–5 substance abuse 31
Richardson, Liz 138 Sugden, Robert 47, 103, 104,
Roland, Martin 26, 89, 91, 95, 97, 98, 118–19
106 Suhrcke, Marc 26, 89, 91, 95, 97, 98,
106
Sabourian, Hamid 50 Sunstein, Cass R. 6, 8, 58–9, 62, 96,
Sanders, Michael 81, 127, 129 109, 111–19, 121, 133–5
Sartori, Giovanni 8 system 1 thinking 51, 117, 122–3,
Scarcity: Why Having Too Little 127–9
Means So Much 64 system 2 thinking 51, 117, 122–3,
Schmidt, Andreas T. 114 127, 129
Schwartz, Barry 44 system-wide changes 14–16
self-nudges 130
Service, Owain 130
tax change 7, 13, 62, 118
Seven Principles for Policy-Makers 57
taxation 52, 97
Shaefer, Tamir 137
Shah, Anuj K. 130 temporary policies of nudge 93–5
Shah, Hetan 57 Thaler, Richard H. 6, 8, 47–9, 51, 52,
Shalizi, Cosma 105 56, 58–9, 88, 105, 109, 111–15,
Shaw, Daniel S. 31 133, 135
Sheffer, Lior 137 ‘t Hart, Paul 137
Sheridan, Margaret 130 thought-provoking nudges 129
Silva, Antonio 96, 129 tools of government 7–8
Simon, Herbert A. 16, 43 Transforming Behaviour Change 66
Simons, Daniel 66 translating nudge into practice
Slovic, Paul 137 behavioural insights 79–80
The Small Big 66 in Britain 74–6
Smith, Adam 64 Behavioural Insights Team 76–9
smoking 24–6, 28 innovation 79–80

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Index /Pg. Position: 6 / Date: 22/1
JOBNAME: John PAGE: 7 SESS: 2 OUTPUT: Fri Jan 26 10:51:11 2018

Index 173

randomised controlled trial 79–80 Walgrave, Stefaan 137

from special unit to international Wallace, George 82
not-for-profit 80–82 weak effects, criticism of nudge
testing and experimentation 91–3
70–74 weak external validity, criticism of
translation to other contexts 82–5 nudge 95
translation phenomenon 63–6 weak knowledge base 97–8
Trump, Donald 82–3, 100 West, Robert 63
Tversky, Amos 48, 58, 131, 137 Whitehead, Mark 103, 104
Why Nudge? (Sunstein) 115–17
UK, behavioural insights 82–3, 85 Widdop, Paul 92
USA, behavioural insights 82–4 Wildavsky, Aaron 62
World Bank 84
Van de Vyver, Julie 96 World Development Report 2015 33,
van Stralen, Maartje M. 63 52, 64, 82

Columns Design XML Ltd / Job: John-How_far_to_nudge / Division: Index /Pg. Position: 7 / Date: 10/1