Oxford Development Studies

ISSN: 1360-0818 (Print) 1469-9966 (Online) Journal homepage: http://www.tandfonline.com/loi/cods20

The Good Governance Agenda: Beyond Indicators

without Theory

Matt Andrews

To cite this article: Matt Andrews (2008) The Good Governance Agenda: Beyond
Indicators without Theory, Oxford Development Studies, 36:4, 379-407, DOI:
10.1080/13600810802455120

To link to this article: http://dx.doi.org/10.1080/13600810802455120

Published online: 15 Nov 2008.

Submit your article to this journal

Article views: 2753

View related articles

Citing articles: 39 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=cods20

Download by: [University of Toronto Libraries] Date: 07 November 2016, At: 08:41
Oxford Development Studies,
Vol. 36, No. 4, December 2008

The Good Governance Agenda: Beyond

Indicators without Theory
MATT ANDREWS

ABSTRACT Effective government matters, but what is it? Good governance indicators go some way
to provide a definition, but how much do they say about what effectiveness is, why this is so, and how
it matters to development? This article argues that much work on the good governance agenda
suggests a one-best-way model, ostensibly of an idyllic, developed country government: Sweden or
Denmark on a good day, perhaps. The implied model lacks consistency, however, seems
inappropriate for use in the development dialogue and is not easily replicated. In short, it resembles
a set of well meaning but problematic proverbs. The good governance picture of effective
government is not only of limited use in development policy but also threatens to promote dangerous
isomorphism, institutional dualism and “flailing states”. It imposes an inappropriate model of
government that “kicks away the ladder” that today’s effective governments climbed to reach their
current states. The model’s major weakness lies in the lack of an effective underlying theoretical
framework to assist in understanding government roles and structures in development. A framework
is needed before we measure government effectiveness or propose specific models of what
government should look like. Given the evidence of multiple states of development, the idea of a one-
best-way model actually seems very problematic.

1. Introduction
Effective government matters, as everyone agrees. However, most people have difficulty
defining what effective government is. Good governance indicators go some way to
provide a definition, typically reflecting what many would consider sensible and attractive
characteristics of effective governments.1 They suggest, for example, that governments
should be fiscally disciplined and decentralized to discern and respond to citizen need,
comprise politically neutral managers, and make and manage business-friendly policies.
Some indicators focus on one of a long list of such characteristics, whereas others marry
30 –40 in single measures. Countless articles have then used these measures to reflect
government effectiveness and test how this matters to development, growth and poverty
reduction. Yet, do the measures really tell us much about what effective government is,
why this is so, and how it matters to development?

*Matt Andrews is Assistant Professor at the Harvard Kennedy School of Government. Contact address: 316
Rubenstein, 79 John F. Kennedy Street, Cambridge, MA 02138, USA. Email: [email protected].
The author gratefully acknowledges support of the Oxford Policy Institute (OPI) to produce this article. The
article benefited from comments at an OPI event in London in January 2008 and by review comments by David
Leonard and Roger Hay and five anonymous reviewers.
ISSN 1360-0818 print/ISSN 1469-9966 online/08/040379-29
q 2008 International Development Centre, Oxford
DOI: 10.1080/13600810802455120
380 M. Andrews

I ask this question here not to criticize the providers of these indicators, but rather to
investigate what they imply an effective government is—and to enquire how valid this
implication is for research and practical application in development. I start off suggesting
that the indicators present a one-best-way model of government—a structural design to
emulate. One might expect better scoring countries to reflect this model: Sweden or
Denmark on a good day, perhaps.2 I then ask whether such a model helps us think about
effective government in development, by posing three basic questions.
. Is the model consistent enough to suggest what an effective government is?
. Is the model correct and appropriate, with its elements reliably explaining
outcomes of relevance to development?
. Is the model replicable, such that its elements are reproduced confidently and with
reliable results?
Unfortunately, I find the model of limited use. Governments scoring well on effective
government measures share strong outcomes and a shell of formalized structure, but differ
in many other senses, especially regarding details of what the government really looks like
and how it manages. The Swedish model differs from the German model, and the British
and American models, for example. The governments emerging as effective are all
examples of what success is, but one does not get a consistent picture of why they are
successful, because of the variation in their structures and designs. Modeling government
effectiveness in this manner is like telling developing countries that the way to develop is
to become developed.3 The effective government models of today are also arguably
inappropriate comparators for use in development. The challenges developing countries
face are different from those faced by countries in which the effective governments are
currently located. Certainly, the effective governments faced much more similar
challenges in the past, but then they also looked very different and they would surely have
scored poorly on the current indicators! The effectiveness of these past governments is
reflected in the progress of their countries over the past century, and the story of how they
muddled through from the past to the present may be of most value for developing
countries today. Elements of the current one-best-way model inform governance reform
initiatives in today’s developing world, which are often unsuccessful. The elements do not
replicate well in different contexts, characterized by different challenges and
environments, and end up resembling proverbs that fit poorly in the wrong context and
contradict one another.
I conclude that our current picture of effective government may be useful in producing
indicators, but is of limited use in development work and also holds some dangers for the
development community. It will yield faulty academic analysis on effectiveness (through
mis-specified variables of effective government), biased donor decisions (that penalize
developing countries for not adopting practices that “work” in developed countries, for
example) and dangerous reform designs (where developing countries adopt practices
embodied in the indicators for external legitimacy even though these might compromise
their internal capacities). In recent parlance, the indicators might end up promoting
dangerous isomorphism, institutional dualism4 and “flailing states”,5 and could impose an
inappropriate model of government that “kicks away the ladder” that today’s effective
governments climbed to reach their current state.6
I identify the lack of an effective framework for thinking about the role of the state in
different contexts as the major problem with these indicators. The emphasis on developing
 The Good Governance Agenda 381

a static, common model of government effectiveness has undermined the importance of

thinking about why states might look different in different developmental contexts—
clarifying what effectiveness might mean and then devising propositions to explain it and
its variation (explaining, inter alia, why different governments achieve effectiveness
differently, how paths of development impact on effectiveness and why different
structural devices facilitate effectiveness differently in different settings). I believe that
pressure to develop best practice models, measure performance and create data has
drowned out clear and consistent thinking about what was being done. Just as Simon
(1947) argued that administrative principles of his day lacked theoretical content,
providing “a great help in persuasion, political debate and all forms of rhetoric” but being
less helpful as a theoretical foundation that explains why something works, when it does
not work, and how it got to work, I argue that we need a framework for thinking about
government that addresses such basic questions before we can properly measure
government effectiveness.

2. What Good Governance Indicators Say About Effective Government

The good governance community has grown in the past decade, producing a host of good
governance indicators. There are many types (Arndt & Oman, 2006; Boivard and Loffler,
2003; Hood et al., 2007; Knack et al., 2003; van de Walle, 2006). A specific variety
spotlights structural and management characteristics of governments and associated
outcomes considered important for development. Common examples are budgeting
system formality, bureaucratic quality, degree of decentralization, level of service
provision in areas such as health, education and infrastructure, and fiscal stability. The
World Governance Indicators (WGI) are the community’s most prominent product. These
combine standalone measures into aggregate indicators of six governance concepts,
including “government effectiveness”. The indicators are widely used in academic
literature as measures of key concepts such as effective government and influence lending
by donor organizations and public sector reform designs. They have been instrumental in
mainstreaming discussion and treatment of corruption, regulatory ailments and
governmental weaknesses in development. All students of government should be grateful
for this effort of data collection.
The indicators, however, arguably underlie what sociologists might call isomorphic
influences on thinking about what effective government is, where isomorphism refers to
the pressure to imitate organizational characteristics from one setting to another.7 The
argument is that academic work, lending engagements and reform proposals gain
legitimacy in the development community when they identify with the “myth” that formal
structures reflected in the indicators provide a rational means to attain desirable ends
(Meyer & Rowan, 1991). The reinforcing character of most indicators strengthens the
isomorphism, especially in developing countries where government seems ineffective and
dependence on external parties such as donors proposing such models is high.8 Conceding
such influence, one hopes that the good governance thinking about effective government is
appropriate for use in development. At least one must ask what this work calls an effective
government.9
The theoretical answer generally comes with reference to key strands of economic and
sociological work, especially Weber’s bureaucracy and North’s institutionalism (Knack &
Keefer, 1995; Rauch & Evans, 2000; La Porta et al., 1999).10 Both strands emphasize
382 M. Andrews

formal rule-bound governments with clear roles, allowing calculated confidence on the
part of the private sector. By extension to other constructs (such as agency theory) the
literature emphasizes limited, disciplined government. Kaufmann et al. (2007, p. 2)
identify the foundation of their good governance work as, “The norms of limited
government that protect private property from predation by the state”.11 The same authors
also note that limited government should be responsible for producing key “inputs” to
growth and development—such as education, health care and transportation infrastruc-
ture.12 Arguments differ on how these should be produced, invoking both Weberian
bureaucracy and New Public Management (NPM) elements. The mix of ideas constitutes
an interesting amalgam that is less than consistent as a theoretical foundation. Thomas
(2006, p. 10) describes this mix as the result of “personal ideas of governance” of people
developing indicators13 and argues that the “underlying [theoretical] construct has not
been defined”. Essentially, the WGI and other products are really the combination of many
different measures with many different underlying theories, normative perspectives and
viewpoints—the aggregated indicators imply these are all valid and their authors provide a
service in combining and “organizing” them.
Indicators reflect the lack of definition. The WGI’s “government effectiveness”
indicator, for example, includes up to 30 elements that seemingly just combine a variety
of better practice characteristics—and other indicators (such as regulatory burden)
incorporate characteristics also vital to the model (see Kaufmann et al. (1999), Arndt &
Oman (2006) and Thomas (2006) for detail of sources and Brinkerhoff & Goldsmith
(2005, p. 203) for a list of “ideal” characteristics):14
. Small government with limited engagement, formalized structures and rules.
. High quality, depoliticized public personnel in formal bureaucracy.
. Efficient and effective program implementation and service delivery (especially
in education, health and infrastructure).
. Disciplined budgets and efficient expenditures.
. Responsive, transparent, participatory and decentralized.
. Stable and credible policies.
. Pro-business orientation.
. Minimal red tape.
Recent critiques focus predominantly on the statistical gymnastics involved in creating
indicators with so many elements. My aim here is not to critique this process or the
indicators, but to identify the picture of government effectiveness it paints, which I suggest
is as follows:

An effective government is small and limited in its engagement, formalized in

mission and process. High-quality personnel devise and implement needed programs
and deliver efficient and effective services via participatory processes and disciplined,
efficient financial management. Responsiveness to the citizenry’s changing needs is
high and effected through transparent, decentralized and politically neutral structures;
consistently, even during political instability, without impeding (indeed supporting)
the private sector.

This picture surely would command instant and widespread approbation and embodies
characteristics one would expect to find in effective governments. Yet it is also demanding,15
 The Good Governance Agenda 383

to the degree that one would anticipate a fit only with governments in developed countries.
This is the case, as reflected in recent country scores shown in Figure 1.16 Ranging between
2.5 and 22.5, scores show the effectiveness of governments relative to others. The space
above zero ostensibly contains effective governments and the space below zero shows
ineffective governments. The three regions to the left are the 12 former Soviet Union (FSU)
countries, a selection of 20 sub-Saharan African (SSA) countries17 and eight South Asian
(SA) countries—all developing and transition areas. Only four governments in this selection
of 40 were comparatively effective (Bhutan at 0.33, Ghana at 0.05, Maldives at 0.03 and
South Africa at 0.78).
The next three regions are Latin America (LA), East Asia (EA), and the Middle East and
North Africa (MENA). There are 46 representatives from these regions and 16 of these
were comparatively effective.18 Two of the most effective governments come from East
Asia (Hong Kong at 1.76 and Singapore at 2.2). Nine out of 15 Eastern European (EaU)
governments scored in the effectiveness section; all are in the European Union (Bulgaria,
Czech Republic, Estonia, Latvia, Lithuania, Slovakia and Slovenia)19 or close to Northern
Europe and in advanced talks about EU membership (Croatia). The final group of 20
OECD countries all score in the effectiveness zone. Thirteen actually score above one.
Nine of the 11 countries with governments scoring above 1.5 are in the OECD.

3. Are One-best-way Models of Effective Government Appropriate for

Development?
One can view these scores as comparative performance against a one-best-way model of
effective government, where I propose a model simply as a structural design for emulation
across settings. The model here could be the picture of government effectiveness outlined
above, which one would expect to see represented in the countries scoring highly on the
index—Denmark, Singapore, Canada, Sweden, Australia and the Netherlands,20 the UK,
Hong Kong, USA, Belgium and Germany.21 As discussed, the dominant characteristic of
these countries is their advanced level of development,22 which raises an obvious concern;
using examples of success to show what success is can be like telling developing countries
that the way to develop is to become developed.23 This concern underlies part of Pritchett
and Woolcock’s worry that thinking based on such examples poses a problem in

Figure 1. Government effectiveness score, 2006 (22.5 to þ 2.5).

384 M. Andrews

development: “Simply mimicking . . . the organizational forms of a particular

‘Denmark’—has in fact been a root cause of deep problems encountered by developing
countries” (Pritchett & Woolcock, 2004, p. 193).
The comment obviously relates to attempts to replicate presumed elements of Swedish
(or other) models in developing countries. Its concern could also be raised with regard
to other uses of such a model in development: Should this approach to thinking about
government effectiveness inform academic work on growth, development and poverty
reduction as much as it does? Should this approach underlie lending decisions by donors?
Similar questions have emerged with regard to the use of best practices in general,
especially in the public sector (Overman & Boyd, 1994). Articles increasingly find best
practices devoid of theoretical and sometimes even practical merit and applied in
inappropriate contexts. Concern has even emerged regarding models of corporate
governance, arguably the private sector parallel of good governance literature. Observers
identify flaws in the one-best-way model of corporate governance propagated by business
schools on the basis of agency theory similar to that informing good governance
indicators. Sumantra Ghoshal links this literature and applications thereof to recent
corporate failures in America.24 Expressing concern at the use of faulty corporate
governance models, Ghoshal (2005, p. 86) reverses Kurt Lewin’s comment that “nothing
is as practical as a good theory” by saying: “The obverse is also true: Nothing is as
dangerous as a bad theory”. Ghoshal claims further that, “Bad management theories are,
at present, destroying good management practices”. The question is whether Swedish
(or other) models of effective government emerging from good governance work are
similarly problematic for the development community.
Answering the question requires analyzing the quality of the good governance model
and its appropriateness in the development context. A basic approach to doing this
involves seeing how well the model—structural design for emulation—meets simple
criteria, raised in the literature on models itself (Mayer, 1989):25 Is the model consistent?
Is the model correct? Can the model be replicated?

3.1 Are One-best-way Models Consistent?

Models are stories by which we connect certain elements and characteristics with certain
realities and often take the shape of structural designs for others to emulate. Consistent
models are characterized by a high correlation between the model’s ideas (its core
elements and implicit causal links, reflected in its structural design) and its examples, and
between the examples themselves. Consistency in the good governance model thus
requires that the high-scoring governments (Sweden and the others) reflect the suggested
picture of effective government and that they resemble each other.
Such consistency is high when one considers the proposed outcome of effective
government, growth and development. The highest-scoring governments all exhibit long
trends of positive growth, dating back to the late 19th Century. The trends include periods
of economic downturn, which are sometimes common and sometimes peculiar.26 Overall,
they are positive and represent sustained and quite common paths, whereby all of the
countries except Singapore and Hong Kong grew at similar rates across periods—
continuing to today.27 Questions about causality between growth and governance suggest
that this may not reflect a result of government effectiveness, however. The provision of
key services can be more firmly attributed to governments (at least at the margin).28
 The Good Governance Agenda 385

Trends in this regard also suggest consistency. Top-scoring governments all produce
highest-level health, education and infrastructure services. Access and performance in
health and education, for example, are in the top percentiles in all nine OECD countries
analyzed. Performance is not just high on key measures such as life expectancy and infant
mortality, but improvements in the past century have seen convergence around these high
levels for the countries, further suggesting consistency.
There is less consistency in the way the different governments produce these services,
however. Although all countries allocate significant shares of their GDP to both sectors29 and
there is a common trend towards more health spending, the role government plays in the
sectors varies significantly. Government in the USA plays a relatively smaller role in both
sectors than in other countries, actually contributing less to health care than the private
sector.30 Countries such as Sweden and Denmark (and even the UK and Canada) stand
in contrast, with relatively small private sector contributions.31 Structural approaches to
delivery also vary greatly, with some governments (such as Sweden) engaged in more quasi-
private activities than others—such as child and elderly care as well as “bakeries, gyms
and garden centers” (Henrekson, 2005). Public–private partnership engagements in health,
education, infrastructure and other sectors also vary across the sample, as do levels of
decentralization (with the UK, Belgium and the Netherlands more centralized in health care
than Sweden, Denmark and Germany (Mosca, 2007)).32 Statistical comparisons suggest that
recent decentralization trends actually run in different directions as well,33 suggesting a lack
of consistency between management structures in the nine OECD governments and the
“government effectiveness” prescription of limited and decentralized government.
There is a further lack of consistency in regard to government size (Handler et al.,
2005). The good governance picture suggests the importance of limited government,
which it measures in terms of legal checks (rule of law) as well as constraints on
government scope and fiscal size.34 Evidence shows, however, that not all of the high-
scoring governments are limited in fiscal size. They differ significantly in this regard, no
matter what measure is used. Public expenditure to GDP ratios ranged from about 35 to
over 55% across the countries in 2004, down from highs of about 37– 73% in the mid-
1990s (Hauptmeier et al., 2007, p. 268).35 Size differences reflect partly on different
approaches to budgeting and financial management as well. Although governments have
followed a common pattern reflected in the recent move to centralize the process across the
entire group, there are still process differences emerging from cultural and political
influences and manifest in different degrees of ex ante and ex post control in budget and
financial management systems, differences in performance management structures and
such (Curristine, 2005; Hallerberg et al., 2007; Joumard et al., 2004).
Differences in key public financial management systems are known to foster
differences in all sorts of management incentives and behaviors. Thus, there is further
reason to question the “consistency” of government models in the countries under
investigation. More reasons arise from a quick glance at revenue statistics. Revenue to
GDP ratios range from just over 30 to 50% in the countries, and tax collections as a
proportion of GDP are also quite different (Tanzi & Schuknecht, 2000). These differences
have implications for the way government is structured and engaged with society; varying
tax burdens imply different linkages between government and the private sector, for
example. High taxes and transfers in countries such as Sweden place an economic burden
on firms that has been blamed for a low-level of entrepreneurship in the country
(Henrekson, 2005).
386 M. Andrews

This kind of observation causes one to rethink whether all of the countries have policies
that good governance indicators might term “pro-business”—with low tax and regulatory
burdens and minimal state interference in the economy. At the very least, evidence
suggests that such “pro-business” orientation of the governments varies, with different
levels of taxation and administrative and economic regulations across the sample (OECD,
2005; Malyshev, 2006). There is a trend towards lowering burdens on business but, as with
the trend towards fiscal centralization and discipline, modalities and administrative
processes still vary considerably. Recent statistics show that businesses must comply with
fewer regulations in these countries than in many others (to start new operations, build
warehouses and such) and that processing times are relatively shorter in these governments
(OECD, 2005, see especially pp. 23 and 31; and World Development Indicators based on
Doing Business surveys36). But Belgium still stands out as having relatively bureaucratic
processes, bucking the trend. Sweden has fewer processes and is quite efficient (now) at
processing administrative requests but has high tax burdens and stringent social,
environmental and economic regulations that some argue hinders new business
development. Both Sweden and Belgium have lower levels of new firm entrants in
areas such as manufacturing and services than the other models (OECD, 2005), ostensibly
resulting from some of these constraints; but these constraints also reflect differences in
both the way government views its role vis-à-vis business and the way it administers itself,
with different types of state intervention considered pro-business in many of these cases,
and low levels of firm entry balanced out by higher levels of new firm success (ostensibly
influenced by state engagement with business).
The high-performing governments certainly do differ in how they manage regulations at
the interface with business. Key regulatory processes vary, written in law in some places
and framed only in lower-level rules in others, resulting in very different levels of
transparency and consultation (Malyshev, 2006). The choice of who manages such
policies also varies, with options including ministerial agencies, independent advisory
agencies and independent regulatory authorities. These are often arms-length agencies,
which the various best practice governments have used in different ways over the past 15
years (Matheson et al., 2007). The size of Sweden’s civil service has decreased in the past
two decades as a result of (in large part) creating these agencies. The civil service
comprised 46% of total public sector employment in 1985 and only 24% in 1996 (OECD,
1999, p. 6). Nothing in the good governance literature helps one to think about what these
entities should look like or how their creation might affect accountability and efficiency.
The “government effectiveness” indicator simply rewards governments for having fewer
civil servants and more “decentralized” structures.
The new structures have implications for core management issues, however. Oversight
procedures are not equally applicable to arms-length agencies in countries such as Belgium
and the USA (Matheson et al., 2007). Personnel procedures for these agencies differ from
those governing the civil service in Belgium, Sweden, the UK and the USA. Rules limiting
political involvement of bureaucrats do not hold for arms-length agencies in Sweden.
Such variations cause one to look carefully at the supposedly strong Weberian
bureaucratic structures that indicators show are in place in the higher-scoring countries.
Certainly all nine governments have merit-based hiring processes in place for the civil
service and all have rules limiting direct political engagement for civil servants; but
creating arms-length agencies where such rules do not hold muddies the picture. So too do
details of the systems in the different countries (Matheson et al., 2007; Rexed et al., 2007).
 The Good Governance Agenda 387

On the subject of political influence, for example, Matheson et al. (2007, p. 15)
commented that, “While there is near universal agreement on the general principle of
political non-partisanship, it is not necessarily equated with an apolitical process for senior
appointments”. Their study of OECD governments reveals a “wide diversity in the level of
involvement by politicians in the appointment and management of senior civil servants”.
Governments such as those of Sweden and the USA allow political engagement in these
kinds of decisions for at least the top four levels of the civil service (and some of the
appointees here are not civil servants). The UK has no such involvement, at any of its
senior levels, which are all merit-based and transparently so.
The meritocratic, Weberian model of bureaucracy may thus not be consistently in place in
the countries that score highly on the government effectiveness indicator. Certainly there is
no consistent bureaucratic model across the countries, just as there is no such model regarding
the processes of engagement with business or the size of government or the methods of
disciplining finances or even the structures of service delivery.37 The most consistent part of
their story is that they are all development successes.38 There is some consistency with regard
to broad formal public management frameworks (e.g. broad budgeting approaches and
bureaucratic structures). Beyond this, however, consistency declines, suggesting a limited
model and problematic basis for thinking about effective government.39

3.2 Are One-best-way Models Correct for use in Developing Countries?

Good models provide an acceptable construct to explain reality. When elements are
combined such models constitute coherent, believable stories explaining the reality under
consideration—structural designs that match the settings in which they will be emulated.
Given the current article’s focus, a correct model would provide a strong story about
how its effective government elements coalesce to foster development and the picture of
effectiveness would constitute a design that seems to match with the development country
setting.
Problems with establishing model consistency obviously complicate any discussion of
correctness. If nine developed country governments assessed as being effective do not
actually resemble one another or the picture of effectiveness implied in the indicators they
commonly score high marks on, there is limited rationale for testing the implied model’s
appropriateness in explaining development outcomes. Beyond this problem, however,
there are further observations about the issue of correctness that warrant brief discussion.
Essentially, government effectiveness scores provide a measure of governments in the
recent past (dating back over a decade at the most, the period in which they score their high
marks). The question is whether these scores and the good governance model of
effectiveness they might reflect have any relevance for governments currently challenged
by early or mid-stage development?
During the last 10 “good governance” years the countries we currently call developing
have struggled with issues that the high-scoring governments have not. Current “effective”
governments are not dealing with aggressive growth spurts or trying to put their many
young citizens to work; they are rather in a slowed growth phase struggling with questions
about fiscal sustainability arising primarily from the looming health care costs of an aging
population. Figure 2 shows differences between current growth rates faced by the effective
governments (onthe right) and other governments (ranging from most ineffective on
the left).
388 M. Andrews

Figure 2. Government effectiveness and economic growth, 2000– 06 (42 governments from the
original sample are represented here. The others did not have data available). Sources: Tanzi &
Schuknecht (2000); World Development Indicators (accessed December 2007); OECD (2007).

The more effective governments were working in economies averaging 11% growth
between 2000 and 2006 (the equivalent of less than 2% annual growth). Less effective
governments watched their economies grow in excess of 20% over the period (on
average).40 Countries with ineffective governments (scoring below zero) actually grew by
an average of 25% in this period (about 4% annually).
The governing context in developing countries differs from that in developed countries
in other important ways. As mentioned, the population dynamics are different, with vast
numbers of younger people in the developing world and a growing older population in
industrialized countries. This suggests different demand-related challenges in key service
areas such as health, education and infrastructural development. One can imagine these
different challenges resulting in different choices of government engagement and in
different resource and administrative needs; governments in developing countries may
need to engage more in programs focused on youth development and build new schools
to accommodate growing numbers of children, for example, whereas those in developed
countries may be focusing less on youth-based policies and more on the aged, closing and
consolidating schools to deal with a declining school-going population and building old-
age homes. Differences like this will also result in demands for different types of people,
and different organizational structures in governments.
Revenue structures also vary significantly—even more than they do in the more effective
governments (as discussed). Figures 3–5 show how the more effective governments have
generally higher domestic revenue sources than less effective governments, are less
dependent on taxes on international trade than less effective governments and are more
dependent on direct domestic income and profit taxes.41 I raise these differences because
revenue-raising abilities and revenue sources arguably have a significant impact on
government focus, structure, responsiveness, and so on. A high dependency on international
trade taxes certainly complicates government’s role in managing trade, for example, and
necessitates having specific regulatory and administrative structures.42 This kind of
dependency often reflects limited formal contracting structures in domestic business
 The Good Governance Agenda 389

Figure 3. How government revenue/GDP ratios differ, 2005. Sources: Tanzi & Schuknecht (2000);
World Development Indicators (accessed December 2007); OECD (2007) (48 governments from the
original sample are represented here; the others did not have data available).

Figure 4. Different dependencies on direct income, profit taxes, 2005. Sources: See Figure 3.

Figure 5. Different dependencies on international trade taxes. Sources: See Figure 3.

390 M. Andrews

(affecting abilities to tax income and sales).43 Such dependency is also expected to influence
what government looks like, as discussed in the emerging fiscal sociology literature, which
argues that “Sources of state revenue have a major impact on patterns of state formation”
(Moore, 2004, p. 297).
Furthermore, small domestic tax bases go with young populations and informal
economies, and create peculiar challenges with regard to fiscal balancing that are different
from those challenging more developed economies, with broader and more formalized
domestic tax bases. The interactive dependency of government as an organization on
domestic business and citizens in more effective governments stands in strong contrast
to the interactive dependency many less effective governments have, contextually, with
foreign entities driving trade, providing aid and buying primary products such as minerals
and oil (often these are the three dominant resource sources).
The differences between government realities in developed and developing countries
manifest in almost every area—growth rates, revenue dependency, access to and ability to
retain skilled people, basic infrastructure (buildings, communications technology, etc.),
and so forth. As a result government challenges are highly divergent as well. Consider the
case of health care demands reflected in Figure 6, which shows how different governments
face different infant mortality problems. Today, more effective governments have to deal
with significantly lower levels of infant mortality than less effective governments. The
differences are large, even though infant mortality rates have been dropping in those
countries with less effective governments.44 Responses to high levels of infant mortality
involve management engagements, structural arrangements and resource allocations that
are not the same as those associated with low infant mortality rate contexts. Given the
differences in governance context and in challenges faced by governments ranked as

Figure 6. Infant mortality rates for various governments and years. Sources: World Development
Indicators (accessed December 2007); OECD (2007).
 The Good Governance Agenda 391

effective and those ranked as ineffective—otherwise known as developed and

developing—one has to ask whether Swedish or other developed country models are
appropriate, or correct, for governments in the Congo or Ghana or Bolivia.45
Figure 6 provides a partial answer in its presentation of longitudinal as well as cross-
sectional information. It shows that, at some time in the past century, the countries with
more effective governments had infant mortality rates similar to those challenging less
effective governments today. The model governments had to deal with high infant mortality
rates in the late 19th Century, which declined significantly between 1870 and 1937. Similar
progress is evident in less effective government settings between 1960 and 2005. By the
numbers alone, the infant mortality challenge in Sweden and the Netherlands in 1937 (a rate
of about 40) is similar to that faced by some developing countries today (including Brazil,
Georgia and Nicaragua).46 The larger challenges Belgium and Canada were still dealing
with in 1937 (rates at 145 and 77) correlate with the problems governments in Angola,
Azerbaijan, Kenya and Pakistan (among others) still struggle with.
Similar longitudinal analysis also reveals important points of connection with regard to
growth, revenues and other variables. Most of the more effective governments faced
economic growth rates similar to those now evident in developing countries (3–4% and
above per annum) between 1870 and 1913, 1920 and 1929 and in the 1960s and 1980s (Tanzi
& Schuknecht, 2000, p. 78). Revenue to GDP ratios were low in these governments in the
latter part of the 19th Century and early 20th Century, generally reaching levels like those
now observed in developing countries sometime in the 1920s and 1930s (Tanzi &
Schuknecht, 2000, pp. 52–53). Revenue composition also changed significantly during the
past century and more. Taxes on international trade amounted to nearly 3% of GDP in Canada
and the USA in 1870 and Canada still drew more than 3% of GDP in these taxes in 1913
(Tanzi & Schuknecht, 2000, pp. 56–57). This exceeds current levels for all but a handful of
less effective governments, which are mostly located in a band between 1 and 2.5%.47 On
average the currently more effective governments in the OECD relied on international trade
taxes for up to 2% of the value of GDP in 1937. In the period since World War II these
governments decreased their dependence on trade taxes and enjoyed increased access to and
(most importantly) cultivated a dependence upon domestic income and sales taxes.48
Even this brief discussion makes it obvious that basic contextual differences complicate
comparisons between current developing country governments and current Swedish (and
other) models. However, developing country governments might be able to learn a lot
from the governments that existed in these countries at points of time characterized by
more similar developmental challenges, and particularly from the way the governments
evolved beyond those contexts. The lessons would be even more pertinent when derived
from the experiences of other developing countries that are perhaps one or two steps
higher on the effectiveness ladder but more contextually connected—adjacent, culturally
and politically similar, for example.49 What did today’s effective governments’ and/or
more effective developing governments’ health sectors look like when faced with infant
mortality rates resembling those we see in developing countries today? What did these
governments’ health sectors look like as they addressed this challenge and produced or
facilitated the much-improved outcomes of today?
I am not suggesting we ask these questions to identify a one-best-way stage-of-
development model (which, incidentally, I do not believe exists), but rather that we try to
understand the evolution of government in contexts where results show effectiveness. This
evolution may not always be applicable in a technical sense—the approach to treating
392 M. Andrews

infant mortality now (via immunizations) contrasts to that adopted in the early 1900s
(confinement)50—and there may be questions about the applicability of yesterday’s
lessons in today’s globalized world. But there are key issues that are surely relevant: Why
and how did the governments identify their challenges, and move from challenge to
challenge? How did they mobilize skills to deal with the challenges? How did they
organize themselves? These are the kinds of questions organizational theorists focus on
when discussing effectiveness. If we ask them, we suggest that effectiveness is about the
way governments identify and deal with problems—through contextually fitting processes
that may evolve and adjust with changing contexts.
One is guaranteed to find that past models adopted in today’s more effective governments
fall short of a high-scoring WGI “effective government”—many of today’s more effective
governments did not have formal budget systems in the 1900s, for example, but were
involved in important developmental endeavors;51 but the effectiveness of the Swedish (and
other) governments in their own development process is evidenced in their improved results
since the 1900s. Could some of those governments currently labeled ineffective be similarly
effective, given their contextual challenges, even though they look really different from the
2006 WGI model? I argue so, because the 2006 model is the incorrect fit for most developing
country governments. It says nothing about how governments come to facilitate development
by “muddling through” their challenges and opportunities—the key storyline in
development—and particularly how they identify and manage problems they face.52

3.3 Can Developing Countries Replicate One-best-way Models?

Replicability is the third criterion I think any good model should possess. The simple idea
is that governments other than the effective government models should be able to adopt
proposed effective government elements and that these elements should be seen to
coalesce to foster development in the new adopters.
Concerns about model consistency and correctness discussed here potentially
undermine any discussion of conveyability. If there is no consistent or correct one-best-
way model to replicate, what is there to examine? These concerns are valid, but a brief
analysis of this issue is still warranted because of the widespread popularity of many of the
ideas and elements embodied in the good governance model of effective government.
In many instances reforms resemble less of a coherent model than a collection of ideas
about what good politics or public management looks like; assortments of popular
perspectives on “what works” drawn from the good governance literature itself or its
various supports. Without examining the replicability of an entire model, therefore, it is
worthwhile to ask whether central ideas pass-on effectively. I shall focus on a few of these
only, starting with prescriptions for limited, business-friendly government that currently
dominates much good governance dialogue about effective government.
Many commentators blamed excessive government engagement for South Korea’s 1997
financial crisis (Brilliantes, 2001). Such an approach melds well with arguments that
effective governments should not impede the private sector with limiting regulations and
“disciplinary” mechanisms.53 The argument ignores some important observations,
however, raised in alternative explanations of the crisis. These show that the Korean
government had been dismantling its state apparatus since the early 1990s. Regulations
had eased, “disciplinary” agencies such as the Economic Planning Board had been closed
and the government was pursuing a more Western model.54 The Korean government was
 The Good Governance Agenda 393

not the dominant force in 1997 that it had been in 1989 (Chang, 1998, 2000; Chang et al.,
1998; Chang & Evans, 1999; Crotty & Dymski, 1998; Sharma, 2004; Weiss, 1999).
A growing chorus thus suggests that the financial crisis was influenced by the new
weakness of the state, the lax regulatory climate and the lack of strong norms and rules in
the new Korean state (rather than the overly engaged state).
It is difficult to evaluate conclusively the validity of this argument, but it does seem that
the reform ideals pursued by Korea may have been an incorrect match to the country and
the ideas of disengaged government may not replicate well from countries such as the
USA to Korea, at least in the short run. The transition from a government-led economy to
one with more limited government was problematic and led to gaps in the social,
institutional and transactional structures.55 Some observers go beyond to claim that the
one-best-way model embodied in these new ideas ignored opinions that East Asia’s tigers
represented another model that looked very different—and was effective in its context.
Korea’s growth under a developmental state up to the 1990s creates problems for
proponents of the good governance model because of the departure from strict principles
of limited government,56 pro-business policies and limited red tape. So does Singapore,
which is cited as an “effective government” in the WGI dataset but beyond its stellar
outcomes and high-quality bureaucracy has traditionally looked very different from the
picture painted earlier. The government controls the media, regulates stringently, and
generally directs the economy in a centralized, non-participatory manner (see Huff (1995)
for a description of government at the time, when it scored 2.28 on government
effectiveness and was an effective government model).
Part of the story here seems to be the existence of an alternative model of effective
government to that presented in the good governance indicators. Another is the failure of
good governance principles to replicate effectively to countries such as Korea.
Deregulation and limiting the state did not lead to effective economic results. Perhaps this
is because of something in the social structure or the environmental context that makes
different approaches fit better into different settings. There are certainly arguments
supporting this perspective from sociology and political science. One is reminded that this
issue is also evident in the public administration literature, where Herbert Kaufman (1977)
contextualized the meaning of red tape in the saying that, “One person’s red tape may
be another’s treasured safeguard”. Perhaps we should contextualize other dimensions
of government engagement as well: one country’s unwanted regulation may be another
country’s steering and stability mechanism. Carrying ideas of deregulation from another
country to the “one” may thus create instability and undermine direction.
This thinking is perhaps useful in explaining some other reform replication problems
seen around the world. Mauritius is one of few African countries categorized as middle
income and having an effective government. It is “often presented as a poster child for how
high-quality institutions bolster development in our era” (Goldsmith, 2007, p. 177). Its
formal bureaucracy receives high marks for being merit based and efficient and for having
facilitated industrial development. However, accounts of actual bureaucratic structures
suggest a different reality. Consider the following from McCourt & Ramgutty-Wong
(2003, p. 609) on the government’s selection processes:

For initial recruitments, suitably qualified candidates sit a test of general knowledge,
so there is an ‘examination’ [in a formal sense] . . . Shortlisted candidates attend an
interview conducted by two commissioners and an adviser . . . [But] There is no
394 M. Andrews

professional selection expertise . . . there is no evidence about whether this selection

approach actually manages to identify able staff, and there appear to be no
innovations in employee selection in recent years.

Complaints of abuse in the hiring process abound, according to McCourt and Ramgutty-
Wong, who cite familial, political and ethnic influences in hiring. There is limited
transparency in the bureaucracy and its functionality is called into question. The Mauritian
Ministry of Economic Development and Regional Co- operation (1997, p. 1.11) is quoted
as complaining about the hindrance such bureaucracy actually posed: “A pervasive and
heavy-handed bureaucracy still rules and it is a miracle, no less, that in these
circumstances, industrial development did take place at all”.
Goldsmith (2007, p. 179) explains why growth did take place, suggesting that many
business contacts were relationally driven and lacked formality. The relational basis of
business may indeed explain growth as well as why a relational government structure was
accepted in the country. Such bureaucratic structures prevail in countries such as Bolivia
as well, where patronage has long been used to facilitate compromise and political
agreement (World Bank, 2000; Montes & Andrews, 2005). Efforts to rationalize and
formalize the civil service, public financial management systems and so forth have not
worked very well and may have undermined government transparency in Bolivia as civil
servants created alternative hiring processes and accounting and financial reporting
structures to avoid new formal systems.57 In these cases formal Weberian bureaucratic
systems did not replicate well, either from colonial sources (in Mauritius) or through more
recent reforms (Bolivia). Governments end up with these systems and processes running
parallel to others, something Brinkerhoff & Goldsmith (2005, pp. 199 –200) call
“institutional dualism”; situations where:

Well intended legal, regulatory and procedural changes often produce a shell of
proper governance that has little bearing on how public decisions are actually made
and implemented. Meanwhile, preexisting and deeply embedded understandings and
practices survive and continue shaping the way people are ruled.

Andrews & Hill (2003, p. 136) observed this dualism in their discussion of why
performance-based budgeting reforms (PBB) are seen to have subdued results at state level
in the USA:

Where PBB is implemented as a mere complement to a pre-existing system,

although PBB-like devices may be in place, the institutions affecting behavior are
more likely to come from the previous budgeting system . . .

The basic idea is that even something like bureaucracy does not reproduce with the same
effect from place to place,58 primarily because every context has prevailing structures that
need to be built upon or given time to evolve. Steps to introduce formal bureaucracy into
Bolivia were complicated by the fact that patronage in Bolivia facilitated political stability
and mirrored social structures. Morris Szeftel found a similar contextual rationale
for patronage in Zambia that has influenced what bureaucracy looks like in his suggestion
of a “link between spoils appropriation and class formation” that can be understood
“as part (perhaps a symptom) of a larger process of social change” (Szeftel, 1982, p. 21).
 The Good Governance Agenda 395

Similar comments center on the spoils system in 19th Century USA, which is criticized
now but played an important role in facilitating political compromise and the emergence
of a middle class in that period. The system took more than 150 years to reform, from the
1830s to the 1978 Civil Service Reform Act (and beyond). Okoli (1980) identifies long
evolutionary processes like this as crucial to the development of model bureaucracies.
He compares these evolutionary processes with reforms in the developing world that
involved quick, forced adoption of new systems that could be adopted only in a dualistic
sense. In these situations the replicated bureaucratic system can have a negative influence
of creating rule-based walls between otherwise relationally connected bureaucrats and
citizens. The type of bureaucracy reproduced in developing countries may look less like
the Weberian form and more like the patrimonial bureaucracy that characterized many of
today’s more effective governments in the past.59 Okoli (1980, p. 4) identifies the time-
evolution dimension as part of the problem, but argues further that necessary evolution
would not actually lead to the passing-on of one-best-way solutions:

Such assertions [about the need for a more evolutionary change process], however,
do not indicate that had the process of bureaucratization been gradual and
incremental, that the primordial ties and traditional values and institutions would
have been eliminated or displaced by the new institutions and values. Neither is it
implied that had the time constraints been taken into account that the new crop of
African leaders would have been completely transformed into anything other than
Africans. Rather the point to be stressed is that had the process been other than what
it was, the traditional norms, values, and institutions would have evolved a “modus
vivendi,” an accommodation with the new institutions and their values.

Replication problems are even more pronounced when looking at interactions of two or
more government effectiveness elements. This is apparent in the recent history of
Armenia’s education system. Struggling to rebound from fiscal imbalance in the wake of
the Russian crisis and influenced by neo-liberal prescriptions, the Armenian government
focused on fiscal control between 1999 and 2005 (World Bank, 2003). The government
also had to improve education and health service delivery, which was still weak following
the Soviet Union’s dissolution. The problems in the education sector were quite
fundamental (World Bank, 2003). Armenia decentralized schooling as a solution, turning
schools (and other entities) into arms-length “non-commercial enterprises as part of a
fiscal decentralization plan” in 2003 (IMF, 2003). The new schools were to be more
participatory, governed by school boards. This decentralization option was rationalized as
a replication of models in countries such as the UK, and “closely tied with international
trends of education development” (Armenia NIE, n.d.). In its wake, however, Armenia
has found the new non-commercial enterprises to be a major problem for fiscal reform.
They receive their money as transfers and do not have to report in any structured manner,
face no controls regarding allocations to important items such as building maintenance and
such, and fall outside all the rationalizing, stabilizing mechanisms (such as budget and
civil service laws). They lack transparency and are difficult to reform.
The basic story is that decentralization and participation (important elements of the
“government effectiveness” model in good governance work) actually created problems
for Armenia when introduced with steps to effect fiscal stabilization, efficiency and
transparency (also important elements of the proposed model). The situation is similar to
396 M. Andrews

that already mentioned in Sweden where efforts to reduce government size by creating
arms-length agencies may have actually undermined some of the strengths of the
bureaucracy (because employees in agencies are not subject to civil service laws and
constraints and may be more open to political influence). Studies on decentralization
and public agency creation routinely identify these tensions and inconsistencies, which
undermine the conveyability of certain packages of effective government elements: Luiz
de Mello (2000) found that coordination problems associated with decentralization create
a deficit bias that complicates fiscal discipline objectives ostensibly characterizing
effective governments; Treisman (2000) cited evidence that decentralization can actually
increase corruption and lead to decreased health and education services; and Stein (1999)
argued that decentralization can stimulate government size.
One can easily think about other problematic packages of effective government
constructs that are likely to lead to conflicting results. Civil servants are meant to be
politically neutral but also to lead the implementation of pro-business policies; policy must
be stable but also highly responsive to changing demands in the citizenry; government must
be formalized and rule-bound but also highly responsive. A common approach to criticizing
new public management involves identifying exactly these kinds of potentially conflicting
elements. Competition, privatization, decentralization, innovation and empowerment are
examples, and also feature in the good governance model as elements of effective
government. All concepts are, however, seen to clash with each other in practice, conveying
poorly in most realities. Williams (2000, p. 523) wrote, for example, that:

They [reform designers] want “competition,” but oppose “duplication.” As a result,

the reinventor is given no useful direction for action. Should she allow government
units and programs to compete with each other in order to provide the best services?
Or should governmental units be rationalized to eliminate duplication? How much
duplication is too much? What considerations should lead to streamlining and which
to competition?

Williams argued that many new public management reform prescriptions resemble the
principles of administration that Herbert Simon decried as problematic proverbs 60 years
ago—quotable and convenient constructs for rationalizing past behavior or justifying
future decisions but defective in providing any explanations or practical advice. Simon
(1947, p. 53) argued that, as with all proverbs, the principles of administration of his
day stood well when applied alone but poorly when considered in tandem with others:
“For every principle one can find an equally plausible and acceptable contradictory
principle”. I believe the observation holds for recent effective government characteristics
as well. Fiscally disciplined governments often provide a strong foundation for economic
growth, for example. Decentralized governments can have better targeted and efficiency-
enhancing service provision structures that also facilitate growth; but governments such as
Armenia (and many others) attempting decentralization find sub-national entities to be a
big threat to fiscal discipline!
Observing such tensions in the principles he analyzed, Simon (1947, p. 53) commented:
“Although the two principles of the pair will lead to exactly opposite organizational
recommendations, there is nothing in the theory to indicate which one to apply”. The
comment is pertinent in development, especially, related to fiscal management and
decentralization reforms and to other effective government characteristics. Having no real
 The Good Governance Agenda 397

basis for identifying exactly what part of the governance proverb-set a government should
attend to leads many countries to apply all of the proverbs at once. This is potentially part
of the reason why many of the development community’s governance solutions end up
becoming problems; most of the ideas they apply do not convey well because of natural
tensions between the ideas themselves. It also potentially explains why studies linking
government effectiveness to growth, poverty reduction and such matters are so frustrating
at present; internally contradictory measures simply fail to reflect an accurate picture of
effective government.

4. Conclusion: Recognizing and Understanding Multiple States of Development

The development community’s current picture of effective government provides a faulty
construct for thinking about what effective government is, especially in the development
context. The construct is based on governance indicators that are useful in provoking
dialogue about key issues in development and may never have been intended to shape a
more theoretical model.60 Applications of the indicators in the development community
go further than dialogue, however, and suggest a model of government effectiveness. The
model is influential, isomorphically determining what the development community sees as
effective government. Meyer & Rowan (1977, p. 340) warn of isomorphism’s dangers:
“Organizations whose structures become isomorphic with the myths of the institutional
environment—in contrast with those primarily structured by the demands of technical
production and exchange—decrease internal coordination and control in order to maintain
legitimacy”.
I believe that isomorphism, arising from the way good governance indicators are used,
is especially dangerous, because of problems with the picture of government this work
portrays. It lacks consistency, correctness and replicability. “Model” governments scoring
well on the indicators are similar only in terms of their commonly strong outcomes
and developed status; the government structures influencing such success vary quite
significantly and we have no understanding of why and no way to identify where variation
is permissible and where it is not.61 Furthermore, governments considered effective today
did not look as they do now when they progressed beyond weak outcomes and facilitated
development. The pre-1990s Swedish model actually referenced in the literature does not
have many features of the good governance picture (such as smallness) but is rather a
bigger government, welfare state model that many saw as successful for most of the 20th
Century (Rojas, 2005).62 Finally, elements of the good governance model do not replicate
well between governments, often creating governance problems rather than solutions. It
would be better to have ideas about the factors that influence what works, when and how,
instead of just a list of what should work.
The main reason for these weaknesses is the lack of any theoretical framework on which
the indicators are based. Such a framework would be characterized by clear thinking about
what effectiveness might mean and propositions that explain observed effectiveness in
different settings and across time. The discussion I provide on the theoretical framework
of the indicators is limited because of the limited references to theory in the good
governance literature. Where theoretical references are made in the good governance
literature they are not consistent, referencing modernist Weberian models at some
junctures, post-modern new public management at others. The only consistent citation is
to North’s new institutionalism and the argument that institutions matter and that different
398 M. Andrews

institutional structures in governments will yield different results. This is simply not
enough to call a theory of effective government, especially from an organizational
perspective.63
An effective government framework must have the kind of detail Simon called for in
1947 and be able to explain effectiveness. As Mintzberg (1979, p. 587) stated:
“Explanation is, of course, the purpose of research”. Some might suggest that such a
framework is not needed to explain the basic issues at hand—how governments identify
what they need to do at any given time, and manage to do it well. However, these questions
are the basis of a long strand in the literature on organizations. This literature shows that
effectiveness criteria change over time; that organizations choose different criteria at
different times—akin to different governments pursuing different policy outcomes at
different times; that many organizations manage to identify these criteria effectively and
others do not; and that systems need to match challenges for effectiveness, but often do
not. Understanding how challenges are identified and addressed is more complex than
saying that this is the basis of effectiveness; understanding and explaining requires a
framework that currently does not exist in the development domain.
There is already much work that could form the foundation of research targeted at these
questions. Musgravian economics could suggest ideas about what more effective
governments might do (providing basic goods that markets cannot provide) and work from
strands as varied as public finance and sociology could provide ideas about why different
governments might identify these challenges differently, for example because of different
political decision-making structures and social logics (Hallerberg et al., 2007; Woolsey-
Biggart & Guillen, 1999). Studies in the public administration tradition and contingency
theorists in organizational literature suggest ways of thinking about the different
processes, institutions and approaches different governments might take to address the
problems they face. The challenge is to draw these ideas together into stories of what
effectiveness might be (in terms of both outcomes and processes) and to see how these
stories explain some of the questions raised here: How do the different contexts of the so-
called effective governments explain why these look different in key areas, and how
should these differences contextualize our perspectives on effectiveness? How should
stage of development influence our picture of government effectiveness? How does
context determine which aspects of the effectiveness the model might convey?
One could imagine this work progressing towards prospective effectiveness measures.
Building on the idea that effective governments do the right things well, these could
center, for example, on the provision of key services in the development domain.
Government effectiveness could be judged according to how well the key outcomes are
provided, compared with others and controlling for contextual factors such as income.64
Using a sports metaphor, effective governments would be those that “box above their
contextual weight” in providing these services (performing better than governments facing
similar contextual challenges such as income levels). The Appendix provides a brief
discussion of what these contextualized indicators might show for a particular outcome
mentioned in this paper—infant mortality rates.65
This approach focuses on outcomes as the basis for assessing effectiveness but proposes
that current governments are not responsible for outcomes in an absolute sense, only
performance at the margin;66 but varying models of “how” governments choose and
address their challenges should be accepted in the approach, which implicitly emphasizes
the importance of political and managerial decisions to effectiveness. The approach does
 The Good Governance Agenda 399

not promote adopting one-best-way models of what governments should do and how, but
rather that governments should develop mechanisms to choose what to do and to choose
the types of structure and process appropriate to ensuring effectiveness in implementation.
Governments “boxing” at or above their “income” weight are assumed to enjoy this
characteristic, whereas those boxing below their weights are assumed not to. They either
have not chosen to address this issue or their choices for implementation are not
facilitating performance.
The lack of a theoretical framework underlying current indicators, discussed in this article,
is partly the result of a general rush to measure effectiveness without a theory of
effectiveness. This has left indicators prompting more questions about effective government
than they answer, and a new approach as suggested above would also raise important
questions. It would be better then to avoid developing new indicators immediately, while
building an understanding of theory to underlie them. If we identify certain services as key
and certain contextual factors as influential, we should be able to understand why they are so,
and we should be able to explain how effective governments adjust their challenge focus and
choose the processes to meet challenges. The focus should be on why different states match
different development contexts. It is unlikely that such a framework will be neat and tidy like
the current good governance literature appears to be.67 Researching organizations and
organizational effectiveness is not popular because it requires getting “involved in actual
organizations” and calls researchers “to muck about with messy data and relationships”
(Kaplan, 1984, p. 415). It will also probably be messy from a moral perspective, raising
questions about the balance between the influence of normative ideals and positive evidence
of what we observe effectiveness to be. Messy, morally demanding work, focused on theory
development is, however, vital if our goals are more focused academic work, accurate donor
engagement and contextual reform design in developing countries.

Notes
1
The good governance indicators come in many shapes and sizes and focus on areas beyond government
effectiveness. My main point of reference is the government effectiveness dimension of the World
Governance Indicators (which is one of six dimensions—the others being Voice and Accountability,
Political Stability and Absence of Violence, Regulatory Quality, Rule of Law and Control of
Corruption).
2
The idea here derives jointly from Pritchett & Woolcock (2004) and Rodrik (2006). The former
identify Denmark as a model, whereas the latter speaks of Sweden or the USA in this way.
3
Again, Dani Rodrik’s comment on what is obviously “hardly useful policy advice!” (Rodrik, 2006,
p. 13).
4
Brinkerhoff & Goldsmith (2005).
5
Lant Pritchett argues that flailing states emerge when governments introduce “better practice” systems
on top of pre-existing systems and end up with a Frankenstein model, where (metaphorically) the head
does not talk to the limbs and the entire structure is dysfunctional.
6
The reference here is to Chang (2003, p. 29), who questions whether open trade is the best path for
developing countries to follow as they build their economies, given that currently developed countries
followed a different path. I cite this argument not because of the application to trade, but rather because
Chang raises an important principle regarding the importance of learning from development paths.
7
The influence carries to other governance concepts as well.
8
DiMaggio & Powell (1983) suggest these and other factors as predictors for isomorphic influence,
especially in its coercive and mimetic forms.
9
Kurtz & Schrank (2007, p. 542) ask a similar question in their analysis of the World Governance
Indicators (WGI): “Are We Sure We Know How Good a Government Is?”
400 M. Andrews
10
This discussion is quite generalized. Hood et al. (2007) point out that there are many different ranking
approaches and that these are associated with a range of different theoretical frameworks. Interestingly, it
appears that the more specialized the indicator the tighter the connection to an explicit theoretical model.
11
Kaufmann, Kraay and Mastruzzi emphasize in their 2007 response to Kurtz and Schrank that North’s
approach relates best to rule of law and control of corruption measures. They do not provide an explicit
rationale for their government effectiveness measure in any of their papers.
12
Especially in the 1999 article describing the indicators.
13
Particularly the WGI indicators.
14
These characteristics are my own constructs. The actual “government effectiveness” index structure
varies for different countries and years due to differing data access. The statistical questions about this
approach are subject to significant debate in other domains. I do not intend to enter such debate here.
15
The demands of such a model led Merilee Grindle (2004) to write her paper, “Good enough
governance”, where she writes: “The good governance agenda, largely defined by the international
development community but often fervently embraced by domestic reformers, is unrealistically long
and growing longer over time” (p. 526).
16
This discussion is based on 2006 scores, which do not differ significantly from other recent years.
17
These are the largest 20 SSA countries, measured by population size. This selection omits some
African governments fitting into the effectiveness zone—Botswana, Cape Verde and Mauritius. These
are notable for their small size.
18
In total there were 17 from LA, 19 from EA and 10 from MENA. Five effective governments came
from LA, six from EA and five from MENA.
19
Hungary and Poland are classified as OECD in this figure.
20
The six that scored above two and make up the top 5%.
21
These scored above 1.5 and rounded out the top 10% in the sample.
22
Some critics suggest that the indicators do not measure government effectiveness or other good
governance elements but rather some other underlying variable, such as economic growth. I would
suggest that this thinking is correct and that the indicators actually capture development broadly—
those performing well are (generally) more developed than others.
23
See Rodrik (2006, p. 13).
24
Ghoshal (2005, p. 81) identifies Enron’s “best practice” structures as an example; the structures
reflected theoretical models and what was considered best practice methods of ensuring agents
maximize principal’s interests, including having stock options for senior executives and independent
directorships. These are now considered key contributors to the problems at Enron.
25
I adapt Mayer’s approach to the use of models in the educational context, which is surprisingly relevant
to development. He identifies the characteristics that make a model useful for assisting people to
understand things and then to act upon such understanding. In development we try to understand why
some countries grow, reduce poverty, and generally advance, while others do not; our understanding
centers on potential future action.
26
Most declined during the Great Depression period of the 1930s, for example.
27
Tanzi & Schuknecht (2000, p. 78) show how the nine OECD countries grew at a fairly brisk pace
between 1870 and 1919, accelerated between 1920 and 1929, slowed or declined in the 1930s,
accelerated significantly following the war and in the 1960s and slowed through the 1980s to 1994.
My own calculations from the World Development Indicators show average growth rates between
1 and 2.5% in the governments for the period 2000–006. Cumulative growth for the period suggests
difference from other country groupings.
28
The literature shows that many outcomes (e.g. infant mortality) are explained by factors such as per
capita GDP. Specific governments can ostensibly have a marginal influence on these figures but are
still not the main influence.
29
Above 5% of GDP with regard to education and above 8% for health.
30
The USA spends more out of GDP on health than the other countries. In 2000 it spent 13% of GDP on
health as against the 9% average of the other countries, and in 2004 the comparative figures were 15
and 10%. Government, however, accounts for less than half of this total (6% of GDP in 2000 and 7% of
GDP in 2004). All figures from OECD (2007).
31
The Swedish government accounted for 7.7 of the 9.1% of GDP allocated to health in 2004 and the
Danish government accounted for 7.8 of 8.6% of GDP allocated to health care in 2004. All figures from
OECD (2007). See Busemeyer (2007) for details on OECD differences in education spending.
 The Good Governance Agenda 401
32
See also Rexed et al. (2007) for discussion on how the details of decentralization differ.
33
Thieben (2003) discusses how some of the governments are decentralizing more and others are actually
centralizing. The paper suggests that they may be converging on an optimal size and level of
decentralization, which would be interesting as a characteristic of effectiveness but is still only a
hypothesis. For further discussion see Stegarescu (2004, p. 11).
34
The WGI “Regulatory Burden” element has as one of its core sources scores on the Heritage
Foundation/Wall Street Journal Index for government intervention in the economy, which is measured
in terms of the following: government consumption as a percentage of the economy; government
ownership of businesses and industries; share of government revenues from state-owned enterprises
and government ownership of property; and economic output produced by the government.
35
The entire group of governments was in fiscal trouble in the early to mid-1990s, the tail of a fiscal
expansion period that led to some significant adjustments in the past 15 years.
36
The World Development Indicators are produced by the World Bank and available at
http://publications.worldbank.org/WDI/
37
Other research making similar observations include Matheson et al. (2007) and Gualmini (2007), who
found that although many governments adopt similar kinds of processes, they do so differently.
38
As already suggested, this seems to be the underlying construct actually measured by the indicator.
39
This conclusion would not surprise Thomas (2006, p. 10), who questions the theoretical constructs in
the WGI indicators: “An examination of the underlying indicators suggests that the WGI construct
definitions are merely summary descriptions of the indicators in the cluster. The authors explain the
clustering as an expression of personal ideas of governance, but these ideas are not articulated . . . If
this is the case, then the underlying construct has not been defined at all and there is no way to check the
validity of these constructs or to explain why we would care about them. No meaningful predictions
can be made regarding the relationship between undefined constructs and observable variables”.
40
This is after adjusting for the presence of oil-rich and other outliers such as Azerbaijan and China.
Including them in the calculation results in an average annual growth rate exceeding 6% for the
sample.
41
A similar observation holds for domestic indirect or sales taxes.
42
Dependency on international trade taxes is actually regularly used to assess governance quality. I am
not, however, ascribing any such value to countries with high international tax dependency here. My
intention is merely to note that such dependency is expected to yield different administrative realities.
43
Governments can only exert taxes on formal transactions that are recorded in institutionalized forms
(e.g. through accounting systems). Many domestic transactions are not subject to such formality in
early stages of development, as societies transact within relational structures. International trade tends
to be formalized earlier because of a lack of inter-personal connections between those engaged in the
transaction. The formalities of trading engagements allow for taxation at an early stage, when perhaps
domestic transactions—wage payments, commercial sales, etc.—are still not open to easy taxation.
44
Similar differences are evident with regard to key outcomes such as life expectancy (health),
educational access (education) and tarred road coverage (transportation infrastructure).
45
Jones & Kettl (2003, p. 10) ask this question with regard to public financial management, noting
“convergence of reforms around general themes” even though “Developing nations have different
management reform problems than developed nations”.
46
All data referenced from Tanzi & Schuknecht (2000) or OECD (2007) or World Development
Indicators (accessed December 2007).
47
One should note that revenues were low in many developed countries in the early 1900s, such that
international revenues making up 3% of GDP accounted for half of total federal revenues. The numbers
I cite here for current developing countries are for trade revenues as a percentage of GDP—not
including donor funding, which accounts for significantly higher numbers in many cases.
48
The average OECD ratio of direct domestic taxes to GDP increased from 3.4% in 1937 to 14% in the
mid-1990s; similar change saw indirect domestic taxes increase from an OECD average of 4.9% of
GDP in 1937 to 13.5% of GDP in the mid-1990s (Tanzi & Schuknecht, 2000, pp. 56–57).
49
I am indebted to an anonymous reviewer for making this observation.
50
Thanks to David Leonard for noting this. One could note many differences between infant mortality
problems in developed and developing countries as well, including the current role HIV/AIDS plays in
weakening health statistics (in countries such as South Africa especially). But there are also many
common challenges: How did Belgium ensure clean drinking water in the 1940s in a way that most
402 M. Andrews

developing countries cannot today? How did Canada manage to set up geographically spread health
outposts financed and resourced by the government 50 years ago?
51
It would be interesting to calculate these scores for the models in the 1920s, 1950s, 1960s and beyond.
This is impossible because the scores are based in large part on collected opinions of interviewees.
Even fact-based elements are not open to assessment, as the questions on which the indicators are
based are not publicly known in most cases, limiting any ability to calculate scores for the past.
52
I borrow from Lindblom, believing government effectiveness of today is largely a function of the
muddling through of the past. The question is why do we not emphasize this muddling through and
learning from such muddling through?
53
Amsden (2001) described government policies in East Asian tigers as “disciplinary” of business.
54
The isomorphic influences of pro-market Western thought are well documented in various accounts of
Korea’s reforms. Echoing DiMaggio & Powell’s (1983) reasons for isomorphism Chang (1998,
p. 1559) cites the rise of this ideology globally and Korea’s “relatively weak intellectual tradition”; the
two together made it difficult to “resist . . . shifts in the world intellectual climate.”
55
Chang (1998, p. 1559) argues that the 1990s reforms reframed government –business relationships.
He notes that the developmental state was characterized by “well publicized ‘rational’ criteria for
intervention” and a “‘generalistic’ state –business relationship”. Reforms in the 1990s eroded the
known rules and led to what Chang calls “particularistic” or “cronyistic” relationships.
56
See Amsden (2001). Also, authors such as Huff (1995) reference the simultaneous development of
institutional capacities, improvement of living standards and economic stability with growing
government control.
57
Creating public agencies, for example, and maintaining multiple sets of books.
58
Olsen (2005, p. 5) notes that “Bureaucratic organization could produce multiple and contradictory
outcomes”, citing different types of entity affected, different cultural contexts, and so on.
59
Chabal (2002) provides a nice discussion of this in the context of African governance, pointing out that
neo-patrimonialism has dominated most African countries since independence. He emphasizes the
importance of thinking about bureaucratic reform in the context of traditional political and social
power structures.
60
In earlier comments on this paper, a reviewer noted that complex indicators that amalgamate multiple
characteristics are unavoidably messy and potentially contradictory, but could still carry weight in
general discussion. I believe he is correct in this regard. The critique here is not of the indicators
themselves, or of a general use of these indicators, but rather of the way many development
practitioners and theoreticians reference the indicators in specifying reform agendas and measuring
specific concepts such as government effectiveness.
61
This variation is a good thing, suggesting there are multiple forms of effectiveness—but it also creates
problems for a dominant one-best-way prescription so often implicit in development work.
62
Rather than prescribing today’s model, it would be really valuable for development experts to
understand the developmental evolution in a country like Sweden.
63
One could defend the indicators as never actually pretending to have a theoretical foundation—which
calls into question the way academics and practitioners use the variables, imputing a theoretical
significance to them.
64
Two ideas lie behind this comment. The first centers on the idea that effectiveness is about
governments doing the right things, which we may be able to boil down to key services that improve
well-being. The second is that effectiveness in providing these services should be comparative, but
contextually controlled. With regard to health care, for example, the argument is that while health
statistics are positively correlated with income, some governments manage their systems so well that
they facilitate statistics better than their income levels suggest.
65
Robert Lawrence (2006) adopts a similar approach to looking at governance in trade regimes in the
Middle East. He adjusts various governance and business environment measures for income levels,
showing how this changes the picture quite significantly.
66
Full variation in outcomes will reflect both historical government effectiveness and other contextual
influences.
67
As stated, this is because the current literature avoids theory; even current debates are almost
exclusively about data manipulation (Kaufmann et al., 1999; Kurtz & Schrank, 2007).
 The Good Governance Agenda 403

References
Amsden, A. (2001) The Rise of “the Rest”: Challenges to the West from Late-industrializing Economies (Oxford:
Oxford University Press).
Andrews, M. & Hill, H. (2003) The impact of traditional budgeting systems on the effectiveness of performance-
based budgeting: a different viewpoint on recent findings, International Journal of Public Administration,
26, pp. 135– 155.
Armenia National Institute of Education (NIE) (n.d.) Education reforms in Armenia. at: http://www.aniedu.am/
ENG/indexeng.html.
Arndt, C. & Oman, C. (2006) Uses and Abuses of Governance Indicators, OECD Development Center Studies
(Paris: OECD).
Boivard, T. & Löffler, E. (2003) Evaluating the quality of public governance: indicators, models and
methodologies, International Review of Administrative Sciences, 69, pp. 313–328.
Brilliantes, A. B. (2001) Public sector reform and poverty reduction. unpublished Research Paper for Asian
Development Bank.
Brinkerhoff, D. & Goldsmith, A. (2005) Institutional dualism and international development: a revisionist
interpretation of good governance, Administration & Society, 37, pp. 199–224.
Busemeyer, M. R. (2007) Determinants of public education spending in 21 OECD democracies, 1980–2001,
Journal of European Public Policy, 14, pp. 582 –610.
Chabal, P. (2002) The quest for good government and development in Africa: Is NEPAD the answer?
International Affairs, 78, pp. 447–462.
Chang, H. J. (1998) Korea: the misunderstood crisis, World Development, 26, pp. 1555–1561.
Chang, H. J. (2000) The hazard of moral hazard: untangling the Asian crisis, World Development, 28,
pp. 775–788.
Chang, H.-J. (2003) Kicking away the ladder: infant industry promotion in historical perspective 1, Oxford
Development Studies, 31, pp. 21– 32.
Chang, H. & Evans, P. (1999) The Role of Institutions in Economic Change. Mimeo, Cambridge University.
Chang, H. J., Park, H. J. & Yoo, C. G. (1998) Interpreting the Korean crisis: financial liberalization, industrial
policy and corporate governance, Cambridge Journal of Economics, 22, pp. 735 –746.
Crotty, J. & Dymski, G. (1998) Can the neoliberal economic regime survive victory in East Asia? The political
economy of the Asian crisis, International Papers in Political Economy, 5, p. 2.
Curristine, T. (2005) Performance information in the budget process: results of the OECD 2005 questionnaire,
OECD Journal of Budgeting, 5, pp. 87– 131.
de Mello, L. (2000) Fiscal decentralization and intergovernmental fiscal relations: a cross-country analysis, World
Development, 28, pp. 365– 380.
DiMaggio, P. & Powell, W. (1983) The iron cage revisited: institutional isomorphism and collective rationality in
organizational fields, American Sociological Review, 48, pp. 147– 160.
Ghoshal, S. (2005) Bad management theories are destroying good management practice, Academy of
Management Learning and Education, 4, pp. 75 –91.
Goldsmith, A. (2007) Is governance reform a catalyst for development, Governance: An International Journal of
Policy, Administration, and Institutions, 20, pp. 165–186.
Grindle, M. (2004) Good enough governance: poverty reduction and reform in developing countries,
Governance: An International Journal of Policy, Administration and Institutions, 17, pp. 525–548.
Gualmini, E. (2008) Restructuring Weberian bureaucracy: comparing managerial reforms in Western Europe and
the United States, Public Administration, 86 (1), pp. 75–94.
Hallerberg, M., Strauch, R. & von Hagen, J. (2007) The design of fiscal rules and forms of governance in
European Union countries, European Journal of Political Economy, 23, pp. 338–359.
Handler, H., Koebel, B., Reiss, P. & Schratzenstaller, M. (2005) The Size and Performance of Public Sector
Activities in Europe, WIFO Working Papers 246, WIFO.
Hauptmeier, S., Heipterz, M. & Schuknecht, L. (2007) Expenditure reform in industrialised countries: a case-
study approach, Fiscal Studies, 28, pp. 293–342.
Henrekson, M. (2005) Entrepreneurhsip: A weak link in the welfare state? Industrial and Corporate Change, 14,
pp. 437–467.
Hood, C., Beeston, C. & Dixon, R. (2007) Rating the rankings: assessing international rankings of public service
performance. International Public Management Journal, 11 (3), pp. 298 –328.
404 M. Andrews

Huff, W. G. (1995) The developmental state, government, and Singapore’s economic development since 1960,
World Development, 23, pp. 1421–1438.
IMF (2003) Republic of Armenia: Letter of Intent, Memorandum of Economic and Financial Policies, and
Technical Memorandum of Understanding (Washington, DC: IMF).
Jones, L. R. & Kettl, D. F. (2003) Assessing public financial management reform in and international contex,
International Public Management Review, 4, p. 1.
Joumard, I., Kongsrud, P., Nam, Y. & Price, R. (2004) Enhancing the Cost Effectiveness of Public Spending:
Evidence in OECD Countries, OECD Economic Studies, No. 37, 2003/2.
Kaplan, R. S. (1984) The evolution of management accounting, The Accounting Review, 59, pp. 390– 418.
Kaufmann, D., Kraay, A. & Mastruzzi, M. (2007) The Worldwide Governance Indicators Project: Answering the
Critics, World Bank Policy Research Working Paper 4149.
Kaufmann, D., Kraay, A. & Zoido-Lobaton, P. (1999) Governance Matters, World Bank Policy Research
Working Paper 2196.
Kaufman, H. (1977) Red Tape: Its Origins, Uses, and Abuses (Washington, DC: Brookings Institution).
Knack, S. & Keefer, P. (1995) Institutions and economic performance: cross-country tests using alternative
institutional measures, Economics and Politics, 7, pp. 207–228.
Knack, S., Kugler, M. & Manning, N. (2003) Second-generation governance indicators, International Review of
Administrative Sciences, 69, pp. 354–364.
Kurtz, M. & Schrank, A. (2007) Growth and governance: models, measures and mechanisms, The Journal of
Politics, 69, pp. 538–554.
La Porta, R., Lopez-de-Silanes, F., Shleifer, A. & Vishny, R. (1999) The quality of government, Journal of Law,
Economics and Organizations, 15, pp. 222–279.
Lawrence, R. (2006) A US–Middle East Trade Agreement: A Circle of Opportunity? Peterson Institute Policy
Analysis in International Economics No. 81 (Washington, DC: Peterson Institute for International
Economics).
Malyshev, N. (2006) Regulatory policy: OECD experience and evidence, Oxford Review of Economic Policy, 22,
pp. 274 –299.
Matheson, A., Webber, B., Manning, N. & Arnould, E. (2007) Study on the Political Involvement in Senior
Staffing and on the Delineation of Responsibilities Between Ministers and Senior Civil Servants, OECD
Working Papers on Public Governance, 2007/6 (OECD Publishing).
Mauritian Ministry of Economic Development and Economic Co-operation (1997) Vision 2020: The National
Long-term Perspective Study (Port Louis: Ministry of Economic Development and Regional Co-operation).
Mayer, R. E. (1989) Models for understanding, Review of Educational Research, 59, pp. 43–64.
McCourt, W. & Ramgutty-Wong, A. (2003) Limits to strategic HRM: the case of the Mauritian civil service,
International Journal of Human Resource Management, 14, pp. 600 –618.
Meyer, J. & Rowan, B. (1977) Institutionalized organizations: formal structure as myth and ceremony, American
Journal of Sociology, 83, pp. 340 –363.
Meyer, J. & Rowan, B. (1991) Institutionalized organizations: formal structure as myth and ceremony,
in: P. DiMaggio & W. Powell (Eds) The New Institutionalism in Organizational Analysis, pp. 340 –363
(Chicago: University of Chicago Press).
Mintzberg, H. (1979) An emerging strategy of “direct” research, Administrative Science Quarterly, 24,
pp. 582 –589.
Montes, C. & Andrews, M. (2005) Implementing reforms in Bolivia: Too much to handle? International Journal
of Public Administration, 28, pp. 273 –290.

1
The approach here is meant for demonstration. A better approach could be to conduct regression analyses of
contextual influences on the outcome in question, and use residuals to show the marginal influence of
government on these outcomes. The calculation used here, however, simply shows the difference between the
percentile position of a given country for infant mortality and the percentile position for real PPP income per
capita. Data come from the World Bank World Development Indicators for 2005. The USA, with an infant
mortality rate of six per 1000 in 2005, sat at the 83rd percentile in a sample of 165 countries. The USA sat at the
99th percentile for income per capita. It thus scores (16 on the proposed effectiveness measure. Sweden is at the
99th percentile for infant mortality performance (three per 1000) and is at the 91st percentile for income per
capita. Sweden thus scores þ 8 on the proposed measure.
 The Good Governance Agenda 405

Moore, M. (2004) Revenues, state formation, and the quality of governance in developing countries, International
Political Science Review, 25, pp. 297– 319.
Mosca, I. (2007) Decentralization as a determinant of health care expenditure: empirical analysis for OECD
Countries, Applied Economics Letters, 14, pp. 511– 515.
OECD (1999) Structure of the Civil Service Employment in Seven OECD Countries (Paris: OECD).
OECD (2005) SME and Entrepreneurship Outlook (Paris: OECD).
OECD (2007) OECD in Figures (Paris: OECD).
Okoli, F. (1980) The dilemma of premature bureaucratization in the new states of Africa: the case of Nigeria,
African Studies Review, 23, pp. 1–16.
Olsen, J. P. (2005) Maybe it is time to rediscover bureaucracy, Journal of Public Administration, Research and
Theory, 16, pp. 1 –24.
Overman, E. S. & Boyd, K. J. (1994) Best practice research and postbureaucratic reform, Journal of Public
Administration Research and Theory, 4, pp. 67–83.
Pritchett, L. & Woolcock, M. (2004) Solutions when the solution is the problem: arraying the disarray in
development, World Development, 32, pp. 191– 212.
Rauch, J. & Evans, P. (2000) Bureaucratic structure and bureaucratic performance in less developed countries,
Journal of Public Economics, 75, pp. 9–71.
Rexed, K., Moll, C., Manning, N. & Allain, J. (2007) Governance of Decentralised Pay Setting in Selected OECD
Countries, OECD Working Papers on Public Governance, 2007/3 (OECD Publishing).
Rodrik, D. (2006) Goodbye Washington Consensus, Hello Washington Confusion?, Kennedy School of
Government Working Paper.
Rojas, M. (2005) Sweden after the Swedish Model: From Tutorial State to Enabling State (Stockholm: Timbro
Publishers).
Sharma, S. (2004) Government intervention or market liberalization: the Korean financial crisis as a case of
market failure, Progress in Development Studies, 4, pp. 47–57.
Simon, H. (1947) The proverbs of administration, Public Administration Review, 6, pp. 53 –67.
Stegarescu, D. (2004) Economic Integration and Fiscal Decentralization: Evidence from OECD Countries, ZEW
Discussion Paper No. 04-86.
Stein, E. (1999) Fiscal decentralization and government size in Latin America, Journal of Applied Economics, 11,
pp. 357–391.
Szeftel, M. (1982) Political graft and the spoils system in Zambia—the state as a resource in itself, Review of
African Political Economy, 9, pp. 4–21.
Tanzi, V. & Schuknecht, L. (2000) Public Spending in the 20th Century (Cambridge: Cambridge University
Press).
Thieben, U. (2003) Fiscal decentralisation and economic growth in high-income OECD countries, Fiscal Studies,
24, pp. 237– 274.
Thomas, M. A. (2007) What Do the Worldwide Governance Indicators Measure? John Hopkins University– Paul
H. Nitze School of Advanced International Studies (SAIS) Working Paper. Available at SSRN:
http://ssrn.com/abstract=1007527.
Treisman, D. (2000) Decentralization and the quality of government, unpublished working paper, Department of
Political Science, UCLA.
van de Walle, S. (2006) The state of the world’s bureaucracies, Journal of Comparative Policy Analysis:
Research and Practice, 8, pp. 437– 448.
Weiss, L. (1999) State power and the Asian crisis, New Political Economy, 4, pp. 317–342.
Williams, D. W. (2000) Reinventing the proverbs of government, Public Administratin Review, 60, pp. 522–534.
Woolsey Biggart, N. & Guillen, M. F. (1999) Developing difference: social organization and the rise of the auto
industries of South Korea, Taiwan, Spain and Argentina, American Sociological Review, 64, pp. 722–747.
World Bank (2000) Bolivia: From Patronage to Professional State. Bolivia Institutional and Governance Review,
Vol. 2 (Washington, DC: World Bank).
World Bank (2003) Armenia Public Expenditure Review (Washington, DC: World Bank).
406 M. Andrews

Appendix
Figure A1 presents a rudimentary picture of how infant mortality statistics might be shown
“contextually” to reveal countries boxing above, at or below their income weight. The
vertical axis shows an effectiveness measure based on the difference between the
percentile position a country recorded for infant mortality rates and per capita income
levels in 2005 (the approach only controls for income, but should address other contextual
variables as well).1

Figure A1. Contextually adjusted effectiveness in infant mortality outcomes.

Governments in countries scoring around zero seem to be “boxing” at their income

weight (ensuring services are provided in line with the comparative income position of the
country). Governments scoring above zero are effectively boxing above their income
weight (ensuring performance above that expected). Governments scoring below zero are
boxing below their income weight (actually failing to ensure service provision at a
contextually appropriate level). The figure is organized in the same way as Figure 1,
allowing one to see the most basic differences: (1) context, and particularly development
context (reflected here in income measures) influence our picture of effectiveness and
must be controlled for; and (2) governments can perform above their income context even
in underdeveloped settings, and below their context even in developed settings. Various
 The Good Governance Agenda 407

developing country governments show up as effective in the provision of health services

(seemingly performing better than their income context would indicate), e.g. Moldova,
Tanzania and Sri Lanka. Some middle income and developed country governments
perform below contextual expectation, including South Africa and the USA.