P1: JzG/JzK

0521836700agg.xml CY409B/Kohli 0 521 83670 0 June 8, 2004 16:10

STATE-DIRECTED DEVELOPMENT
Political Power and Industrialization
in the Global Periphery

ATUL KOHLI
Princeton University

iii
P1: JzG/JzK
0521836700agg.xml CY409B/Kohli 0 521 83670 0 June 8, 2004 16:10

published by the press syndicate of the university of cambridge

The Pitt Building, Trumpington Street, Cambridge, United Kingdom

cambridge university press

The Edinburgh Building, Cambridge cb2 2ru, uk
40 West 20th Street, New York, ny 10011-4211, usa
477 Williamstown Road, Port Melbourne, vic 3207, Australia
Ruiz de Alarcón 13, 28014 Madrid, Spain
Dock House, The Waterfront, Cape Town 8001, South Africa
http://www.cambridge.org


C Atul Kohli 2004

This book is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press.

First published 2004

Printed in the United States of America

Typeface Sabon 10/12 pt. System LATEX 2ε [tb]

A catalog record for this book is available from the British Library.

Library of Congress Cataloging in Publication Data

Kohli, Atul.
State-directed development : political power and industrialization in the
global periphery / Atul Kohli.
p. cm.
Includes bibliographical references and index.
isbn 0-521-83670-0 (hb) – isbn 0-521-54525-0 (pb)
1. Industrial policy – Developing countries. 2. Industrialization – Developing
countries. 3. Developing countries – Economic policy. 4. Developing
countries – Politics and government. I. Title.
hd3616.d452.k64 2004
338.9 009172 4–dc22 2003069755

isbn 0 521 83670 0 hardback

isbn 0 521 54525 0 paperback

iv
P1: JzG/JzK
0521836700agg.xml CY409B/Kohli 0 521 83670 0 June 8, 2004 16:10

Contents

List of Tables and Figures page ix

Acknowledgments xi

Introduction: States and Industrialization in the Global

Periphery 1

PART I GALLOPING AHEAD: KOREA

1 The Colonial Origins of a Modern Political Economy:
The Japanese Lineage of Korea’s Cohesive-Capitalist
State 27
2 The Rhee Interregnum: Saving South Korea for Cohesive
Capitalism 62
3 A Cohesive-Capitalist State Reimposed: Park Chung Hee
and Rapid Industrialization 84

PART II TWO STEPS FORWARD, ONE STEP

BACK: BRAZIL
4 Invited Dependency: Fragmented State and Foreign
Resources in Brazil’s Early Industrialization 127
5 Grow Now, Pay Later: State and Indebted Industrialization
in Modern Brazil 169

PART III SLOW BUT STEADY: INDIA

6 Origins of a Fragmented-Multiclass State and a Sluggish
Economy: Colonial India 221
7 India’s Fragmented-Multiclass State and Protected
Industrialization 257

vii
P1: JzG/JzK
0521836700agg.xml CY409B/Kohli 0 521 83670 0 June 8, 2004 16:10

viii Contents

PART IV DASHED EXPECTATIONS: NIGERIA

8 Colonial Nigeria: Origins of a Neopatrimonial State and
a Commodity-Exporting Economy 291
9 Sovereign Nigeria: Neopatrimonialism and Failure
of Industrialization 329
Conclusion: Understanding States and State Intervention
in the Global Periphery 367

Select Bibliography 427

Index 447
P1: JzG/JzK
0521836700agg.xml CY409B/Kohli 0 521 83670 0 June 8, 2004 16:10

List of Tables and Figures

Tables
1 Some Relevant Indicators of Korea, Brazil, India, and Nigeria page 24
2.1 Economic Growth during the Rhee Period 75
3.1 Savings and Investment Ratios, 1961–1980 103
7.1 Some Basic Growth Data, 1950–2000 258
7.2 Patterns of Capital Formation, 1980–1998 282

Figures
1 Pattern of Per Capita GDP Growth: Korea, Brazil, India,
and Nigeria, 1950–2000 24

ix
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

Introduction
States and Industrialization in the Global Periphery

Legitimate states that govern effectively and dynamic industrial economies

are widely regarded today as the defining characteristics of a modern nation-
state. Ever since Western countries developed such political economies a few
centuries back, those left behind have sought to catch up. Among late de-
velopers, countries such as Japan and Russia avoided being colonialized by
consolidating their respective states and adopting alternative strategies of in-
dustrialization, with varying results. The search for development among late-
late-industrializers of Asia, Africa, and Latin America intensified mainly after
the Second World War, when numerous activist states emerged as sovereign.
It is clear from the vantage point of the end of the twentieth century that
state-led development efforts have been more successful in some parts of
the global periphery than in others. This book looks at the role states have
played in fostering different rates and patterns of economic development,
especially via deliberate industrialization.
States in most peripheral countries of Asia, Africa, and Latin America
are important, active economic actors, engaged in varying patterns of state
intervention. In some developing countries the state’s economic role has
come to be associated with both rapid industrial transformation and en-
hanced equity. In other cases, by contrast, governments and bureaucrats
have pilfered the economic resources of their own societies, failing to stim-
ulate economic growth and facilitating transfer of wealth into the hands of
unproductive elites. In yet other cases, state intervention is associated with
mixed outcomes: States have helped to solve some important economic
problems, while ignoring other problems and creating new ones.
This study undertakes a comparative analysis of the state as an economic
actor in developing countries. Why have some of these states been more
successful at facilitating industrialization than others? This question really
has two components: What features distinguish state intervention in the
more successful cases from intervention in the less successful cases? and
How does one explain varying state capacities to choose and implement
1
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

2 Introduction

economic decisions? The first question of patterns of state intervention fo-

cuses both on the state’s policy choices and on its relationship with such
key economic actors as business and labor. By contrast, the second ques-
tion, concerning state capacities, looks to the institutional character of
the state itself, an identity often assumed well before the political elite
initiated deliberate industrialization. This book then is about patterns of
state construction and patterns of state intervention aimed at promoting
industrialization.
It is mainly an inductive study that seeks a general understanding of the
state as an economic actor in developing countries via detailed analyses
of four major developing countries of the twentieth century: South Korea,
Brazil, India, and Nigeria (see Table 1 and Fig. 1, below). These cases provide
a range of variation in state capacities to pursue economic transformation,
from a fairly effective, growth-promoting state in South Korea to a rather
ineffective and corrupt Nigerian state, with Brazil and India providing mixed
cases. What helps to explain these variations? Indeed, this is a key puzzle in
late-late-development.
The main argument of the book, drawn from comparative historical anal-
ysis of these four countries, is that the creation of effective states within the
developing world has generally preceded the emergence of industrializing
economies. This is because state intervention in support of investor profits
has proved to be a precondition for industry to emerge and flourish among
late-late-developers. Patterns of state authority, including how the politics
of the state are organized and how state power is used, have decisively influ-
enced the economic context within which private economic decisions are
made. They are thus important, nay, critical, for understanding varying rates
and patterns of industrialization. Patterns of state authority, in turn, often
exhibit long-term continuities. Colonialism in the first half of the twentieth
century, especially, was defining of the state institutions that emerged in de-
veloping countries, and that in turn molded their economies in the second
half of the century.

I. Some Clarifications
The focus of this study requires, if not a justification, at least some prelimi-
nary clarifications. Three brief caveats are in order.
First, it is clear that development involves a lot more than economic
growth and that variations in economic growth reflect more than under-
lying variations in industrial growth. Moreover, any study of the role of
the state in industrialization must attend to issues of agricultural growth
and income distribution. Nevertheless, the primary focus of this study is
on the political and policy determinants of industrialization, because indus-
trial growth is a key determinant of any country’s overall economic growth,
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

States and Industrialization in the Global Periphery 3

and economic growth remains a core element of any understanding of

development.
Second, a central concern with the economic role of the state in develop-
ment really does not require any justification. It is more than an idiosyncratic
assertion to hold that states are important economic actors in developing
countries and thus worthy of serious scholarly interest – even if rates of
economic growth reflect a host of other factors. In this regard one can note
numerous variables that influence a given country’s economic performance
over a specific period, including world economic conditions, resource en-
dowments, differing starting points, demographic factors, national price
regimes, patterns of savings, levels of technology, and entrepreneurship.
For their part, however, historians of economic processes repeatedly em-
phasize the significance of institutions, especially the role of government.
Lloyd Reynolds, for example, concluded his major study of economic growth
in the “Third World” by observing that the analysis of underlying economic
factors does not fully reveal the “mystery” of “sources of sustained growth.”
What is missing from the economic models, according to him, “can be la-
beled as political. Government matters.”1 W. Arthur Lewis similarly noted
in a presidential address to the American Economic Association that one
building block of any full understanding of “the engine of growth” would be
“a theory of government, where government would appear to be as much
the problem as the solution.”2 More recently, a number of scholars have
gone further, suggesting that an activist state has been a key ingredient of
rapid development, especially in parts of East Asia, and that, conversely, mal-
functioning states have contributed heavily to developmental failures, say,
in sub-Saharan Africa.3 The prima facie case for a broad focus on state and
development is thus strong.

1 See Lloyd G. Reynolds, Economic Growth in the Third World, 1850–1980 (New Haven, Conn.:
Yale University Press, 1985), 413–14.
2 W. Arthur Lewis, “The State of Development Theory,” American Economic Review 5 (March
1984): 8.
3 See, for example, Peter Evans, Embedded Autonomy: State and Industrial Transformation
(Princeton, N.J.: Princeton University Press, 1995); Stephen Haggard, Pathways from the
Periphery: The Politics of Growth in the Newly Developing Countries (Ithaca, N.Y.: Cornell University
Press, 1990); Robert Wade, Governing the Market: Economic Theory and the Role of Government
in East Asian Industrialization (Princeton, N.J.: Princeton University Press, 1990); and Alice
Amsden, Asia’s Next Giant: South Korea and Late Industrialization (New York: Oxford University
Press, 1989). A recent assessment of this “developmental state” literature is Meredith Woo-
Cumings, The Developmental State (Ithaca, N.Y.: Cornell University Press, 1999). For the role
of the state in African cases, see Richard Sandbrook, Politics of Africa’s Economic Stagnation
(Cambridge: Cambridge University Press, 1985). A recent collection in which a group of
economists emphasize the role of institutions in economic growth, including the role of
public institutions, is Dani Rodrik, ed., In Search of Prosperity: Analytic Narratives on Economic
Growth (Princeton, N.J.: Princeton University Press, 2003). This volume came to my attention
only as the present study was going to press.
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

4 Introduction

Finally, there is the issue of case selection: Why have Korea, Brazil, India,
and Nigeria been chosen?4 Two brief and preliminary justifications for case
selection will suffice for now. First, these four are large and significant de-
veloping countries: Nigeria, for example, is the most populous country in
sub-Saharan Africa, as is Brazil in South America (see Table 1, below).5
Any meaningful generalizations about state and development would nec-
essarily have to apply at least to these cases. Moreover, the general ideas
developed below are actually examined against more cases than a focus on
four countries may suggest. For example, detailed analysis of countries over
time provides other cases within a single country, and some further check on
propositions will be provided by examining them briefly against other cases,
especially within each region. Thus, for example, I return in the conclusion
to a consideration of how well the propositions developed with reference,
say, to the Korean case apply to other successful East Asian cases or, similarly,
the extent to which the propositions developed with reference to Nigeria
help to explain other African failures. And second, as already noted, these
four cases provide a range of variation – from effective state intervention in
Korea to an ineffective state in Nigeria – that helps to frame the analytical
puzzle: What explains such variation in developmental efficacy?

II. The Intellectual Context

As an investigation of both state construction and state intervention in the
developing world, this study addresses several interrelated scholarly debates.
One of the more heated debates concerns the role of states versus that of
markets in the process of late-late-development. A less well developed debate
that this study engages concerns how it came about that some parts of the
developing world acquired more effective developmental states than did
others.
The systematic political and economic study of developing countries be-
gan only after the Second World War and grew rapidly thereafter. For some
two to three decades the organizing framework was to seek generalizations

4 Much can be said on this issue, and I return to some related concerns in due course, especially
in the conclusion. Some basic development indicators of these four cases are summarized in
Table 1.
5 Other ready candidates for inclusion on the grounds of “large” and “significant” would
have been China, Indonesia, Turkey, Egypt, and Mexico. Given its communist past, China
was too much of an “outlier”; the Chinese case is nevertheless important for a discussion
of “state capacity” and I consider it briefly in the conclusion. Egypt has been so mired in
regional conflicts that it is atypical. Indonesia, Turkey, and Mexico, however, were all worthy
of inclusion. The main excuse for not including them is limited scholarly energy. One useful
comparative analysis that goes over some of the same thematic concerns as this study but
with reference to Mexico, Egypt, Turkey, and India is John Waterbury, Exposed to Innumerable
Delusions: Public Enterprise and State Power in Egypt, India, Mexico, and Turkey (Cambridge:
Cambridge University Press, 1993).
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

States and Industrialization in the Global Periphery 5

about the developing world as a whole, with an emphasis on determin-

ing how this world, the “Third World,” was distinct from the one com-
posed of the more advanced and industrialized political economies. This
approach held as much for political scientists and sociologists associated
with the modernization framework as it did for radical dependency crit-
ics and for economists trying to carve out a niche for economic develop-
ment studies within economics. Thus modernization scholars worried about
the prospects for stability and democracy in the developing world, depen-
dency scholars examined why dependent political economies were not likely
to replicate the dynamism associated with early capitalism, and economic
development specialists often argued that market imperfections so preva-
lent in the developing world necessitated state intervention and import
substitution.6
But this early development literature seldom asked why some developing
economies performed better than others. This is because in an important
sense, it was simply too early to ask this question in the 1950s and the 1960s,
as only in the 1970s did the dramatic variations in performance across de-
veloping countries, especially in rates of industrialization, start to become
apparent. Only from then on did these variations become a central con-
cern for development scholars. By contrast, insofar as the earlier literature
shed some light on these issues of comparative political economy, it was
only indirect: Modernization literature could be construed as suggesting
that variation in economic performance may have something to do with
the role of regime type: democratic versus authoritarian; dependency liter-
ature proposed that the same variations may instead reflect greater or lesser
dependency on global capitalism; and it was implicit in the economic de-
velopment literature that varying economic performance was considered to
reflect the relative success of the state’s efforts to rectify market imperfec-
tions. These were important insights, some of which are incorporated into
the empirical analyses below. We also revisit more specific hypotheses asso-
ciated with these literatures in the conclusion. Suffice it to note for now that
none of these earlier frameworks is up to the task at hand because they were
not designed for the explicit purpose of developing sound political economy

6 These are large bodies of literature and this is no place to discuss them. For one good review of
the modernization literature, see Samuel Huntington, “The Change to Change,” Comparative
Politics 3 (April 1971): 3, 283–322. Dependency literature is well reviewed in Gabriel Palma,
“Dependency: A Formal Theory of Underdevelopment or a Methodology for the Analysis of
Concrete Situations of Underdevelopment?,” World Development 6 (1978): 881–924. And the
literature on economic development is usefully assessed in Albert O. Hirschman, “Rise and
Decline of Development Economics,” in Hirschman, Essays in Trespassing: Economics to Politics
and Beyond (Cambridge: Cambridge University Press, 1981); and Paul Krugman, “Toward a
Counter-Counter-Revolution in Development Theory,” World Bank Economic Review, Supple-
ment (Washington, D.C.: Proceedings of the World Bank Annual Conference on Develop-
ment Economics, 1992, 1993).
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

6 Introduction

explanations of variations in rates and patterns of economic development

across the developing world.7
The political economy debate on why some parts of the developing world
have had faster-growing economies than others polarized in the last two
decades of the twentieth century around a more neoliberal, promarket po-
sition, on the one hand, and a statist argument, on the other hand. The
promarket position emerged nearly hegemonic, especially among some
economists and development policy practitioners.8 For its part, the statist
argument, often articulated by interdisciplinary scholars of development,
provided a cogent scholarly dissent.9 This book hopes to advance this de-
bate further yet by developing the existing statist position, though not by
adopting an aggressive antimarket positon. On the contrary, one of the
themes of this study is that state intervention in rapid industrializers was of-
ten characterized by market-reinforcing behavior, understood in the sense
of supporting profitability for private investors – and not as strengthening
competitiveness or openness or, even less, as a state’s self-limiting proclivity.
The state versus market mind-set thus is simply not very helpful for under-
standing how the interaction of states and markets has served to produce a
range of economic outcomes.
Two claims at the heart of the promarket position on economic devel-
opment are examined. First, there is the outward-orientation claim, which
suggests that greater openness and greater competitiveness in the econ-
omy generate higher rates of production growth via more efficient allo-
cation of scarce resources. And second, the laissez-faire claim holds that
state intervention in the economy necessarily generates distortions that hurt
economic growth. Although the outward-orientation position is the more
compelling of the two claims, it nonetheless requires a number of qualifi-
cations about the importance of state intervention for export promotion
and about the problem of identifying the causal direction between exports
and growth. Moreover, the preponderance of the evidence indicates that

7 I have critiqued the modernization and dependency literatures for their neglect of the au-
tonomous and varying role of the state in development and the development economics
literature for assuming that states will play a “welfare maximizing” role without a good un-
derstanding of when and how states might take on such a role. See “Introduction,” in Atul
Kohli, ed., The State and Development in Third World (Princeton, N.J.: Princeton University Press,
1986); Atul Kohli, The State and Poverty in India: The Politics of Reform (Cambridge: Cambridge
University Press, 1987), chap. 1; and Atul Kohli, “State, Society, and Development,” in Ira
Katznelson and Helen Milner, eds., State of the Discipline (New York: W. W. Norton, 2002),
84–117.
8 See, for example, Anne Krueger, “The Political Economy of the Rent Seeking Society,”
American Economic Review 64 (1974): 291–303; Deepak Lal, The Poverty of “Development” Eco-
nomics, Hobart Paperback no. 16 (London: Institute of Economic Affairs, 1983); and World
Bank, Development Report (New York: Oxford University Press, 1991).
9 Among others, see Wade, Governing the Market; Amsden, Asia’s Next Giant; and Evans, Embedded
Autonomy.
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

States and Industrialization in the Global Periphery 7

late-late-industrialization has always commenced under conditions of pro-

tection. As to the laissez-faire claim, there is a stunning lack of evidence for
the proposition that less government facilitates more rapid industrializa-
tion in the developing world. On the contrary, the evidence shows that state
intervention aimed at boosting investor profitability is strongly associated
with rapid industrialization. This study investigates validity of these claims
via detailed case materials, and the main findings – that there are serious
doubts about these claims – are summarized in the conclusion.
The main dissent from the promarket position in contemporary devel-
opment studies is offered by statist scholars who emphasize the constructive
role played by state intervention in select, successful cases. This literature
has carefully documented how numerous such state interventions – tariffs,
subsidies, credit control, manpower training, technology promotion, and
bureaucratic cooperation with the private sector and oversight – have con-
tributed to industrialization and rapid growth, especially in East Asia. Other
scholars have gone further, asking what sort of states are capable of such
interventions and often emphasizing such variables as leadership priori-
ties and/or the quality of bureaucracy.10 The present study builds on this
literature; indeed, it could not have been written without the foundation
laid by these bold and original formulations. One argument that is central
in this statist literature informs the thinking here as well – that successful
state intervention often involves close cooperation between the state and
private investors. At a popular level, this idea is often expressed in such
characterization of East Asian political economies as “Korea Incorporated”
or “Taiwan Incorporated.” At a more scholarly level, the idea is expressed as
“embedded autonomy” being a precondition for successful development, an
idea that suggests not only close cooperation between the state and business
but also a measure of insulation for the bureaucratic elite so as to minimize
corruption and “state capture” by private interests.11
Although the present study builds on the statist literature, it also departs
from that literature in important ways. First, this study is as interested in the
question of why some parts of the developing world have acquired more
effective states than others as it is in the more widely discussed question
of what characterizes effective, developmental states and what exactly such
states do to foster rapid industrialization. This shift in focus demands a more

10 A highly influential early such study, not of a developing country, but on reinterpreting the
Japanese experience, was Chalmers A. Johnson, MITI and the Japanese Miracle: The Growth
of Industrial Policy, 1925–1975 (Stanford, Calif.: Stanford University Press, 1982). The most
significant recent contribution along these lines is Evans, Embedded Autonomy. See also the
citations in n. 3.
11 See Evans, Embedded Autonomy, passim but esp. chap. 1. There is some striking overlap be-
tween Evans’s book and this one. However, as explained immediately below, differences in
the questions posed, modes of analysis, and the underlying theoretical orientation also loom
large.
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

8 Introduction

historical orientation because the core character of these states was often
acquired long before they started intervening in the economy to promote
development. A comparative historical analysis, in turn, leads to a consider-
ation of the role of varying types of colonialisms in molding different types
of states across the developing world. Second, this study, in examining is-
sues of state effectiveness, is as concerned with the political organization
of the state as it is with bureaucratic capabilities. A central concern then is
with power, as distinct from competence or information. And finally, when
analyzing what exactly interventionist states do to promote rapid industrial-
ization, I focus as much on issues of mobilizing basic factors of production as
on improving the efficient use of such factors via upgrading technology or
“learning by doing.” This shift, too, leads to a more detailed consideration
of such issues as (1) the state’s capacity to collect taxes and thus to make
public investments and (2) the state’s role in disciplining the labor force.
In sum, the present study, while clearly a part of the scholarly statist dis-
sent from the more neoclassical accounts of development, at the same time
departs from many existing statist accounts by providing a more historical
and political analysis of the state’s role in both promoting and hindering
industrialization.

III. The Argument

Industrialization involves social change. While its narrow outcome is an
increase in industrial production from existing or new factories, a broader
set of societal changes have also generally accompanied, if not preceded,
industrial development. These include a situation of political stability, the
availability of experienced entrepreneurs and of a capable urban work force
and mobilizable capital, the emergence of a market for industrial goods, and
the presence of a growing body of technical knowledge. It is not surprising
that among the earliest, “spontaneous” industrializers, such as England, the
process occurred slowly, over centuries, and was “caused” not by any single
development but by the merging of several streams of underlying changes.
It was Gerschenkron who first argued persuasively that follower countries
within Europe did not reproduce England’s “spontaneous” model. Instead,
he held, they needed a more organized initiative from banks or states to
help to generate “a movement on a broad front” to help industry to take
off by mobilizing capital, creating a work force, and facilitating technology
transfer.12 What was true for late-developers within Europe or Japan was,
of course, doubly true for late-late-industrializers of the developing world.
Since the mid-twentieth century, states have sought to promote industry in
most countries of Asia, Africa, and Latin America. What is also clear from

12 See Alexander Gerschenkron, Economic Backwardness in Historical Perspective (Cambridge:

Harvard University Press, 1962), esp. chap. 1.
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

States and Industrialization in the Global Periphery 9

the vantage point of the end of the century, however, is that this half-century
of effort has seen considerably more success in some parts of the developing
world than in others. Explaining these divergent pathways is the main task
of this book.
Based on detailed comparative analysis of four countries that vary from
success to failure, I argue that the way state power is organized and used has
decisively influenced rates and patterns of industrialization in the global
periphery. A core analytical task involves identifying different patterns of
state authority and then tracing the impact of those variations on economic
outcomes and probing the origins of the varying state types. The role of
concentrated political power in helping states to set the agenda for their
economies also needs to be understood. No quick summary of the argument
can substitute for detailed empirical analysis. This is not only because the
devil really is in the details but also because summary arguments may be
more persuasive after digesting the empirical details. I return to the big
picture again in the conclusion. At this time, it is useful to note four of the
arguments underlying the book’s comparative analysis.

State Types
Referring to ideal types, I identify three historical patterns of how state
authority is organized and used in the developing world, neopatrimonial
states, cohesive-capitalist states, and fragmented-multiclass states. Although
these labels are less than fully satisfactory, they are better than most others
in use. In any case, the focus ought to be more on the patterns described by
the categories than on the labels themselves.
In addition to centralized and coercive control over a territory, a defining
characteristic of all modern states is a well-established public arena that is
both normatively and organizationally distinguishable from private interests
and pursuits. Unfortunately, for a variety of historical reasons, this distinc-
tion between the public and the private realms was never well established in
a number of developing country states, especially African states. As a result,
a number of distorted states emerged with weakly centralized and barely le-
gitimate authority structures, personalistic leaders unconstrained by norms
or institutions, and bureaucracies of poor quality. These states are labeled
here as neopatrimonial because, despite the façade of a modern state, pub-
lic officeholders tend to treat public resources as their personal patrimony.
These are therefore not really modern, rational-legal states. Whether orga-
nized as a nominal democracy or as a dictatorship, state-led development
under the auspices of neopatrimonial states has often resulted in disaster,
mainly because both public goals and capacities to pursue specific tasks in
these settings have repeatedly been undermined by personal and narrow
group interests. Of the cases analyzed in this study, Nigeria best exemplifies
this ideal-typical tendency.
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

10 Introduction

Cohesive-capitalist and fragmented-multiclass states are two of the other

ideal-typical states to be found in the contemporary developing world. The
more effective modern rational-legal states in the developing world tend to
vary mainly along two dimensions: cohesion of state authority and the state’s
class commitments. Cohesion of authority is manifest at both the intraelite
and the elite-mass levels, and variations in patterns of authority demar-
cate the more cohesive states from the more fragmented ones. Developing
country states may also be narrowly committed to working with capitalists
or may rest their power and goals on a more multiclass base.13
The cohesive-capitalist states, sometimes called developmental states,
are situated opposite neopatrimonial states on the political effectiveness
continuum.14 These states are characterized by cohesive politics, that is, by
centralized and purposive authority structures that often penetrate deep
into the society. For a variety of historical reasons these states have tended
to equate rapid economic growth with national security and thus defined
it as a priority. In their pursuit of rapid growth, cohesive-capitalist states
have carved out a number of identifiable links with society’s major eco-
nomic groups and devised efficacious political instruments. Especially no-
table among the social links is a close alliance with producer or capitalist
groups. An important corollary of this political arrangement is a tight con-
trol over labor. The main political instrument of these states is, of course, a
competent bureaucracy. Since a narrow elite alliance between the state and
capital is difficult to hold together, politics within these units has often been
repressive and authoritarian, with leaders often using ideological mobiliza-
tion (e.g., nationalism and/or anticommunism) to win acceptance in the
society. Cohesive-capitalist states in developing countries, such as in South
Korea under Park Chung Hee and in Brazil during both Estado Novo and
the military dictatorship, thus share some organizational and class character-
istics with fascist states of interwar Europe and Japan.15 (Obviously, however,

13 Careful readers may wonder whether such other combinations as cohesive-multiclass and
fragmented-capitalist are also possible. The short answer is yes, but generally not in the
developing world. A well-organized social democracy, such as Sweden, may illustrate the
cohesive-multiclass category, and the United States may be a good example of a fragmented-
capitalist state. A fuller discussion of these state types is beyond the scope of this study, though
some such issues are discussed in due course, especially in the conclusion.
14 It will become clear in due course, especially in the section on Korea, why I prefer the
concept of cohesive-capitalist states over “developmental states.” Suffice it to note here that
the idea that developmental states facilitate development strikes me as too obvious for both
analytical and normative comfort. A label such as cohesive-capitalist states (in an earlier draft,
I had used the term “neofascist states”) instead both captures better a state’s independent
political characteristics – by which I mean the ideology, organization, and the underlying
class alliances – and cautions observers to take note of the costs incurred by the type of
“development” these states promote.
15 Generalizing about “fascist-style regimes” during the interwar years, Claudio Segre notes that
“within wide national variations, fascist states had certain characteristics and aspirations in
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

States and Industrialization in the Global Periphery 11

these states do not draw explicit comparisons with discredited fascism, and
they tend to shy away from the politics of mass mobilization and ethnic
cleansing.) For better or for worse, these states have also proved to be the
most successful agents of deliberate state-led industrialization in peripheral
countries.
In between the two extremes of political effectiveness defined by neopat-
rimonial states on the one end and cohesive-capitalist states on the other
lie fragmented-multiclass states. Unlike neopatrimonial states, fragmented-
multiclass states are real modern states. They command authority, and a
public arena within them is often well enough established that leaders
are held accountable for poor public policies and performance. Unlike
in cohesive-capitalist states, however, public authority in these states tends
to be more fragmented and to rest on a broader class alliance – meaning
that these states are not in a position to define their goals as narrowly or
to pursue them as effectively as are cohesive-capitalist states. Leaders of
fragmented-multiclass states thus need to worry more about political sup-
port than do leaders of other types of developing country states. For exam-
ple, they must typically pursue several goals simultaneously, as they seek to
satisfy multiple constituencies. Industrialization and economic growth may
be an important state goal, but it is only one among others: agricultural
development, economic redistribution, welfare provision, and maintaining
national sovereignty. Policy formulation and implementation, moreover, is
often politicized, either because of intraelite conflicts or because state au-
thority does not penetrate deep enough down in the society to incorporate
and control the lower classes. When confronted by mobilized opposition,
fragmented-multiclass states typically become obsessed with issues of legit-
imacy and often find themselves promising more than they can deliver.
While not all fragmented-multiclass states are necessarily democracies, all
developing country democracies with plebiscitarian politics and weak in-
stitutions constitute a special subset of fragmented-multiclass states. The
cases of India and Brazil in several periods exemplify this type of state.
Attempting to pursue a complex state-led agenda with limited state capaci-
ties, then, fragmented-multiclass states tend to be middling performers on
numerous dimensions, including the promotion of industrialization and
growth.

common. In their political systems, they created police states, one party systems led by a
charismatic dictator. Their economic systems aimed to develop some form of national so-
cialism. The government was to play an active role in controlling the economy, but unlike
Marxian socialism, the state was not to take over the means of production. Fascist socialism
was directed at the interests of the nation. . . . Fascism also aspired to some form of the cor-
poratist state. . . . Fascist regimes mobilized and disciplined societies to transform themselves
far more rapidly than would have been the case under a laissez-faire system.” See his entry
on “fascism” in The Oxford Companion to Politics of the World (New York: Oxford University
Press, 2001), 274–76.
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

12 Introduction

Neopatrimonial states and the two more modern, rational-legal types of

states, namely, cohesive-capitalist and fragmented-multiclass states, are ideal
types of developing country states and are not found in a pure form in any
of the countries discussed below. Instead, countries in specific periods ex-
hibit more of one tendency than another. When comparing and analyzing
state types empirically, one generally needs to focus on some such state
characteristics as leadership goals, degree of centralization of public au-
thority, downward penetration of public authority, political organization of
the mobilized political society, scope of state intervention in economy, and
quality of the economic bureaucracy. It may be easy in an abstract fashion
to suggest, as I have, that cohesive-capitalist states are characterized by the
top leadership equating rapid economic growth with national security, a
highly centralized and penetrating public authority, state-controlled polit-
ical society (though in close alliance with capitalist groups), and a highly
interventionist state, with a good quality economic bureaucracy. Neopatri-
monial and fragmented-multiclass states have been similarly characterized
in their pure form. Real historical records of actual countries, however, sel-
dom reveal state types in their ideal-typical form; states instead tend more
toward one set of characteristics than another, opening the way for more
complex analyses.

State Types and Patterns of Industrialization

If authority structures in the developing world can be variously categorized
as neopatrimonial, cohesive-capitalist, and fragmented-multiclass states, the
first relevant question for this study concerns how these states influence
economic outcomes. The nearly exclusive focus in the literature on appro-
priate policy choices is incomplete, even misleading. Policy choices matter,
of course, but these choices must be explained. More important, the impact
of the same policy applied in two different settings may vary because of the
contextual differences, some of the more obvious being varying global con-
ditions and different initial conditions of an economy. The import of these
issues will become clear in due course, but now the contextual difference
that deserves emphasis – because it is significant and generalizable across
cases – concerns the varying political and institutional conditions in which
economic policies are chosen and pursued.
More specifically, identifying variations in how states are organized and
in the institutionalized relationship of the state to the private sector is key
to understanding the relative effectiveness of state intervention in the econ-
omy. In the cases examined in this study, this relationship varies along a
continuum stretching from considerable convergence in goals to mutual
hostility between the state and the private sector. I argue that, other things
being equal, the setting that has proved to be most conducive (i.e., serves
as a necessary but not a sufficient condition) to rapid industrial growth in
the developing world is one in which the state’s near-exclusive commitment
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

States and Industrialization in the Global Periphery 13

to high growth coincided with the profit-maximizing needs of private en-

trepreneurs. The narrow ruling coalition in these cases was a marriage of
repression and profits, aimed at economic growth in the name of the nation.
Cohesive-capitalist states have generally created such political economies.
Turning their countries into state-guided corporations of sorts, they have
tended to be the fastest growers in the developing world.
Growth-oriented cohesive-capitalist states pursued their commitment to
high growth by developing trade and industry with well-designed, consistent,
and thoroughly implemented state intervention. Specific policy measures
varied but were generally aimed at easing supply-and-demand constraints
faced by private entrepreneurs. Some of these interventions were direct,
and others, indirect. On the supply side, for example, we find that cohesive-
capitalist states helped to facilitate the availability of capital, labor, technol-
ogy, and even entrepreneurship. Thus supply of capital was boosted at times
by superior tax collection and public investment, at other times by using pub-
licly controlled banks to direct credit to preferred private firms and sectors,
and at yet other times by allowing inflation to shift resources from both agri-
culture and urban labor to private industrialists. Repression was also a key
component in enabling private investors to have a ready supply of cheap,
“flexible,” and disciplined labor. Examples of less-direct interventions on
the supply side include promotion of technology by investing in education
and research and development and/or by bargaining with foreign firms to
enable technology transfer.
On the demand side, too, cohesive-capitalist states have pursued a vari-
ety of policies to promote their growth commitment. These have included
expansionist monetary and fiscal policies, and tariffs and exchange-rate poli-
cies aimed at boosting domestic demand. And when domestic demand was
not sufficient, these states have just as readily adopted newer policies that
shift the incentives in favor of export promotion or, more likely, that help
to promote production for both domestic and foreign consumption.
There was thus significant variation in the specific policy measures un-
dertaken by cohesive-capitalist states. Only some policies, such as labor dis-
cipline, necessitated a repressive state. But what most policies adopted by
cohesive-capitalist states reflected instead was a single-minded and unyield-
ing political commitment to growth, combined with a political realization
that maximizing production requires assuring the profitability of efficient
producers but not of inefficient ones. Sometimes this required getting prices
right, but just as often it required “price distortions,” such as undervalu-
ing exchange rates, subsidizing exports, and holding wages back behind
productivity gains. The central issue concerned the state’s goals and ca-
pacities, expressed in the institutionalized relationship between the state
and the private sector. Cohesive-capitalist states in successful industrializers
have thus been pragmatically – and often ruthlessly – procapitalist, much
more than they have been purely and ideologically promarket. Among the
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

14 Introduction

cases in this study, South Korea under Park Chung Hee and Brazil during
Estado Novo and under military rule most clearly fit this political economy
model.
Perfect coincidence between the goals of the state and those of private
elites has been rare in the developing world, depending as it does on the
difficult-to-acquire political precondition of cohesive state power and a nar-
row alliance between the state and the capital-owning elites. Instead, many
ruling elites governed states with fragmented political institutions and de-
fined the public good more broadly. The elites pursued (or, at least, debated)
several crucial goals simultaneously: economic growth, redistribution, legit-
imacy, and national sovereignty. Policy intervention in these fragmented-
multiclass states was aimed not only at promoting growth but also at en-
hancing legitimacy and short-term welfare provision.
Mixed political goals of fragmented-multiclass states had several conse-
quences for choosing and pursuing development policies. First, ruling elites
were less focused in these cases on assessing state intervention strictly from
the vantage point of growth consequences. Diffuse goals, in turn, enabled
various groups and individuals to capture state resources for short-term,
consumption-oriented benefits. Second, the relationship of the state to the
private sector in such contexts was considerably more complex than in
cohesive-capitalist states, sometimes cooperative but just as often conflictual.
And third, both policy making and implementation were more politicized,
diluting their unidirectional effectiveness.
Fragmented-multiclass states are thus actually more “normal” than the
other two ideal-typical cases being discussed here. But because the choice
of economic strategy and of policy tools in these cases reflected the logic of
both growth and politics, the institutional setting of fragmented-multiclass
states was seldom conducive to achieving hypergrowth in industry. The case
of India supports such a general contention, as do the cases of Brazil and
South Korea in select periods.
Let us consider specific examples of the political economy dynamics
of fragmented-multiclass states. Fragmented-multiclass states were neither
more nor less interventionist than cohesive-capitalist states, but they were
generally less effective at alleviating the supply-and-demand constraints
faced by their investors. Again, for example, when it came to mobilizing cap-
ital in many fragmented-multiclass states, tax-collecting capacities were lim-
ited, public-spending priorities included numerous goals other than growth
promotion, attempts to direct credit easily evolved into cronyism, and infla-
tion as a tool of resource transfer could readily become a liability for political
leaders concerned about their legitimacy. Periodic hostility on the part of
the state elite toward private investors made the latter, both domestic and
foreign, reluctant to invest. Repression of labor was also not a ready alter-
native in fragmented-multiclass states, thus making it difficult for investors
to mobilize a cheap and docile labor force.
P1: JZH/JzK P2: JzG
0521836700int.xml CY409B/Kohli 0 521 83670 0 May 14, 2004 21:54

States and Industrialization in the Global Periphery 15

On the demand side, monetary and fiscal policies seldom reflected a

consistent growth commitment but fluctuated instead with political cycles
characterized by greater or lesser legitimacy. And finally, tariff and exchange-
rate policies adopted to protect the national economy, and thus to promote
demand for indigenous goods, often created powerful interest groups. As
these groups were difficult to dislodge, fragmented-multiclass states found
themselves more rigidly committed to a particular development path. In
sum, fragmented-multiclass states, like cohesive-capitalist states, sought to
promote industrialization, but they did so less effectively because their goals
were more plural and their political capacities less developed. In other
words, varying patterns of state authority decisively influenced developmen-
tal trajectories.
According to this line of argument, the worst setting for industrialization
on the periphery was the states that had no clear public goals and whose
leaders reduced the state to an arena for personal aggrandizement. These
neopatrimonial states have unfortunately constituted a significant subset
of the developing world. State intervention in these cases has often been
motivated either by the need to build short-term political support via pa-
tronage or by personal greed – or sometimes by both. The relationship of
the state and the private sector in such contexts has just as often been mutu-
ally corrupt: Political instability, inconsistent policies, and pilfering of public
resources for personal and sectional gains have all hurt state-led efforts to
promote industry and growth. The case of Nigeria provides a striking in-
stance of such a development path, though elements of the same are also
evident elsewhere.
We will see that neopatrimonial states, like cohesive-capitalist and frag-
mented-multiclass states, have also intervened heavily in their economies,
but with disastrous results. Neopatrimonial states have often emerged in so-
cieties with weak private sectors, but instead of strengthening the private sec-
tor, these states have appropriated scarce economic resources and diverted
them everywhere but toward productive investment. Inconsistent economic
policies, failure to support indigenous capitalists, poor-quality but activist
labor, and political instability have all reinforced the existing weakness of
the national private sector in manufacturing and industry.
Given this profound weakness of domestic capitalism, neopatrimonial
states have sought to undertake economic activities directly or invited for-
eign goods and producers to fill the vacuum. Given the states’ nondevelop-
mental proclivities and organizational weakness, efforts to produce goods
in the public sector have generally failed. The remaining alternative of im-
porting goods or attracting foreign investment makes sense only if there are
alternative sources of income and demand. For a country such as Nigeria, oil
exports provided a ready source of income and demand, which was met by
foreign goods and producers; this is less true for other neopatrimonial states.
Commodity booms, however, seldom last forever. The political incapacity to