Ha Joon Chang The Political Economy of Industrial Policy Palgrave Macmillan 1993

The Political Economy
of Industrial Policy
Ha-Joon Chang
Faculty o.f Economics and Politics
University of Cambridge
Published in Great Britain by
MACMILLAN PRESS LTD
Houndmills, Basingstoke, Hampshire RG2 l 6XS
and London
Cornpanies and representatives
throughout the world
First edition 1994

Reprínted (with alteratíons) 1996
A catalogue record for this book is available
from the Britísh Líbrary.
ISBN 0-333-58862-2 hardcover

ISBN 0-333-67890-7 paperback
First publíshed in the United States of Ameríca 1994 by

ST. MARTIN'S PRESS, INC.,
Scholarly and Reference División,
175 Fifth Avenue,
New York, N.Y. 10010
ISBN 0-312-10294-l (cloth)
Líbrary of Congress Catalogíng-in-Publícation Data
Chang, Ha-Joon.
The política! economy of industrial policy I Ha-Joon Chang.
p. cm.
"First published in Great Brítaín ... by the Macmillan Press Ltd"-
-T.p. verso.
Revision of the author's thesis (doctoral-Cambridge)
Jncludes biblíographical references and index.
ISBN 0-312-10294-1
l. Industrial policy. 2. Industrial policy-Korea (South)
l. Tille
HD361 I.C42 1994
338.9--dc20 92-14247
CIP
© Ha-Joon Chang 1994. 1996
All rights reserved. No reproduction, copy or transmission of

this publication may be made without wrítten permission.
No paragraph of thís publication may be reproduced, copied or
transmítted save with written permíssion or in accordance wíth
the provisions of the Copyright, Desígns and Patents Act 1988,
or under the tcrms of any licence permítting limited copying
íssued by the Copyright Licensing Agency, 90 Tottenham Court
Road, London W l P 9HE.
Any person who does any unauthorised act in relatíon to thís
publícatíon may be liable to criminal prosecution and civil
claíms for damages.
10 9 8 7 6 5 4 3 2 l
05 04 03 02 01 00 99 98 97 96
Printed in Great Britain by

lpswich Book Co Ltd, Ipswich, Suffolk
To my parents
This page intentionally left blank
Contents
Preface to the 1996 Reprint xi
Acknowledgements xvi
Introduction: The Search for New Patterns of Economic
Management
1 Theories of State Intervention: A Review of the 7

Literature
Introduction 7
1.1 Efficiency: the market-failure literature 7
1.1. l Public goods 8
1.1.2 Non-competitive markets 9
1.1.3 Externalities 10
1.2 Morality: paternalism and contractarianism 12
l.2.1 Paternalism and contractarianism 12
1.2.2 What should the state do? 14
1.2.3 Sorne reflections 15
1.3 Intention: the political economy literature 18
1.3.1 The autonomous-state approach 18
1.3.2 The interest-group approach 19
1 .3.3 The self-seeking-bureaucrats approach 22
1.3.4 Summary 24
1.4 Ability: the governrnent-failure literature 25
1.4 .1 The information problem 26
1.4.2 Rent-seeking 27
Conclusion 31
2 A New Institutionalist Theory of State Intervention 33

Introduction 33
2.1 Remedying government failures 1: the information 34
problem
2.1.1 Substantive rationality vs. procedural 34
rationality
2.1.2 Mitigating the information problem 35
2.2 Remedying government failures U: the rent-seeking 38
problem
vii
VIII Contents
2.2.1The vulnerability of the state 38

2.2.2The problem of collective action 40
2.2.3Competitiveness of thc rent-seeking proccss 41
2.2.4Sorne reflections: competitive markets and 44
competitive politics
2.3 A new institutionalist theory of state intervention 45
2.3.1 The nature of economic costs 46
2.3.2 Transaction costs and state intervention 48
Conclusion 53
3 The Political Economy of Industrial Policy 55

Introduction 55
3.1 The industrial policy debates 56
3.1.1 Does manufacturing matter? 56
3.1.2 What is industrial policy? 58
3.2 The logic of industrial policy J: the static dimension 61
3.2.1 The nature of the coordination problem 61
3.2.2 Industrial policy as a device of coordination 65
3.2.3 Concluding remarks: credibility, fairness 70
and flexibility
3.3 The logic of industrial policy 11: the dynamic 71
dimension
3.3.1 Knowledge, change, and evolution 72
3.3.2 Industrial policy as a device to 74
promote change
3.3.3 Concluding remarks: the socialisation of risk 78
3.4 Possible problems of industrial policy 79
3.4.1 Problems of information 79
3.4.2 Problems of rent-seeking and 82
entrepreneurship
3.4.3 Political problems: legitimacy and 85
democratic control
3.4.4 The problem of supporting institutions 87
Conclusion 89
4 Industrial Policy in Action - The Case of Korea 91

Introduction 91
Contents ix
4.1 A brief overview of the Korean economic 92

performance
4.2 Explaining the Korean experience 97
4.2.1 A free market? 97
4.2.2 Market-preserving state intervention? 101
4.3 Major themes of Korean state intervention 108
4.4 The economics of industrial policy in Korea 113
4.4.1 The evolution of Korean industrial policy 113
4.4.2 State-created rents and industrial 117
development
4.5 The politics of industrial policy in Korea 123
Conclusion 128
Conclusion: The Market, the State and Institutions 131
Notes and References 137
Bibliography 157
lndex 173
Preface to the 1996 Reprint
While one writes a book because one thinks that one has something to
say that will interest more than just a handful of people, for an
unknown first-time author, it was still a wonderful surprise to find out
that The Political Economy of Industrial Policy has been doing suf-
ficiently well to justify a paperback edition. Since the book was
published, 1 have received a surprising number of responses from many
people through various channels. There were sorne predictable
reactions, be they positive or negative, but there were also sorne which
I did not expect when I wrote the book. And for sorne of these reactions
I felt the need to respond in order to clear up the misunderstandings.
Moreover, as sorne of the issues I discussed in the book have seen a
flurry of debate in recent years, 1 felt that I shou]d add a word or two on
the more recent developments.
First of ali, my use of the concept of transaction costs in analysing
the relative costs of different forrns of coordination invited a lot of
justified and unjustified criticism. One criticism was that transaction
costs are perhaps not as important as the costs of coordination failure
themselves, and therefore that my emphasis on transaction costs was
diverting attention from a more important issue. This was a response
that I anticipated at various places in the book, but perhaps I did not do
enough to prevent such misunderstandings. 1 was by no means suggest-
ing that the costs from coordination failures were less important than
the transaction costs involved in correcting them, but simply that
correcting coordination failures is not without cost, as it is assumed in
many economic theories.
Toe other criticism that I faced due to my adoption of the term
'transaction costs' was that I was advocating the purely 'self-seeking'
view of individuals, as many who use the concept tend to be, and that
partly as a consequence of this I failed to see other forrns of
coordination such as 'networks' which are at least partly based on non-
selfish, cooperative behaviours. This was, again, an unfortunate
misunderstanding. While I certainly did not explore the network as an
altemative coordination mechanism to the market or state intervention,
this was not because I believed that human beings are so selfish they
are incapable of spontaneous cooperation. As I keep emphasising at
various places throughout the book, especially in the first two chapters,
xi
12 Preface to the 1996 Reprint
12
12
I do not subscribe to the pure, 'self-seeking' view of individuals. The
reason why I did not deal with mechanisms of coordination based on
horizontal cooperation was partly because I did not have enough
expertise in the area, and partly because the book was on 'policy' by
the state. Indeed, whether the state can make its policy implementation
more effective through promoting cooperation among the relevant
agents is one interesting issue that we should look at in future.
In choosing to use the term 'transaction costs', and also the term
'New lnstitutional Economics', I had sorne misgivings myself. My own
definition of transaction is somewhat broader than that used by many
other authors writing in the tradition of the New lnstitutional
Economics, and refers to the costs of running the economy or, as I also
call in the book, the 'costs of coordination' (see Chapter 2, Section
3.1). And when the term also carried certain ideological baggages to
which I do not fully subscribe, such as its emphasis on opportunistic
individual self-seeking, 1 was not sure whether I should use the term.
On the other hand, 1 did not want to create another term in our jargon-
infested profession when there is an already accepted term which could
convey the meaning sufficiently, if not perfectly, well. Hence my
decision to use the term. Unfortunately, although somewhat predictably,
this brought about unnecessary misunderstandings amongst many readers,
but I still feel that my decision to use the term
'transaction costs' was, on balance, justified.
When one writes on arguably one of the most contentious issues in
modern economic debate, namely, industrial policy, one is bound to
face the problem that by the time the book gets known to people, the
debate has moved on. The book was finalised, except for minor
editorial changes, at the end of 1992, and there has been a good deal of
controversy in the area since then.
The most prominent of the recent debates was due to the World
Bank's East Asían Miracle Report, which was published in late 1993.
The so-called 'Miracle Report' stirred up a good deal of controversy on
industrial policy by arguing that such policy attempted in the East
Asían countries was largely ineffective, perhaps except in Japan, and
that it is not something that other developing countries can try anyway
because they do not have the administrative and other institutional
capabilities to carry out such a demanding policy. While there is no
need to repeat the detailed criticisms by many authors, including
myself, of the methodologies and empírica! validities of the Miracle
Report' s view on industrial policy, 1 feel that I should present a few
salient points in summary form (for details, see the articles published in
Preface to the 1996 Reprint xiii
the special symposium in World Development, 1994, no. 4; essays in

Fishlow et al., 1994; and Chang, 1995).
First of all, the World Bank claims that the industrial structural
transformation that happened in these countries was 'market-
conforming' in the sense that their industries became capital-intensive
more or less in line with the rise in their income and thereby capital
stock. It then argues that this means these changes would have
happened even without industrial policy. One problem with this
argument is that with the rise in incorne, which the Bank takes as
something that drove the transformation in industrial structure, was
actually also the result of the industrial structural formation of the
economy, which was shaped by industrial policy. Also, the structural
transformation of the East Asían economies may be 'market-
conforming' in the very abstract sense that they started from labour-
intensive industries and only later moved to capital-intensive industries.
Yet there was little 'market conformity' in the process, as the particular
industries that developed over time were often selected and nurtured by
the state, rather than by spontaneous market forces. Moreover, even if
one makes the biggest concession and accepts that the transformation
of the East Asian industrial structure was market-conforming, this still
does not mean that industrial policy in East Asia had no impact because
the speed of such a transformation was certainly accelerated by
industrial policy. Anyhow, in the absence of actual examples of
market-based progress towards an advanced industrial economy during
this century (barring the very exceptional case of the colonial city-state,
Hong Kong), the burden of proof is really on those who believe that
such a progress is feasible.
Secondly, the World Bank rejects the effectiveness of industrial
policy in East Asia, especially in Korea, on the grounds that total factor
productivity in the 'promoted' industries such as machinery did not
grow as fast as that in the 'neglected' industries such as textiles. Even
ignoring the fact that the textile industry was in fact one of the most
heavily promoted in Korea (as its ability to earn foreign exchange was
highly prized by the policy-makers; see Chapter 4, Section 4.1), the
total factor productivity growth figures provided by the Miracle Report
should be taken with more than a pinch of salt. There are enough con-
tradicting evidences presented elsewhere, as one would expect given
the well-known problems of measuring total factor productivity growth.
It should also be added that the Miracle Report ignores other important
indicators of industrial policy performance such as balance of payments
contributions or spill-over effects beyond the two-digit industry level.
14 Preface to the 1996 Reprint
14
14Toe Miracle Report's last line of defence is that industrial policy
requires a high institutional capability - such as a capable bureaucracy
and a well-developed forum for government-business dialogue - and
therefore is not applicable to countries without such capabilities. At one
leve!, 1 could not agree with this statement more. lt is certainly
important to take into account one's administrative and institutional
capabilities when one is designing a policy. Moreover, 1 regard this as a
welcome improvement over the standard mainstream practice of
recommending an essentially same economic model - namely, an
idealised version of the Anglo-Saxon economic system - to a
bewilderingly wide range of countries regardless of their political and
institutional conditions, which naturally often produces very
disappointing results. Although one gets somewhat suspicious when the
same people who used to ignore política! and institutional conditions
when applying the Anglo-Saxon model suddenly become concerned
about them when it comes to applying the East Asían model, 1 take this
as a very positive sign that our profession is finally realising the
institutional diversity of capitalism emphasised in this book.
However, what is missing in the World Bank's new position
regarding the question of institutional capability is a discussion of how
a country that does not have an adequate institutional capability to
administer 'complex' policies can construct such capability. As I point
out at various places in the book, the East Asían countries did not start
their industrialisation efforts with a competent bureaucracy and good
institutions but consciously built them over time. Moreover, one should
not forget that, as in production, there is also learning-by-doing in
administration, and therefore that a bureaucracy which is intent on
eventually administering a complex set of policies should try to build
up its capabilities over time by engaging in sorne degree of complex
policies from the beginning. So if a country starts from the assumption
that it does not have an adequate ability to implement an intelligent
industrial policy and therefore gives up any attempt, it will never learn
to do it.
As I repeatedly argue in The Political Economy of Industrial Policy,
learning from other countries does not mean exactly copying their
institutions and policies - which is not going to be successful anyway -
but requires first an understanding of the principies that lie behind these
successful policies and institutions, and then constructing their
'functional equivalents' (which may or may not serve the given
function better) with due attention to local conditions. There is an old
Korean proverb that 'he who does not have teeth chews with his gums',
Preface to the 1996 Reprint XV
which nicely sums up the spirit of my argument. To develop this insight

further in the modero context, one could say that if one hasn't got teeth,
one should think about buying a set of false teeth or even a food
processor. To drive this point closer to home, I can do no better than
refer the readers to the quote in the third chapter of the present book
from the eminent scholar of Japan, Ronald Dore, on the replicability of
sorne Japanese institutions on British soil.
References
Chang, H-J. (1995) 'Explaining "Flexible Rigidities" in East Asia' in T. Killick

(ed.), The Flexible Economy (London: Routledge).
Fishlow, A., Gwin, C., Haggard, S., Rodrik, D. and Wade, R. (1994) Miracle or
Design? - Lessons from the East Asian Experience (Washington, D.C.:
Overseas Development Council).
World Development (1994) no. 4.
Acknowledgements
This book is a revised version of my PhD dissertation at the Faculty of

Economics and Politics, University of Cambridge. In writing the disser-
tation, and subsequently the book, 1 owe my biggest debt to my
supervisor Bob Rowthom. He has taught me how to analyse economic
problems with logical rigour without losing sight of history, politics and
institutions. He was always willing to listen to my half-baked ideas and
discuss them until I could clarify them. Most of ali, he has provided me
with much more moral support than I deserve. No amount of words can
express my gratitude to him.
Ajit Singh, Peter Nolan and Gabriel Palma have always been willing
to discuss any problem I had in writing the book and have provided me
with continued encouragement. 1 also want to thank John Sender for
teaching me the importance of balancing theory and evidence when
carrying out research and Ugo Pagano for introducing me to many of
the ideas that form the theoretical foundation of the book.
Mushtaq Khan deserves a special mention for bis continued intellec-
tual and personal friendship. Without him the book would never have
looked like it does now. 1 also want to thank Jong-il You, interaction
with whom in the last phase of the dissertation was extremely important
in clarifying the arguments. I also benefited a lot from intellectual inter-
action with Richard Wright, whose wideranging knowledge of econo-
mic history helped me to pul my arguments into perspective. Terence
Moll read the whole of the first draft and provided valuable theoretical
comments and superb editorial advice. Jon Di John helped me to clarify
my ideas in the final stage of the dissertation and provided editorial
ad vice.
Nathan Rosenberg, while he was visiting Cambridge for a year, read
about a half of what became the base material for the book and made
sorne important suggestions to strengthen my arguments, with which he
did not always agree. Alice Amsden, through her book on Korea and
personal contact, provided me with sorne very important theoretical
insights. Robín Matthews read parts of the manuscript and provided
sorne enlightening comments.
XVI
Acknowledgements xvii
Ashish Arora, Anuradha Basu, Pranab Bardhan, William Baumol,

Jagdish Bhagwati, David Canning, Hendrik du Toit, Geoff Harcourt,
Trina Haque, Ho-In Kang, Sandeep Kapur, Paul Kattuman, Zeljka Kozul,
Michael Landesmann, Jay-Min Lee, David Lehrnann, Sudipto Mundle,
Ramaswamy Ramana, Raja Rasiah and Helen Shapiro read parts of the
earlier versions of the manuscripl and made helpful comments.
In turning the dissertation into a book, 1 have benefited a lot from dis-
cussions with Christopher Freeman, Alan Hughes and John Toye. Al-
though I have not been able to incorporate ali the important points they
raised, their suggestions and critícísms will prove important in
developing my ideas in the future. 1 also want to thank Wendy Carlin,
Chris Edwards, Anclrew Glyn, Kyun Kim and Adrian Wood, who read
the dissertation and provided advice which proved valuable in the final
conversion process. 1 also wish to thank Ali Cheema, Anjali Mody and
Rathin Roy for their help with proof-rcading and indexing.
My cousin, Hawon Jang, and cousin-in-law, Chaewha Lim, who is not
with us any more, have provided me with a lot of love as well as intellec-
tual stimulation during my PhD years. Terry Fry and Amy Klatzkin, who
were my 'adopted family' during my first year abroad, provided me with
immense emotional and intcllectual support, which continues till these
days. Byong Kon Kim not only provided me with another adopted family
but also taught me about the world of philosophy and political science.
My brother Hasok has been a long-time intellectual and emotional
companion. Without him, my PhD would have been an extremely lonely
trip. My sister Yon Hee, brother-in-law Soo Bin Im, and their son Jay
Hyuck also provided me with a lot of love. Without them it would have
been impossible to survive sorne of those difficult days. Hee Jeong Kim,
my fiancée, walked into my life at the final stage of preparing the book
and made the job much more bearable. Finally, 1 want to thank my
parents, whose great love and intellectual support made me what I arn
now. 1 dedicate the book to them.
Cambridge, England HA-JOON CHANG

lntroduction: The Search for
New Pattems of Economic
Management
The quarter century after the Second World War was truly the 'golden
age' of capitalism (see Marglin and Schor (eds), 1990, and Armstrong
et al., 1991 ). Ali the major industrial countries in this period experi-
enced unprecedented growth with no serious inflation and near-full
employment, which resulted in a spectacular rise in general living
standards. This achievement was impressive enough to establish
Keynesian economic management, which dominated the economic
policy-making of the major industrial powers in this period, as the
ultimate way to runa viable capitalist system, or the 'mixed economy' .1
However the collapse of rapid growth and full employment in the
mid-1970s in most of the leading industrial economies was a rude
awakening to the fact that certain fundamental structural changes in
national and world economies occurred during the 'golden age'. The
rapid absorption of the agricultura) labour force into industries and
services eliminated surplus labour and created tight labour markets,
often with inflationary consequences. Investments in manufacturing by
multinational corporations emerged as an important form of ínter-
natíonal capital flow, and this, together with the lowering of barriers to
other forms of capital movements, made national economic manage-
ment much less effective, Japan became one of the world's leading
industrial countries and the East Asían newly industrialising countries
(henceforth NICs), that is, Hong Kong, Singapore, Taiwan, and South
Korea (henceforth Korea), becamc major actors in sorne world markets,
such as textiles, electronics, shipbuilding and steel. Together these
changes meant massive dislocations in the old heartlands of world
industrial production, that is, the 'rust belt' of the USA and the old
industrial centres of Western Europe, resulting in a fundamental
transformation in the map of wotld industries.
The rise of new-right ideas and the election of the Reagan and
Thatcher governments in the last years of the l 970s and the early
2 The Political Economy of Industrial Policy
1980s, partly prompted by the failure of the US and British

economies to adjust rapidly to these fundamental structural changes,
seemed to herald a 'brave new world' of monetarism and supply-side
economics.2 The apparent failure of the French socialist experiment
of unilateral Keynesianism in the early l 980s seemed to strengthen
the new right's belief in its new project of 'liberalising' economic
systems from the fetters of state intervention and other rigidities
created by the advance of collectivism in the guise of the welfare
state.
However, by the early 1990s, new-right experiments in the USA and
the UK were rapidly retreating, although they will have long-lasting
influences in sorne, if not ali, areas of the economy and society. 3 The
declining competitiveness of the UK economy (manifested in its
massive trade deficits) and the threatened position of the US economy
as the world's greatest economic power (manifested in its massive
budget and trade deficits) were probably enough to produce various
new attempts to examine alternative ways to run the capitalist econo-
my, or alternative patterns of economic management, especially in the
latter half of the 1980s.
One important common theme for these new attempts is the recogni-
tion that different capitalist countries have in fact long had distinctive
patterns of capital accumulation and economic policy-making. The
differential performances of different systems of economic rnanage-
ment, it may be argued, were not manifest in the goiden age when ali
the industrialised countries were in good shape, but the need for mas-
sive economic restructuring since the late 1970s has caused so much
strain that the differential abilities of different systems of economic
management to adapt to changes seem to have resulted in differential
economic performances.
One of the attempts to search for alternative patterns of economic
management is the literature on social corporatism.4 This literature
starts with the observation that sorne, notably Scandinavian, advanced
capitalist countries have not experienced the massive unernployrnent
that has characterised many advanced capitalist economies since the
late l 970s. These countries combined near-full employment with
relatively low inflation and constantly rising levels of income and
consumption for most members of society. Still more impressive, and
at the same time contrary to the conventional wisdom, was the fact that
these countries have restructured their economies successfully in a
relatively painless and less divisive manner despite strong trade unions
Introduction 3
(often accused of creating rigidities in the labour market which obstruct

structural change) and high taxation (often accused of being a dis-
incentive to hard work and entrepreneurial activity).
The basic explanation for this irnpressive performance is found in the
tripartite bargaining between the centralised ernployers' organisation,
the centralised trade unions and the state, through which the levels of
investment, wage and employment in the economy are coordinated to
ensure an effective structural change without causing high inflation or
mass unemployment.5 In addition, egalitarian wage policy (especially
in Sweden and Denmark) and 'active labour-market polícy' (retraining
of and migration assistance to the dislocated workers) are identified as
the rneans with which these countries have avoided inhurnane and
divisive forms of industrial restructuring in which a substantial rninority
of society bears the whole burden of restructuring, as happened in sorne
other advanced capitalist countries with very high and regionally
concentrated long-term unernployrnent (Rowthorn, 1990).
Another rnajor atternpt to find alternative patterns of econornic

managernent is represented by the so-called industrial policy debates.?
Industrial policy, despite its irnportance in rnany advanced capitalist
countries (notably Japan and France) and sorne East Asian NICs
(notably Korea, Taiwan and Singapore) during the post-war period, was
largely ignored until the late J 970s in the English-speaking acadernic
world. The concept of industrial policy does not seern to be recognised
as a legitirnate topic of acadernic discussion arnong econornists -
for example, The Palgrave Dictionary of Economics, despite its
wideranging interests, does not have 'industrial policy' asan entry.
Yet the
1980s saw a rapid rise in interest in this issue following the destruction
of old industrial heartlands in the USA and Western Europe, on the one
hand, and the spectacular industrial performances of Japan and the East
Asian NICs (especially South Korea as the biggest and rnost diversified
among thern), on the other.
The supporters of industrial policy argue that the succcss of East
Asian countries is prirnarily due to the active role of the state in
formulating a vigorous econornic systern that prornotes capital accumu-
lation, innovation and productivity growth.7 They argue that state inter-
vention in these countries is characterised by direct state involvernent
in the fate (especially investrnent and technical change) of individual
industries. They also argue that the failure of manufacturing in many
advanced capitalist economies, especially the USA and the UK, is not
just the result of macroeconomic mismanagement but of lagging

investment, innovation and productivity growth. They argue that the
manufacturing sector in these countries can only effectively be revived
through industrial policy, whose direct and particularistic nature stands
out in sharp contrast to the indirect and general nature of macro-
economic policy, which was the main pillar of the Keynesian pattern of
economic management.
The industrial policy literature, in contrast with the social corporatism

literature where a lot of theoretical advance has recently been made, has
suffered from the lack of a sound theoretical foundation. This is partly
because its proponents have been concerned largely with practica) policy
issues, but mainly because there was very little in conventional economic
theory that could justify such policy. In the absence of theoretical tools to
explain why industrial policy works, even its most ardent supporters
could do no more than either describe in detail how it works, say, in
Japan, without explaining why, or conclude that it works because of the
idiosyncratic cultures of the countries in which it is practiced. Fortunately
there have been sorne recent developments in economic theory that can
irr.prove this undesirable situation.
The recent rise of 'new institutional economics' has reminded us that
the market is not the only, or even predominant, way in which our econo-
mic life is organised.8 It has shown that non-market institutions are
integral parts of socio-economic life, and not necessarily unfortunate
rigidities that are best eliminated. Another important development in
economic theory has been the growing interest in technical change, which
has not been properly dealt with in conventional economic theory.9 The
recent literature on technological change has fundamentally changed our
view of the capitalist economy. Now we are increasingly inclined to see
the capitalist economy as in constant flux where learning plays an
important role. The conventional image of the capitalist economy as a
system fundamentally in equilibrium and subject only to extemal shocks
(which are responded to by instant and painless adjustments) is becoming
less and less appealing. The rise of 'government failure' literature has
played an important role in dispelling the 'welfarc economics' image of
the state as the selfless social guardian.l'' Toe literature has made us think
of how and why the state may fail to do what it is supposed to in textbook
welfare economics, and many other interventionist theories.
These theoretical developments, are of course, not totally new in their
ideas. Adam Smith, Karl Marx and Alfred Marshall were ali, in their own
lntroduction 5
ways, concemed about the development of the division of labour in the

capitalist economy and its implication for the organisation of economic
and political life. Adam Srnith, Karl Marx, Joseph Schumpeter and
Friedrich Hayek are only a few of the many great economists who
grappled with the idea that the capitalist economy is in a permanent flux
of technological development and institutional change. From Karl Marx
on the left to George Stigler on the right, generations of economists have
been concemed with the role of the state and the logic of interest-group
politics in the capitalist economy.
What is new is that these hitherto scattered ideas are now coming
together to produce an altemative to the conventional economic theory
whose elegant but overly formalistic models have ali but eliminated our
ability to deal with issues of economic change and the concomitant
conflicts among different groups in society. The application of the
recent theoretical developments in economics mentioned above to the
issue of industrial policy is the new and ambitious aim of the present
book.
STRUCTURE OF THE BOOK
The book is organised into four parts. The first chapter reviews the
Iiterature on state intervention. Without claiming to be exhaustive orto
be doing justice to ali individual theories, four groups of Iiterature are
reviewed, namely market-failure literature, contractarian Iiterature,
political-economy literature and government-failure literature. It is
argued that, even ignoring the moral and political reasons put forward,
respectively, by the contractarian literature and the political-economy
literature, we cannot fully accept the conventional case for state inter-
vention based on the existence of market failures because the very
process of correcting such failures may incur costs greater than the
efficiency gains from it, as the government-failure literature has
correctly pointed out.
In the second chapter we develop a new institutionalist theory of
state intervention. New institutional economics has largely concen-
trated on the theory of firm and contracts, but, as we wish to demon-
strate in this chapter, it can provide sorne interesting insights into the
question of state intervention. In this chapter we interpret the costs of
state intervention as transaction costs incurred ( 1) in policy decision
and implementation, and (2) as a result of the increase in the incentives
for the prívate sector to spend resources to change the property-rights
system by influencing the state. The standard recommendation from the

government-failure school in the face of such costs is to get rid of state
intervention altogether, but we argue that it is often better that these
costs are reduced through institutional and organisational innovations.
We then go a step further and argue that, while state intervention incurs
certain transaction costs, it can bring about benefits over and above the
efficiency gains from the correction of market failures by reducing the
transaction costs in the private sector.
The third chapter, which is the central chapter of the book, applies
the theory of state intervention developed in the previous chapter to the
issue of industrial policy. After briefly reviewing the industrial policy
debates, two dimensions of industrial policy are discussed, drawing on
Austrian, Schumpeterian and new-institutional economics. On the static
dimension, we discuss how industrial policy can overcome the limita-
tion of the market as a coordination mechanism. We discuss how issues
like credibility, fairness and flexibility have to be taken seriously for
the design of a successful industrial policy. On the dynarnic dimension,
the role of industrial policy in promoting industrial change is examined.
It is argued that changes often require coordination between agents and
therefore a certain degree of socialisation of risk may be beneficia) in
promoting industrial development. The chapter closes with a discussion
of the possible problems of industrial policy and remedies to them.
Problems of information, entrepreneurship, politics and supporting
institutions are discussed.
The last chapter looks at industrial policy in Korea as an example of
successful state intervention. We argue that, contrary to popular belief,
the rapid development of Korea owes a great deal to the efficiency of
state intervention, and especially of industrial policy. A detailed ac-
count of Korean industrial policy is given, and an explanation is
provided as to how Korea was able to maximise the benefits and
minimise the costs of industrial policy. Finally we examine the political
and institutional foundations of Korea's successful industrial policy.
While recognising the importance of historical and cultural factors, we
emphasise the role of conscious political innovation and institution
building in the construction of such foundations.
1 Theories of State
Intervention: A Review of
the Literature
INTRODUCTION
The role of the state in the capitalist econorny has been one of the rnost
controversial issues in econornics since the birth of the discipline (Deane,
1989). Almost everyone agrees that the state has a role to play, but there
is little agreernent as to when and how it should act. Perhaps the reason
why there is such little agreement is that state intervention is a complex
phenomenon involving many contentious issues such as efficiency,
morality, power, liberty and legitimacy, to name justa few.
In this chapter we organise our discussion into four parts to rnake this
complex issue more tractable. In the first section we examine analyses of
market failure which are mainly concemed with the possible failure of the
market mechanism to achieve Pareto efficiency and with the state's role
in overcorning such failure. Then we turn to the politico-philosphical de-
bate on whether the state as a political entity should or should not inter-
vene to correct the market outcome, be such correction efficient or not in
some sense. Thirdly we consider the political econorny - both right-wing
and left-wing - literature, which asks whether it is correct to assurne that
the state serves sorne 'public' or 'social' purpose rather than individual or
group interests. Lastly we examine the govemment-failure literature,
which asks whether the state has the ability to intervene effectively,
whatever its intention is.
1.1 EFFIClENCY: THEMARKET-FAILURELITERATURE
The most developed literature on state intervention is that of market fail-

ure or welfare econornics. The literature is primarily concemed with the
failure of the market mechanism in equating prívate and social costs and
benefits and with the possible correctives to such failures through state
intervention.1 We will examine three groups of arguments in this tradi-
tion, that is, public goods, non-competitive markets and extemalities.
7
1.1.1 Public Goods
In the economics literature, goods are usually classified into two

categories, that is, private and public. A private good is something that
can only be consumed by those who have paid for it, whereas a public
good is something that can be consumed by those who did not pay for it
as well as those who have paid for it. In other words, supplying a public
good to somebody means supplying it to others, not simply because it
has to be 'jointly-supplied' or it has 'non-rivalness in consumption',
but more importantly because it is not economically feasible to exclude
the non-payers. That is, the defining characteristic of a public good is
its 'non-excludability' (Olson, 1965, pp. 14-15, 38-40).2
Since it is possible for one to consume the good without paying for it
once somebody else has paid for it, there is always an incentive to
understate one's preference for a public good (Schotter, 1985,
pp. 57-63). The possibility of such strategic behaviour means that
public goods are Iikely to be underprovided due to the free-rider prob-
lem or the problem of collective action (see Olson, 1965; Hardin, 1982;
Elster, 1989). In such a situation, individual rationality leads to collective
irrationality, because, through individuals' attempts to maximise their
own net benefits (paying as little as possible, given the benefit),
everybody ends up suffering from the underprovision. There- fore, in
order to provide the optimal amount of a public good, it is argued,
the state needs to intervene by taxing people and providing public goods
with the revenue (see, for example, Cullis and Jones,
1987, p. 19).3
One of the criticisms of the public-good argument is that non-
excludability, the fundamental condition for a good to be a public good,
cannot be regarded as pennanent. According to Peacock ( 1979a),
_technological innovation can eliminate the 'publicness' of sorne goods
by solving the problem of non-excludability. He argues that 'even in
the most farnous "polar case" of a public good, ways and means can be
found ... by which the problem of non-excludability can be solved'
(p. 133). Although this is a valid point, it should not be taken as
implying that technical progress will eventually solve ali public-good
problems (for example substitution of lighthouses with radio signals),
because excludability is more a property-rights problem than a
technological problem (see Note 2).
Another important criticism of the argument for state intervention
based on the public-good consideration is that it does not necessarily jus-
Theories of State lntervention 9
tify state intervention.4 That is, even if there is a public (collective) good,
it is not clear whether the 'collectivity' to provide it should be the state.
As Olson (1965) argues, in a small group, optimal public-good provision
may be achieved without state intervention because in a small group it is
likely that there are individual members who gain so much benefit from
the public good concerned that they are better off providing thc good uni-
laterally (pp. 43-52). And even in a large group where this condition
does not obtain, state intervention is not always necessary. Public goods
may be optimally provided even in a large number setting, if sorne
'selective incentives' (Olson, 1965) in the form ofprivate goods provided
by 'political entrepreneurs' (Popkin, 1979) can overcome the free-rider
problem by bringing individual cost/benefit structures in line with the
social (or group) cost/benefit structure. Of course these arguments do not
allow us to conclude that the public-good problem can always be solved
by prívate initiatives. In many cases, the use of coercion by the state (for
example taxation) may be the only possible way to resolve the problem.t
1.1.2 Non-Competitive Markets
The existence of scale economies and/or collusive behaviour can result

in non-competitive market structures, where individual producers' deci-
sions can affect the quantity and price in the market. When monopoly or
oligopoly prevails in a market, the quantity of goods supplied is smaller
than in the competitive context, as firms face negatively sloped demand
curves.6 Thus, in a non-competitive market, sorne consumer surplus will
be transferred to firms in the form of 'monopoly profit', and such a
transfer will impose a 'deadweight loss' to society (the Marshallian
triangle). In this case, it is argued, it is justifiable for the state to inter-
vene to guarantee the optima! output, that is, the output which would
have been provided in a competitive setting.7 Anti-trust Iegislation,
which may involve the regulation of pricing and the breaking up of
existing monopolies, is the most frequently used interventionist measure
in this regard, but public ownership is another commonly practised and
probably more powerful measure.f
A powerful argument against state intervention to 'correct' non-

competitive markets is that based on the theory of 'second best' (Lipsey
and Lancaster, 1956), which points out that rectifying monopolistic situ-
ations (or any other price distortion) in sorne, but not ali, markets may
not necessarily improve the efficiency of the economy. On the basis of
this theory, anti-interventionists argue that no gain is guaranteed by
anti-trust intervention which does not eliminate 'distortions' from ali

markets (Peacock and Rowley, l979a). However this argument does
not discredit fully the interventionist argument because the theory of
second best does not imply that there can be no gain from such inter-
vention, only that it is not guaranteed. Actually the theory of second
best may be used as an excuse for wholesale state intervention, because
no global optimality is attainable without letting ali the other markets
depart from their local optima] conditions.
Another anti-interventionist argument related to non-competitive mar-
kets is that state intervention is one major source of such outcomes. For
example, Friedman (1962) argues that, '[m]onopoly frequently, if not
generally, arises from government support of collusive agreements
among individuals' (p. 28). The implication, then, is that the state should
stop meddling with the market if it is serious about correcting distortions
in non-competitive markets. In the words ofMises (1979), '[i]t is absurd
to see the government ... point its finger at business, saying: "There are
cartels, therefore government interference with business is necessary", It
would be much simpler to avoid cartels by ending the government's
interference with the market - an interference which malees these cartels
possible [such as protectionism]' (p. 52). Although there can be no doubt
that state-administered entry barriers are often sources of monopoly, it
should be pointed out that many, if not ali, initially competitive markets -
through business cycles, structural changes or even sheer luck - have
been transforrned into non-competitive ones without collusion or state
mtervention.? lndeed, if the assumption of self- perpetuating competitive
markets is unrealistic, it may be pointless from the public-policy point of
view to argue for or against state intervention in monopolistic markets,
since there is no criterion according to which they are to be corrected
other than the fictitious perfectly competitive market (Demsetz, 1964).
1.i.3 Extemalities
One of the assumptions in textbook economics is that each individual

has only to consider his/her own means and ends. Technically speak-
ing, individual preference systems (or utility functions) and production
functions are independent from those of others. Externalities exist
where there are sorne spill-over effects from an individual's activities
to those of others, leading to a discrepancy between the prívate cost/
benefit structure and the social cost/benefit structure. Of course, inter-
dependence among individual activities cannot be a problem in itself,
Theories of State Intervention 11
but where its effects are not properly compensated for there arises an
'untraded interdependence' (Nath, 1973, p. 43), namely extemaiity.
At least in principie, it may be possible to overcome this problem by
defining property rights more precisely and having negotiations be-
tween the parties affecting and affected (for exarnple the owner of the
smoke-spewing factory compensating housewives living nearby who
have to spend extra time and money on laundry). However, in many
cases, it is economically impossible to do so, because of the transaction
costs involved in infonnation acquisition, negotiation and contract en-
forcement (Coase, 1960; Stigler, 1975, pp. 106-7; Dahlman, 1979). In
the absence of such a 'property-rights' solution, the state is justified in
ensuring the provision of goods with externalities in socially optima)
amounts through other means. State provision of goods with positive
externalities at subsidised prices (for example education, health, social
infrastructure), subsidisation of those who create positive externalities
(for example, subsidies to R&D), and taxation of those who create
negative externalities (for example pollution tax), are examples of state
intervention on extemality grounds.
Sorne anti-interventionists try to dismiss the extemality argument for

state intervention by assuming that the magnitudes involved are negli-
gible. However, as Baumol ( 1965) and others have pointed out, the Iist of
extemalities can be extended almost infinitely. Most goods create sorne
negative extemalities in their production processes in the fonn of pollu-
tion, except in those few cases where proper compensation is made.
When considering linkage effects (Hirschman, 1958, ch. 6) or pecuniary
externalities (Scitovsky, 1954), many goods may additionally be
classified as having positive extemalities. Sorne economists even argue
that sorne goods that have conventionally been treated as lacking ex-
ternalities, say basic foodstuff, can be seen as creating extemalities when
they are not consumed in the proper amount and therefore induce crime
(Schotter, 1985, pp. 68-80). Moreover, there exists interdependence
between individual preferences. Por example, people have what Elster
(1983, ch. 2) calls counteradaptive preferences - 'the grass is always
greener on the other side of the fence'. Toe psychology of luxury-good
consumption - part of one's pleasure derives from the fact that one con-
sumes what others do not - is another example of interdependent con-
sumer preference. Indeed once we begin to accept the pervasiveness of
externalities, it seems questionable whether we are justified in having
market transactions at ali. The important issue here is not whether
extemalities exist or not, but to explore under what conditions market
12 The Política/
TheoriesEconomy
of State of
lntervention
Industrial Policy 12
12 12
transactions will be (or should be) adopted and under what conditions
non-market institutions, including state intervention, will be (or should
be) adopted, as new institutional economics has recently tried.
Sorne other anti-interventionists argue that correcting one set of
externalities leads to another. Friedman (1962), for example, has grave
reservations about state intervention based on the externalities argu-
ment - the 'neighbourhood-effects' argument as he calls it - pointing
out that: (a) it will in part introduce an additional set of neighbourhood
effects by failing to charge orto compensate individuals properly; and
(b) it creates externalities in the form of 'threatening freedom' (p. 32).
This is unconvincing since: (a) whether the gains from eliminating exist-
ing externalities are smaller than the losses from the newly-created
externalities cannot be determined a priori; and (b) unsolved external-
ities also mean the limiting of someone's freedom - the affecting in the
case of positive externalities, the affected in the case of negative ex-
ternalíties - because one party has the desire to trade but cannot
(recall that externalities are untraded interdependences).1 º
1.2 MORALITY: PATERNALISM AND CONTRACTARIANISM In
the previous section we discussed the market-failure argument, the

most important element in the traditional argument for state interven-
tion. Another important element in the traditional interventionist argu-
ment is the moralistic argument that the state, as a representative of the
members of society, may intervene in the market, if necessary, at the
cost of efficiency. Can this argument be justified?
1.2.1 Paternalism and Contractarianism
The moralistic argument for state intervention usually takes two forms.
Firstly, it is argued that the state may intervene in the provision of 'rnerit
goods', which are 'goods the provision of which society (as distinct
from the preferences of the individual consumer) wishes to encourage
or, in the case of demerit goods, to deter' (Musgrave and Musgrave,
1984, p. 78). Secondly, state intervention may also be justified if society
believes that market-type transactions are not morally acceptable in
sorne areas, for example blood donations or police services. In this case,
the argument goes, the state, as the social guardián, should remove such
activities from the domain of the market and conduct them itself.
13 The Política/
TheoriesEconomy
of State of
lntervention
13 13
The above argument is often branded as paternalism by those who

believe in methodological individualism and its politico-philosophic
counterpart, contractarianism (for an exposition of contractarianism,
see Nozick, 1974). From the methodologically individualistic point of
view, the above argument is flawed in that it attaches an independent
will to society, which is no more than a collection of individuals. From
this politico-philosophic point of view, the belief that the state should
decide on what individuals should produce and consume and in what
ways is a first step on 'the road to serfdom' (Hayek, 1972).
Those who believe in individualism-contractarianism argue that 'the
individual and not the group should be the basic repository of rights and
obligations' (Schotter, 1985, p. 18). Thus it is believed that 'individuals
should be allowed, within defined limits, to follow their own values and
preferences rather than somebody else's .... lt is this recognition of the
individual as the ultimate judge of his ends, the belief that as far as
possible his own views ought to govern his action, that forms the essence
of the individualist position' (Hayek, 1972, p. 59). Of course, this
position 'does not assume, as is often asserted, that man is egoistic or
selfish or ought to be' (Hayek, 1972, p. 59). However, no other person or
authority can impose his/her or its own ethical judgement on the
individual, since the individual knows best what his situation is and what
his best option in that situation is. Any interference with the making of
individual decisions is seen as violating the innate right to freedom of the
individual (Hayek, 1972, 1988; Mises, 1929, 1979; Friedman, 1962).
lf we do not, or rather should not, introduce any exogenous ethical
code other than that of the individual being the judge of his/her own
destiny, contractarianism, whether of the Hobbesian or the Rawlsian
variety, becomes the only consistent view of the state, or more gen-
erally, of politics. As Buchanan says, 'if politics is to be interpreted in
any justificatory or legitimatising sense without the introduction of
supra-individual value norms [emphasis added], it must be modelled as a
process within which individuals, with separate and potentially differ- ing
interests and values, interact for the purpose of securing individually
valued benefits of cooperative effort. lf this presupposition about the
nature of politics is accepted, the ultimate model of politics is Con-
tractarian [emphasis original]' (Buchanan, 1986, p. 240). The ideal
state is then the product of voluntary contracts between free individuals
who found sorne potential gains in restricting the unfettered exercise of
the individual free will of other individuals and of their own. Contract-
arianism in its most consistent form should be based on the unanimity
14 The Politícal
TheoriesEconomy
of Stateof
lntervention
14 14
rule. Otherwise, however fair the outcome of a politico-economic system

may appear from sorne other point of view, it cannot be justified from the
contractarian viewpoint. Imposing a decision on individuals (which is
bound to happen except under the unanimity rule) violates the funda-
mental principie that individuals should be free to malee contracts and
not be coerced into any transaction, however beneficial it may appear to
an outsider.
1.2.2 What should the State do?
On the basis of individualism-contractarianisrn, it is argued that any state

intervention other than sorne (rather ill-defined) minimal functions is
illegitimate because it violates individual freedom as the ultimate value in
human society. Indeed, according to Peacock and Rowley (1979b), what
they tenn 'Iiberalism' is to be 'prepared to trade off economic efficiency
for individual freedom where such a policy conflict becomes apparent'
(p. 26).11 .
Mises (1979) categorically states that 'the government's only legit-
imate function is ... to produce security' (p. 40). To him, any interven-
tionist attempt is doomed to invite in more and more intervention, if the
state is serious about achieving its original purpose, leading inevitably to
socialism. There is no such thing as 'the third way' (Mises, 19Z9, 1979).
Hayek (1972) argues that state intervention, except in areas that can
be justified on contractarian grounds plus sorne 'even-handed' interven-
tions in non-exchange economic activities (for example production
activities), is bound to erode individual freedom. Competition (and íts
prerequisite, freedom of entry) is seen as the best means of coordinat-
ing the economy 'not only because it is in most circumstances the most
efficient method known but even more because it is the only method by
which our activities can be adjusted to each other without coercive or
arbitrary intervention of authority' (Hayek, 1972, p. 36). That is, state
intervention is objectionable not only because it is inefficient but mainly
because it violates the fundamental values of individualism and hence
of contractaríanism.P
Buchanan (1986) gives a typology of the state functions that can be
justified on contractarian grounds. According to him there are three
levels of collective action. At the first level are activities involving
enforcement of the law. The role of the state here is that of an umpire
of a game. The second level involves collective action within the lirnits
of existing laws. Here the role of the state is the financing and provision
15 The Politícal
TheoriesEconomy
of Stateof
lntervention
15 15
of public goods and services. The third leve) involves changes in the
law itself, namely changes in the rules of the game. Unless there is a
unanimous case for changes in the basic rules of the game of society,
the role of the state should be either the enforcement of the rules of the
game - namely the protection of property (and human?) rights and the
enforcernent of voluntary contracts - and the provision of special goods
and services whose private provision will be suboptimal from the social
point of view - namely the provision of public goods.
Friedman, usually known as the anti-interventionist economist, states
that '[t]he role of government ... is to do something that the market
cannot do for itself, namely, to determine, arbitrate, and enforce the rules
of the game' (Friedman, 1962, p. 27). This view is similar to that of
Buchanan, but in fact Friedman is much more generous, if more vague,
about state intervention than the strict contractarians cited above. His list
of legitimate functions of the state is as follows: maintenance of law and
order, definition of property rights, service as a means whereby people
modify property rights and other rules of the economic game, adjudica-
tion of disputes about the interpretation of the rules, enforcement of con-
tracts, promotion of competition, provision of a monetary framework,
engagement in activities to counter technical monopolies and to over-
come neighbourhood effects widely regarded as sufficiently important to
justify government intervention, supplementation of private charity and
the private family in protecting the irresponsible, whether madman or
child (Friedman, 1962, p. 34).
1.2.3 Sorne Reflections
Denouncing any moral judgment other than those based on narrowly

defined individualism, as contractarian economists tend to do, is as
meaningless as citing moral reasons for state intervention without dis-
cussing the role of morality in our social and economic life (McPherson,
1984). Individualism is not a 'scientific' point of view which can do
without morality, as is frequently contended (for example Friedman,
1962), it is no more than a particular form of morality.
Methodological individualists assume that each individual knows
best his/her interest and the constraints he/she faces. This view is not
without its problems. First of ali there are individuals who even con-
tractarians do not regard as wholly responsible (for example, madmen
and children), but the borderline between 'normal' and 'abnormal' is
ambiguous, as even Friedman admits (Friedman, 1962, pp. 33-4). And,
16 The Political
TheoriesEconomy
of Stateof Industrial Policy
lntervention 16
16 16
more importantly, who has the right to decide who is responsible and
who is not, if there should be no supraindividual value? Secondly, there
is 'no intrinsic reason why individuals should always pursue their own
good or why they will always do so better than others can do it for
them' (Freeden, 1991, p. 89). As Goodin (1986) notes, people make
decisions on incomplete information, ignorance of their future prefer-
ences, ignorance of the full consequences of their own actions, deceptive
decision frameworks, the desire to avoid responsibility for risks. The
existence of implicit preferences for preferences (for example reckless
drivers and drug addicts are not to be seen as acting in their best inter-
ests) is another case in point. In such situations it is not clear whether
we should regard individual decisions as the manifestation of their
preferences, and therefore argue against ali intervention.
In contractarian philosophy it is argued that the state cannot be
regarded as being 'above' individuals, since it is a product of free con-
tracts between independent individuals. Contractarians hypothesise a
'state of nature' where ali individuals are free to make contracts but are
involved in a state of war against everybody else, which leads to the
need for the imposition of an extra-individual authority, through volun-
tary contracts, in the form of the state.
This 'state-of-nature' scenario is of course a fiction. During the his-
tory of mankind, the choice has been the one between one form of
authority and another, and not the one between anarchy and authority,
as the contractarians put it (for example Nozick, 1974). Even a cursory
look at the history of the last few centuries reveals that the building of
the modern state was largely initiated by rulers, and not by freely-
contracting indi viduals (Poggi, 1990). Moreover it is in contradiction to
historical truth to argue that market-type transactions brought about the
state - the opposite view is more correct. The market in its present form
is a newer form of social institution compared with other forros, includ-
ing the state. As Polanyi (1957) persuasively puts it, historical
experience shows us that:
[t]he road to the free market was opened and kept open by an
enormous increase in continuous, centrally organised and controlled
interventionism [emphasis added]. To make Adam Smith's 'simple
and natural liberty' compatible with the needs of a human society was
a most complicated affair. Witness the complexity of the provisions in
the innumerable enclosure laws; the amount of bureaucratic control
involved in the administration of the New Poor Laws which for the
17 The Political
TheoriesEconomy
lntervention 17
17 17
first time since Queen Elizabeth's reign were effectively supervised

by central authority; or the increase in govemmental administration
entailed in the meritorious task of municipal reform .... Admin-
istrators had to be constantly on the watch to ensure the free working
of the system. Thus even those who wished most ardently to free the
state from ali unnecessary duties, and whose whole philosophy de-
manded the restriction of state activities, could not but entrust the
self-same state with the new powers, organs, and instruments required
for the establishment of laissez-faire (p. 140).
Of course most contractarians are not so silly as to believe in the 'state

of nature' as a historical reality. Buchanan argues that the contractarian
argument is an ex post conceptualisation or legitimisation of the
política) process as a complex exchange relationship and notan ex ante
moral justification of the existing political order (Buchanan, 1986,
p. 247). He thus admits that the contractarian argument is not based on
actual history but on sorne arbitrary belief.13 Nozick (1974) tries to
defend the state-of-nature type explanation by saying that '(w]e learn
much by seeing how the state could have arisen, even if it did not arise
that way. If it did not arise that way, we also would learn much by
determining why it did not; by trying to explain why the particular bit
of the real world that diverges from the st.ate-of-nature model is as it is'
(Nozick, 1974, p. 9). This is peculiar logic. Since it is definitionally
impossible that two or more different end-states arise through the same
process, it is hardly justifiable to introduce one of them as a 'potential
explanation' (as Nozick calls it) of the other. Of course there may be
some heuristic value in the exercise, but this advocacy based on
heuristic value can only be fully justified when the 'as if explanation is
complemented by the 'as it was' one, unless it can be shown precisely
why the 'as if explanation is superior to the 'as it was' one.
More importan ti y, contractarianism does not necessarily guarantee a
minimal state. For example, if there is a unanimous belief among the
members of society that market-type transactions are not morally
acceptable in sorne areas (for example blood donation, defence), taking
such activities out of the market may be justified even from the contract-
arian point of view. That is, what if the individuals in the society gather
together to write a new social contract that endorses an interventionist
state?l4 It is not satisfactory to argue for a mínima) state on the
assumption that individuals will opt for a minimal state if they are given
a chance, unless we can show that individuals are wealth-maximisers (a
18 The Political
TheoriesEconomy
lntervention 18
18 18
common assumption in economics, but nothing more than an assurnp-

tion) and that the free market will ensure them the maximisation of their
wealth (a proposition that has been proven problematic by welfare
economics).
1.3 INTENTION: THE POLITICAL ECONOMY LITERATURE The
political-economy viewpoints, both on the left (for example the

Marxists) and the right (for example the Chicago school) of the politi-
cal spectrum, have for long criticised the market-failure approach for
too readily assuming that the state will act like Plato's Philosopher K.ing.
That is, is it the intentíon, or the objective, of the state to serve the public?
Below, we will examine three arguments questioning such assum-
ptions, namely the autonomous-state approach, the interest-group
approach and the self-seeking-bureaucrats approach.
1.3.1 The Autonomous-State Approach
The view that the state should be regarded as 'a dynamic independent
force' (Findlay, 1990, p. 195) with its own objective function that is
distinct from that of the society as a whole is not new. A stream within
the Marxist tradition, originating from Marx's Eighteenth Brumaire of
Louis Bonaparte (Marx, 1934), has recognised that a certain state may
acquire an 'autonomy' from society, if no class is powerful enough to
impose its will on the state (for example, Alavi, 1972).15 One strand in
the recently popular 'neoclassical political economy' goes a step further
and characterises the state as a 'predator', which, acting as a discrimi-
nating monopolist, develops a property-rights structure and a tax
system which maximise its 'profit' or net revenue (tax minus expend-
iture), if necessary at the expense of social productivity.16 Of course
neoclassical political economy recognises that revenue-maximisation
by the state is an exercise in constrained maximisation, since the threat
of takeover by an altemative ruler from within or without the country
imposes a competitive constraint (see North, 1981, ch. 3; Findlay,
1988; Eggertsson, 1990, ch. 10).
The view that the state may actas an entity with its own will (and greed)
is a useful antidote to the naive assumption of welfare economics that it
will correct market failures as soon as it finds them. Moreover, when the
traditional interest-group approach has treated the state as a black box in
19 The Political
TheoriesEconomy
lntervention 19
19 19
which interest groups feed their policy inputs, resulting in sorne

disappointingly sirnplistic policy irnplications, there seerns to be sorne
value in seeing the state as an autonornous entity (Skocpol, 1985; also
see 1.3.2 below). However the approach is not without its problerns.
First of ali, the approach treats the 'autonornous' or 'predatory' state
as a unified entity. As recognised e ven by one of the earlier proponents
of the predatory-state approach, the richer institutional context of the
rnodern polity with a bureaucracy and (frequently) a working legis-
lature makes it difficult to apply this sirnplistic rnodel to an analysis of
real-life examples (North, 1990a). In fact the rnost distinctive charac-
teristic of the modern state is the development of institutional bounds
on arhitrium (Poggi, 1990, pp. 74-6). In particular, the 'necessity of
developing agents (a bureaucracy) to monitor, meter, and collect
revenue' (North, 1990a, p. 190) introduces a complication in the form
of a conflict of interests between the ruler, who wants to maximise net
revenue, and the bureaucrats, who want to maximise the budgets of
their own bureaux (see 1.3.3 below for further discussion).
Secondly, talking of state autonomy in the abstract is not very help-
ful for the understanding of real life problems. Whether we call a state
autonomous or not should depend on what issues we are interested in. First,
one may wish to investigate the effect of their different degree of
autonomy on the actions of different states (for example the Taiwanese
state is more autonomous than the Indian state) or the same state at
different points of time (for example, the Japanese state in the late nine-
teenth century was more autonomous than it is now; the Singaporean
state may become less autonomous in the future than it has been under
Li Kwan Yew). Second, different states may have different degrees of
autonomy in different areas. For example the Swedish state may be less
autonomous than the Korean state in influencing investrnent decisions
of capitalists, but may be more autonomous in taxing them. Whether or
not one should assurne autonomy of the state depends on the country
one wants to Iook at, the time period one wants to study and the areas
of policy one is interested in.
1.3.2 The Interest-Group Approach
Another group of arguments which question the intention of the state is

what we call the interest-group approach. This approach sees the state
as 'an arena within which econornic interest groups or normative social
rnovements contended or allied with one another to shape the making
20
of public policy decisions' about 'the allocations of benefits among
demanding groups' (Skocpol, 1985, p. 4).17 And, the argument goes,
since the most powerful groups will be most able to affect the decisions
of the state, state economic policies will be inevitably biased towards
them.
Toe most representative of these theories is the 'regulatory-capture'
theory of the Chicago school. According to Stigler (1975), the pioneer
of this theory, 'regulation is acquired by the industry and is designed
and operated primarily for its benefit' (p. 114) through subsidies, entry
restrictions, restrictions on substitutes and subsidies to complements,
and price-fixing. Political control over this process is limited by the
infrequency of voting and the high cost of acquiring infonnation on the
side of the voters (Laffont and Tirole, 1988). Introducing the problem
of collective action (due to Olson, 1965) in a more explicit manner,
Peltzman (1976) argues that the reason why producers - rather than
consumers - capture the regulatory agency is that their smaller number
makes collective action easier. This version of the interest-group
approach then prescribes that the best way to avoid the possibility of
regulatory capture is to deprive the state of the power to regulate.
Sorne Marxists have argued along similar lines, although with
different political connotations. They argue that the state, whose exist-
ence depends on the reproduction of the dominant mode of production
in society, has to serve the interests of the economically dominant class
in that society - that is, the capitalist class in the capitalist society
(Miliband, 1969; O'Connor, 1973; Gough, 1979). In the simpler version
of the theory, the state is seen as defending the capitalist class interest, if
necessary at the cost of individual capitalists' interests. In this view the
solution to the problem of divergence between the 'public interest' and
the objectives of the state is to overthrow capitalism and therefore get
rid of class divisíons and relations of domination, because the public
whose interest the state serves will then become the whole of society.
Toe interest-group approach is useful as a broad framework to under-
stand politics, since it enables us to see how the public is not an homo-
geneous entity but is made up of diverse groups which struggle with
each other to affect the decisions of the state. However, apart from
sorne sophisticated Marxist versions, this approach has the following
shortcomi n gs.
First of all, many versions of the approach do not adequately discuss
the problem of collective action. As was pointed out by Olson (1965),
Hardin (1982) and Hindess (1987), the traditional theories of interest-
Theories of State Intervention 21
group politics (both the pluralist and the Marxist varieties) have too
readily assumed that all existing interests will be represented. However,
even if people have identical preferences they are not necessarily able to
act them out due to the difficulty of collective action. As many studies in
the tradition of the theory of collective action have shown, which groups
can assert their interests depends on how large the group is and whether
it is feasible to devise sorne selective incentive/sanction scheme. In
addition, the frequency and duration of the interaction among the mem-
bers of the group (Axelrod, 1984; Taylor, 1987) and ideology (North,
1981) also matter for the possibility of successful collective action.
Moreover there is the problem of 'agenda formation'. It is not true that
all interests can be organised once the configuration of the group is such
that the free-rider problem can be overcome. How interests are
represented also depends on which issues are more easily put on the
public agenda. In a country at any point of time, there exist social norms
(or ideologies) according to which sorne issues simply cannot be put on
the agenda for public action, and therefore it is almost impossible to
organise an interest group around such issues.18 'Certain kinds of
argument, powerful though they may be in private deliberations, simply
cannot be put in a public forum' (Goodin, 1986, p. 87), or 'although both
individual material interests and the interest of particular groups and
classes are still the essential forces at play in the politics ... the form in
which those interests are publicly expressed and argued over allows for a
real "distance" to open up between political and public debate and those
interests' (Kitching, 1983, p. 61 ). The motivation and the capacity of the
state also become important variables in determining what kinds of policy
alternatives can be discussed, adopted, and implemented, because
government Ieaders and bureaucrats often take initiatives well beyond the
demands of social groups (Skocpol, 1985). In other words, the existing
forms of politics, the structure of the state apparatus, the prevailing
ideologies (or social norms) and so on affect what kinds of interest groups
can and would be formed (see Skocpol, 1985, for a detailed discussion).
Thirdly, most versions of the interest-group approach do not ad-
equately deal with the process of interest-group politics. Rather, their
analysis is 'structuralist' in the sense that the outcome of interest-group
politics is seen as predetermined by systemic parameters. Given such a
static view of politics, it is more than natural that the proponents of the
approach think that the problem of 'biased representation' could never
be resolved without destroying the existing social structure (for example
a return to the minimal state, the socialist revolution). However, what
22
rnatters in reality is not only whether a particular social structure allows
sorne or other group to dominate, but also the nature of the process of
gaining such dominance - in other words, how surplus is appropriated is
as, if not more, irnportant as who gets the surplus (Khan, 1989). For
example, in sorne South Asian countries, mobilisation of clientelist
groups has become the rnajor channel for surplus appropriation, with
sorne detrimental consequences for capital accumulation (Bardhan,
1984, for India; Khan, 1989, for Bangladesh). Another example is Latín
America, where the fractured nature of society produces a volatile
pattern of capital accumulation where a growth cycle starts only to be
quickly disrupted by hyperinflation and then needs a major social
upheaval (for example military dictatorship, austerity programmes) to
resume itself (Fishlow, 1990; Amadeo and Banuri, 1991 ). Interest-group
politics may be properly understood only when the particular process of
contest for political and economic rights in the society concerned are
analysed in detail (for a seminal work in this vein, see Khan, 1989).
1.3.3 The Selí'-Seeking-Bureaucrats Approach
Another important critique of the 'benevolence' assumption in welfare

economics is what we call the theory of self-seeking bureaucrats. Toe
theory is based on the postulate that bureaucrats are in no sense different
frorn other individuals in pursuing their own intercsts. It is absurd, the
argurnent goes, to believe that one and the same individual will behave
altruistically in the office and egoistically after office hours. It is assumed
that the bureaucrats are budget-maximisers, following Niskanen's argu-
ment that '(a]mong the severa! variables that may enter the bureaucrats'
motives are: salary, perquisites of the office, public reputation, power,
patronage, output of the bureau, ease of making changes, and ease of
managing the bureau. AII except the last two are a positive function of
the total budget of the bureau during the bureaucrat's tenure' (Niskanen,
1973, p. 22). Since bureaucrats derive utility from higher salaries and
greater power of their bureaux, it is rational for them to maximise the
budget of their bureaux rather than to optimise the social output.
Although the vote-rnaxirnising behaviour of politicians rnay impose
sorne constraints on the size of the budget (because the politicians, who
have to be re-elected, do not want too high laxes), the outcome is likely
to be in favour of the bureaucracy. This is dueto the fact that the politi-
cians are atan infonnational disadvantage concerning the cost functions
of the bureaux, not only beca use they lack the expertise to estímate such
functions but also because the state bureaux are in monopolistic posi-
tions and therefore there is no criterion to gauge their efficiency
(Mueller, 1979, p. 157). The politicians, even when they are public-
minded, have no more power than to monitor the bureaucracy according
to the crude criterion that the total costs of state expenditure should not
exceed total benefits (Cullis and Jones, 1987, p. 127). It is thus argued
that the bureaucrats, acting as rational and selfish agents, will produce
the goods and services under state provision in more than a socially
optima! quantity.19
The self-seeking-bureaucrats approach can be criticised on the following
grounds. First of ali, the scope for the realisation of bureaucratic self-
interest through overprovision of public goods and services differs
according to the institutional setting and the political process around the
bureaucracy. For example, if bureaucrats are recruited through written
sitting tests (as in Japan or Korea), it is more difficult to expand bureaux
(a good way of maximising the budget) than when a higher official can
easily recruit anybody he/she wants (as in sorne developing countries).
Also, if the state acts as a 'predator' (see above), it has an incentive to
under- rather than over-provide public goods and services (Eggertsson,
l 990, pp. 235-6), thus mitigating the tendency of overprovision due to
burcaucratic self-seeking, Even in societies where the 'principal' is the
diffused public, bureaucrats are not totally free to do whatever they
want:. Given its claim to be a 'public' agency, the bureaucracy is more
vulnerable to 'voice-type' checks - say, media criticism - than private
º
firrns are (on the concept of 'voíce', see Hirschman, 1970).2 Indeed,
historically there have been institutional developments within modero
states that keep bureaucrats from wielding arbitrary powers, such as:
competitive exams for appointments; the auditing of expenditure; the
development of the principie of equality of citizens before the Iaw; the
development of the expectation that office-holders will operate on the
basis of the law, their superiors' directives and their own 'science' and
'conscience': and the subordination of the bureaucracy to ultimate
political decisions (Poggi, 1990, pp. 75-6).
More importantly, bureaucrats can and do act in a fashion that is not
solely self-interested, Often bureaucrats think of themselves, rightly or
wrongly, as the guardians of the public (or national) interest, however
defined, and act to promote it. One such reason is that 'public-
spiritedness', altruism, and so on are often held as a genuine principie,
and not as a thin veil to disguise self-interest. As McPherson (1984)
puts it, '[i]f whatever moral concerns people have are simply rede-
24 The Política/ Economy of Industrial Policy
scribed as peculiar forms of self-interest - she tells the truth because

she'll feel guilty if she lies; she voted against the farrn bill because it
made her feel good to defy the "interests" - then the self-interest
hypothesis becomes empty' (p. 77). The second reason is that bureau-
crats (like the members of any other organisation) are constantly asked
to conform to the organisational objectives of the state, which is always
under sorne degree of pressure to promote the national or public
interest.21 Moreover, as demonstrated by the psychology literature (for
example Tversky and Kahneman, 1986), decision frameworks do
influence people's decisions. Bureaucrats usually face questions put in
terms of public interest, which will invoke a preference-ordering that is
different from that used in prívate decision-making situations (Goodin,
1986). And through the process of endogenous preference formation,
they will develop organisational loyalty and other attitudes that differ
from what is supposed in the theory of self-seeking bureaucrats (Schott,
1984,pp. lll-17;Simon, 1991).
1.3.4 Summary
As the interest-group approach emphasises, groups in society engage in

struggles to ensure property rights over productive assets, to claim
more resources from the state budget, and to restrict property rights and
the distributive claims of other groups, and in this process, they try,
and sometimes succeed, to influence the state to their advantage. Of
course it would be problematic to interpret every policy as an outcome
of interest-group politics because the state not only responds to interest-
group demands but often takes initiatives, for good or bad reasons. In
this respect the autonomous- (or predatory-) state approach, which
treats the state as an entity with its own objectives, is a useful antidote.
However, as was emphasised by the theory of self-seeking-bureaucrats,
the state apparatus itself is made up of bureaucrats who act as agents of
the ruler (the sovereign in earlier periods and the public in modern
democracies). The principal-agent problem here makes it impossible to
assume that the state is a unified entity without looking at individual
cases in terms of the strength of the hierarchy within the bureaucracy,
the independence of high-ranking bureaucrats, the prevalent ideology
within the bureaucracy, the recruiting method, and so on.
There can be no presumption that the state will act in the public
interest, as is usually assumed in standard welfare economics. However
it is equally inadequate to employ another sweeping assumption as to
Theories o/ State lntervention 25
the objective of the state, such as net-revenue-maximisation or budget-
maximisation. What kind of objective function the state operates with
will depend on: what kinds of interests can be formed and acted out as
a pressure on the state; how resistant the state can be to these demands
(or how 'hard', in the Myrdalian sense, it can be); what the objectives
of the top politicians are; how strongly they control the bureaucracy;
how strong hierarchies within the bureaucracy are; how bureaucrats are
recruited; what the prevalent ideology within the bureaucracy is and
how it is formed; and so on.
1.4 ABILITY: THE GOVERNMENT-FAILURE LITERATURE
In many theories of state intervention it is (implicitly) assumecl that the

state knows everything and can do everything. Welfare economics is an
extreme case of this tendency. In welfare economics it is assumed that
the state has ali the relevant information for social-welfare-maximising
intervention and is able to achieve what it sets out to do. The political-
economy literature suffers from the same problem, albeit to a lesser
degree. For example, in the predatory-state approach, although the
competitive constraint put on the state is recognised, there is a strong
presumption that the outcome will be in favour of the state as a
'predator'. The interest-group approach concentrates on whose object-
ives are imposed on the state but does not properly discuss whether the
state can achieve such objectives. The same applies to the theory of
self-seeking bureaucrats, where it is assumed that the bureaucrats can
assert their own objectives without too many difficulties.
However, recently, the assumption ofthe 'omnipotent' state has been
questioned by arguments which hold that, even if it is 'benevolent' and
genuinely tries to improve the efficiency of the economy (although the
opposite tends to be assumed among those who hold this view), the
state may fail to achieve its objectives. These arguments, which we
would call the government-failure literature, have two major strands.
One is the informational argument, which points out that the state may
be able to collect and process ali the information relevant for the
correction of market failures only at costs that are greater than the
benefits from such correction. The other is the theory of rent-seeking,
which argues that state intervention creates additional 'wastes' that may
more than offset the benefits it produces. In the following we examine
these two arguments, which question the ability of the state.
26
1.4.1 The Information Problem
When the state contemplates a policy, it needs to spend resources to

collect and process information in order to formulate the possible
alternatives and make a decision. Even after the policy is decided on it
needs to collect and process the information necessary to monitor the
compliance of lower-level bureaucrats, on the one hand, and the groups
and individuals at whom the policy is targeted, on the other hand.
Part of the information problem is that of insufficient information.
According to this argument, the state simply 'does not know better' about
the future course of events and such informational deficiency can only be
corrected ata prohibitive cost. This is a point which was already made by
the Austrian school in the central-planning debate in the 1930s (see
Lavoie, 1985, for a discussion of the debate).22 According to the
Austrians, even if it is theoretically possible for the planning authority to
'simulate' textbook welfare-economics prescriptions (as Oskar Lange and
others have argued), the amount of information required to do so is so vast
that we cannot possibly expect it to collect and process ali the relevant
information. Indeed the practice of central planning in socialist countrics
shows that the amount of information that could be processed in time for
the writing of the plan is too meagre to allow a plan of even a reasonable
sophistication (see Brus, 1972; Dobb, 1974; Brus and Laski, 1989).
A more important dimension of the information problem is the
existence of informational asymmetry, or, in the language of modern
economics, the principal-agent problem (see Stiglitz, 1987). There exist
two types of informational asymmetry in relation to state intervention.
Firstly, there exists informational asymmetry between the top decision-
makers and the lower-level bureaucrats within the state apparatus. 23
The classic example is the prevalence of shortages in socialist
economies due to the attempts of managers of state-owned enterprises
to secure sufficient inputs by understating their capabilities (sec
Dobb, 1970, 1974; Ellman, 1989, chs 2-3). Secondly, there exists
informational asymmetry between the state and the policy 'target'
entities (for example firms, income groups, individuals). A good
example is the existence of firms under infant-industry protection
which persistently fail to grow out of their 'infancy' in many develop-
ing countries (see Bell et al., 1984). Whatever its source, informational
asymmetry means that the state may not be able to implement effect-
ively its policies unless it spends enormous resourccs to overcome thc
asymmetry.
Although the information problem is a serious handicap on any attempt

for the state to intervene effectively, we think that the problem can be
exaggerated.
Problems of information collection and processing exist for state
intervention because the rationality of individuals who make up the
state apparatus is 'bounded' (Simon, 1983; also see Chapter 2.2.1).
Indeed, overcoming the limitations of individual rationality is the very
raison d'étre of human organisations, including the state apparatus
(Stinchcombe, 1990). If this is the case, it is unreasonable to criticise
the state for having insufficient information while assurning that decision-
makers in prívate organisations (for example firms) know everything
they need to know (see Chapter 3.4.1).
Concerning the problem of asymmetric information within the state
apparatus, note that the problem exists in any organisation of reason-
able size, and not just within the state apparatus. The fact that large
organisations, including the state, develop and function reasonably
well shows that there are ways and means to mitigate the principal-
agent problem, for example by designing an appropriate organisa-
tional structure and promoting organisational loyalty (see Chapter
2.1.2). Moreover, asymmetric information may exist between the state
and the policy target groups, but it also exists between the parties in
prívate contracts. And, again, the fact that prívate transactions that
involve high informational asymmetry are conducted routinely shows
that there are ways to control this problem through organisational
innovations.
1.4.2 Rent-Seeking
The theory of rent-seeking argues that state intervention incurs not

only traditional deadweight losses but also costs when resources are
diverted into unproductive activities by prívate agents in order to
capture rents generated by state intervention (for a survey, see
Tollison, 1982; also see essays in Buchanan et al. (eds), 1980, and
Colander (ed.), 1984).
Rent is defined as 'that part of the payment to an owner of resources
over and above that which those resources could command in any
alternative use' (Buchanan, 1980a, p. 3), that is, the receipt in excess of
the opportunity costs of the resources. An attempt to capture rents, the
argument goes, is perfectly rational at the individual level and socially
productive in a certain context. For example, entrepreneurs innovate to
28
capture the monopoly rent. This entrepreneurial activity is socially
productive in a competitive setting where there is free entry, since once
rents are created in an industry people will move into the industry and
thus drive prices down to competitive levels.24 However, when the state
intervenes to create artificial rents, the resources spent to capture them,
the theory of rent-seeking insists, may be worth expending from the
individual point of view but are wasted from the social point of view,
since they are spent in resource reallocation rather than resource
creation. Buchanan (l 980a) states:
Rent seeking on the part of potential entrants in a setting where

entry is either blocked or can at best reflect one-for-one substitution
must generate social waste. Resources devoted to efforts to curry
[the authority 's] favour might be used to produce valued goods and
services elsewhere in the economy, whereas nothing of net value is
produced by rent seeking. In the competitive market, by compar-
ison, resources of potential entrants are shifted directly into the
production of the previously monopolised commodity or service, or
close substitutes; in this usage, these resources are more productive
than they would have been in alternative employments. The unin-
tended results of competiti ve attempts to capture monopoly rents are
'good' because entry is possible; comparable results of attempts to
capture artificially contrived advantageous positions under govern-
mentally enforced monopoly are 'bad' because entry is not possible
(p. 8).
Once government restrictions on entry are introduced, rent-seeking

activities (and the resulting 'waste' from them) cannot easily be
eliminated and sorne apparently plausible means of eliminating them
may merely shift rent-seeking to another leve) (Krueger, 1974; Posner,
1975). For example the auctioning of monopoly rights (franchise
bidding), while eliminating monopoly profit, will transfer the rents to
the state, which in turn is likely to lead to higher salaries for bureau-
crats. Consequently people will devote 'excessive' time and resources
to becoming bureaucrats (say, by investing 'excessively' in acquiring
educational qualifications), whose compensation exceeds the real
opportunity cost (the compensation for prívate-sector jobs requiring
comparable ability). Even when there is no rent component in
bureaucratic compensation, it is likely that people will strive to capture
the rent extracted by the state in the forms of tax cuts or subsidies
Theories of State Jntervention 29
(Buchanan, 1980a, pp. 12-13). Moreover, according to the proponents
of the theory, the existence of rent-seeking activities affects people's
perceptions of the legitimacy of the economic system, inviting further
state intervention. 'lf the market mechanism is suspect, the inevitable
temptation is to resort to greater and greater intervention, thereby
increasing the amount of economic activity devoted to rent seeking'
(Krueger, 1974, p. 302).
On these grounds, the theory of rent-seeking contends that state
intervention, be it good-willed or ill-willed, is doomed to generate fonns
of inefficiency that have not been generally recognised in the standard
welfare-economics literature - that is, the 'cost of creating monopoly'
(Posner, 1975, p. 823). The policy prescription is that the state should
not intervene in a manner which restricts entry. And on a more
fundamental level, there is the need for a 'constitutional revolution'
(Buchanan, 1980b) to establish a new set of efficient institutions.25
The rent-seeking argument provides sorne interesting insights con-
cerning the interaction between individual behaviours and the insti-
tutional setting. lt raises the important point that the combined results
of individual maximisations can differ sharply according to the institu-
tional settings. However the theory has a lot of problems that may not
be obvious at first sight.
First of ali, the nature of rent-seeking costs is not clearly defined in
the literature, resulting in conceptual confusions (also see Samuels and
Mercuro, 1984 ). In the standard rent-seeking literature, it is customary
to regard ali resources expended in rent-seeking as wasteful. However
rent-seeking may not involve real expenditure of resources but only a
transfer of wealth (for example bribery), which does not constitute a
social cost.26 The real costs involved in rent-seeking are costs involved
in transfcrring property rights, that is, transaction costs, and not the
transfer elements (Littlechild, 1981; Varían, 1989).27 Of course, as
noted above, the proponents of the theory argue that the transferred
rent (for example bribes or the proceeds from franchise bidding) will,
at sorne level, be competed for, for example through excessive
investment in education or lobbying for tax cuts. However, such an
argument is based on the assumption that within the social system
there is at least one area into which entry is unrestricted, be it
education or the political market of lobbying, which we argue may not
be the case (see Chapter 2.2.3 for a more detailed discussion). In
addition it should also be noted that, when we relax the assumption of
full employment implicit in the theory of rent-seeking, even an
30
apparently unproductive activity can be useful for society (recall the
famous Keynesian example of the 'hole-digging' exercise).28
Moreover, the wastes from rent-seeking may not even be the major
costs from state-created rents. Rent-seeking cost is often of a once-and-
for-all nature, because once a rent is granted an entry barrier into the
rent market is likely to be erected, which will discourage potential en-
trants from spending resources to dislodge the incumbent. A more seri-
ous danger is that state intervention may protect or even encouragc
inefficient producers or production methods, with a long-lasting effi-
ciency consequence (see also Littlechild, 1981 ). This problem is not
explicitly discussed in the rent-seeking literature, which normally
assumes that ali rent-seeking agents are identical and use optima)
production methods. However, in the real world, there is no guarantee
that someone who is competent (or even lucky) at seeking rents is
equally competent as a producer, although this may well be the case if
rent-seeking takes the form of franchise bidding.
Thirdly, in the rent-seeking literature it is usually assumed that re-
strictions on entry will only be created by state intervention. This is a
strongly biased view. Firms are always anxious to deter potential
entrants. Anything from the secrecy of production technologies, to ex-
cess capacity, to the brand loyalty of consumers can be used asan entry
barrier. The implicit belief behind the theory of rent-seeking - that
competitive markets are self-reproducing - is unwarranted (see 1.1.2
above). .Krueger's (1974) assertion that the existence of state ínter-
vention, and hence rent-seeking, will erode people's faith in the market
mechanism, leading to calls for more and more intervention, is also
based on this dubious belief in self-perpetuating competitive rnarkets.
Most importantly, rent-seeking may be directly unproductive, but
'indirectly productive'. Rent-seeking is only unambiguously harmful
for society when it can be assumed that 'the initial institutional creation
of an opportunity for rent seeking [creation of entry barriers] ensures a
net destruction of economic value' (Buchanan, I 980b, p. 359). How-
ever the costs of rent-seeking may well be more than offset by the
dynamic gains of productivity growth which the rent allows, say, by
enabling firms to increase R&D expenditure (Littlechild, 1981) - this is
precisely the reason why we have, for example, patent systems (given
the public-good nature of technological knowledge) and infant-industry
protection (given the possibility of learning by doing). Of course, the
theory of rent-seeking is correct in arguing that free entry is necessary
to guarantee the beneficia) effects of rent, but the theory is far too
reluctant to acknowledge that the creation of a monopoly by the state
may be beneficia! for productivity growth if the state can withdraw the
rent when necessary (see Chapter 4.4.2).
The refonn proposal for a rent-seeking society is also unconvincing. If
we can achieve a 'constitutional revolution' and start afresh, then we
may talk of non-intervention on a rent-seeking ground.s? However any
realistic person will recognise that the possibility of a successful
constitutional revolution is almost nil. In the first place this proposal
assumes that, once created, competitive markets will perpetuate them-
selves. We have repeatedly pointed out that this is not the case.
Secondly, the proposal underestimates the power of vested interests,
which may not agree to reform. Even Buchanan admits that the
majority of individuals are usually losers in rent-seeking games, and
therefore the winners will have strong incentives to defend their
positions (Buchanan, 1980b). When there is no feasible political
platform to achieve such a change, providing such a refonn proposal
amounts at best to an evasion of responsibility and at worst to an
apparently populist rhetoric that amounts to a refusal to refonn.
CONCLUSION
Welfare economics elegantly and convincingly spells out why markets

may fail and what kinds of interventionist measures the state can
employ to correct them.30 However, as many economists from across
the política! spectrum argue, the theory is based on a naive set of
assumptions about the nature and ability of the state. Even if we do not
accept the contractarian argument that the state, being the product of a
social contract among free individuals, has no right to intervene in the
market, we still have two thorny questions: <loes the state really serve
the public interest?; and can it achieve what it sets out to do?
In relation to the first question we examined three approaches - the
autonomous-state approach, the interest-group approach and the self-
seeking-bureaucrats approach - and concluded that none of them on its
own can provide generalisable assumptions about the objectives of all
states, regardless of the time and space where they exist. The question
is more of an empírica! than a theoretical one. It was suggested that, in
order to establish a reasonable set of hypotheses concerning the
objectives of a particular state, we should look more carefully at the
process of interest group fonnation and collective action as well as
32
the operation of the bureaucracy in the particular system of political
economy in which the state is operating. In doing so, the insights from
ali the three approaches reviewed can fruitfully be used.
In relation to the second question, we examined two approaches -
the informational argument and the rent-seeking argument - which we
grouped together as the government-failure literature. lt was argued
that the importance of the information problem can be exaggerated and
that there are ways to mitigate the problem. In relation to the rent-
seeking argument, it was pointed out that despite sorne important
contributions, the argument suffers from a few major theoretical defi-
ciencies. Common to these two approaches is their inability (or
unwillingness?) to suggest how govemment failures may be remedied
other than by non-intervention. Does this mean that we are condemned
to accept failing markets in favour of failing govemments as the lesser
of the two evils? We will attempt to answer this question in the next
chapter.
2 A New lnstitutionalist
Theory of State
lntervention
INTRODUCTION
The early debate on the role of the state was conducted in the tenns set
by welfare economics. The degree of market failure may have been
questioned, along with the moral legitimacy and the political intention
of state intervention, but no question was asked as to whether well-
informed (with welfare economics) and well-intentioned state interven-
tion can actually improve the efficiency of the economy by correcting
market failures. With the advent of the govemment-failure school the
terms of debate have considerably changed. Although the possibility of
improvement in efficiency through state intervention is still accepted
(except by sorne staunch free-marketeers who deny the existence of
market failures), now it is being asked whether the net outcome of such
intervention is efficiency-improving when state intervention itself
carries certain costs.
We think that the govemment-failure literature has provided a valu-
able corrective to the naive belief about the state held by sorne welfare
economists that, once we can somehow have a 'benevolent' state, it
will solve ali problems. However, as we pointed out in Chapter 1, we
are still left with one important question: Is there any way to correct
govemment failures, as there are ways to correct market failures'? The
first two sections of this chapter look at the infonnation and the rent-
seeking arguments exarnined earlier and suggest how govemment
failures may be remedied, if not completely eliminated. The last section
moves a step further and develops what we call a 'new institutionalist
theory of state intervention', which incorporares both the market-failure
and the govemment-failure perspectives and suggests sorne roles of the
state that have hitherto been neglected in existing literature.
33
34 Thelnstitutionalist
A New Political Economy of Industrial
Theory Policy
of State lntervention 35
34
2.1 REMEDYING GOVERNMENT FAILURES 1: THE

INFORMATION PROBLEM
2.1.1 Substantive Rationality vs Procedural Rationality
In the standard model of decision-making, the decision-making agent

has a well-defined objective function and is aware of an exhaustive set
of alternatives (and their respective likelihoods). In this model of what
Simon (1983) calls 'Olympian rationality', there may be inefficiency as
a result of individual decisions which lead to collective irrationality (for
example see Sen, 1982), but there is no inefficiency dueto the costs of
making decisions because everyone knows everything.1 When we adopt
this model of decision-making, the question boils dowrr to securing
'correct' decisions, or in Simon's terminology, substantively rational
decisions (on different notions of rationality, see Hargreaves Heap,
1989, and Langlois, 1986b).
However, given the limited cognitive ability - or 'bounded ratíonality'
in Simon's terminology - of individual agents, devising a rational
decísíon-making procedure becomes a more important question than
ensuring that each decision is as substantively rational as possible
(Simon, 1975). After all, 'the elaborate organisations that human beings
have constructed in the modero world to carry out the work of production
and govemment can only be understood as machinery for coping with the
limits of man's abilities to comprehend and compute in the face of
complexity and uncertainty' [emphasis added] (Simon, 1979, p. 501).
Of course this <loes not mean that substantive rationality <loes not
matter. Especially when we are dealing with complex phenomena, ex-
pertise plays an important role, not least because such knowledge is not
easily transferable.2 However the point should not be over-emphasised.
As Schelling (1984) argues, '[t]o expect an organisation to reflect the
qualities of the individuals who work for it or to impute to the indi-
viduals the qualities one sees in the organisation is to commit what
logicians call the "fallacy of composition... (p. 32). In other words, what
matters more in a complex and uncertain situation is the rationality of
the decision procedure (determined largely by the organisational
structure) rather than the substantive rationality of individual decisions
(determined largely by the quality of individual members of the
organisation) (Simon, 1975, and 1983).
That procedural rationality may matter more than substantive ration-
ality is also suggested by the fact that those East Asian states that have
Theory Policy
35
successfully intervened in the economy do not in general have well-

trained economists as their economic bureaucrats. In his superb study
of the famous Japanese Ministry of International Trade and Industry
(henceforth MITI), Johnson (1982) points out that the bureaucrats who
have been orchestrating successful economic development have usually
been lawyers by training (also see Hadley, 1989, p. 300). The same is
true for Korea (see Chapter 4.5). More interestingly, Wade (1990)
points out that the economic bureaucracy in Taiwan has traditionally
been manned by engineers, and not economists.
2.1.2 Mitigating the Information Problem
2.1.2.1 lmproving the decision-making capacity of the state

As the government-failure literature points out, the limited ability of the
state to collect and process information means that successful state
intervention can only be achieved at an informational cost. In this regard
the resource capability of the state may matter, because an organisation
with a larger resource capability will, ceteris paribus, be able to collect
more information. And indeed in many developing countries the Iack of
government resources to collect information (for example understaffed
statistical agencies) often poses a serious problem. However, as our
argument above indicates, the mere fact that we have more information
does not mean much when there is a bottleneck in our ability to process
it - that is, when there exists bounded rationality.
As the now-extensive literature on organisation and decision-making
(for a survey, see Stinchcombe, 1990) demonstrates, the ability of the
state (or any organisation) to process information is, more than anything,
determined by its decision-making structure, the most important
dimension of which is the degree of centralisation (Simon, 1975; March
and Simon, 1958). Ceteris paribus more Iayers of autonomous decision-
makers result in more time and resources spent on decision-making (for
example communication and negotiation costs), and therefore centralisa-
tion of decision-making can save costs by promoting quick decision-
making.
Of course a centralised decision-making structure runs the danger of
resulting in substantively Iess rational decisions, because centralised
decisions can be overly influenced by the information possessed by the
top decision-makers, which is Iikely to be Iess complete when compared
with that possessed by the operative units. Again, however, this does
not mean that the most decentralised decision-making structure will be
Theory Policy
36
able to secure the best decision. Whether this is the case depends upon
the nature of the decision.
People's perceptions of the world, which are bound to be incomplete,
tend to become biased according to their personal experiences. In more
practica) tenns, people tend to develop 'sub-goal identification'
(Simon, 1991). For example Simon (1979) found that '[tjhe business-
men's perceptions oí the principal problems facing the company ...
were mostly determined by their own business experiences - sales and
accounting executives identified a sales problem, manufacturing execu-
tives, a problem of interna) organisation' (p. 501). And when there
exists subgoal identification, it is not necessarily true that leaving the
decision to those al the locality will enhance global efficiency. When
the problem concemed is of a global nature, the top decision-maker, be
he/she the central executive of a firm or a minister of the govemment,
may identify the problem (and hence its solution) more correctly, not
because he is a superior being but simply because he does not have the
subgoal-identiñcation problem. lf the decision is about the global
efficiency of the economy (which is likely to be the case when a state is
contemplating an intervention), a centralised decision-making structure
may be a blessing rather than a curse.
One more point to be made is that the top-decision makers in the
state decision-making hierarchy should be assigned only strategic tasks
- that is, tasks which relate to the long-tenn direction of the economy
as a whole - and be allowed to delegate more routinised day-to-day
tasks to the lower-level decísion-makers.' In the organisational litera-
ture it is found that, when day-to-day tasks and more analytical tasks
were vested in the same decision-making unit, '[t]he pressure of regular
deadlines and the tasks of supervising clerical personnel usually [gives]
the day-to-day activities priority over equally important, but postpon-
able, analytic tasks' (Simon, 1982, p. 81). Given that any individual's
ability to process infonnation is limited, freeing the top-decision
makers from routinised decisions will enhance the overall ability of the
state to process infonnation.
2.1.2.2 Reducing informationa/ asymmetry
lf designing a decision-making structure that can secure rational de-
cisions at the lowest costs can alleviate the infonnation problem that
results from the limited ability of the state to collect and process informa-
tion, how can the infonnation problem due to asymmetric information
be reduced? As we discussed earlier, there are two kinds of infonnational
A New Institutionalist Theory of State lntervention 37
asymmetry involved in state intervention. One is that which exists with-

in the state apparatus, between the top decision-makers and the lower
Jevel bureaucrats, and the other is that which exists between the state
and other agents.
Concerning informational asymmetry within the state, ít should be

noted that when the person in charge of a subordinate unit (and hence
accountable to the top decision-maker) does not have a substantial
measure of control over variables influencing the performance of the
unit, he/she can always blame others for bad decisions (Simon et al.,
1955, p. 78). In this case, the top decision-makers have to spend time
and resources on measuring the performance of their subordinates to
verify such arguments. To reduce such costs, the degree of discretion
available to an organisational participant should be linked to the
characteristics of activities performed by him/her. For example, if it is
easy to observe job activities and job outputs and to relate activities to
output, less discretion should be allowed, and vice versa (March and
Simon, 1958, p. 147).
lmproving organisational design is not the only way to reduce the
informational asymmetry and the resulting 'agency costs' in the state
apparatus. As we discussed in relation to the self-seeking-bureaucrats
approach, organisational loyalty plays an important role in reducing
selfish behaviour among the organisational participants (see Hirschman,
1970, and Simon, 1991, on the role of organisational loyalty). Pro-
moting organisational loyalty among the bureaucrats, therefore, would
reduce the agency costs within the state apparatus to a great extent.
Informational asymmetry between the state and other agents also can
be reduced through institutional design. For example, the willingness of
Japanese firms to reveal the necessary information to the state, which
reduces informational asymmetry, is often attributed to the tradition of
a close government-business relationship (Okimoto, 1989, p. 156).
Nevertheless it should be remembered that this tradition can only be
fully exploited because of the role of state-promoted industry associations
as information-clearing houses (Magaziner and Hout, 1980, p. 37). The
'plan contracts' between the ministries and the public (and sorne
prívate) enterprises in France and the establishment of laws for regular
reporting from targeted industries in Korea are other examples of
reducing informational asymmetry through conscious institutional
design (on France, see Hall, 1987, chs ó-8, and Hayward, 1986; on
Korea, see Chapter 4.4.1 ). 4
38 The Poluica! Economy of Industrial Policy
2.2 REMEDYING GOVERNMENT FAILURES 11: THE RENT-

SEEKING PROBLEM
In the standard model of rent-seeking, wherever the state restricts

entry economic agents (individuals and groups) will try to capture the
rent arising from such restrictions and, in that process, will spend an
amount of resources equivalent to the amount of the rent. In order for
this result to obtain, a few rather stringent conditions need to be
satisfied.5 First of ali, it should be possible for prívate agents to
influence the state. If the state is not open to influence there is no point
in spending resources to affect its decisions. The second condition is
that there are agents who will engage in rent-seeking activities. This
condition may not be met when the rents are granted to an interest
group (for example industry) but the individual agents of the potential
interest group fail to organise themselves due to the collective action
problem. Thirdly, competition should exist in the rent-seeking process,
because if there is no need to compete for the rents it is not necessary
to spend resources to acquire them. Let us examine these conditions
one by one and explore their implications for the reduction of rent-
seeking costs.
2.2.1 Tbe Vulnerability of the State
As Congleton (1980) argues, 'artificial scarcity' alone <loes not create

rent-seeking. There should be sorne means for rent-seekers to influence
the outcome of the contest. A11d since rent-seeking involves state-
created rents, the vulnerability of the state becomes a crucial issue,
since if the state is invulnerable to outside influence, there will be no
rent-seeking however big the rent may be.6
One reason why a state may be invulnerable is that its key decision-
makers are not dependent on electoral support. A good example is
provided by the 'bureaucratic' régimes of Gaullian France, Japan and
Korea (for a more detailed discussion of these countries, see Chapters 3
and 4). In these régimes most of the decision-making power concerning
economic policies is entrusted to the bureaucrats, who are less
vulnerable to influence because of their job security, and not to the
politicians whose need for re-election makes them vulnerable to outside
influence. Invulnerability of the bureaucracy may be intensified if the
state bureaucrats think of themselves as guardians of social interests
and refuse to be affected by outside influence (see Chapter 1.3.3).
A New /nstitutionalist Theory of State Intervention 39
A state may also be invulnerable because it has voluntarily abdicated

its power to malee decisions (for this reasoning, see Schelling, 1960,
ch. 1, and Elster, I 984, pp. 411-22). If the state Iegislates that picking the
winner will be by methods such as a genuine Iottery, equal sharing,
rotation, or will be totally arbitrary (although none of these methods is likely
to be used to any extensive degree), people will not seek rent for they
have no chance of affecting the outcome of the contest. Other policy tools
may also reduce, if not eliminate, the vulnerability of the state. Sorne,
though not ali, forros of queuing will be among such methods. Unless
queuing requires physically waiting in line (for example, the queue
can be formed by submitting written applications), forcing people to
forego opportunities for other productive activities, the costs incurred in
the process of queuing will not be Jarge.7 Another case of 'abdication' will
be the 'indicative-planning' exercise in Japan and France, which can be
understood as an attempt by the state to bind itself to a policy and thus
malee it difficult for individual ministries and bureaucrats to significantly
deviate from the announced priorities (see Johnson, 1982, on Japan, and
Cohen, 1977, on France). This is a good contrast to the US industrial policy,
where the employment of ad hoc measures meant that there was plenty
of scope to bend the rules (Johnson, 1984).
More interestingly, the state might be vulnerable to pressure but de-
cides on the method of picking the winner so as to generate sorne bene-
ficia! by-products. Picking the winner by observing each contestant's
past contribution to social productivity or his/her future prospect for
such contribution will have, to borrow from Congleton (1980), a bene-
ficial 'by-stander effect'. For example, contestants in the rent-seeking
process may be required to use the wealth effectively to generate more
wealth through innovations - in products, production processes and
organisations - or to demonstrate benevolence or altruism (Congleton,
1980, pp. 172-3). Thus if the state malees itself vulnerable to the in-
fluence of someone who demonstrates his/her productive capabilities,
sorne cost will be incurred to secure the rents (for example R&D costs),
but they may be more than offset by the benefits from what the rent-
seekers have created in that process (for example better products).
It should of course be remembered that an invulnerable state is not

necessarily a good state. For example, when a state is pursuing object-
ives that are objectionable from the social point of view, as in the
predatory-state approach (see Chapter 1.3.1) oras in regulatory-capture
theory (see Chapter 1.3.2), we may want it to be vulnerable to pressure
40 TheInstitutionalist
Theory Policy
of State Intervention 40
40 40
to change its objectives. Moreover, as will be clear from our discussion,
different forms of invulnerability will have different efficiency implica-
tions, although they may be equally effective in reducing rent-seeking
costs. Por example, in choosing the recipient of the rent, if the state
makes itself invulnerable to influences other than someone's productive
contribution, the outcome is likely to be much more positive than when
it simply abdicates its power and chooses the recipient at random.
2.2.2 The Problem of Collective Action
Rents are often granted to a group of agents (for example industry,

firms in a region) rather than to individuals. This is partly because
'influencing activities' (the phrase owes to Milgrom and Roberts, 1990)
requires a large set-up cost (for example one may need an industry
association in order to lobby) but also because the principie of
'fairness' often demands that agents with identical attributes receive
equal treatment. If a particular contest for rent, for whatever reason,
requires organising an interest group, such a contest may not happen
due to the collective-action problem. That rent-seeking might suffer
from the collective-action problem (since it usually involves organising
an interest group) has already been pointed out by Tullock (1984).
Tullock, however, does not clearly specify that the problem of
collective action exists only for sorne kinds of rent, that is, the ones
generated from a public source of rent (named after a public good) - or
from property rights that allow open access to those within the relevant
group.8
There are sorne rents which simply cannot be shared with others.
Rents from industrial licensing and patents are cases in point. The
sources of these rents are private goods in the sense that the sharing of
them with others reduces one's benefit. In this case the problem of
collective action simply does not exist, and therefore Tullock's
argument does not hold. The problem of collective action does exist
when the rent is generated from a public good. Tariff protection is a
case in point, since the fact that others are getting tariff protection does
not reduce one's benefit from the protection. When the source of the
rent is a public good (although the rent itself is a prívate good - an
analogy is that a fishery might be a public good but the fish caught from
the fishery are prívate goods) and when the potential rent-seekers fail to
organise themselves, rent-seeking may not arise and therefore there
may be no waste from rent-seeking, as Tullock argues.
Theory Policy
41 41
lf this is the case, the magnitude of rent-seeking costs is affected by the

types of policy tools used. Certain types of policy tools (for example
tariff) are less prone to rent-seeking than others (for example industrial
Jicensing) because they have the collective-action problem, although
both types of tools may create comparable amounts of rent. This means
that, ceteris paribus, by using more public tools of intervention (and
thus deliberately creating the collective-action problem among potential
rent-seekers), the state may reduce rent-seeking costs.
2.2.3 Competitiveness of the Rent-Seeking Process
One basic assumption of the theory of rent-seeking is 'that getting a

monopoly whether by influencing the government or by getting it pri-
vately is essentially a competitive industry' (Tullock, 1984, p. 228). The
assumption of competitiveness is important because if there is no
competition, there is no need to spend resources to influence the state's
decision to one's favour and hence there is no waste from rent-seeking.
Contrary to what is usually assumed in the standard literature, however,
rent-seeking does not always, and not even predominantly, involve a
large number of equally positioned competitors. Especially in many
developing countries, the oligarchy, which has an exclusive access to
state-created rents, is so deeply entrenched that the rent-seeking industry
can hardly be described as competitive.
The Jimiting case of non-competitive rent-seeking will be monopolistic

rent-seeking where there is only one contender for the rent. For example
an íncumbent firm in, say, the petrochemícal industry may become the
sole contender for the rent from tariff protection because it happens to
have a monopoly position dueto the state's industrial licensing policy or
due to scale-economy reasons (natural monopoly). Or Mr Brown may be
the sole contender for the rent from the Jicense for the new automobile
plant because every other potential seeker of this rent knows that he is a
brother-in-law of the prime minister and is certain to get the license. In
such situations the sole contender for the rent does not have to spend
anythíng to get the rent, and therefore no rent-seeking cost is incurred,
both from the individual and from the social point of view.
The more general case of non-competititive rent-seeking is that of
oligopolistic rent-seeking. Small numbers, of course, do not necessarily
guarantee less rent-seeking. There can be cut-throat competition even in
a rent market with a small number of participants. However a small-
Theory Policy
42 42
number setting is inherently different from a large-number one. In the
former case an individual's best response is affected by the response of
ali others (and the common knowledge about this fact), whereas in the
latter case one can act as if one's action does not affect, and is not
affected by, those of other individuals. In a small-number situation,
therefore, one cannot take the environment as given, since one is a part
of the environment. This means that there exists strategic uncertainty
resulting from interdependence between individual decisions.9
More interestingly, when communication is possible players in the
small-number setting may be able to overcome strategic uncertainty by
devising a binding contract between themselves. Here there arises the
possibility of collusion. The competitors for a particular rent might
agree, and make a binding contract between themselves, that only one
will bid for the rent at a price that is far lower than the rent, and that
they will split the spoils (the rent minus the price), rather than compete
the rent away. Let us elaborate on this point with an example.
If two firms are contending for a monopoly position that will yield a
rent of 100, and if they have an even chance of getting the position, we
can suppose a payoff matrix as shown in Figure 2.1. In this example, if
both A and B bid for the rent, each has an expected payoff of O, because
each runs the risk of spending the sum equal to the expected value of the
rent (that is 50) but not getting the rent, one out of two cases (0.5 x [ 100
B
Bid Stay Out
Bid o.o 100, O
Stay Out O, 100 o.o
payoffs to (A, B)
Figure 2.1 An example of duopolistic rent-seeking

Theory Policy
43 43
- 50) + 0.5 x [O - 50] = 0). However, if the two firms can write a
binding contract such that one of them will stay out of the contest, they
can acquire the rent without spending any resource, In this case the
social cost involved may be nil. Of course this simple example ignores
the fact that the possibility of such collusion will depend on the
feasibility of making credible commitments, the possibility of side
payments, and other conditions related to bargaining (for a more
detailed discussion of these conditions, see Schelling, 1960), but it
shows that non-competítive rent-seeking is likely to incur less rent-
seeking costs than a competitive one due to the inherent likelihood of
collusion.
The above discussion shows that when the rent-seeking process is less
competitive, it is more likely that the total costs arising from the
process are lower. This happens basically because there exist entry
barriers into the rent market. Of course, one may argue that resources
will still be wasted even in the case of non-competitive rent-seeking
because people will spend resources to build up entry barriers to the
rent market. We agree that this is sometimes the case - for example I
may spend resources in building personal ties with the industry minister
in order to build barriers against future rent-seeking contests. However,
in a world with imperfect foresight, the scope for such action may be
seriously limited - 1 cannot be sure how helpful the minister will be in
a later, and unforeseen occasion when I want to bid for a rent.I?
Even if the entry barriers to the rent market are sufficiently high that
there are few rent-seeking costs, there may be costs from 'second-tier'
rent-seeking if the state distributes the rent extorted through franchise
bidding or bribery (see Chapter 1.4.2). For example, if the rent extorted
by the state is distributed to bureaucrats in the form of higher salaries,
and if anyone can become a bureaucrat with a certain educational
qualification, people may invest an excessive amount of resources in
education.11 Likewisc, if the state extorts sorne part of the rent in the
form of, say, bribery and distributes it in the form of tax cuts, which are
up for grabs by any interest group, interest groups will spend resources
to secure such rents. These examples of second-tier rent-seeking, how-
ever, still assume that there is unlimited entry al at least one leve) of the
political economy. However, if entry is contained at sorne level and if,
as a result, competition does not overflow into other levels, the rent
extorted by the state may not be tota11y dissipated. And although this
outcome may have undesirable política! and distributional conse-
quences, it may be a less wasteful outcome. For example, if entry into
Theory Policy
44 44
the bureaucracy is open only to a tiny minority, or if only the most
powerful group is able to secure a new tax cut, there may be no second-
tier rent-seeking and hence no rent-seeking cost to society, although
income distribution may become skewed as a result.
2.2.4 Some Reflections: Competitive Markets and Competitive

Politics
Fundamentally, what causes rent-seeking is the existence of political

competition, and not the absence of economic competition (which causes
rent but not rent-seeking). If this is the case, restricting political
competition seems to be a solution to the problem of rent-seeking. Above
we suggested sorne ways in which this can be done. However the
proponents of the theory of rent-seeking would ask: why not attack the
root cause of the problem, that is, why not eliminate all the rents simply
by freeing the markets from state interference? Why bother about politics
when separating economics from politics will solve the problem?
This view, however, is based on the assumption that there is no
reason to restrict competition (and thus create rents) in markets. As
Schumpeter ( 1987) and Richardson ( 1960) have persuasively shown,
however, there is a need for restrictive practices (which create rents), in
order to achieve innovation and long-term investment, which are the
causes of dynamic growth of the economy. Even if the state does not
intervene, the market in a dynamic economy is bound to create a lot of
rents, and people are bound to spend resources to capture those rents.
Toen why are the market-generated rents not subject to wasteful rent-
seeking whereas the state-created ones are?
The point is not that people do not seek rents in markets. As our
quote from Buchanan (1980a) in the previous chapter indicates,
innovation is rent-seeking par excellence. However, rent-seeking in
markets is regarded as productive because we assume that it creates
more resources than it uses up in the process, and not because it does
not use up any resource. For example, firms do spend resources on
R&D to win the competition, but produce more resources (for example
better products) or reduce costs (for example better processes) as a
result - although, contrary to the conventional presumption, there is a
possibility that such competition will result in waste (see Chapter
3.2.1). Política] competition is wasteful only because it is implicitly
assumed that it does not create any value. However political
competition does create (if intangible) values. First of all, political
Theory Policy
45 45
activities may involve organisational innovations, which may

ultimately reduce the costs of political competition in the future - for
example it wiJI be cheaper to have a political party than to let millions
of people bargain with each other. Moreover political activities are
often ends in themselves and people may derive value from the
activities per se as well as from the products of such activities (see
Hirschman, 1982, pp. 85-6).12
Then, whether competition, be it economic or political, is wasteful or

not - that is, whether it creates more resources than it spends - is a
moot point. Especially in the case of political competition, there is no
unambiguous way to measure the value of its output. Even if we follow
the proponents of the theory of rent-seeking in assuming that political
competition creates no value whatsoever, there are ways and means to
reduce the costs from rent-seeking, essentially by restricting competi-
tion at one level or another. The real question, then, is at what level of
the overall system of political economy - for example education,
bureaucratic recruiment, interest-group activities, market competition
- should competition be restricted and at what level should it be en-
couraged? The answer to this question can only be determined with
reference to the concrete political economy of the society concemed,
and not on the basis of the simplistic model of political economy envis-
aged by the proponents of rent-seeking in which there exists no entry
barrier other than the ones created by the state.13
2.3 A NEW INSTITUTIONALIST THEORY OF STATE

INTERVENTION
Up to now we have discussed how the costs of state intervention may

be reduced, or in other words, how govemment failures may be cor-
rected. We argued that there are ways and means to devise institutional
schemes to reduce the costs of state intervention while not foregoing the
possible benefits from such intervention - for example by improving
efficiency through correction of market failures or by encouraging
productivity growth by granting infant-industry protection. In this sec-
tion we generalise our discussion and suggest a 'new institutionalist'
theory of state intervention, which a'llows us to look at the problem of
state intervention in a more balanced manner and reveal sorne import-
ant roles of the state that have hitherto been neglected.
Theory Policy
46 46
2.3.1 The Nature of Economic Costs
In neoclassical economics, of which welfare economics is a branch, the

only 'economic' problem is that of achieving allocative efficiency
between the production units. As the famous definition of economics by
Robbins goes: '[e]conomics is the science which studies human behavi-
our as a relationship between ends and scarce means which have alter-
native uses' (Robbins, 1932, p. 16). What is curious here is that solving
this economic problem does not require any cost and the only cost in
our economic life is the cost involved in the material production, which
is seen as an engineering problem, and notas an economic problem.
In this framework markct failures are inefficient in the sense that the
economy does not achieve the degree of social welfare that is theoreti-
cally attainable, or what Demsetz ( 1964) calls the 'nirvana' state, but
they still do not impose any real cost to society. And better still, such
inefficiency in resource allocation (market failure) can be remedied by
the state and the remedy costs nothing to society, because what the state
does, so to speak, is announce new prices at no real cost.
Recently, however, there have been arguments criticising this frame-
work on the ground that the achievement of efficiency may be a costly
activity - independent of the hypothetical costs arising from the state of
not achieving efficiency.
In relation to production, two groups of argument have been pul
forward against neoclassical orthodoxy. Firstly, the Iiterature on
technical change points out that technology is not a blueprint but that it
contains a Iot of tacit knowledge which cannot be pul to work without a
costly process of learning (for example training costs).14 Secondly, the
Marxian labour-process Iiterature, the efficiency-wage argument, and
the theory of team production all argue that production is not merely an
engineering process but is also a Iabour process in which workers have
control over their effort, and therefore it rnay be costly to organise
production (for example monitoring costs).15
In relation to resource allocation also, two kinds of arguments have
been raised. Firstly, the transaction-cost branch of new institutional
economics argues that market exchange (or market transaction),
through which resource allocation is achieved in the market economy,
is not costless as assumed in neoclassical economics because bounded
rationality requires us to spend resources to set up safeguards against
the opportunistic behaviour of the trading partner (for example the cost
of writing and enforcing the contract).16 Secondly, the government-
47 A New
Thelnstitutionalist Theory
Political Economy of State lntervention
of Industrial Policy 47
47 47
failure literature argues that the process of resource allocation by the

state is not costless as assumed in welfare economics, if only because
the state is not an omniscient and omnipotent being (for example costs
of information processing).l?
Thus seen, the recent theoretical developments emphasise that
achieving efficiency - or coordinating the behaviour of interdependent
but independent agents - is costly, independent of the hypothetical
costs arising from actually falling short of what Dernsetz calls the
'nirvana' state. From this point of view, the costs of actually
producing goods (which are inevitable dueto the law of physics, so to
speak) are not the only costs of our economic life, and a significant
portion of economic costs are made up of the costs of actually
allocating resources between production units and managing the
production process within the production units. These costs of
coordination between and within economic units, which are to be
differentiated from pure engineering costs, have recently come to be
called transaction costs. ts
Transaction costs exist because economic activities (be it the produc-
tion process or the allocation process) involve not only interaction
between people and things but also interaction between people.l?
For econornic interaction between people to start at ali, the rights of
individuals need to be defined. Property rights have to be assigned and
theír boundaries defined - 1 may own my knife but I may not be allowed
to kill people with it. Rights over other entitlements (for example social
security, working conditions, and product safety) also have to be nego-
tiated and defined. These are ali costly processes that require information-
collection and processing, bargaining, and contract writing. After the
rights structure is defined, it is necessary to maintain the given rights
structure, and this will also require resources for policing and enforce-
ment, Moreover resources will continually be spent by those who want to
redefine the rights structure - for example, through enclosing open fields,
spending on R&D to acquire patents and lobbying to secure tariff
protection - and by those who want to defend the existing rights structure.
Even within the existing rights structure it will be costly to acquire
and process the inforrnation necessary to arrange the contract - say,
concerning the contracting partner and the likelihood of each possible
futurc event - and bargain over it. lt will often be necessary to incur
costs to actually write the contract. Once a contraer is made, the adher-
ence of the contracting partners to the contract needs to be monitored.
This involves measurement costs (Milgrom and Roberts, 1990) and
48 A New
Thelnstitutionalist Theory
Political Economy of State lntervention
48 48
policing costs. When a breach of contract is suspected it may be neces-

sary to incur costs to settle the disputes (for example the cost of lawsuits).
Once the breach of contract is ascertained, enforcement of the contract
will become necessary and this will also incur costs.
2.3.2 Transaction Costs and State Intervention
If transaction costs are the costs incurred for the purpose of defining
and redefining the property (and other) rights of economic agents, on
the one hand, and of writing, monitoring and enforcing contracts within
the existing rights structure, on the other hand, the costs of state inter-
vention may also be reinterpreted as transaction costs. Informational
costs, which are costs necessary for the state to decide on a policy and
effectively implement it, are obviously of this kind (analogous to such
costs for private contracts). Costs of rent-seeking, as costs involved in
the process of redefining the property rights structure, can also be class-
ified as transaction costs (see Chapter 1.4.2; see also Varian, 1989;
º
Milgrom and Roberts, 1990).2 Reinterpreting the costs of state ínter-
vention as transaction costs allows us to make sorne important observa-
tions in relation to the theory of state intervention.
In neoclassical economics the allocation of resources is a costless process

whether it is achieved through the market (that is, market transactions)
or through the state (that is, state intervention). What counts is only
the difference between the allocation actually achieved and the ideal
allocation in the 'nirvana' state. Given such a scheme, it is no wonder
that the government-faílure literature could criticise welfare economics
on the rather obvious grounds that state intervention incurs certain
(transaction) costs. However many in the government-failure school are
misleading the argument by implicitly comparing the ideal market, which
is costlessly run, with the real state, whose activities are costly.
In the real world, both state intervention and market transaction are
costly, Therefore the comparison must be between the costs of allocat-
ing resources through market transactions (or the transaction costs of
rnarket allocation) and the costs of allocating resources through state
intervention (or the transaction costs of state allocation). Thus seen, the
real question is whether the state can achieve the same allocative effi-
ciency at a lower cost than the market can do, and not whether state
intervention is costly per se. Introducing transaction costs into our
scheme allows us to introduce a role of the state which has hitherto
49 The Instinaionalist
Theory Policy
49 49
been neglected in conventional economic theory, that is, the role of

lowering transaction costs in the economy. Let us discuss exactly what
the state can do in this regard.
Various authors in the tradition of new institutional economics have
suggested that one important role of the state is to institute a well-
defined property-rights system and effectively enforce it (Coase, 1960;
Matthews, 1986; North, 1981). A clear and effectively enforced set of
property rights is seen as saving transaction costs, because otherwise
people whose decisions are interdependent may have to spend an
inordinate amount of resources in order to resolve the extemalities pro-
blem. Of course instituting a well-defined property rights structure and
enforcing it, in theory, need not be done by the state, but the monopoly
on the legitimate use of violence by the state meaos that it is often, if
not necessarily, the most cost-effective agent in doing so (Matthews,
1986, p. 910). Historically, too, transaction costs in market exchanges
have been substantially reduced only with the involvement of the state
in establishing and enforcing property and other rights (Polanyi, 1957;
Coase, 1988; North, 1990b).
The role of the state in reducing transaction costs does not stop at
instituting and enforcing a property-rights (and other rights) system.
Another important role is to reduce the macroeconornic instability of the
economy through aggregate demand management. Given the necessity
of stability in the environment for agents with bounded rationality
to make any calculation (Simon, 1983, pp. 19-20), increased
macroeconomic instability will lead people to divert resources into
activities that are intended to create microenvironments that will make
rational calculation possible. Spending resources on writing, monitoring
and enforcing long-term supply and subcontracting contracts - instead of
going into spot-market transactions, which do not require high transaction
costs - is the most obvious example of such activities (Richardson,
1972). Financia) hedging and inventory holding are equally important
examples. Toe costs involved in these activities are ali transaction costs
in the sense that they are not necessary from a purely engineering point of
view. With reduced instability of the system, there will be less of these
activities and consequently less transaction costs in the economy.
There are also arcas other than macroeconomic stabilisation to which
our insights about the role of the state in economising transaction costs
apply. Macroeconomic instability is essentially a case of coordination
failure between the activities of disparate individuals (Leijonhufvud,
50 The Instinaionalist
Theory Policy
50 50
1981). According to Koopmans (1957), coordination failure may occur

due to secondary uncertainty, which exists due to the 'lack of
communication, that is from one decision maker having no way of
finding out the concurrent decisions and plans made by others (or
merely of knowing suitable aggregate measures of such decisions or
plans)' (p. 163). If this is the case, we can extend our argument to the
coordination problem in general.
A special case of the coordination problem is that of 'pure' coordina-
tion, which exists when there are multiple equilibria that are equally
preferred by every party involved (see Figure 2.2). In this case, although
the agents are indifferent between different outcomes (that is, there is
no clash of interests), there is no guarantee that all of them will go for
one particular equilibrium. Deciding which side of the road people
should drive on is an obvious example of the pure-coordination pro-
blem. No one would ex ante mind driving on the left or on the right as
long as everybody else does likewise, but it is still necessary to decide
on which side of the road people will drive. Although a spontaneous
norm may emerge to resolve the coordination problem (Schotter, 1981),
reaching a coordinated outcome may involve high transaction costs (for
example the cost of bargaining), especially when many agents are
involved. And in this case, the state may intervene to provide the norm
for coordination, consequently reducing the transaction costs involved
in the process.
B
Left Right
Left (1, 1) (0, O)
A
Right (0, O) (1, 1)
Payoffs to (A, B)
Figure 2.2 An example of the 'pure' coordination problem

A New lnstiuaionalist Theory of State Intervention 51
A more generalised problem of coordination exists 'if the payoff

space of the game it defines is such that at any equilibrium point, not
only does no player have any incentive to change his behaviour, given
the behaviour of the other players, but no player wishes that any other
player would change either' (Schotter, 1981, p. 22). In this case, every-
one wou]d prefer a coordinated outcome to an uncoordinated outcome,
but it may be that each may prefer a particu]ar coordinated outcome.21
A schematic example of such a case is given in Figure 2.3, where two
VCR producers, A and B, are using two different types of techno]ogy,
say, VHS and Beta techno]ogies, but want to introduce a national
product standard, which will bring about network externalities or what
David ( 1985) caJls 'system sca]e economies'. In Figure 2.3, A (B),
being a VHS (Beta) producer wiJJ prefer the standard to be VHS (Beta)
technology, but both will prefer either of the coordinated outcomes to
the uncoordinated ones.
B
VHS Beta
VHS (2, 1) (O, 0)
A
Beta (O, O) (1, 2)
Payoffs to (A, B)
Figure 2.3 An example of the generalised coordination

problem
The transaction costs invo]ved in resolving this generalised coordina-

tion problem through prívate negotiations are Jike]y to be higher than in
the case of the pure coordination prob]em, because more intense bargain-
ing may be necessary dueto different people's preferences for different
outcomes (that is, there is a partial clash of interests). Such costs will be
particularly high when many agents are involved, which means a multi-
tude of bargaining over, writing, monitoring and enforcing the con-
52
tract.22 And indeed sorne in the tradition of new institutional econo-
mies (for example North, 1990b) have suggested that superseding
prívate attempts at coordination with state intervention may greatly
reduce transaction costs in the economy (although they do not put it in
this particular way).23 State imposition of a system of weights and
measures (North, 1990b) and of technological standards (see Porter,
1990, ch. 12, for examples) are good examples of this.
In order to achieve the coordinated outcome, the state need not com-
pletely supersede market transactions, as in our examples above. lndeed
such an exercise may be prohibitively costly, as the examples of central-
planning practice in the socialist countries show (Coase, 1988; for
further references, see Chapter 1.4. l ). There exist ways in which the state
can reduce the transaction costs in the economy without necessarily
superseding all market transactions.
First of all, the state can reduce the transaction costs within the
economy by changing the institutional configuration of society. The
state can provide legal backing to organise agents into large groups and
thus reduce the number of necessary bargainings (and consequently the
transaction costs involved). Social corporatism in Scandinavia or
Austria, whereby the workers and capitalists are respectively organised
into encompassing organisations and conduct nationwide bargaining, is
a good example of this - although the transaction-cost aspect of social
corporatism has been little discussed (see the Introduction to this book
for references).24
Secondly, the state can intervene, through its influence on the educa-
tion system and the mass media, to promote a national 'ideology' or
value system, which will help to reduce the costs involved in the ex-
change of infonnation and bargainíng.P As Arrow (1974) observed, it
is 'easier to communicate with other individuals with whom one has a
common approach ora common language' (p. 42), and therefore con-
tracts between agents sharing the same ideology will reduce the
bargaining, monitoring and enforcement and other transaction costs
required (North, 1981, ch. 3).26 This is obviously what people have in
mind when they say that a homogeneous society like Japan or Korea is
easier to manage than a heterogeneous one like the USA or India.
Although this role of the state may be objected to on the ground that it
can be used to foster 'false consciousness' and other undesirable
qualities among the people, it is undeniable that an ideological
campaign can pJay a tremendously important role, as can be seen in its
A New Institutionalist Theory of State Intervention 53
role in the process of rapid industrialisation in many late-developing

countries such as Japan, the Soviet Union, China and Korea (see
Chapter 4.5 for the Korean case).
Most interestingly, the state may save transaction costs in the economy
by providing a focal point, or consensus, around which decisions can be
coordinated (on the concept of 'focal point', see Schelling, 1960). For
example, investrnents in complementary projects may be made only at
high transaction costs when they are agreed upon through private
bargaining (Richardson, 1971, 1972). lf this is the case, the provision of a
focal point by the state may save the transaction costs involved in these
bargaining processes. A good example of this is the Japanese and the
French 'indicative-planning' exercises, where the state provides a
'vision' for the future economy and induces private agents to work
toward the same goal (see Johnson, 1982, on Japan; Cohen, 1979, on
France).27 For example, in the Japanese case, 'a sense of overall direction
to the overall evolution of the economy has been provided by the annual
economic white papers of the Economic Planning Agency with their
thematic tilles and by the forward looks or 'visions' published every two
or three years by the Mm, via the Structure Council . . .. . In practica!
terms, these documents provide criteria or orientation against which
countless individual decisions by private and public officials can be
tested and hence given order and coherence which could otherwise only
be achieved by a muen higher degree of centralisation' [emphasis added]
(Renshaw, 1986, p. 144).28
CONCLUSION
While recognising that state intervention to correct market failures

inevitably incurs certain costs, which may indeed exceed the benefits
from improved allocative efficiency, we argued that such a possibility
does not warrant the laissez-faire policy prescription of the govern-
ment-failure school. Information costs can be reduced through appro-
priate changes in the organisational structure of the bureaucracy and in
the values held by the members of the bureaucracy and society. Rent-
seeking costs may also be reduced through changes in the rules of
political competition and the tools of state intervention.
Drawing on the insights of new institutional economics, which
emphasises that the market is not the only viable coordination
mechanism, we constructed a theory of state intervention that sees
54
the state - along with the market, the firm and other economic
institutions - as an important device of coordination. We emphasised
that the state may resolve the coordination problem at a lower cost
than the market (and other economic institutions) and thus reduce
transaction costs, which are the costs of coordination, in the
economy. Institution of an effective property-rights system, macro-
economic stabilisation, organising society into large groups, promo-
ting national ideologies, and coordinating complementary investment
decisions are examples of such a role. This type of intervention is
particularly attractive because it is relatively cheap compared with
other types, which may indeed incur large transaction costs (for
example central planning).
Our theory of state intervention provides a framework that can
incorporate insights from the various theories of state intervention we
examined in Chapter J. Our theory considers both the benefits and the
costs of state intervention, in contrast with the market-failure approach
and the government-failure approach which deal, respectively, with
on]y the benefits or the costs of state intervention.I? Although our
theory is still far from satisfactory, we think it can serve as a good
starting point for the construction of a sophisticated theory of state
intervention that allows us to look at the problem in a more balanced
and realistic way. In the next chapter we apply our theory of state
intervention to the area of industrial policy and demonstrate how it
improves on the existing theories of state intervention.
3 The Political Economy of
Industrial Policy
INTRODUCTION
Until recently there existed a moderate consensus on the agenda of the

debate on the role of the state, although there were intense debates con-
cerning how best to achieve the individual items on the agenda. The
items on the agenda included an improvement in income distribution,
the achievement of macroeconomic stability, the provis ion of public
goods (for example infrastructure, education and environmental
protection) and, more controversially, anti-trust activities. State inter-
vention in industry was, except anti-trust activities, looked at with sus-
picion as opening the window of opportunity for business interests to
loot the state exchequer. This suspicion seemed more than natural when
state intervention in industry - or industrial policy - did not make theo-
retical sense according to the conventional framework. Nevertheless,
as we mentioned earlier, the recent rise ofEast Asian economies where
the state has implemented strong industrial policy measures with great
success has aroused interest in industrial policy, as manífested in the
ongoing debate on the applicability of industrial policy in other, notably
Anglo-Saxon, countries.
In this chapter we argue that industrial policy not only makes sense
but can sometimes provide a better alternative both to the unregulated
market and to other fonns of state intervention (for exarnple central
planning). After reviewing the industrial policy debates, we introduée
sorne recent theoretical developments in the studies of econornic
institutions and technical change and spell out the logic of industrial
policy, both from the static and the dynamic points of view. Regarding
the fonner, we discuss why the rnarket mechanism may lead to
coordinatíon failures and why such failures can be costly. Then we
discuss the role of industrial policy in overcorning such failures.
Regarding the latter, we discuss the nature of econornic change and see
what role industrial policy can play to promote it. We then move on to
discuss possible problems of industrial policy. Problems of
infonnation, rent-seeking, politics and institutions are exarnined.
55
3.1 THE INDUSTRIAL POLICY DEBATES
Despite the fact that industrial policy, far from being a novelty of East
Asia, has been an integral part of economic policies of many advanced
capitalist countries during the post-war period, it has become an
important issue only recentiy. In the English-speaking world the OECD
has been the pioneer in this area (see the series of country studies
published by OECD in the early 1970s). In the UK industrial policy
became a controversial issue with the (not hugely successful) intro-
duction of industrial policy programmes by the Labour government in
the late 1970s. l The famous UK deindustrialisation debate also, to a
degree, discussed industrial policy as a possible way to halt deindus-
trialisation and revive the economy.2 During the 1980s, studies of
various European countries' policy responses to the industrial crisis of
the late 1970s also emerged.' The issue of industrial policy, however,
has probably been most hotly debated in the USA, especially in the
early 1980s, with the Harvard Business Review as the major forum.4
The recent rise in strategic-trade-policy literature has also been heavily
influenced by (and has influenced) the industrial-policy debates.5
3.1.1 Does Manufacturing Matter?
One of the central points made by the proponents of industrial policy is

that manufacturing does matter, although a pro-manufacturing attitude
<loes not necessarily imply an endorsement of industrial policy.f The pro-
ponents of industrial policy argue that the UK and the US economies
are deindustrialising (the shrinking share of industrial output and
employment in the national economy) due to neglect of thc manu-
facturing industries, and that, given the vital importance of the manufac-
turing sector for a prosperous economy, this is a dangerous sign (Cohen
and Zysman, 1987). They think that macroeconomic measures, albeit
important, are not sufficient for a vigorous development of manufac-
turing since allocation of capital is more important than aggregate capital
formation for productivity growth (Reich, 1982, p. 75). The conclusion,
then, is that the state should intervene to promote industrial development,
if necessary using industrial targeting (Reich, 1982; Johnson (ed. ), 1984).
Many opponents of industrial policy argue that the advanced capitalist
economies are moving towards becoming post-industrial economies,
where service activities become the centre of economic life (for example
Bhagwati, 1988, pp. 110-14). Given that the tendency to move towards
The Political Economy of Industrial Policy 57
service activities is dictated by market forces (that is, demand for services
increases as income rises), the argument goes, favouring manufacturing
is not only unnecessary but also harmful. That is, favouring manu-
facturing would block the natural-selection mechanism of the market by
hampering the necessary reallocation of resources towards service
activities, and therefore damage the long-term viability of the economy
(for example Burton, 1983). Therefore, it is argued that we need not, and
indeed should not, have policies that favour manufacturing, not to
mention industrial policies geared to the needs of specific sectors (for the
most sophisticated version of this argument, see Bhagwati, 1989).
Confusion about the very concept of deindustrialisation and lack of

understanding of the logic of long-term structural change (away from
manufacturing towards services) seem to have produced many ill-
informed discussions on the deindustrialisation issue. Fortunately sorne
recent studies have spotted the source of confusion and clarified sorne
of the major theoretical issues (Rowthorn and Wells, 1987; Baumol
et al., 1989). The conclusions emerging out of the 'manufacturing-
matters' debate are the following.
First of ali, the long-term structural shift towards a service economy
does not happen solely because people want more services as they grow
richer, as was believed by sorne proponents of the theory of post-
industrial economy. The major reason for such a structural shift in
employment towards services seems to be the (relative) cost-inflation of
services due to their lagging productivity growth (compared with that
of manufacturing), rather than a real shift in demand towards services
as incomes rise.7
Secondly, deindustrialisation, defined as the decrease in the share of
manufacturing employment in total employment (and the decrease in
the share of manufacturing output in total output in current prices), is
an inevitable long-term result of differential productivity growth rates
between manufacturing and services, and is not necessarily related to
the declining competitiveness of the economy's manufacturing sector.
Even successful exporters of manufactures, such as Japan and West
Germany, have experienced deindustrialisation in this sense, which
implies that, contrary to what was believed by those who condemned
deindustrialisation as a sign of industrial decline, deindustrialisation
and industrial decline are not one and the same thing, although
industrial decline can affect the timing and scale of deindustrialisation
(see Rowthorn and Wells, 1987, ch. 1). Therefore it is wrong to argue
that an economy's manufacturing sector is in trouble solely on the

ground that it is deindustrialising in terms of the above definition.
Thirdly, the fact that deindustrialisation is an inevitable long-term
trend <loes not necessarily mean that a country can ignore manufacturing
completely and rely fully on services. This is largely because many
services are either basically non-tradable (for example govemmental
services, legal services, child care, elementary and secondary education)
or have a large non-tradable component (for example transportation,
distribution), although there are other services that have become, or are
rapidly becoming, tradable (for example financia! services, management
consultancy, higher education). With a growing share of services in
national income, compensating productivity growth in manufacturing is
needed - on the reasonable assumption that no dramatic increases in
productivity in agriculture and services are likely in the foreseeable
future - if a country wants to maintain its income level without running
into balance-of-payments problems.
3.1.2 What is Industrial Policy?
A major problem with industrial policy issues is that the very concept of
industrial policy is not clearly defined, resulting in heated but often
fruitless debates. A good example of this is the discussion on the postwar
Japanese experience, which inspired many of the industrial-policy
debates. Opponents of industrial policy point out that subsidies and
govemmental loans to industries in Japan are small (in relative terms),
even smaller than in many European countries, and on this ground claim,
as the title of one article goes (Trezise, 1983), that 'industrial policy is
not the major reason for Japan' s success'. Proponents of industrial policy
argue that the non-quantifiability of the famous Japanese 'administrative-
guidance' system makes people underestimate the success of Japanese
industrial policy (Boltho, 1985). Unless we define what we mean by
industrial policy, we cannot judge who is correct and who is not.
Reich (1982), the most prominent proponent of industrial policy in the

USA, includes the following policy measures in his definition of indus-
trial policy: favouring promising industries; creating skilled workforces;
developing infrastructure; regional policy (p. 75). Pinder (1982), a
British proponent of industrial policy, goes a step further and regards ali
of the following as components of industrial policy: general industrial
support policies such as manpower policy; fiscal and financia) incentives
59
for investment; public investment programmes; public procurement
policies; fiscal incentives for R&D; firm-level policies such as specific
R&D support; antitrust policy; merger policies to create 'national
champions'; support for small firms; regional policies such as the
development of physical and social infrastructure and the establishment
of industrial complexes; generalised trade protection; sectoral policies
such as the organisation of recession cartels in depressed industries;
product upgrading in labour-intensive industries (pp. 44-52).
Toe tendency to adopt an encompassing definition exists even among
those who oppose industrial policy. Donges (1980), an ardent European
critic of industrial policy, categorically states that industrial policy
'embraces ail government actions which affect industry' (p. 189).
Corden ( 1980) also implicitly adopts this definition when he states that
'the best industrial policy may be to provide an adequate infrastructure,
sorne limits on the powers of monopolies and cartels, an education sys-
tem that helps to generate the human capital for industrial success, indi-
cative guidance about industrial prospects (without compulsion or
subsidies), stability and simplicity in the system of taxation, a free and
flexible capital market and a steady movement towards zero sectional
protection, whether direct and indirect' (pp. 182-3).
Despite the fact that all the above policies would have implications
for industrial development, we do not think that classifying every
policy that affects industrial development as industrial policy is a useful
way to proceed. In the above examples, industrial policy is used as a
catch-all term for policies affecting industrial performance, that is,
effectively, every economic policy. Such a practice overloads the
concept of industrial policy, rendering the concept meaningless.
Johnson (1984) provides a more focused definition of industrial policy

by defining itas 'a summary term for the activities of governments that
are intended to develop or retrench various industries in a national
economy in order to maintain global competitiveness' (p. 7), but falls
into the sarne trap of overloading the concept when he includes not
only what he calls 'micro' policy of 'industrial targeting' but also such
policies as 'governmental incentives for prívate saving, investment,
research and developrnent, cost-cutting, quality control, maintenance of
competition, and improvements in labour-management relations' (p. 9)
into the category of industrial policy.
As Johnson (1984) rightly points out (p. 9), targeting or micro-
industrial policy cannot succeed without favourable macroeconomic
conditions. However, why should ali policies that constitute precon-

ditions for the success of another policy be treated as components of
the latter? If one adopts this logic, one can argue that targeting should
be a component of macroeconomic policy because, under certain
conditions, it is possible for targeting to have an impact on such
macroeconomic variables as savings and investments. For example,
targeting sorne big projects and financing them through inflationary
means may increase ex post savings. However, does this make such
industrial targeting a macroeconomic policy? We think not. In our
opinion the best way of defending industrial policy is not to include in
it everything that is good for industrial development, but to narrow its
definition and demonstrate that its benefits are bigger than its costs.
Landesmann (1992) makes an important contribution by emphasising
the particularistic, or discriminatory, nature of industrial policy.
According to him, industrial policy is 'designed to be specific, i.e.,
directed towards particular industries, firms, regions, groups in the
labour market, etc., rather than general ... Implicit in industrial policy
formulation and execution are therefore always trade-offs between
different groups, regions, industries, etc.' (p. 245). According to this
definition, we may exclude such general policies as creating skilled
workforces or improvements in labour-management relations from the
realm of industrial policy, making the concept more focused.
However Landcsmann' s concept of industrial policy is still some-
what overloaded, because it includes policies designed to affect both
particular regions and particular groups in the labour market. True,
industrial policy affects different regions and different groups differ-
ently, but its effects on particular regions and groups are better viewed
as by-products than as aims of the policy. Likewise regional and group-
oriented policies may affect particular industries (for example setting
up an industrial park for the garment industry in a high-unemployment
region), but this does not make them industrial policies.8
Toe existing definitions of industrial policy, then, tend to be too over-

loaded to be useful in practice. We propose to define industrial policy as
a policy aimed at particular industries (and finns as their components)
to achieve the outcomes that are perceived by the state to be efjicient for
the economy as a whole. This definition is close to what is usually called
'selective industrial policy' (for example, by Lindbeck, 1981).9
In our definition, first of ali, we emphasise the words particular
industries, and therefore implicitly exclude policies designed to affect
61 The Political
Economy of
ofIndustrial
Industrial Policy
Policy 61
61
industry in general (for example educational investment, infrastructural

development) and policies aimed principally at categories other than
industry (for example regional policy, 'group-oriented' policy) from the
domain of industrial policy. Secondly, we emphasise the word efficient
to stress that the guiding principie of industrial policy in its purest forro
is efficiency, and not other aims (for example equity), Following our
discussion in Chapter 2, efficiency is defined more broadly than in
conventional economics and includes transaction-cost economising as
an important dimension. Thirdly, we emphasise the phrase the economy
as a whole to stress that, although it is directed at specific industries,
industrial policy ultimately aims to improve the efficiency of the
economy as a whole and not just that of particular industries. Therefore,
in an industrial-policy régime, whenever the efficiency objective of an
individual industry and that of the whole economy clash with each
other, the latter is permitted to dominate.l" Lastly, we emphasise the
phrase perceived by the state, to stress that the perception of the state
may not necessarily be corrector justifiable to everyone.11
3.2 THE LOGIC OF INDUSTRIAL POLICY I: THE STATIC

DIMENSION
We defined industrial policy as a policy intended to affect particular

industries to achieve outcomes that are perceived by the state to be effi-
cient for the economy as a whole. More concretely, it means that there is
a case for the state 'selectively monitoring entry, establishing
mechanisms to malee possible more ex ante coordination than is possible
through market mechanisms alone, and for govemmental regulation or
overview to constrain or supplement profit incentives' (Nelson, 1981, p.
l 09). However, what is the logic behind opting for ex ante coordination
by the state instead of ex post, or 'spontaneous' (in Hayek's words),
coordination by the market? Broadly we can say that this is because
markets fail, but this seems hardly enough. To answer this question we
need to look more closely at the nature of the coordination problem the
market mechanism is supposed to sol ve but often fails to do.
3.2.1 The Nature of the Coordination Problem
In the model of perfect competition, upon which mainstream industrial

economics is based, there is no need for ex ante coordination of the
62 The Political
Economy of
ofIndustrial
Industrial Policy
Policy 62
62
plans of different agents regarding production and pricing decisions

(Pagano, 1985, ch. 8). In this 'perfect decentralisation model', as
Demsetz (1982) aptly calls it, there is no need for ex ante coordination
because assumptions are made to ensure that the actions of individual
agents are negligible - infinitesimally small in the limiting case -- in
the sense that a unilateral action of a single agent is unable to change
the aggregate outcome (Khan, 1987, pp. 831-4). When individual
agents are negligible, there is no interdependence among individual
agents and hence no need to coordinate their activities ex ante.
One crucial assumption to guarantee the total absence of interdepend-

ence supposed in the neoclassical model of perfect cornpetition is that
production technology is characterised by decreasing returns to scale
(DRS) - at an infinitesimally small output level in the limiting case.
Even under the widely-used assumption of constant returns to scale
(CRS) - not to mention the disturbing case of increasing returns to scale
(IRS) - the perfect-competitíon model does not guarantee a solution to
the coordination problem, as was pointed out long ago by Richardson
(1960, pp. 31-2). When CRS technology prevails in a large-number
setting, ex ante, firms may behave as if the demand curve is horizontal
(that is as if they are individually negligible), but ex post there is no
guarantee that the market will clear, since an individual tirm, not being
bound by production technology, can produce as much as it wants. In
other words there is no way to determine the number of the firms and
their respective outputs in an industry characterised by a CRS produc-
tion technology, as is recognised e ven by standard neoclassical text-
books (for example Varian, 1984, p. 88). Therefore even with CRS
technology there may be so few finns in a market as to give rise to inter-
dependence and consequently to the need for ex ante coordination.12
Of course the coordination problem will not exist except in the case of
IRS technology if ali the finns can correctly predict how much the other
tirms will produce, as implicitly assumed in textbooks. However, as
Hayek (1949a) remarks, '[t]he statement that, if people know everything,
they are in equilibrium is true simply because that is how we define
equilibrium' (p. 46). lf everybody knows everybody else's plan, then why
do we need a price system or any other coordination mechanism? In other
words, equilibrium in the perfect-competítíon model is attained only
because the coordination problem is assumed away from the beginning!
As Hayek (1978) somewhat derogatorily says, 'a state of affairs
which economic theory curiously calls "perfect competition'", that is,
63 The Political
Economy of
ofIndustrial
Industrial Policy
Policy 63
63
'a situation in which ali the facts are supposed to be known' leaves 'no
room whatever for the activity called competition' (p. 182).13 He argues
that '[t]he peculiar nature of the assumptions from which the theory of
competitive equilibrium starts stands out very clearly if we ask which
ofthe activities that are commonly designated by the verb "to compete"
would still be possible if those conditions were ali satisfied ... Advert-
ising, undercutting, and improving ("differentiation") the goods or
services produced are ali excluded by definition - "perfect" competition
means indeed the absence of ali competitive activities' (Hayek, 1949b,
p. 96).
Models of oligopoly in the ncoclassical tradition recognise the

coordination problem arising from the indetenninacy of the market out-
come in a small-number setting. However their solutions to the
coordination problem are not entirely satisfactory. The usual solution to
the problem is to employ the concept of the mixed (or randomised)
strategy (for example patent race in Rasmusen, 1989, pp. 295-8). How-
ever a mixed strategy does not guarantee an optima) solution except in
the probabilistic sense that, if the situation occurred an infinite number
of times, randomising one's actions would yield the highest average
payoff. When the situation is not recurrent, employing the concept of
probability is less than meaningful (for a classic discussion, see Knight,
1921, Part III), and it is therefore dubious to describe the mixed-
strategy equilibrium as optima!. For example, how can finn A's strategy
regarding its investment in production capacity for a 4Mb memory chip
be 'randomised' in any meaningful sense, when, given the speed of
technical progress, it is clear that the next round of investment will be
in a 16Mb memory-chip capacity?
One way of avoiding the difficulty of employing probabilistic beha-
viour by individual agents in non-recurrent situations is to intcrpret
the mixed strategy as an 'evolutionarily stable strategy' (ESS),
whereby individual agents do not randomise their actions but there are
sufficient different types of agents in the population for the aggregate
outcome to be the same as when individual agents randomise (for the
concept of ESS, see Maynard Smith, 1982). However, even in the
biological world where the concept originated, the ESS equilibrium
holds only approximately, because '[geneticJ heterogeneity and
changing conditions must mean that often populations are not perched
at adaptive peaks. Even when the conditions are constant, selection
becomes progrcssively weaker towards the peak of a continuous
64 The Political
Economy of
ofIndustrial
Industrial Policy
Policy 64
64
fitness function; infinite.time and infinite populations would be needed

to achieve the peak itself [emphasis added] (Parker and Maynard
Smith, 1990, p. 31 ). And the intuitive meaning of ESS becomes even
less clear in many industrial markets where the conditions change so
rapidly that the selection mechanism does not have time to work to its
full extent and where the agents (being humans) learn and change not
only their 'genes' (behavioural characteristics) but also the selection
mechanism (the environment) and consequently the ESS itself (see
3.3.1 below).
Toe deficiency of the unregulated market as a coordination device was

already recognised by Marx, who saw firms as islands of planned
economy in the capitalist sea of anarchy. According to him, '[t]he same
bourgeois consciousness which celebrates the division of labour in the
workshop, the lifelong annexation of the worker to a partial operation,
and his complete subjection to capital, as an organisation of labour that
increases its productive power, denounces with equal vigour every
conscious attempt to control and regulate the process of production
socially, as an inroad upon such sacred things as the rights of property,
freedom, and the self-determining "genius" of the individual capital-
ist. ... [l]n the society where the capitalist mode of production prevails,
anarchy in the social division of labour and despotism in the manu-
facturing division of labour mutually condition each other ... ' (Marx,
1976, p. 477).14 Marx saw an enormous waste of resources in the fail-
ure of the market as a coordination device (what he called the anarchy
of the social division of labour) and hoped to extend the ex ante
coordination that already existed in the firm to the economy-wide level
- what he called the despotism in the manufacturing division of labour
or what Williamson (l 975) calls the 'hierarchy' - through central plan-
ning or at least sorne form of central coordination of individual
activities (Pagano, 1985, ch. 3).15
Why are coordination failures 'wasteful'? There are coordination
failures, the counterargument may run, but are not such failures correct-
ed via the competitive process whereby firms perish unless they correct
their mistakes? If so, why should any ex ante coordination be neces-
sary? For example, it may be argued that even when an industry is
characterised by IRS technology, it would still not require ex ante
coordination, because if more than the optimal number of firms enter
the industry, sorne will inevitably go bankrupt through competition,
thereby finally achieving the optima! outcome.
65 The Political
Economy of
ofIndustrial
Industrial Policy
Policy 65
65
The above reasoning assumes that resources invested in the bankrupt

firms can be instantaneously and costlessly shifted to other activities.
Nevertheless it is only in the financiers' world (or the economists'
world?), where every asset is 'general' and 'liquíd' (as in Keynes'
concept of liquidity preference), that any investment, if found unprofit-
able, can be instantly withdrawn at no, or at most little, cost.16 How-
ever, in modern industrial economies, assets are often specific to
investments and therefore cannot be redeployed without a loss in its
value (for the concept of 'asset specificity', see Williamson, 1975 and
1985). In a world with asset specificity, ex post coordination through
the market can be wasteful, as Marx argued, because a coordination
failure that involves specific assets means a net reduction in the amount
of resources available to the economy .17
lf the market fails to sol ve the coordination problem and if such failure
can produce waste, there is a case for non-market, or ex ante, coordina-
tion (Pagano, 1985, ch. 8). As new institutional economics demon-
strates, the firm (or the hierarchy, in Williamson's words) is the most
representative form of non-market coordination, but other diverse
forms of non-market coordination mechanisms exist. As Winter (1988)
puts it, '[m]arkets appear and disappear; firms expand in scope and then
turn back toward specialisation; quasi-firms and quasi-markets prolif-
erate' (p. 168).18 Central planning is also an institutional device to
solve the coordination problem (Richardson, 1971), and industrial
policy is another such device.
3.2.2 Industrial Policy as a Device of Coordination
One characteristic of modem industrial economies is the use of produc-

tion technologies that require large fixed-investments, mainly in the form
of machinery.19 Large fixed-costs mean a decreasing average cost curve,
or scale economies. Moreover, a large part of these fixed assets are
specific or 'sunk' in the sense that their costs cannot be fully recovered
when sold elsewhere. Scale economies often force firms to produce at a
scale that will allow no more than a few firms in an industry, because, by
producing at the most efficient scale, a firm can undercut its competitors
and force them out of the industry. Out of fear of extinction, other firms
have to adopt the same (or more efficient) technology or perish. The
outcome is an oligopolistic industry in which strategic interdependence
among the decisions of the firms exist.
66 The Political
Economy of
ofIndustrial
Industrial Policy
Policy 66
66
Under certain likely conditions, strategic interdependence may Iead

to inefficiency (see below; also Telser, 1987 and Yamamura, 1988),
providing a case for state intervention. The intervention needed here is
not necessarily an antitrust-type policy, because the benefits from
breaking up the oligopolistic firms (that is reduction in the deadweight
Ioss) may be more than offset by higher production costs due to sub-
optimal scales of production. Below, we show the role of industrial
policy in this context.
3.2.2.1 lnvestment coordination

An industry whose cost structure is characterised by significant scale
economies is likely to experience a price war - firms selling at long-
term losses (or at prices that do not cover fixed costs) to undercut
competitors. Under adverse demand conditions, which might occur due
to factors like externa) shocks (for example a rise in energy prices) anda
slower demand growth than was expected at the time of the investment
decision, firms in the industry might prefer to engage in a price war
rather than forego sales, and hence incur heavier Iosses due to their
inability to recover fixed costs. This makes an industry with scale
economies subject to the dangers of underinvestment or overinvestment,
which may not be easily resolved through the market mechanism.
In a new industry (oran expanding industry) with scale economies, if
many of the potential entrants expect that enough others would enter
the industry to start a price war, there may be insufficient entry, result-
ing in a suboptimal Ievel of output - a case of underinvestment. On the
other hand, if they expect that not many competitors would enter, too
much investment may be undertaken, because then they have an incent-
ive to install as much capacity as possible to reduce their unit costs
(which would provide a distinct advantage in future competition) - a
case of overinvestment. However, in Richardson's words, 'over-
investment, by causing a collapse of prices, will penalise ali suppliers'
(Richardson, 1971, p. 441). And if sorne firms go bankrupt in this
process, the resources put into their investments will have been wasted,
insofar as they involve specific assets.
Since under- and over-investment are essentially problems of strategic
uncertainty (each potential entrant not knowing the intentions of
others), the state can intervene in this industry to assure optima) entry
by guaranteeing potential entrants that there will not be more than
optima} entry. lt can do this through arbitrating private bargaining
The Política/ Economy of Industrial Policy 67
among potential entrants, but also by superseding private dealing and
thus reducing the transaction costs involved in such bargaining. Licens-
ing entry and regulating capacity expansion are the most common
forms of state-imposed investment coordination. The negotiated indus-
trywide investment plans (the so-called 'investment cartels') in Japan
during the l 960s for industries like steel, vinyl chloride, synthetic
fibres, pulp, paper, cement, petroleum, petro-chemicals, cars, machine
tools and sorne branches of electronics are classic examples of invest-
ment coordination achieved through state-led private negotiation (Dore,
1986; Magaziner and Hout, 1980). An interesting variety of investment
coordination is 'conditional entry' whereby the state links the number
of entrants (or the scale of new capacity) to changes in demand
conditions. An example of conditional entry is the Korean passenger-
car industry, where, faced with a lagging demand growth, the state
forced one of the three existing firms to exit on the condition that it
would be allowed back when demand expanded (see Chapter 4.4.2).
3.2.2.2 Recession cartel

Even industries with optima! capacity may experience price wars if .
there are unforeseen fluctuations in demand due to, say, downturns in
the business cycle, a sudden import penetration, changes in raw-
material prices or world recession (in the case of export-oriented in-
dustries). If the fall in demand is temporary, it may be desirable to
organise a 'recession cartel', whereby individual firms limit their
production for a limited period of time, rather than allow a price war.
In the conventional wisdom, cartel arrangements are strongly opposed
because they are seen as creating allocative inefficiency (that is
deadweight loss) in the process of transferring consumer surplus to
producers (that is the process of creating monopoly profit). However the
costs of cartels (that is the deadweight loss) should be weighed against
their possible benefits. First of ali, as we have already discussed, when
there are specific assets involved price wars can lead to bankruptcy and
therefore social waste. Recession cartels may allow such waste to be
avoided. Moreover, even assuming that there is no waste involved in
bankruptcy, prohibiting recession cartels may increase allocative
inefficiency in the Ionger run. As Okimoto (1989) points out, without a
recession cartel stronger firms will survive at the cost of weaker ones,
eventually extracting more monopoly profit after the recovery. Thirdly,
and most importantly, letting firms engage in a price war may have
disastrous consequences for long-term productivity growth if firms have
68 Economy of
The Political Economy ofIndustrial
Industrial Policy
Policy 68
68 68
lo reduce their investment levels in order lo malee up their losses from

the price war.
As was common in the interwar period in many advanced capitalist

countries, recession cartels may be organised lhrough prívate ini-
tiatives. Nevertheless such cartel arrangements may be costly to
organise, for example, when a large number of finns are involved,
due to the cost of overcoming the collective-action problem. For
another example, if the sellers do not publicly quote prices and make
separate deals with separate buyers (say, because the number of the
buyers is small), it will be difficult to detect breaches of the cartel
agreement (Tirole, 1988, p. 241). In this case the cartel may have to
spend a lot of resources on monitoring. Or if the history of the
industry is such that firms do not trust each other, working out an
arrangement may incur a large bargaining cost, If', for whatever
reason, it is costly to make a recession cartel work on prívate
initiatives, the state may intervene and organise a more credible
arrangement, thus cutting down the transaction costs involved (for
examples of state-led recession cartels in Japan, see Magaziner and
Hout, 1980, and Dore, 1986).
3.2.2.3 Negotiated exit/capacity scrapping

lf the demand downturn turns out to be of a long-tenn nature, the
cumulative costs of a recession cartel may exceed its benefits. In this
case, apparently, there is a need for market forces to weed out the
weak. However leaving the adjustment to the new long-tenn demand
situation to market forces can also be costly. A pennanent fall in
demand requires sorne finns to exit, but this may cause a war of attri-
tion, whereby no finn wants to exit first because it will benefit by stay-
ing if others exit first (Ghemawat and Nalebuff, 1985; Fudenberg and
º
Tirole, 1986).2 A war of attrition can lead to a protracted price war,
leaving everybody worse off than they would have been with timely
exits (Tirole, 1988, p. 313 ).
Of course, if there is no specific asset involved the fonn of adjust-
ment in this situation may not matter because exit (and the consequent
redeployment of physical and human capital) does not cost anyone any-
thing. However when the assets involved are specific there is a case for
orderly exit or capacity scrapping. Obviously, if there are no transac-
tion costs, the parties involved may work out a contract with side-
payment schemes. However the existence of transaction costs hinder
69 Economy of
The Political Economy ofIndustrial
Industrial Policy
Policy 69
69 69
such contracts, and there is a case for state intervention. Orderly exit or
capacity-scrapping arrangements organised or assistcd by the state can
take the following forros.
First of ali, sorne firms can exit altogether in return for sorne side-
payments. Side-payment may take the form of direct compensation by
the remaining firms, as seen in the Japanese textile industry in the early
1980s (Dore, 1986).21 State subsidies may quicken the process, espe-
cially when negotiations over side-payments prove difficult. Mergers
can also make ít easier to devise side-payment schemes, as seen in the
reorganisation of the French chemical industry (Hall, 1987, pp. 208-9).
Side-payment can also take the form of an increase in a firm's share in
other markets in return for exit from one market. This option may be
feasible if the firms concerned belong to larger entities simultaneously
operating in multiple markets (for example conglomerates) - the 1980
industrial reorganisation programme in Korea is a good example of this
(see Chapter 4.4.2).
Secondly, ali firms can scrap sorne of their capacities according to
sorne norm, for example according to each firm's share in total
industrial capacity or according to its market share. The best examples
are seen in the capacity-scrapping arrangements in the Japanese
aluminium, shipbuilding, textile, petrochemical and steel industries in
the 1970s and the 1980s (Dore, 1986, p. 142; Okimoto, 1989, p. 110).22
The advantage of capacity scrapping over an exit arrangement is that it
can improve the vintage structure of capital, thus raising overall
productivity (on the vintage effect, see Salter, 1960). A capacity-
scrapping arrangement may need state intervention more than an exit
arrangement does because it is more difficult to monitor the compliance
of the parties involved. lt is fairly obvious whether a firm is in
operation or not, but it is difficult to observe whether a firm really has
scrapped its capacity. The presence of government inspectors in
capacity-scrapping processes, as in sorne Japanese capacity-
scrapping arrangements, may help to solve this problem (see Dore,
1986).
Thirdly, there is the interesting practice of 'mothballing', defined as
stripping equipment down and concreting in the mountings so that it
requires a good deal of time and effort to rehabilitate it, as practiced in
Japan (Dore, 1986, p. 142). This mitigates the problem of credibility
that is inherent in recession cartels by making cheating costly. However
it keeps open the option of returning to the former levels of production
if necessary (although ata cost) and therefore avoids the risk of scrap-
70 The
The Political
Political Economy
Economy of
ofIndustrial
Industrial Policy
Policy 70
70 70
ping too much capacity due to an unduly pessimistic forecast of future

demand. As a hybrid between recession cartels and capacity-scrapping
arrangements, mothballing may be appropriate when it is uncertain
whether the demand downturn is permanent, while recession cartels
(capacity scrappings) are appropriate when the demand downturn is
certain to be temporary (permanent).23
Fourthly, with state arbitration or even decree, firms can divide a
market into segments and exit from sorne segments in return for the
exit of others from the segments where they are given permission to
stay. Such a segmentation arrangement may be a good idea when the
industry can easily be divided into segments (for example, ships over or
under a certain tonnage). One example of such market-segmenting or
specialisation arrangement is given by the reorganisation of the Korean
electronic telephone-switching-system industry in 1980, when faced
with serious overcapacity, each of the four incumbent firms was forced
by the state to specialise in a different product (see Chapter 4.4.2). The
market-segmentation arrangement in the Japanese industrial machinery
industry in the late 1960s is another such example (Dore, 1986, pp.
137-8).
3.2.3 Concluding Remarks: Credibility, Fairness and Flexibility
In the first part of this section we discussed why the market mechanism
may fail to solve the coordination problem, and why coordination fail-
ures can be costly. In the second part we discussed how state ínter-
vention can prevent and/or redress coordination failures. lnvestment
coordination, recession cartels and negotiated exit/capacity scrapping
were examined.
Common to all these forms of industrial policy is the problem of
strategic uncertainty. Of course the existence of strategic uncertainty does
not necessarily mean that state intervention is the optima! solution. After
ali many non-market institutions enable long-range planning by reducing
strategic uncertainty (Schumpeter, 1987, pp. 102-3; Eatwell, 1982,
p. 210).24 Long-term supply contracts, technological cooperation and
vertical integration between firms all fall into this category (Richardson,
1972). In a situation of strategic uncertainty, making one's commitment
credible is vital in working out a coordinated outcomc. And, as we
argued, state intervention can help overcome the problem of credibility in
such situations. Investment coordination by the state is a way of avoiding
overinvestment and underinvestment due to the difficulty of making
71 The
The Political
Political Economy
Economy of
ofIndustrial
Industrial Policy
Policy 71
71 71
credible commitments concerning one's investment decision. A state-led

recession cartel is a way of overcoming mistrust inherent in a private
recession-cartel arrangement. Toe presence of govemment inspectors in
capacity scrapping or mothballing arrangements can also help to make
the commitments credible.
Another problem common in all the arrangements we discussed is
that of devising a scheme that is considered fair among the participants.
Decisions regarding the quota of each finn in recession cartels, which
finns should exit, which finn should cut how much capacity, and so on,
ali in vol ve the question of fairness. The-capacity or the market share of
each finn may provide focal points for such fairness, but they need not
do so necessarily, especially when the firms involved are hetero-
geneous. For example, in the case of Japanese shipbuilding industry,
'[t]he large companies and efficiency-oriented civil servants wanted to
see the big companies cut capacity, and man y of the small companies to
close down. The small companies wanted the large ones to take ali the
cuts. Companies which had newly invested in up-to-date berths ...
wanted special exemptions' (Dore, 1986, p. 145).25 Although the state
may not necessarily be better situated to work out a 'fair' norm, it can
help this process by representing the national interest, which may serve
as a focal point in negotiations.
The third problem is the question of flexibility. The limits of human
cognition (or bounded rationality) mean that the demand (and other)
forecasts on which investment decisions are based can prove wrong.26
For example, even if demand has fallen substantially, we do not know
whether this situation will last long enough to justify exit or capacity
scrapping. And if it is not implausible that demand might improve in the
future, it may be wise to bear certain short-term costs, say through a
recession cartel, in order to keep open the option of exploiting improved
demand in the future.I? Conditional entry is one device to maintain
flexibility in an expanding industry. Mothballing is a device to maintain
flexibility in declining industries where demand is unlikely to improve.
3.3 THE LOGIC OF INDUSTRIAL POLICY 11: THE DYNAMIC

DIMENSION
In the previous section we paid little heed to endogenous technical

change, considering mainly changes in demand. However we cannot
possibly ignore this issue, as the very strength of the capitalist system is
72 The
The Political
Political Economy
Economy of
ofIndustrial
Industrial Policy
Policy 72
72
its ability to generate endogenous technical change. By its nature tech-

nical change is an unpredictable process, and no one, including the state,
can claim superior knowledge of its future course. Moreover, it is often
argued, technical change is an evolutionary process, whereby only those
who develop better technology survive. Therefore sorne opponents of
industrial policy argue (for example, Burton, 1983) that, however well
industrial policy may sol ve the 'static' problem of coordination, it will
do more harm than good in the long run because it hampers the
workings of the natural-selection mechanism of the market economy.
After ali, does not the failure of central planning prove that the
coordination problem (which it solved at least to a degree) may be far
less important than the attainment of dynamic efficiency? lndeed the
advocacy of central planning is usually based on the proposition that a
centrally-planned economy can solve the coordination problem as well
as, if not better than, a market economy, and not that it generates more
dynamism (see essays by Lange and Taylor in Lippincott (ed.), 1938;
see also Lavoie, 1985, ch. 4).28 As Rosenberg and Birdzell (1986)
argue, '[t]he failure of planning can be attributed in part to its concep-
tion of an economy system as a lifeless machine, without the interna)
capacity to change, adapt, grow, renew, reproduce itself and shape its
own future. Plans ... do not ordinarily provide for creating extensive
classes of people with capacity to engage in independent economic
activities not envisioned by the plan. But a growth system is like a
living organism with impulses of its own. The result of planning for
growth is to produce an economy that is, if nota wholly lifeless statute
of the real thing, at best a tame zoo-bred shadow of the natural animal'
(p. 331). How then can industrial policy cope with the problem of
change? Before we answer this question we need to look more closely
at tae nature of economic change.
3.3.1 Knowledge, Change and Evolution
According to Hayek and the Austrian school, the esse�ce of our eco-
nomic problem is that those variables treated as data by orthodox (that
is neoclassical) economics keep changing (Hayek, 1949b, pp. 93-4).
The market, far from already embodying ali the information necessary
for coordination, can only gradually reveal them through a competitive
process (Kirzner, 1973, ch. 1). Hayek (1949c) argues that '[t]he various
ways in which the knowledge on which people base their plans is com-
municated to them is the crucial problem for any theory explaining the
economic process, and the problem of what is the best way of utilising
72 The
The Political
Political Economy
Economy of
ofIndustrial
Industrial Policy
Policy 73
73
knowledge initially dispersed among all the people is at least one of

the main problems of economic policy - or of designing an efficient
economic system' (pp. 78-9).
According to Hayek, human knowledge can never be fully codified,
and therefore the crucial question in economics becomes: '[h]ow can
the combination of fragments of knowledge existing in different rninds
bring about results which, if they were to be brought about deliberately,
would require a knowledge on the part of the directing mind which no
single person can possess' (Hayek, 1949a, p. 54).29 In particular, as
Svennilson puts it, when the knowledge involved is technical know-
ledge, 'only a part, and mainly the broad lines, of [such] knowledge is
codified by non-personal means of intellectual communication or com-
municated by teaching outside the production process itself (quoted in
Rosenberg, 1976, p. 155).30
The virtue of the market mechanism, according to this argument, is
that it acts as the most economical mechanism through which dispersed
agents cxchange information without explicit coordination. If this is the
case, the market mechanism may need be preserved to promote eco-
nomic change, because it is 'highly conducive to the achievement of
many different individual purposes not known as a whole to any single
person, or relatively small group of persons' (Hayek, 1978, p. 183).
If we recognise the importance of competition in generating change,

should we not understand the market process as an evolutionary process,
whereby natural selection operates to pick the winners? And does this
not mean that industrial policy is harmful because it attempts to tinker
with the natural-selection mechanism, which is beyond any human
comprehension? We think likening economic process toan evolutionary
process is a helpful analogy.U However an analogy is an analogy, and
therefore should not be taken too literally.
First of ali, biological evolution is characterised by the lack of con-
scious planning (Gould, 1983; Dawkins, 1986), whereas economic
evolution is characterised by the human ability to learn consciously
(from one's own and others' experience) and accordingly change one's
behaviour. That is, mutation at the genetic level is essentially a
random process, whereas economic mutation - or 'industrial mutation'
in Schumpeter's language (Schumpeter, 1987, p. 83) - is often subject
to intentional changes.P That this is the case is potently demonstrated
by the examples of sorne late-developing nations which forged ahead
by overcorning initial disadvantages through conscious leaming,
74 74
despite the existence of 'cumulative causation' - where the initial

(dis)advantage leads to further (dis)advantages - in modem industrial
economies (Abramovitz, 1986; Nelson, 1991).33
Secondly, biological evolution is essentially a Darwinian process in
which only hereditary characteristics can be transmitted, whereas eco-
nomic evolution is essentially a Lamarckian process in which acquired
characteristics can also be transmitted (Hodgson, 1988, p. 143). This is
because human beings have the ability to codify knowledge (for
example, Janguages and signs), and, more importantly, store it (for
example, books, computer memories), however limited such ability may
be. And that acquired characteristics, and not just hereditary character-
istics, can be transmitted means that Jearning plays an important role in
the process of economic evolution.
Thírdly, natural selection in the biological world, while systematic, is
índependent of the actions of the units of selectíon, that is, the indi-
vidual organisms. However the selection mechanism in economic Jife is
not 'natural' in the sense that it is totally out of reach of the conscious
attempts by the objects of selection (in this case, firms) to change it to
their advantages. The participants in economic life enhance their ability
to survive not only by changing themselves (the genes) but also by
changing the environment (the seiection mechanism). For example, a
firm operating in an industry with network externalities (for example
typewriters, computers, telecommunications) can change its chance of
survival by spreading its own technology - for example by encourag-
ing other smaller firms to produce clones of its products or by provid-
ing Joans to its customers. Advertising is another example whereby
firms change their possibility of survival by changing the selection
mechanism - that is, consumer preference.
3.3.2 Industrial Policy as a Device to Promote Change
In the previous section we examined the nature of change in the

capitalist economy. In particular we examined the evolutionary
argument, which likens the process of change in the capitalist
economy to biological evolution. We argued that the fundamental
difference between biological and economic evolution is that, in the
latter, the units of selection have the capacity to intentionally 'mutate'
and change the selection mechanism itself, at least to a degree. This is
essentially due to the human ability to Jearn, especially from others,
and to the ability to pass on (at Jeast part of) the knowledge acquired
75 75
through codification (for example by writing a book on Japanese

business management) and institutionalisation (for example by
introducing sorne elements of Japanese business management). How
then can industrial policy be used as a means to promote economic
change and learning?
3.3.2. l Economic change, coordination and industrial policy

One important point not addressed by the opponents of industrial policy
who employ the (misunderstood) evolutionary argument is that eco-
nomic changes may require coordination to be successful. In a world of
interdependence, the existence of a better alternative does not neces-
sarily mean the advent of a change. For example, there exist more
efficient alternatives to the QWERTY typewriter (and computer) key-
board, but an agent (or even a group of agents) who unilaterally shifts
toan alternative keyboard will be penalised unless others also opt for it
(David, 1985). More generally, when interdependence prevails between
economic agents, changes would not automatically be made without
the guarantee that complementing changes would also be made
(Richardson, 1960, ch. 2).
For example, if a successful computer industry depends on a strong
semiconductor industry, people will be reluctant to invest in the computer
industry unless there is a credible commitment for adequate investment
by the potential investors in the semiconductor industry, and vice versa.34
As Abramovitz ( 1986) argues, if 'the capital stock of a country consists of
an intricate web of interlocking elements', then 'it is difficult to replace
one part of the complex with more modero and efficient elements without
a costly rebuilding of other components' [emphasis added] (pp, 401-2).
Now, '[t]his may be handled efficiently if ali the costs and benefits are
interna! to a firm', but when the capital stock is 'interdependent in use but
divided in ownership', and thus the accompanying costs and benefits of
change are divided among different firms and industries, 'the adaptation
of old capital structures to new technologies may be a difficult and halting
process' [emphasis added] (p. 402).35
Although it is possible that potential investors in complementary

projects may devise a contract between themselves, such a contract
may be costly to draw up, particularly when there is asymmetry in asset
specificity of investments between different investors (the failure of the
complementary investments to materialise can be more damaging to the
investor with greater asset specificity). State intervention in this case
76 76
may cut sharply the transaction costs involved in such contracts. Such
intervention need not involve financia! resources such as subsidies. As
we discussed earlier, govemmental announcements (for example the
French and East Asian 'indicative-planning' exercise) may suffice if
they can provide obvious focal points for coordination between com-
plementary investments (see Chapter 2.3.2).36 Financia! incentives
provided by the state, say, for cooperative research in new industries,
although not necessary, may make the state's commitment to its
announcement more credible by serving as a signalling device (Porter,
1990, ch. 12). Thus seen, industrial policy that coordinates comple-
mentary investment decisions may be essential for economic change in
a world of interdependence, rather than be an obstacle to it.37
3.3.2.2 Codifiability of knowledge, product cycle and industrial

policy
Toe limited codifiability of technical knowledge requires that we have to
incorporate the problem of knowledge generation into our industrial-
policy framework. Toe problem of knowledge generation is captured in
more practica! terms by the theory of the 'product cycle' (see Vemon,
1987), which is known to be incorporated into Japanese industrial-policy
practice (Okimoto, 1989, ch. l; Magaziner and Hout, 1980, ch. 4).38
According to this theory, a young market is characterised by a phase
of experimentation in which different ways of doing the same thing vie
with each other. As the market matures, most technical knowledge
becomes codified and easily transmittable. When a few technologies
emerge as the best-practice ones, they are adopted across the industry,
as firms leam from the experiences of others. As a market matures and
finally becomes senile, the discovery potential in that rnarket gradually
diminishes (as knowledge becomes more codified) and the role of
'competition as a discovery procedure' (Hayek, 1978) is accordingly
reduced.39 Let us examine how the idea of the product cycle can be
incorporated into the practice of industrial policy.
In the infant stage of an industry, where experimentation is necessary to

generate new knowledge, industrial policy should encourage it. More
aggressive experimentation and learning can be encouraged by
providing firms with a more stable environment (through, say, patent
systems, subsidies, tariffs or other types of entry barrier), as the familiar
infant-industry argument goes. At this stage it would also be necessary
for the state to set up institutional arrangements that can cope with the
77 77
new externalities generated by this new industry (Nelson and Soete,

1988, pp. 633-4). Moreover industrial policy has a valuable co-
ordinatory role to play at this stage. lntroducing national product and, if
necessary, process standards, coordinating competing investment deci-
sions to prevent underinvestment and overinvestment, and ensuring that
complementary investments are made will ali be useful at this stage.
As the industry matures, experimentation becomes less important.
According to Hayek, when 'we have a highly organised market of a
fully standardised commodity produced by many producers, there is
little need or scope for competitive activities because the situation is
such that the conditions which these activities might bring about are
already satisfied to begin with. Toe best ways of producing the commod-
ity, its character and uses, are most of the time known to nearly the
same degree to ali members of the market' (Hayek, 1949b, p. 103).
When the technology has been stabilised and codified, on the one hand,
and the institutional arrangements necessary to cope with with new
configurations of asset specificity and the resulting uncertainty have
been set up, on the other hand, the 'static' dimension of industrial
policy, which we discussed earlier, becomes more important.4 º
As an industry enters its senile stage, production shrinks, labour is
shed and capacity is scrapped. The material and human resources
employed in an industry may be highly specific to that industry so their
redeployment may be extremely difficult, or even impossible. In the
face of a possible loss in value of specific assets, the owners of the
assets will resist change, and this may result in a considerable waste of
resources if a war of attrition among the firms concerned or protracted
labour disputes take place. Toe role of industrial policy in this phase
will be to encourage private negotiations regarding exit and capacity
scrapping between the relevant agents, or even to impose a centralised
solution when negotiations reach a stalemate. Retraining and relocation
programmes organised by the state will also greatly assist the process
of negotiation by reducing the would-be displaced workers' resistance
to the firm's decision to exit or scrap the capacity.
3.3.2.3 Diversity of innovatory sources and industrial policy

Nelson ( 1981 ), in discussing the innovation mechanism of the capitalist
economy, argues that the waste that is bound to be generated by
competitive innovative attempts (say, due to duplication) may be a
price worth paying to avoid the dangers oí relying on a single mind for
innovation (that is monopoly). Oras Abramovitz (1986) puts it, 'in the
78 78
uncertainty that obscure early effort to explore new fields, it would be

quite unwise to concentrate ali effort on a single approach to a still
cloudy goal' (p. 41 ). This is because innovation is basically a chase
after a moving target, a job at which nobody can claim absolute superi-
ority. That is, 'were man omniscient and omnipotent, he would not
choose to organise his R&D activities through prívate enterprise [given
the wasteful nature of competitive R&D activities]. The case for prívate
enterprise as an engine of progress must be posed in recognition of
bounded rationality' (Nelson, 1981, pp. 108-9).
This is a powerful argument. Unless human rationality is unbounded,
there will be a pressing need to preserve a diversity of the sources of
knowledge in an ever-changing world - although this statement should
not necessarily be interpreted as an advice against ali 'collusive' beha-
viours amongst firms (Jorde and Teece, 1990, pp. 81-2).41 However,
does the state have any role to play, if this is the case? Should it not just
leave things to evolve on their own?
One point against the apparent laissez-faire implication of this
'diversity' argument is that imperfections in the capital market put a
follower firm in a disadvantaged position if there is a high fíxed-
investment requirement.42 In this situation the state can actas a surrogate
capital market and subsidise a potential entrant that is deemed to he at
least equally capable as the incumbent firm except in its financial ability.
State-organised venture-capital schemes conducted in countries such as
Korea, France and the United Kingdom are good examples of this.
The state can contribute to increasing the diversity of innovatory
sources in a more direct manner. For example it may expand the pool
of potential entrants into an industry with high R&D components by
subsidising related R&D activities by firms who operate in similar
lines. Or, altematively, it may encourage related basic research in uni-
versities or public laboratories, which will publicise the results of their
research. This of course carries sorne possibility of duplication, but it
may be a price worth paying to preserve and develop diversity.
3.3.3 Concluding Remarks: The Socialisation of Risk
In the first half of this section we discussed the nature of change in the
capitalist economy. We argued that the process of change in the capital-
ist economy is best characterised as a quasi-biological evolutionary
process whereby the agents can and do change both their own 'genes'
(behavioural characteristics) and the selection mechanism (or environ-
79 79
ment). Drawing on this argument, we discussed the dynamic dimension

of industrial policy in the second half of the section. It was pointed out
that coordinating changes, encouraging experimentation and preserving
diversity are the most important roles industrial policy can play.
The crucial theme emerging from our discussion in this section is that
of the 'socialisation of risk', whereby risks involved in economic
changes are borne by society rather than by individuals. In the models
in the orthodox tradition, where individuals make decisions in an
atomised fashion, risks involved in changes are necessarily borne by
the individuals. To those who subscribe to this view, the socialisation
of risk opens doors to the moral hazard of excessive risk-taking by
those individuals whose risks are borne by society (see 3.4.1 below).
However in the real world many changes involve interdependent
decisions. If the risks involved in these situations have to be solely
borne by the individuals, necessary changes may not come about, The
socialisation of risk through state intervention is a means of promoting
changes that involve interdependence.
Contrary to what is implicitly assumed in mainstrearn economics, the
capitalist economy has developed on the basis of the growing social-
isation of risk. As Rosenberg and Birdzell (1986) argue, '[t]he West
has grown rich, by comparison to other economies, by allowing its eco-
nomic sector the autonomy to experiment in the development of new
and diverse products, methods of manufacture, modes of enterprise
organisation, market relations, methods of transportation and communi-
cation, and relation between capital and labour' (p. 333). And in such a
process, institutional arrangements that, by 'internalising benefits and
externalising costs of private investment' (North, 1981, p. 62), allow
experimentation and risk-taking beyond a scale whose risk can
personally be borne out by the experimenter (for exarnple systems of
limited liability) have played an important role. The socialisation of
risk through state intervention, then, may be seen as but one extension
of these already existing institutional arrangements.
3.4 POSSIBLE PROBLEMS OF INDUSTRIAL POLICY
3.4.1 Problems of Infonnation
One common objection to industrial policy is based on problems of

information (for example Burton, 1983; Grossman, 1988). There are
80 Political Economy of
The Political of Industrial Policy
Policy 80
80 80
two major elements in this argument (also see Chapters 1.4.1 and 2.1 ).
First of ali, it is argued that the state does not possess enough informa-
tion to decide correctly on the future industrial structure of the econo-
my. This is, according to our earlier classification, the problem of
'insufficient information'. Secondly, it is argued that the state is at an
informational disadvantage vis-a-vis the firms that are subject to indus-
trial policy. The firms, the argument goes, may use their informational
advantage to extract more than they deserve on social grounds (a moral
hazard problem). This is the problem of 'asymmetric infonnation'.
With the exception of sorne staunch free-marketeers (for example,
Burton, 1983), those who employ informational arguments (for exam-
ple, Caimcross et al., 1983; Grossman, 1988; OECD, 1989) support a
generalised industrial policy targeted at certain types of activities rather
than at particular industries (for example, investment, R&D), against
selective industrial policy - the type of policy we defined as industrial
policy proper at the beginning of the chapter. If the state has ali the
relevant information, the argument goes, particularistic interventíons
may work, but since this is unlikely to be the case, the state should
support productive behaviour in general rather than pick the winners on
the basis of incomplete information (Price, 1980; Líndbeck, 1981 ).
3.4.1.1 lnsufficient information

Concerning the insufficient-information argument, note first that insuf-
ficient information does not prevent us from planning our future eco-
nomic life. Actually the uncertainty of the future is exactly the reason
why we plan for the future. Overcoming uncertainty is one of the most
important functions of business management, especially in large modern
corporations (Richardson, 1960; Williamson, 1975; Stinchcombe,
1990). A firm chooses its production technology, capacity, liquidity
position, inventory level and so on, to minimise the potential loss in case
of abrupt changes in environmental factors such as market demand,
macroeconomic conditions and the state of technological development -
in other words, to overcome parametric uncertainty. A firm goes into
long-term binding contracts concerning its purchases of raw materials,
labour power, parts and equipment, on the one hand, and its sales of
products, on the other, to minimise a potential loss in the event of
opportunistic behaviour by its business partners - in other words, to
overcome strategic uncertainty. It is inadequate to argue that the state
should not attempt to plan the future of the national economy because of
insufficient information, when firms can and do plan their own future
despite - or rather, precisely because of - insufficient information.
Policy 81
81 81
Secondly, the infonnational requirement for intelligent state inter-

vention is not always so great as to disallow state intervention alto-
gether. Entrepreneurs themselves often operate on the basis of infonned
guesses or of 'animal spirits' in making investment decisions. Fre-
quently much of the infonnation used by the finn to make investment
decisions - for example estimates of present and future demand, the
availability of best-practice technology, the financial situation of the
finn, the distribution network of the finn - are readily available to any-
body, and not just to the finn itself. Moreover, a large part of the
infonnation used by the finn is acquired from extemal sources such as
consultants, research institutes and state agencies (for exarnple a central
statistical bureau). Thus seen, it is not necessarily true that the state
suffers from insufficient infonnation whereas the finns do not. In fact
one of MITl's resources in dealing with the private sector has been its
'superior information' (Okimoto, 1989, p. 145), thanks to the more
extensive infonnational network in the hands of the state (also see the
Korean case discussed in Chapter 4).43
Thirdly, in the context of late development (on the concept of late
development, see Gerschenkron, 1966), the problem of identifying
desirable industrial structures is far less serious. This is because late-
developers can have the 'second-rnover advantage', by which they can
watch the countries on the frontier of economic development and learn
from their experiences.44 Even in a country like Japan, which was
pretty close to the frontier of industrial development, it is recognised
that '[s]etting priorities, picking the next likely winners, has not been
difficult throughout the post-war period when the objectives of policy
were primarily "catching up" objectives' (Dore, 1986, p. 135). The
insufficient-infonnation argument loses most of its force in the case of
late-developers (that is almost ali countries) where the scope for
'conscious mutation' is great (also see Dore, 1989).
3.4./.2 Asymmetric information
Conceming the asymmetric-infonnation argument, note first that asym-
metric infonnation is not confined to the relationship between the state
and finns, but is a ubiquitous problem in economic life, as we have
repeatedly pointed out. Infonnational asymmetries exist between finns
and lending institutions and between managers and shareholders. It
exists, moreover, within the finn itself, that is, between headquarters and
subsidiaries ( or other component parts of the finn). If the asymmetry of
infonnation is always so severe as to disallow state intervention, neither
financing somebody else's investment projects nor managerial planning
Policy 82
82 82
can be justified. After ali, large modem corporations carne into being
despite the dangers of the principal-agent problem, because there are
ways of controlling managerial excesses (see Chandler, 1962). And
likewise there are ways and means to reduce informational asymmetry
between the state and firms (see Chapter 2.1.2).
Secondly, the problem of asymmetric information is not unique to
industrial policy as defined by us, but it applies to other policies too.
Moreover, general industrial policy, which its supporters assume to
have no moral-hazard problem due to asymmetric information (for
example Corden, 1980, pp. 182-3; Balassa, 1985, p. 319), may suffer
even more acutely from such a problem. As contracts become more and
more general, the contingencies to be considered become more and
more numerous, resulting in prohibitive transaction costs in drawing up
effective contracts. This means that general industrial policy can be
compromised by unforeseen contingencies. An interesting example is
the US 1981 tax-code provisions, which were originally intended to
boost industrial R&D but ended up subsidising advertising firms
(Lawrence, 1984, p. 140, n. 45). Industrial policy, as defined by us,
being particularistic in its nature, tends to involve contracts that are
more custom-designed and hence allow fewer unforeseen contingencies
and less moral hazard. The use of plan contracts with specified targets
between the state and individual firms in France is the best example of
preventing moral hazard through the use of custom-designed policy
(see Hall, 1987, p. 207).
Thirdly, the asymmetric-information argument assumes that local
information is always better than global information because it is more
finely-meshed.O However, as we pointed out earlier, people with
localised information may make a substantively less rational decision
due to the subgoal-identification problem (see Chapter 2.1.2). lf the aim
of industrial policy is to improve the efficiency of the economy as a
whole, it may actually be better, under certain circumstances, not to be
affected by the localised information possessed by the firm. Especially
when the decision involves extemalities that are not borne out by the
firm, the state can make a better decision solely due to the more global
nature of its information, and not because it is a superior being.
3.4.2 Problems of Rent-Seeking and Entrepreneurship
The ever-changing nature of the capitalist economy - or Schumpeter's

'gales of creative destruction' (Schumpeter, 1987) - and the consequent
Policy 83
83 83
pervasiveness of uncertainty gives entrepreneurship a vital role to play

in the capitalist development process as the generator and/or finder of
new knowledge (Schumpeter, 1987; Dobb, 1925; Kirzner, 1973;
Nelson, 1986). However, as Baumol (1990) argues, depending on the
incentive structure in the economy, the existing stock of entrepreneurial
talents may be diverted away from productive purposes into unproduc-
tive or even destructive ones. According to him, entrepreneurial talents
can be used for destructive purposes when rents are granted to those
who are best at destroying existing assets (for example warmongering
in Europe in the Middle Ages). They can be used for unproductive
purposes when rents are granted to those who are best at transferring
existing assets, as in rent-seeking (also see North, l 990b, ch. 8).
The state has a major role to play in preventing the diversion of
entrepreneurial talents into unproductive venues by reducing pro-
fitable opportunities in those areas. As argued in Chapter 2, the nature
of unproductive activities is that they are transaction-cost-generating
activities, and many such activities exist because of strategic uncer-
tainty. As we discussed, the reduction in macroeconomic instability
through appropriate macroeconomic management would reduce the
need for activities that are specifically designed to deal with it (for
example financia! hedging), and therefore limit the diversion of
entrepreneurial talents into these activities (see Chapter 2.3.2). The
same applies to activities intended to <leal with more specific coordi-
nation problems. Investment coordination through indicative planning
and the establishment of national product standards are areas where
the state can intervene to reduce strategic uncertainty inherent in the
coordination problem and therefore the scope for unproductive
entrepreneurship.
Another crucial condition for entrepreneurship to be a productive activ-

ity is that the rent accruing to the entrepreneur should be durable but
not permanent, as already pointed out by Marx (1981) and Schumpeter
( 1987). In general people pursue rents primarily because they value the
quiet life that follows from the acquisition of a monopoly position, and
not because they are devoted to the cause of productivity growth
(although this may be the case in certain contexts). Therefore, if the
monopoly position (and the accompanying rent) is too quickly eroded,
there will be little incentive to innovate (Schumpeter, 1987, pp. 104-5).
However, if the monopoly position lasts too long, the cumulative dead-
weight loss due to its existence will ultimately cancel out the initial
productivity gains made in the process through which such a monopoly

position was established. Thus the important question is: how can it be
ensured that the rent lasts long enough to motivate people to capture it
but short enough to force people to keep improving productivity
(Richardson, 1960, ch. 3)? In employing industrial policy measures,
which are bound to create rents, this question becomes particularly rel-
evant. Toen how can the state ensure that the rents are durable but do
not become permanent?
One obvious way is to use the patent system. The patent system, by
guaranteeing a monopoly position, frees the innovator from the fear of
being caught up with, but at the same time, by limiting the length of
such a guarantee, ensures that the cumulative deadweight loss will not
ultimately cancel out the initial productivity gains. However, when it is
notjust a particular productor process that enhances social productivity
but a whole investment project (say, through spillover effects), the
patent system may not be used. And in this case, the necessary incent-
ive, the rent, needs to be created in other ways such as subsidies, import
protection and industrial licensing.
If the rents are created by means cther than the patent system, the
crucial question becomes whether the state is able to withdraw the rent
whenever necessary. This means, first of ali, that when it is contem-
plating an industrial policy, the state should set strict performance
criteria so that the rents would not go on regardless of the performance
of their recipients. For example, in France the provision of state aid to
ailing industries dependent on their performance is not unrelated to the
relative effectiveness of those aids (Hall, 1987, p. 210). Secondly, the
state has to ensure that it has the power to punish the firms if they resist
the elimination of the rents. It is nota coincidence that industrial-policy
exercises were more successful in France, whose state has control over
the banking sector, than in, say, the UK, whose state has had only a
limited control over the flow of financia) resources in the economy. 46
In addition, sorne industrial policy measures that may be used as
means to preserve diversity - for example, venture-capital schemes and
subsidisation of related commercial R&D and basic research - can also
be used as means of keeping rents from becoming pennanent. Even
when new entry does not add to diversity (that is when the entrants
have the same technology and organisational structure), prohibiting the
incumbent firm from expanding its capacity beyond what is justified by
scale-economy considerations and allowing new entrants can prevent
pennanency of the rent.
85 85
3.4.3 Political Problems: Legitimacy and Democratic Control
As we discussed in the lntroduction, an industrial-policy régime is not

merely a technical means to achieve efficiency, it is fundamentally a
régime of poli ti cal economy. This means that a discussion of industrial
policy cannot be satisfactory without discussion of the political prob-
lems associated with it. There are severa) political problems related to
industrial policy, but we shall discuss only the two which are the most
relevant in this context - legitimacy and democratic control.
3.4.3.1 Legitimacy
Sorne may argue that industrial policy should not be used because it
undennines the legitimacy of the state. First of all, by opening the door
for special interests, industrial-policy practice can erode the image of
the state as a social guardian and therefore make people question its
intentions.47 Secondly, industrial policy gives bureaucrats the power to
allocate property rights and hence creates scope for bureaucratic
corruption. In addition to its efficiency consequences (for example an
industrial license may go to an inefficient producer), corruption may
have consequences for the legitimacy of the political system (Krueger,
1990, p. 18). lf industrial policy may endanger the legitimacy of the
political system, should we not refrain from it, whatever its efficiency
gains may be?
First of ali, it should be pointed out that legitimacy is concerned with

the socio-economic system as a whole, of which the political system is
only a part. People may be generally disenchanted with the outcome of
the socio-economic system - for example, high income-inequality -
even if the state is impartial and honest. For example Dobb (1925)
argues that monopoly 'may give occasion for a psychological tendency
to antagonism and distrust on the part of dependent groups and classes
towards those in a position of advantage .... If this happens ... the
society may cease to have the "general will" which is supposed to exist
in a hannonious democratic communíty; its sections may not respond
to the same idealistic appeals, and their latent antagonism may prevent
them from subordinatíng theír own sectional interests to the success of
the whole' (pp. 157-8).48 The problem of legitimacy is much more
fundamental than whether or not a particular government, or a specific
typc of policy, is open to corruption.
86 86
Moreover, although the erosion of legitimacy is a serious danger

when conducting industrial policy, it is by no means a possibility
confined to industrial policy. Other, more general, policies may also
suffer from the problem of legitimacy. For example monetary policy
may primafacie appear to be immune to interest-group activities, but it
is well known that industrialists often lobby for expansionary monetary
policies whereas financiers usually lobby for tight monetary policies.
And there is no guarantee that such lobbying would not involve
corruption and therefore endanger the legitimacy of the state.
3.4.3.2 Democratic control

The fact that an industrial-policy régime apparently requires an élite
bureaucracy has often raised concerns about democratic control.
Especially among those who believe in parliamentary democracy of the
Anglo-Saxon variety, the weakness of legislature in 'industrial-policy
states' such as Japan and France (not to mention Korea, which has been
non-democratic for the major chunk of its modern history) has been a
great concern. To them the fact that bureaucrats, who are not subject to
popular mandate, are powerful means that the whole political process is
rigged to ensure that efficiency dominates democratic values.i? They
believe that industrial policy is less subject to democratic control
because it is open to bureaucratic discretion, in contrast with other
'even-handed' or general policies.
Against this view, it should firstly be pointed out that sorne degree of
bureaucratic control is necessary for any society of reasonable sophis-
tication, because 'many decisions have to be taken in response to rapid-
ly changing situations and cannot, except at the cost of total stasis and
chaos, be "left" until a highly democratic decision-rnaking process has
been completed. Almost immediately then, in any real situation it
becomes necessary to delegate powers from larger, more democratic
bodies ... to smaller, more "efficient" bodies. However, once such
delegation has occurred, a great deal of the real day-to-day decision-
making power is taken out of democratic channels and placed in the
hands of small minorities which may then be beyond the effective
control ofthe larger bodies' (Kitching, 1983, p. 39). That is, there may
be a certain trade-off between democratic control and efficiency in
decision-making. However no a priori criterion can tell us which mix
of democratic control and efficiency - including the one existing in an
industrial-policy régime - is the most desirable.
87 87
Secondly, it is notjust industrial policy that suffers from the problem

of democratic control. For example, those who criticise industrial
policy usually support an independent central bank, but it is not clear to
us why the democratic credentials of an official in the Japanese MITI or
the British DTI should be viewed with suspicion whereas those of a
Gennan Bundesbank official should be accepted without question. No
policy is free from the personal discretion of the policy-maker. More-
over other policies may be even less subject to democratic control
than industrial policy dueto their less transparent nature (Dore, 1987,
pp. 199-201 ). Industrial policy usually clearly reveals the beneficiaries
of the policy, whereas other policies (for example monetary policy)
often do not clearly revea) who is benefiting from them. Such trans-
parency may make it easier to exercise democratic control over indus-
trial policy than in the case of other policies, if there is a will to do so.
3.4.4 The Problem of Supporting Institutions
Opponents of industrial policy often point out that 'industrial-policy

states' have a particular set of institutional arrangements, especially an
élite bureaucracy with a wide discretionary range and a cooperative
government-business relationship.S? They argue that it is difficult to
change institutions and therefore that industrial policy cannot be a
realistic option for other countries, that lack such institutional
arrangements, no matter what merits it may have.
Although an effcctive industrial-policy régime does require an appro-

priate set of supporting institutions, the difficulty of building it should
not be exaggerated. Countries learn from their own past experience and
from other countries and engage in institutional innovations. For
example in Japan many of the institutions that are often said to have
arisen because of Japan's unique culture are actually products of con-
scious institution building. The fact that Japanese labour, product and
financia! markets were vastly more volatile in the early 1950s than they
have been since shows that the renowned 'collectivist' characteristics
of the Japanese are not 'just a "hangover" of ancient (feudal) cultural
traditions' (Dore, 1986, p. 250) but also products of conscious institu-
tional innovation (see also Magaziner and Hout, 1980, p. 2).
Moreover, institutional innovation does not necessarily take a long
time. The famous Japanese lifetime employment is basically a postwar
creation (Johnson, 1982, p. 14). The French state, which is renowned for
88 88
its interventionist and 'modemising' attitude, was famous for its laissez-
faire and 'anti-modern' attitude before the Second World War (Cohen,
1977; Kuisel, 1981). Toe well-known Swedish labour-capital consensus
emerged in a relatively short period of time out of one of the most
contested industrial relations in Europe of the 1920s (Korpi, 1983).
Moreover learning from other countries with different institutions
does not necessarily mean that a country has to exactly copy their
institutions. lt is often possible to create functional equivalents of
foreign institutions. For example the Swedish 'active labour-market
policy' and Japanese lifetime employment are very different institu-
tional arrangements, but they are functionally equivalent in creating a
positive attitude among workers toward technological change by
guaranteeing them jobs. In this regard, the following quotation from
Dore ( 1986) is well worth consideration:
[Learning from the Japanese experience] need not mean that we [the
British] have to become Japanese, absorb the Confucian ethic, or
raise our sense of national identity to the Japanese levels. What it
does mean is that we should ask ourselves whether there are not
other ways in which sorne of the things which Japanese institutions
and traditions achieve for the Japanese might be obtained by other
methods, other institutional arrangements, more consonant with our
own tradition. If close co-operation and consultation between mana-
gers and workers seems to be a precondition for rapid innovation in
manufacturing firrns, and if it is difficult to achieve this, given our
adversaria! traditions, what forrns of industrial democracy or work-
place decision-sharing might substitute for the easy acceptance of
bureaucratic hierarchy which facilitates co-operation in Japanese
firrns? If we cannot have, and do not want, lifetime employment to
be the norrn, if we want to preserve a more mobile system with the
greater personal freedom which that provides, can we at the same
time devise schemes which would give British employers the same
incentive to invest in training their employees as the lifetime
employment expectation gives Japanese employers? If the crucial
aspect of the Japanese system of financing industry seems to be the
way in which it facilitates long-term planning and investment, and
reduces preoccupations with next year's bottom line, is there any
way in which our own financia! institutions could be mended to
achieve the same effect, without necessarily modeliing our stock
exchange on Japan's? If inflation control in Japan crucially depends
89 The Political
The Política/ Economy
Economy of
ofIndustrial
Industrial Policy
Polícy 89
89 89
on institutionalised wage leadership and a nationally simultaneous

pay settlement date, does that not suggest the wisdom of re-
examining the many suggestions that have been made for introducing
synchro-pay in Britain? (p. 252).
CONCLUSION
The common reaction to the argument for industrial policy has been
one of suspicion and incredulity. The opponents of it regard industrial
policy either as a bureaucratic meddling that is at best irrelevant - for
example, 'Industrial policy is not the major reason for Japan's success'
(Trezise, 1983, tille) - or as a peculiar form of state intervention that
works only in countries with a particular culture - for example, 'Indus-
trial policy: It can 't happen here' (Badaracco and Yoffie, 1983, title).
Such reactions are more than understandable when thinking that
orthodox economic theory hardly recognises any form of coordination
other than the idealised perfect market and ignores the role of
endogenous technical change and learning.
However, as we have tried to show, industrial policy is a policy
practice that can be firmly anchored in economic theory if we incor-
porate recent developments in economic theory that take seriously the
issues of institutional diversity and technical change. As a coordination
mechanism, industrial policy can be most efficient in a context where
interdependence and asset specificity are important. In this context,
coordination through the market would incur high bargaining costs and
coordination through central planning high information costs, while
industrial policy is likely to incur little of both types of cost. When we
take the issue of technical change into account, industrial policy also
emerges as a superior way to promote it. Industrial policy does not kili
off the profit motive - which is the most irnportant, if not the only,
driving force behind technical progress - as central planning would,
and, through the socialisation of risk, it can promote changes that are
additional to what the market can produce on its own.
Industrial policy, needless to say, is no panacea. Like any other
policy, or any other form of economic coordination, it has its own costs
and benefits. Its benefits seem to have more than offset its costs in
success stories like thosc of Japan and Korea, but we have plenty of
other examples that show that its costs may overwhelm its benefits. The
90 The Political
The Política/ Economy
Economy of
ofIndustrial
Industrial Policy
Polícy 90
90 90
real question is not whether industrial policy can work or not (because
it does), but how it can be made to work. In this chapter we have tried
to provide sorne theoretical grounds for identifying the economic,
political and institutional conditions under which industrial policy
would work and have suggested sorne ways and meaos to achieve
them. The next chapter, which looks at the Korean experience,
demonstrates how they have been achieved in practice.
4 Industrial Policy in
Action - The Case of
Korea
INTRODUCTION
That industrial policy has played an important role in the postwar

economic development of countries such as France and Japan is widely
accepted. However the role of industrial policy in another irnportant
industrial-policy state, Korea, is less well known, despite the fact that
the Korean state has, as we shall see below, much more actively em-
ployed industrial-policy measures than its French or even its Japanese
counterparts. This rather unsatisfactory state of affairs is partly due to
the ideological climate of the Cold War era. For those economists who
were on the side of the 'free world', it was unthinkable that a successful
non-socialist economy such as Korea may not be a free-rnarket econ-
omy. And consequently they turned a blind eye to the active role of the
state in Korean development. Mote important than the ideological
climate, however, was the inability of neoclassical economics to
accommodate other institutions (or devices of coordination) than the
perfect market and the perfect state.
By applying the theories of state intervention and industrial policy
developed in the previous chapters, which acknowledged the diversity
of economic institutions, the present chapter provides an account of
Korean industrial policy that analyses why things work in Korea as
they do. After presenting a brief overview of Korea's economic
performance, we critically review sorne of the mainstream interpreta-
tions of the Korean experience that deliberately downplay the role of
the state, and particularly of industrial policy. The next section
explores sorne major themes behind Korean state intervention in
general. The chapter concludes with an extensive discussion of the
economics and politics of industrial policy in Korea.
91
92 The Industrial Policy in of
Political Economy Action - Korea
92 92
4. J A BRIEF OVERVIEW OF THE KOREAN ECONOMIC

PERFORMANCE
Since the early 1960s Korea has shown a truly remarkable economic
performance in terms of growth and structural change. Between 1965 and
1986 Korea's annual per capita real GNP growth was 6.7 per cent
compared with that of 2.9 per cent for the developing world as a whole
(World Bank, 1988, Table 1 ). Table 4.1 compares the growth perform-
ance of Korea since i 950 with that of other leading developing countries
- Taiwan, China, India, Argentina, Brazil, Chile and Mexico - and of
three OECD countries that were at relatively low levels of development
in the immediate post-war years - Japan, Italy and Austria. Although the
Korean growth performance was respectable before 1964, it was not
particularly outstanding. However from 1964 its growth accelerated and,
especially in the inter-oil-shock period (1973-9), its growth performance
exceeded that of all other countries in our sample, including Taiwan, the
world's best growth performer in the post-war period.
Table 4. 1 Comparative growth performance of Korea, 1950-87 (average

annual growth rates)
1950-64 1964-73 1973-9 1979-1987 1950-87 1964--87
Korea 6.1 9.6 9.0 7.0 7.6 8.5

Taiwan 8.3 11.0 8.4 7.4 8.8 9.1
China 5.2 6.9 5.0 9.3 6.5 7.2
India 4.3 2.7 3.4 4.6 3.8 3.6
Argentina 3.0 4.9 2.3 --0.4 2.6 2.4
Brazil 5.9 8.1 6.5 3.5 6.0 6.1
Chile 4.2 2.8 2.3 1.6 3.0 2.2
Mexico 6.2 6.6 6.1 1.7 5.3 4.7
Austria 5.5 5.1 2.9 1.7 4.2 3.3
Italy 5.7 5.1 2.6 2.2 4.3 3.4
Japan 9.5 8.9 3.6 3.8 7.1 5.7
Source: Calculated from Maddison (1989), tables B-3,B-4.B-5.
Table 4.2 compares the manufacturing growth rates of sorne selected

developing countries. Between 1963 and 1972, the first decade of
Korea's rapid development, Korea's manufacturing sector grew at an
annual rate of 18.3 per cent. This performance exceeded that of ali other
major NICs, such as Singapore (17.0 per cent), Brazil (6.7 per cent),
93 93
Mexico (8.7 per cent), India (4.5 per cent), Greece (9.7 per cent) and
Spain (10.8 per cent). Thanks to the country's ambitious heavy and
chemical industrialisation (HCI) programme launched in 1973 (see 4.4.1
below for details), the growth of the Korean manufacturing sector
accelerated even further between the two oil shocks (1973-8), a period
when most other major NICs experienced a substantial slowdown in
their manufacturing growth. In the ten-year period following the second
oil shock (1979-88), Korea's manufacturing growth slowed down to
11.7 per cent, due to reasons such as the slowdown in world-demand
expansion after the end of the 'golden age', the growing maturity of the
economy (reduced catching-up effect), and the substantial restructuring
of the economy in the ear!y half of the l 980s. However this was still an
Table 4.2 Manufacturing growth in sorne selected developing countries
(average annual growth rates)
1963-72 1973-8 1979-88

Korea 18.3 24.7 11.7
Brazil 6.71 n.a. 1.5
China t---- 9.52 -t 12.63
Chile 4.1 -2.9 2.7
Greece 9.7 4.3 0.4
India 4.5 4.3 8.3
Malaysia n.a. n.a. 7.3
Mexico 8.7 7.4 o.o
Singapore 17.04 7.1 6.8
South Africa 6.8 1.3 1.6
Spain 10.8 3.3 1.5
1 . For 1 963-69
2. For 1 965-80
3. For 1980-87
4. For 1966-72
Sources: UN, The Growtb of World lndustry, 1973; UN, Yearbook of
Industrial Statistics, 1979, 1988; World Bank, World
Development Report, 1988 (for China).
extremely good performance by international standards, and was second

only to that of China ( 12.6 per cent) in our sample during this period.
The rapid expansion of the manufacturing sector transformed the
structure of the Korean economy with a speed that was second to that
of no other country in the world. Table 4.3 compares Korea's perform-
94 94
anee in structural change with that of other developing countries and

sorne poorer European economies (Greece and Spain) between 1965
and 1986. The production structure of Korea, which looked more like
that of low-income countries such as India and Kenya in the 1960s,
began to look more like that of upper-middle-income countries such as
Argentina, Brazil and Spain by the mid- l 980s. The Korean manu-
facturing sector increased its share of GDP by two thirds over a period
of 20 year (from 18 per cent to 30 per cent), and the agricultura!
sector's share of GDP contracted to one third of its size over the same
period (from 38 per cent to 12 per cent). In our sample, no other
country, including Taiwan - the best growth performer in the post-war
world - increased its manufacturing share in GDP by more than one
third during the 20-year period.
Not only was there a structural shift from agriculture towards manu-
facturing, but, more importan ti y, there was a rapid structural change
Table 4.3 Structural change in selected developing countries, 1965-86
Percapita Population Production structure Production structure

GNP (dollars) (millions) (1965) ( 1986)
( 1986) (1986) (as percentage (as percentage
ofGDP) ofGDP)
A l M s A I M s
India 290 781.4 47 22 15 31 32 29 19 39
China 300 1054.0 39 38 30 23 31 46 34 23
Kenya 300 21.1 35 18 11 47 30 20 12 50
Colombia 1230 29.0 30 25 18 46 20 25 18 56
Chile 1320 12.2 9 40 24 52 n.a. n.a. n.a. n.a.
Brazil 1810 138.4 19 33 26 48 11 39 28 50
South Africa 1830 32.3 1042 23 48 6 46 22 49
Mexico 1860 80.2 14 31 21 54 9 39 26 52
Argentina 2350 31.0 1742 33 42 13 44 31 44
Korea 2370 41.5 38 25 18 37 12 42 30 45
Taiwan 3580 19.4 n.a.
n.a. 221 n.a. n.a. n.a. 292 n.a.
Greece 3680 10.0 24 26 16 49 17 29 18 54
Spain 4860 38.7 15 36 n.a. 56 6 37 27 56
Notes: A=Agriculture; Ieíndustry (Mining, Manufacturing, Construction,

Electricity, Water and Gas): M=Manufacturing; S=Services.
l. For 1960
2. For 1985
Sources: World Bank, World Development Report, 1988: Wade (1990), tables
2.2, 2.6; Maddison (1989).
95 95
Table 4.4 Korean manufacturing performance, 1963-88 (annual growth rates

of production)
/SIC lndustry 1963--72 1973-8 1979-88

Light lndustry* 17.9 18.0 7.6
311/2 Food products 14.7 17.0 8.6
313 Beverages 14.6 18.8 6.0
314 Tobacco 12.5 11.0 3.4
321 Textiles 25.3 16.3 7.2
322 Apparel 22.11 26.71 9.0
323 Leather and Ieather products 17.9 60.4 10.7
1 1
324 Footwear 6.6
331 Wood products Furniture 21.0 11.J --0.4
332 and fixtures Paper and 5.8 29.3 13.9
341 paper products Printing 13.0 17.3 11.5
342 and publishing Pottery 14.2 l 1.3 6.2
361 and china -7.4 14.0 8.9
362 Glass and glass products 17.4 12.6 10.5
369 Non-metal products 19.2 13.0 8.5
390 Other industries 9.3 12.8 9.8
Chemical lndustry* 17.6 18.1 8.1

351 Industrial chemical 25.7 21.5 7.9
352 Other chemicals 19.7 24.3 11.8
353 Petroleum refineries 33.12 9.5 4.2
354 Petroleum and coa) products 12.1 10.9 8.1
355 Rubber products 9.2 19.7 11.6
356 Plastic products n.e.c. 50.33 32.3 7.5
Heavy lndustry " 13.9 39.5 17.4
371 lron and steel 16.9 34.9 10.5
372 Non-ferrous metal 6.2 33.8 17.8
381 Metal products 4.8 48.6 10.3
382 Machinery n.e.c. 9.4 25.4 18.6
383 Electrical machinery 22.4 44.7 21.8
384 Transport equipment 20.6 46.5 17.5
385 Professional goods 18.5 37.9 13.4
3 Manufacturing * 17.0 22.l 12.1
* Weighted averages: the weights are, respectively, output at producers'

prices (including indirect taxes and excluding subsidies) for 1967, 1975
and 1983.
1. Major group 324 is included in 322.
2. Annual growth rates for 1964--72.
3. Annual growth rates for 1965-72.
96 96
Source: UN, The Growth of World lndustry, 1969, 1973; UN, Industrial
Statistics Yearbook, 1975, 1978, 1979, 1987, 1988.
97 97
within the manufacturing sector. As we can see in Table 4.4, the heavy
industries - International Standard Industrial Classification (ISIC)
numbers 37 and 38 - which grew much more slowly than the manu-
facturing sector as a whole between 1963 and 1972 (13.9 per cent per
annum as opposed to 17.0 per cent per annum), started to grow rapidly
with the launch of the HCI programme and showed an incredible growth
performance of 39.5 per cent per annum between 1973 and 1978 - as
opposed to around 18 per cent per annum for other industries. Even after
the slow down caused by the advent of the 1979-80 recession (for
details, see Chang, 1987), heavy industry continued to grow at a much
faster pace than the chemical and the light industries (17.4 per cent per
annum as opposed to 8.1 per cent per annum and 7 .6 per cent per annum
respectively between 1979 and 1988). As a result the share of heavy
industry in total manufacturing output (in current prices) rose from 16.4
per cent in 1963 to 42.0 per cent in 1987 (calculated from UN Yearbook
of Industrial Statistics, 1969, and Industrial Statistics Yearbook, 1988).
Since the launch of the HCI programme, chemical industry has also
shown slightly faster growth rates than that of light industry, increasing
its share in total manufacturing output from 17.l per cent to 20.l per
cent during the same period.
Toe magnitude of Korea's intramanufacturing structural change can
be put into perspective by comparing it with that of Taiwan. Despite the
fact that the Korean economy grew more slowly than that of Taiwan
(see Table 4.1), Korea experienced an intramanufacturing structural
change of a greater magnitude than that of Taiwan. Table 4.5 compares
Korea's intramanufacturing structural change with that of Taiwan
between 1965 and 1984. In this period, the share of heavy and chemical
industry in total manufacturing output rose by 48.7 per cent in Korea
(from 38.2 per cent to 56.8 per cent), whereas that in Taiwan rose by
only 17.5 per cent (from 49.8 per cent to 58.5 per cent). Even when
Table 4.5 Intramanufacturing structural change in Korea and Taiwan,
1965-84 (percentages)
lndustry Country 1965 1971 1975 1981 1984

Light Taiwan 51.2 50.7 46.7 43.4 41.5
Korea 61.8 54.7 51.6 47.2 43.2
Heavy and Chemical Taiwan 49.8 49.3 53.3 56.6 58.5
Korea 38.2 45.3 48.4 52.8 56.8
Source: Adapted from Wade (1990), table 2.8.

98 98
considering that Korea was in a 'catching-up' position in relation to

Taiwan -Taiwan's industrial structure was more advanced than that of
Korea in the early J 960s - this is a striking differencc in performance.
From our overview of the Korean experience during the last three
decades we can see that its most remarkable feature has been the
magnitude of structural change. The fact that Korea's structural change
has been much faster than that of the fastest-growing country in the
world, Taiwan, seems to suggest that there has been something more than
just the 'pull' from market-led demands behind the country's remarkable
economic transformation. Especially given that Taiwan is by no means a
free-market economy (see Amsden, 1985; Wade, 1990), this seems to
suggest that the 'push' factor in Korea must have been extremely strong.
We argue that this 'push' was achieved by the strong industrial policy of
the state. Before we discuss the conduct of industrial policy in Korea,
however, Jet us critically examine sorne mainstream arguments that
deliberately attempt to downplay the role of industrial policy, and more
generally of the state, in the Korean developmental experience.
4.2 EXPLAINING THE KOREAN EXPERIENCE
4.2.1 A Free Market?
With the accumulation of studies that revea) the important role of the
state in Korea, the fonnerly common interpretation of the Korean
developmental experience as a free-market and free-trade economy is
rapidly losing its popularity.1 However it is useful to discuss briefly the
argument because the more recent mainstream interpretations examined
below (4.2.2) can be seen as atternpts to rescue the conclusions of the
carly interpretation.
According to the free-market view, the Korean economy stagnated in
the late I 950s after it had depleted the possibility of 'easy irnport-
substitution' in non-durable consumer goods. This inward-looking
strategy was trammelled with inefficiencies due to distortions generated
by excessive state intervention in various markets. Such chaotic import
substitution with multiple exchange rates and across-the-
board protection, largely by the use of discretionary quantitative
restrictions rather than universal tariffs, allowed inefficient firms to
survive and discouraged export activities. This in tum added to the
foreign-exchange shortage and hence to pressure for more import
res trictions.
99 99
The phenomenal growth of the economy, the argument goes, started

with the transition from an inward-looking, or an import-substitution,
industrialisation strategy to an outward-looking, or an export-led
growth, strategy. The tuming point in this transitíon was a series of
policy reforms around 1965, whose most important ingredients
included: (i) the introduction of a unified, realistic exchange rate régime;
(ii) trade liberalisation involving cuts in tariffs and the abolition of most
quantitative restrictions; (iii) a substantial increase in real interest rates.
These policies are regarded as having radically improved the perform-
ance of the economy for the following reasons. Firstly, realistic ex-
change rates, by making export activities as profitable as they 'should
be', allowed Korea to follow its comparative advantage in labour-
intensive industries, and therefore to reap the gains from foreign trade.
Secondly, trade liberalisation improved the efficiency of the economy
by exerting competitive pressures on domestic producers. Finally, the
rise in interest rates enabled the economy to invest more by mobilising
more savings, on the one hand, and to use scarce capital more efficiently
by restoring the relative price of capital near to its 'realistic' level.
In the words of Ranis and Fei (1975), '[s]tabilisation plus dismant-
ling of the various existing direct control measures, on trade, interest
rate and foreign exchange, thereby created a more market oriented
economy most conducive to access for large numbers of domestic
entrepreneurs seeking efficient utilisation of the economy's relatively
abundant resources via embodiment in labour-intensive industrial
exports' (p. 56).2
There is abundant empirical evidence, however, that runs counter to

this explanation. Conceming trade-related reforms, Luedde-Neurath
( 1986) meticulously shows that such reforms were not as thorough as is
usually presented and were implemented rather half-heartedly.3 Firstly,
tariffs were still quite high after 'Iiberalisation', and the bureaucracy
retained the power to impose emergency tariffs (for items with
'excessively' fast import growth) without changing the relevant laws.
In the second place, quantitative restrictions, usually applied under
various special laws and import area diversification regulations, were
pervasive even after liberalisation. As late as 1982, 93 per cent of total
imports (in value terms) were subject to one or more such restrictions
(Luedde-Neurath, 1986, p. 156, table 14.4). Thirdly, prohibitive inland
laxes were often used virtually to ban the importation of luxury
consumer items that were subject only to non-prohibitive tariffs.4
100 100
Fourthly, there were many state supports for import substitution, for
example subsidised credits to import-substitutors and the purchasers of
sorne domestic products (especially machinery), which in effect acted
as import restrictions. Lastly, and most importantly, there was
widespread foreign-exchange rationing, which often meant that
importation of a certain ítem was impossible not because it was illegal
to do so, but because it was impossible to obtain the foreign exchange
to pay for it.
Table 4.6 Real interest rates in Korea, 1960-84 (percentages)
Period Unregulatedfinancial market! Deposits? Export loans
1960-4 31.1 -6.7 n.a.

1965-9 44.4 26.9 n.a.
1970--4 28.2 -0.2 -16.3
1975-9 24.0 --4.5 -12.5
1980-4 19.7 2.4 1.3
J. Nominal interest rates less consumer price inflation.

2. Nominal interest rates less inflation of the GNP deflator.
Source: Calculated from Dornbusch and Park (1987), p. 419, table 14.
The effect of financia) refonn also has to be reinterpreted (for a

detailed discussion, see Harris, 1987, and Dornbusch and Park, 1987).
As can be seen in Table 4.6, the post-reform high interest rates did not
Iast for long, and real interest rates had been negative until the 1980s due
to rather high inflation. Nevertheless, savings as a proportion of GNP
have shown a rising trend, up from less than 1 O per cent in the 1950s and
the early 1960s to more than 30 per cent in the late 1980s ( see Table
4.7). That is, 'Korea's saving responds little to interest rates. Overall, the
Korean experience suggests that there is no need for high positive real
ínterest rates to mobilise saving through the financia! system; as long as
large negative real interest rates are avoided, the real interest rate is
relatively insignificant' (Dornbusch and Park, 1987, pp. 418-19).5
Even if liberalisation in Korea was as comprehensive as the pro-
ponents of the free market insist, it is not necessarily true that a 'more
market oriented economy' is more efficient, as the above quote from
Ranis and Fei implies. As the theory of second best tells us, the removal
of market distortions in sorne, but not ali, markets does not guarantee
that the economy will achieve greater allocative efficiency (Lipsey and
Lancaster, 1956). Moreover, even if it is true that the move to a more
100
Table 4.7 lnvestment and its financíng in Korea, 1953-86 (as percentage of
GNP at current prices)
Gross National Private Government Foreign

Year investment savings savings savings savings
1953 15.4 8.8 11.2 -2.4 6.6

1954 11.9 6.6 9.3 -2.7 5.3
1955 12.3 5.2 7.6 -2.4 7.1
1956 8.9 -1.9 1.0 -2.9 10.9
1957 15.3 5.5 8.6 -3.1 9.8
1958 12.9 4.9 8.0 -3.1 8.0
1959 11.1 4.2 6.9 -2.7 6.9
1960 10.9 0.8 2.9 -2.1 8.6
1961 13.2 2.9 4.7 -1.8 8.6
1962 12.8 3.2 4.8 -1.6 10.7
1963 18.1 8.7 9.1 --0.4 10.4
1964 14.0 8.7 8.3 0.4 6.9
1965 15.0 7.4 5.7 1.7 6.4
1966 21.6 11.8 9.1 2.7 8.5
1967 21.9 11.4 7.3 4.1 8.8
1968 25.9 15.1 9.0 6.1 11.2
1969 28.8 18.8 12.9 5.9 10.6
1970 26.8 17.3 10.8 6.5 9.3
1970 24.6 16.2 11.0 5.2 9.1

1971 25.) 14.5 10.0 4.5 10.5
1972 20.9 15.7 13.8 1.9 5.0
1973 24.7 21.4 18.4 3.0 3.7
1974 31.8 19.3 17.3 2.0 11.9
1975 27.5 16.8 14.5 2.4 10.3
1976 25.7 22.2 17.9 4.3 2.6
1977 27.7 25.4 20.9 4.5 1.2
1978 31.9 27.3 21.9 5.3 4.3
1979 36.0 26.5 20.0 6.5 8.9
1980 32.1 20.8 15.4 5.4 11.5
1981 30.3 20.5 14.9 5.6 9.8
1982 28.6 20.9 14.8 6.1 7.0
1983 29.9 25.3 18.1 7.2 4.7
1984 31.9 27.9 20.9 7.1 4.0
1985 31.1 28.6 21.7 6.9 3.1
1986 30.2 32.8 26.2 6.6 -·2.8
Note: Toe series between 1953 and 1970 and the series between 1970 and
1986 are not fully compatible due to changes in the accounting method.
Source: BOK, National Accounts, 1987.
Jrulustrial Policy in Action - Korea 101
market-oriented economy improved the allocative efficiency of the

economy by moving it closer to its comparative advantage, it does not
explain why Korea grew [aster after the reform in the mid-1960s than
before. There is very little economic theory to support the view that
conforming more closely with comparative advantage (or with any other
form of static efficiency) leads to higher growth, as even the leading
neoclassical trade theorists admit (for example see Krueger, 1980).
4.2.2 Market-Preserving State Intervention?
Evidence shows that state intervention was pervasive in Korea during its
rapid industrialisation period. As Bhagwati (1987) correctly argues, '[t]he
key question then is not whether there is govemmental action in the Far
Eastern economies, but rather how have these successful economies
managed their intervention and strategic decision making in ways that
domínate those of the unsuccessful ones' (p. 285). And, naturally, sorne
neoclassical economists, including Bhagwati hirnself, have put forward
explanations of the Korean experience that try to reconcile the existence
of an interventionist state with the rapid growth of the economy.
One such argument is the theory of virtual free trade régime, which
suggcsts that various measures of statc intervention in Korea cancelled
out each other to produce a neutral incentive structure (Little, 1982;
Lal, 1983; World Bank, 1987). Another is the theory of prescriptive
state intervention, which argues that state intervcntion in Korea does
not hinder growth because it leaves room for prívate initiatives
(Bhagwati, 1985, 1987, 1988). In effect these theories argue that,
whatever state intervention there may have been in Korea, it did not
affect the workings of the market mechanism because it was either self-
cancelling (virtual free trade) or porous (prescriptive state intervention).
Below, we will examine these arguments in turn, and argue that they
are neither theoretically convincing nor empirically correct.
4. 2. 2. 1 Self-cancelling state intervention?

According to the proponents of the theory of virtual free trade régime,
in Korea there existed widespread price distortions due to one set of
state interventions (for example import protection), but these were
cancellcd out by another set of interventions (for example export
subsidies), producing a neutral incentive structure between production
for export and production for the domestic market. They also em-
phasise the fact that Korean exporters had free access to imported
102 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 102
102 102
inputs at world market prices, and consequently bypassed various

import restrictions. The economic crisis in the early 1980s following
the HCI drive during the 1970s is presented as proof that a departure
from incentive neutrality was a disaster (La!, 1983 ). 6
On the theoretical level, it is not clear how meaningful it is to call the

import-substitution-cum-export-incentives régime a virtual 'free trade'
régime because there is no reason why the structure of relative prices
under this trade régime should be the same as the one under genuine
free trade (Yusuf and Peters, 1985, p. 18, n. 49). And if the relative
price structures under the two régimes are different, we cannot say that
the incentive structure under the former is neutral, because what
matters in determining the relative attractiveness between export and
production for the domestic market is the relative price structure, and
not the average incentive (for a more detailed discussion, see Wade,
1990, ch. 5).
Moreover, on the empirical leve!, it is not true that Korean
exporters could freely obtain imported inputs (raw materials and
machinery) at world-market prices. Luedde-Neurath (1986, ch. 4)
shows that only raw materials can be described as relatively, but not
absolutely, freely importable in the Korean trade régime. The
importation of machines was heavily controlled in order to prornote
the domestic machinery industry, which was seen as the vital
ingredient in building a well-integrated economy (see 4.3 below).
Credits were usually refused to those who wanted to irnport
domestically available machines, and instead subsidised credits,
which sometimcs amounted to 90 per cent of the product value, were
provided to the purchasers of domestic machincry (KDB, 1981, pp.
473-4).7
Moreover it is quite incorrect to assert that the promotion of hcavy and

chemical industry in the 1970s through a departure from incentive
neutrality was a failure. As we have already seen in the previous
section, since the launch of the HCI programme the growth perform-
anee of heavy, and to a lesser degree chemical, industry outstripped
that of light industry. By international standards the performances of
Korean heavy and chemical industries are extremely impressive. Table
4.8 compares the performance of Korean industries between 1979 and
1988 with those of sorne countries at comparable levels of development
and, except for Brazil, of a similar size. While almost ali individual
Korean industries performed far better than their counterparts abroad,
103 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 103
103 103
Table 4.8 Korea's comparative industrial performance, 1979-88 (annual

growth rates of production)
South
/SIC lndustry Brazil Chile Greece Korea Mexico Africa Spain
[:: l [ :::i
Light lndustry* 1.5 2.6 0.3 7.8 2.7
31 1/2 Food products 2.2 3.83 0.5 8.6 2.9
313 Beverages 1.8 1.5 4.0 6.0 6.4
314 Tobacco 3.0 -1.1 0.2 3.4 4.2
321 Textiles o.o 2.1 o.o 7.2 --0.16 1.2 --0.3
322 Apparel --0.41 2.1 -2.5 9.0 _6 0.1 -3.61
323 Leather and n.a. -5.4 -3.3 10.7 n.a. ....:1.3 0.4
leather products
324
331 Footwear
Wood products _1 -1.44
-3.7 -4.7 6.6
--0.4 n.a. --0.5
332 Fumiture and
n.a.
n.a. 12.1
-5.5
-8.4 13.9 [1.1 J 0.4
-1.0 n.a.
fixtures 0.3
341 Paper and paper 4.2 2.3 5.6 11.5 2.7 5.1 3.29
r, [�¡]
products
publishing
342 Printing and n.a. 8.0 1.1 6.2 n.a. 2.7 _9
361 Pottery and china [ ] 16.2 -4.0 8.9

362
369
Glass and glass
products
Non-metal
products
0.4
2.2
4.7
--0.4
1.5s
10.5
8.5
t0.6
7.2
1.3
390 Other industries n.a. -4.8 0.6 9.8 1.7 6.9 n.a.
Chemical industry* 2.6 0.6 2.2 7.5 3.41 1.3 0.8
35! Industrial 3.62 1.9 1.4 7.9 3.41 0.8 1.12
chemical
352 Other chemicals _2 1.0 3.9 11.8 _7 2.2 _2
353 Petroleum 1.6 0.3 2.0 4.2 _7 n.a. 0.3
refineries
354 Pctroleum and n.a. n.a. --0.2 8.1 _7 n.a. n.a.
coal products
355
356
Rubber products
Plastic products
2.7
n.a.
1.8
n.a.
1.6
2.3
11.6
7.5
_7
_7
1.6
4.8 [ 1.7 J
n.e.c.
371
372
381
Iron and steel
Non-ferrous me
Metal products
[ l 3.6 --0.4 10.5 0.3 J
104 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 104
104 104
Heavy industry* [ 2.7

l.2J [ 2.1
0.8
tal 0.6
1.4 1.6
2.0 --0.8
--0.6 17.2
17.8 0.4
3.8
2.0 --0.1 10.3 2.6 -1.3 0.9
382 Machinery n.e.c. 0.1 -1. 7 -5.1 18.6 _8 -1.3 2.0
383 Electrical 2.1 3.9 --0.8 21.8 _8 2.8 3.6
machinery
384 Transport -1.5 -5.5 -1.3 17.5 _8 --0.8 2.8
equipment
105 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 105
104
Table 4.8 (cont.)
South
Brazil Chile Greece Korea Mexico A/rica Spain
!SIC Industry
385 Professionai n.a. n.a. --0.8 13.4 _8 5.3 7.4

goods
3 Manufacturing 1.5 2.7 0.4 11.7 2.1 1.6 1.5
1986 per capita 1810 1320 3680 2370 1860 1830 4860
GNP (dollars)
* Weighted averages: the weights are 1980 (except for South Africa, which is
for July 1980-June 1981) output at producers' prices (including indirect
taxes and excluding subsidies), except for Greece, South Africa and Spain,
which are at factor values (excluding indirect taxes and including subsidies).
3. Excluding slaughtering.
4. Excluding sawmills.
5. Magnesite roasting is included in mining.
7. Major groups 352 to 356 are included in 351.
8. Major groups 383 to 385 are included in 381 and 382.
Source: UN, Industrial Statistics Yearbook, 1983, 1984, 1990.
the difference was especially pronounced in heavy industry, where the

Korean performance was truly spectacular.
The advance of the heavy and chemical industries has not been con-
fined to the domestic market. As we see in Table 4.9, which shows the
trade performance of commodities at 1- and 2-digit levels between
1977 and 1987, almost ali the items promoted through HCI - SITC
numbers 5, 67, 68 and 7, which roughly correspond to ISIC numbers
35, 37 and 38 - show (often dramatic) improvements in trade balances
during this period. The export-import growth differentials (which
shows the rate at which the sectoral trade deficit [surplus] decreases
[increases]) of these items are mostly higher than that between overall
exports and imports, although, again, the performance of chemical
products is not particularly outstanding. This suggests that it was these
industries which improved the country's balance of payments during
the 1980s. A detailed cost-benefit analysis of the HCI programme is
beyond the scope of this chapter, but the above evidence suggests that
106 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 106
106 106
the HCI programme, far from being a failure, produced impressive

growth and trade performance, especially in heavy industry.
107 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 107
107 107
Table 4.9 Trade performance by commodity group in Korea, 1977-87
Trade Trade Export lmport Differ-

balance balance growtb growth ential
SJTC Commodity ( 1977) (/987)
o Food and live animals 1.1 0.5 8.2 8.5 -0.3

1 Beverages and tobacco 0.4 n.a. -1.7 n.a. n.a.
2 Crude materials, -7.9 -6.2 4.3 11.7 -7.4
excluding fuel
3 Mineral fuel (e.g., -9.9 -6.0 20.4 10.7 9.7
petroleum, petroleum
products)
4 Animal, vegetable n.a. n.a. n.a. 5.0 n.a.
oíl and fat
5 Chemicals and -3.7 -3.7 19.6 16.6 3.0
related products
51 Organic chemicals -1.9 -2.0 16.0 16.0 o.o
52 Inorganic chemicals n.a. n.a. n.a. 16.3 n.a.
53 Dyes, tanning colour n.a. n.a. n.a. 16.3 n.a.
products
58 Plastic materials -0.8 -0.5 40.2 17.0 23.0
59 Chemical materials n.e.c. n.a. n.a. n.a. 20.4 n.a.
6 Basic manufactures 7.2 4.5 12.9 15.2 -2.3
61 Leather, dressed fur, etc. -0.5 -0.4 23.7 15.4 8.3
62 Rubber manufactures n.a. 0.6 16.3 n.a. n.a.
n.e.c,
63 Wood and cork n.a. n.a. -14.1 n.a. n.a.
manufactures n.e.c.
(e.g., plywood)
64 Paper products 0.2 0.1 16.5 19.8 -3.3
65 Textile yarn, fabric, etc. 3.5 3.0 14.1 15.2 -0.9
66 Nonmetal mineral 1.0 0.2 8.5 22.2 -13.7
manufactures n.e.c.
(e.g., cement, glass)
67 lron and steel -1.1 0.5 19.6 11.7 7.9
68 Non-ferrous metals -0.8 -0.8 27.3 18.4 8.9
69 Metal manufacture 2.1 1.2 10.7 15.5 -4.8
n.e.c. (e.g., tools,
cables, cutlery)
7 Machines and transport -3.6 3.2 24.7 16.9 7.8
equipment
71 Power generating -1.5 -0.7 31.7 12.0 19.7
equipment
72 Machinery for special -2.0 -2.1 22.6 17.0 5.6
industries (e.g.,
textile machinery)
73 Metal working machinery n.a. n.a. n.a. 11.2 n.a.
108 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 108
108 108
Table 4.9 (cont.)
Trade Trade Export lmport Dijfer-

balance balance growtk growtk ential
SITC Commodity ( 1977) (1987)
74 General industrial -2.4 -1.9 35.6 15.5 20.1

machinery n.e.c.
(e.g., furnaces, pumps)
75 Office machinery -O.O 0.8 39.3 28.5 10.8
(e.g., computers)
76 Telecommunication 1.2 4.4 25.4 15.3 10.1
and sound equipment
(e.g., TV, phone)
77 Electric machinery, --0.2 -O.O 23.6 22.6 1.0
n.e.c. (e.g., trans-
formers, microchips)
78 Road vehicles --0.4 3.1 48.2 18.5 29.7
79 Other transport 1.2 0.4 35.3 10.3 25.0
equipment (e.g.,
ships, aircraft)
8 Miscellaneous 14.7 14.9 16.2 19.6 -3.4
manufactured goods
84 Clothing and accessories n.a. n.a. 13.8 n.a. n.a.
87 Precision instruments --0.4 --0.8 18.9 22.6 -3.7
n.e.c. (e.g., measuring
equipments)
88 Photo equipment 0.1 --O.O 15.6 17.6 -2.0
89 Miscellaneous 1.8 2.9 20.5 18.2 2.3
manufactured goods n.e.c.
(toys, arms, wigs)
9 Goods not classified n.a. n.a. n.a. n.a. n.a.
by kind
Total -3.8 7.0 16.8 14.3 2.5
1. Ali figures are percentages.

2. Data for each commodity group appears only if the value in each year is
greater than or equal to 0.3 per cent of the total trade for that year.
3. Trade balance for commodity i is (X; - M¡)ff,(Xj + Mj), where X; (M;) is
exports (imports) of commodity i and X (M) is total exports (imports).
4. Export and import growth rates are average annual growth rates between
1977 and 1987 in current value terms.
5. The differentials are the differential between the export and import growth
rates.
Source: UN, Yearbook of lntemational Trade Statistics, 1980 and
lntemational Trade Statistics Yearbook, 1987.
109 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 109
109 109
4.2.2.2 Porous state intervenuon?

The theory of prescriptive state intervention proposed by Bhagwati
(1985, 1987, 1988) tries to resolve the (self-imposed neoclassical)
dilemma of an interventionist state in a rapidly growing economy by
characterising the Korean state as a 'do' (or a prescriptive) state, in
contrast with a 'don't' (ora proscriptive) state, such as India. According
to this theory, state intervention in Korea does not hinder growth
because it is less stifling. According to Bhagwati, 'although a prescript-
ive government may prescribe as badly as a proscriptive government
proscribes, a proscriptive government will tend to stifle initiative,
whereas a prescriptive govemment will tend to leave open areas (outside
of the prescriptions) where initiative can be exercised' (pp. 98-9). That
is, state intervention works in Korea because it is porous, or allows the
prívate sector to circumvent it.
On a superficial level, it is hardly objectionable that an obstructive state
will not be very helpful for business, and, by implication, economic
growth. However on closer examination we find the theory of porous
state intervention theoretically and empirically unconvincing.
In a world with scarce rcsources (and therefore with opportunity
costs), doing something means not doing something else. In this world,
saying 'do A' is often equivalent to saying 'don't do not A'. And there
can be no presumption that saying 'do A' ( = 'don 't do not A') will
allow more initiatives than saying 'don 't do A' (= 'do not A'). A
prescriptive state can be as stifling as a proscriptive one, since it can
force prívate enterprises to do so many things against their will that
they are left with few resources to do what they want, even if these
activities are not explicitly forbidden. Likewise a proscriptive state may
allow a lot of initiative if it proscribes only a few things. lf we adopt ·
the liberal concept of negative freedom ('freedom from': see .also
Chapter 1.2), we may say that one has less freedom under a prescriptive
state than under a proscriptive one, because prívate enterprises with a
state prescription are coerced to execute the prescription, whereas
prívate enterprises with a state proscription are not coerced into any
particular action and therefore can choose the best option from what-
ever is not forbidden by the state, and thereby exercise initiative.
The Korean state's prescriptions were certainly stifling in many ways.
The Korean state prescription for private finns to invest in heavy and
chemical industries in the 1970s was a proscription against investing in
less risky and often more profitable (partly due to higher protection)
110 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 110
110 110
consumer-goods industries. The best example in this regard is the ship-

building industry, which has, since 1973, grown literally from scratch to
become the world's second largest in less than a decade.8 The Korean
shipbuilding industry was set up in direct response to a personal com-
mand from the then president, Park Chung Hee, against the will of the
Hyundai group, the boldest of Korean business groups, who are ali, in
any case, famous for their boldness (Jones and Sakong, 1980,
pp. 119-20, 357-8). Moreover, if prívate firms could not be made to do
something, the Korean state did not hesitate to set up public enterprises,
making the share of public enterprises in GDP almost egua! to that of
India, whose state is regarded by Bhagwati as the classic example of an
obtrusive proscriptive state (Jones and Mason, 1982, pp. 22-3). In a
country like Korea where private firms depend almost totally on the state-
run banking sector for their investment funds (see 4.5 below), the state's
channelling of money into public enterprises can have a very visible
impact on private initiative.
Moreover, when a knowledgeable magazine observes that 'most
business activities are prohibited unless expressly approved' (FEER, 30
May 1991, p. 54 ), we begin to wonder whether the degree of proscriptive
intervention by the Korean state has been so minimal as to warrant the
description prescriptive without serious qualification. In addition to
restrictions on the prices of foreign exchange and credit (that is exchange
rates and interest rates), there was a legally implemented direct price
control over ali marketed products up to 1973. And even after 1973 the
state reserved the right to impose a price ceiling when necessary. No price
change is conceivable without formal and/or informal state permission,
except in sorne unimportant markets.? In addition to price controls almost
ali important Korean industries have had restrictions on entry, capacity
and technology (see 4.4.1 below). Frequently the Korean state reorganises
industries which it thinks have too many firms, through state-Ied mergers
and market-sharing arrangements (see 4.3 and 4.4.2 below). It is not clear
to us how anything could be more stifling, if so many firms have so little
freedom to decide what and how to produce and at what price.
4.3 MAJOR THEMES OF KOREAN STATE INTERVENTION
If the remarkable growth and structural-change performance in Korea

was not the result of free play of market forces but was orchestrated by
the state as we have suggested, exactly how has state intervention
111 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 111
111 111
worked in Korea? Before we look at details of the policy measures
employed in Korea, we need to look at the general direction in which
the Korean state was driving the economy. In this section we presenta
stylised account of the major themes of state intervention in Korea. The
account draws both from the revealed and, especially, the stated policy
objectives, which are to be found in various policy documents but
which have hardly been investigated by other researchers.l''
The basic theme of state intervention in Korea has been the making of an
'independent economy' (Jarip Gyongjé) (see various Five-Year-Plan
[FYP] documents and EPB, 1982).11 Until recently the balance-of-
payments constraint has been the main concem of Korean policymakers,
and, as Michell (1982) points out, even exports were regarded more as a
means to reduce the unfavourable externa! balance than as the engine of
growth (p. 196). Policymakers have regarded the ultimate solution to the
problem of dependence on foreign savings for financing of investments
(Table 4.7) to be the construction of an economy with the degree of
technological capability that would permit a reasonable living standard
without a chronic balance-of-payments deficit. It was believed that the
cause of the balance-of-payments problem lay in the underdevelopment
of the capital and intermediate goods industries, and therefore that 'a shift
towards heavy and chemical industries is imperative in order to increase
the independence of the Korean economy' (WP, 1970, p. 340) - a
principie known in Korea (and Japan) as 'upgrading' the industrial
structure (also see, Second FYP, pp. 9-10; Third FYP, p. 1).
To Korean policymakers, industrial upgrading required giving priority
to investment, which was essential for growth (WP, 1968, p. 48).
Therefore, macroeconomic policy was geared towards the need to
create an expansionary environment - if necessary through inflationary
measures - which was seen as vital for a sustained high leve] of invest-
ment through its effect on investors' confidence.
Until the late 1980s, of course, there existed a persistent savings gap,
which had to be filled by foreign savings (see Table 4.7). Although the
filling of the savings gap was believed to depend ultimately on the rise in
income leve! (a Keynesian savings assumption), serious attempts were
also made to repress consumption demand through policy measures,
expressed in unashamedly patemalistic terms like 'the need to establish a
sound consumption pattem' [emphasis added] (Fourth FYP, p. 27). The
banks, which are mostly owned by the state, were instructed not to make
consumer loans. The heavy reliance on indirect taxes was also justified -
112 The Political
IndustrialEconomy
Policy inofAction
Industrial Policy
- Korea 112
112 112
against the accusation that they are less equitable than income taxes - in
tenns of their discouraging effect on consumption (Third FYP, p. 16).
Toe control was even stricter when it carne to consumption that involved
foreign-exchange expenditure. For example, foreign holidays were
banned until the late 1980s when the country finally became a net
exporter of capital, and imported luxury goods were either banned or
were subject to high tariffs and inland taxes (see 4.2.1 above). One
outcome of such anti-consumption policies was low passenger-car
ownership, which was discouraged by high taxation and restrictions on
consumer loans until very recently. Despite Korea being a major exporter
of passenger cars, Koreans have owned far fewer passenger cars than
people in other developing countries with a comparable income level, In
1985 there were 73.5 people per passenger car in Korea, whereas the
corresponding figures in 1983 were 27.0 in Taiwan, 21.8 in Chile, 16.3 in
Malaysia and 15.2 in Brazil (NRl, 1988, p. 190, table 9-8). Given such a
clear (stated and revealed) anti-consumption bias, Korean macroecono-
mic policy may be more appropriately understood as 'investment
management' rather than as 'aggregate-demand management'.
Maintaining a high investment leve) through investment management,

however, was seen by Korean policymakers as necessary but not suffi-
cient to upgrade the industrial structure in a short period of time.
Macroeconomic policy measures were seen as ineffective in rapidly
upgrading the industrial structure due to their uncertain impact on spe-
cific sectors, and consequently they took second place to industrial
policy. lt was explicitly stated that 'the market mechanism cannot be
entirely trusted to increase the competitiveness of Korean industries [in
the world market]' (WP, 1984, p. 123), and therefore that sectors with
high-productivity growth potential had to be identified by the state and
designated as 'promising strategic industries' or 'priority sectors', and
given custom-designed financia), technical and administrative supports
(see 4.4 below for details).
Although macroeconomic constraints often set severe limits on the
conduct of industrial policy, industrial policy has been actively used
whenever deemed necessary and practica!. When the aim of macro-
economic stability clashed with the aim of upgrading the industrial
structure, the latter was usually allowed to domínate, as testified to by
the fact that preferential (subsidised) loans directed to the priority
sectors increased faster than general (non-subsidised) loans during
recession periods, when the availability of financing can be a matter of
industrial Policy in Action - Korea 111
111
life and dcath for finns (Ito, 1984; Chang, 1987). That is, even when111 the
incrcase in the overall money supply was contained, the priority sectors
were guarantced financing at the cost of non-priority sectors. The unfair
nature of such policy has been widely criticised inside and outside
Korea, but the dominant attitude in policy-making circles has been that
being unfair in the short run is justified in the long run by the greater
benefits generated by the priority sectors in the fonn of faster growth,
In moving towards high-productivity sectors, the biggest concern for

Korean policy-makers was that these industries are often charactcrised by
large scale economies (for example WP, 1968, p. 174). The strong
emphasis on scale economies in Korean economic policymaking is
exemplified by EPB (1982), which diagnoses the causes oftroubles in the
heavy and chemical industries in the early l 980s - an example that is
often thought to be the classic case of overly ambitious investments by
economists of neoclassical persuasion - as 'the lack of scale economies
due to the participation of too many finns in each industry' (p. 222), that
is, in our terminology, the failure of investment coordination.
The prevalence of scale economies in many priority sectors posed two
challenges to Korean policymakers. One was that individual finns in
these sectors needed to be large in order to obtain the minimum efficient
scale of production.P Finns were often instructed by the state to build
plants of efficient production scale, which, given the small size of the
domestic markets, had the beneficia) side-effect of compelling them to
start exporting as soon as possible in order not to incur losses dueto low
capacity utilisation. And whenever finns were thought to be smaller than
the minimum efficient scale, state-initiated or state-subsidised mergers
were implemented.P The most dramatic example of this was the 1980
reorganisation of six major industries (see 4.4.2 below for details). The
merger of two automobile producers in 1965, of five PVC producers in
1969 (see KDB, 1981) and the mergers within the fertiliser, shipping
and overseas-construction industries in the 1980s (see Leipziger, 1988)
are other examples. The second challenge from the presence of large
scale economies was the high possibility of 'excessive competition', a
term used by Korean (and Japanese) policymakers to describe the well-
known propensity of industries with large sunk costs to engage in price
wars. As a result serious attempts were made to restrict entry and
regulate capacity expansion in such industries.
The apparent antitrust implication of the above policies (rnerger,
entry restriction and so on) has been regarded as secondary, because the
112 Theindustrial Policy in Action
Political Economy - Korea
112
Korean policymakers thought that excessive competition could result
112 in
social waste (for example, WP, 1968, p. 173). In traditional textbook
economics where it is believed that large numbers guarantee com-
petition and small numbers hamper it, the notion that there can be
excessive competition and that it can result in social waste may not be
readily accommodated. However, as we argued earlier, competition is
nota costless process in the presence of specific assets (See Chapter 3).
Korean policy-makers have regarded competition as a means to
achieve efficiency rather than as an end in itself. This view is ex-
emplified by the 6th FYP document, which states that collusive
behaviour should be allowed, and even encouraged, in 'promising
industries' that need to 'increase R&D, improve quality, attain efficient
production scale' and in 'declining industries' that need to 'scale down
their capacities' (p. 79). Likewise, the antitrust law (the Law for the
Regulation of Monopoly and for Fair Trade), which carne into being in
1981, after four abortive attempts at legislation (in 1964, 1966, 1969
and 1971), claimed to be concerned mainly with restricting anti-
competitive behaviour, rather than market concentration itself, although
the growing criticism of the concentration of economic power into the
hands of conglomerates brought about an amendment (in 1986) with
stronger restrictions on cross-investments between members of the
same conglomerates (Paik et al., 1988, pp. 28-9, 40-2).
Another important theme of Korean state intervention is the

policymakers' attitude towards foreign firms. Korean policy-makers
have regarded assimilation of advanced technology by domestic firms
as a vital condition for effective industrial upgrading.I+ To them, this
meant tight state control over foreign direct investment.15 Of course the
persistent savings gap had to be filled, but Korean policy-makers tried
to avoid foreign direct investment and contracted (all state-guaranteed)
foreign loans to do the job, if they could afford it. As a result the share
of foreign direct investment in total foreign capital inflow (except
foreign aid) during 1962-83 was a mere 5 per cent (Amsden, 1989, p.
92, table 5).
Although restrictions on foreign direct investment have been
weakening recently, even the latest version of the Law for Importation
of Foreign Capital (amended in 1988) - which is regarded as a liberal
one by Korean policy-makers - specifies that foreign direct investment
should be restricted in priority industries, infant industries, industries
using large quantities of imported raw material, consumer (especially
113 IndustrialEconomy
The Political Policy inof
Action - Korea
113 113
113
Juxury) goods industries, polluting industries, and agriculture and fishery

- which can mean practically ali industries, if the state so wishes.Is
Even when foreign dircct investment was allowed, foreign majority
ownership was practically banned, with sorne rare exceptions, outside
the free trade zones (FfZs). The fact that only 6 per cent of multi-
nationals in Korea (including the ones in the FrZs) are wholly-owned
subsidiaries, compared with 50 per cent in Mexico and 60 per cent in
Brazil, suggests a substantial degree of state control over foreign direct
investment in relation to ownership (Evans, 1987, p. 208). Even techno-
Jogical licensing, which was preferred to foreign direct investment
whenever feasible, was subject to heavy restrictions. For example, the
Jatest version of the Law for Importation of Foreign Capital clearly
states that technological licensing is banned in industries where local
technological capability is deemed to be promising - which, again, can
effcctively mean any industry.
4.4 THE ECONOMICS OF INDUSTRIAL POLICY IN KOREA
4.4.1 The Evolution of Korean Industrial Policy
Throughout the last three decades the Korean state has chosen severa)
industries at a time as priority sectors and has provided massive support
to them. Most of Korea's major industries have been designated as
priority sectors at sorne stage and were developed through a com-
bination of massive support from and heavy control by the state. The
designated industries had priority in acquiring rationed (and often
subsidised) credits and foreign exchange, state investment funds,
preferential tax treatments (for cxample tax holidays, accelerated de-
preciation allowances) and other supportive measures, including import
protection and entry restrictions.17 In return for this support, they
became subject to state controls on technology (for example production
methods, products), entry, capacity expansion and prices.
The practice of giving priority to certain industries identified as

important is a common practice in industrial-policy states such as
France and Japan. In Korea the practice originated in the very early
years of economic development, with the designation of cement,
fertiliser, and oil refining in the lst FYP (1962---6) as 'basic' indus-
tries.18 In the Second FYP (1967-71), chemicals, steel and machinery
were designated as priority sectors. And during the Third and Fourth
Action - Korea
114 114
114
FYP periods (1972-8 J ), especially through the HCI prograrnme

(announced in 1973), non-ferrous rnetals, shipbuilding and elcctronics
were added to the Second FYP's list of priority sectors. The practice
continued in the Fifth and the Sixth FYP periods (1982-91), during
which rnachinery, electronics, autornobile, chernical, shipbuilding and
various high-tech industries (semiconductor, new rnaterials, biotechnol-
ogy) were designated priority sectors.
Details of such support and control rneasures can be seen in Table
4.1 O, which surnrnarises the rneasures ernployed in sorne selected
prornotional laws, enacted in the late l 960s and early 1970 (except for
textiles) to provide legal backing for support for and controls over
priority sectors. Korean policyrnakers' concern for 'excessive corn-
petition' and the resulting 'social waste' is reflected in the laws in the
form of entry restrictions and regulations on capacity expansion.
Violators of such restrictions could be heavily punished with the
revocation of Iicenses, fines and, in sorne serious cases, prison sen-
tences. Another interesting feature of these laws is the tight
perforrnance-rnonitoring systern. The rnonthly export-performance
rnonitoring by the Korean state is already farnous (see, for exarnple,
Jones and Sakong, 1980, p. 97), but ali finns in promoted industries
were required to report not just on their export performance but also
on their performance in other areas. Failure to report regularly and/or
false reporting could result in the irnposition of fines and prison
sentences. Such a systern provided the Korean state with up-to-date
and detailed infonnation concerning the state of businesses in priority
sectors, sornething which is essential if the asyrnrnetric inforrnation
problern is not going to weakcn the effectiveness of industrial policy.
More recently, various prornotional laws were integrated into the
Industrial Developrnent Law (enacted in 1986). The novelty of the
Industrial Developrnent Law (IDL), cornpared with the prornotional
laws, is its ernphasis on rationalisation prograrnmes with limited, albeit
extendable, Iifetimes (usually 2-3 years). The rationalisation pro-
grarnmes are custorn-designed to the needs of individual industries and
airn to provide ternporary boosts to industries that need import
substitution, capacity upgrading and irnprovernent in international
cornpetitiveness, on the one hand, and temporary protection to declin-
ing industries that need a smooth phasing-out, on the other hand.
Rationalisation programrnes based on IDL rnay be irnplernented upon
application frorn the industry, but they rnay also be irnplemented by the
governrnent without such applicationP
Action - Korea
115 115
115
Table 4.10 Major content of promotional Iaws
Non-
Major Content Mach- Ship- Electr- Petro- lron & ferrous
(year of inery building onics chemical steel metals Textiles
enactment} ( 1967) ( /967) (1969) (1970) (1970) (1971) (1979)
REGULATIONS
Entry Restriction X X X X X X X
Capacity Regulations
Setting up Facility X X
Standard
Capacity Expansion X X X
Approval
Incentives to use X X
Domestically
Produced Facilities
Production
Regulation
Regulation of X X
Material Imports
Production X X X X X
Standard and
its Inspection
Restrictions on X X
Technology lmports
Price Control X X
Reporting and X X X X X X X
Inspection
RATIONALIS-
ATION
Rationalisation X X X X X
Programmes
R&DSUPPORT
Subsidies to R&D X X X X
Joint R&D Projects X
FINANCIAL
SUPPORT
Special Purpose X X X X X X
Fund
Financia) X X X X X X
Assistance
Subsidies
Direct Subsidy X X
Reduced Public X X
Utility Rates
Action - Korea
116 116
116
Table 4. JO Major content of promotional laws (cont.}
Non-
Major Content Machin- Ship- Electro- Petro- /ron & ferrous
(yearof ery building nics chemical Steel metals Textiles
enactment} (1967) (1967) (1969) (1970) (/970) (1971) (1979)
Tax Preferences
Special Deprecia- x X
tion
Tax Reduction/ x X X X X X
Exemption
SPECIAL x X X X
INDUSTRIAL
COMPLEX
ADMINISTRAT-
IVE
ASSISTANCE
Facilitating O X X
verseas Activities
Purchase of
Raw Materials X X
PRODUCERS' x
ASSOCIATION
X X X
Sources: Kim (1989), p. 34, table 3.1.; S. H. Lee et al. (1989), pp. 52-9.
The measures employed by the IDL can be divided into three groups.20
First, there are protective measures to ease the adjustment process, which
inc1ude import restrictions on competing products, rcductions in tariffs
on raw materials, priee controls, and outright subsidies. Secondly, there
are measures related to the attainment of an optimal production scale and
the prevention of excessive competition. These include restrictions on
entry and capacity expansión, state-initiated mergers, coordinated
capacity scrapping and/or exit, rnarket-sharing arrangements (that is
subdividing markets into non-overlapping segments). Thirdly, there are
measures aimed at raising productivity: These inc1ude the provision of
subsidised credits for such activities as capacity upgrading (or capacity
scrapping for declining industries), import substitution of inputs (for
Action - Korea
117 117
117
example machine parts), subsidies for expenditure on R&D and training
programmes, and joint research programmes between the prívate firms
and government-funded research institutes.
Action - Korea
118 118
118
As we can see, except for the introduction of limited lifetimes for

rationalisation programmes, absent in the promotional laws, the major
characteristics of Korean industrial policy changed very Iittle with the
introduction of the IDL. The policy measures used are virtually the
same, and the discretion of the bureaucracy remains as great as before
(because the eligibility criteria are deliberately made vague enough for
any industry to qualify). The custom-designed nature of individual
programmes also remains strong.
Between 1986 and 1989 automobiles, coal-mining, dyeing, ferro-
alloys, fertilisers, heavy construction machinery, heavy electrical equip-
ment, naval diese) engines and textiles underwent rationalisation
programmes (for details, see S. H. Lee et al., 1989, pp. 64-72; Kim,
1989). As mentioned before, the rationalisation programmes were
custom-designed to meet the different needs of different industries. For
industries with a need for technology upgrading and involving large sunk
investments - automobiles, ferro-alloys, heavy construction machinery,
heavy electrical machinery and naval diesel engines - the emphasis was
given to creating more stable environments for major new investments
and R&D activities through measures Iike state-led market-sharing
arrangements, entry restrictions and subsidies on investment and R&D.
For industries with satisfactory technological capabilities but aging
capital stocks - textiles and dyeing - the priority was capacity upgrading,
and therefore subsidies were given to producers for scrapping old
machines and installing new ones. For the (largely state-owned) fertiliser
industry, where the local technological capability was already substanti-
ally developed, the programme aimed to introduce more competition in
the product market by granting sales Iicenses to more distributors and by
reducing tariffs on fertiliser imports. Coal-mining, which was identified
as a declining industry, was under a phasing-out programme, which
involved restrictions on entry and capacity expansion, subsidised
capacity scrapping, price controls and import restrictions.
4.4.2 State-Created Rents and Industrial Development
If the world operates according to the model of perfect competition, the

governments of developing countries should not deliberately seek to
develop new industries because this means, ceteris paribus, a less
efficient use of resources. And this was perhaps what was on their
minds when various international lenders, including the World Bank
and KISA (Korea International Steel Associates) - a consortium
fonned by steelmakers from the USA, West Germany, the UK and ltaly
Action - Korea
119 119
-119 turned down Korea's three applications for loans to build an
integrated steel mili in the l 960s (Amsden, 1989, p. 295; Watanabe,
1987, p. 69) or when the Bank of Japan fiercely opposed MITI's effort
to develop the automobile industry in the l 960s (Magaziner and Hout,
1980, p. 55). However the fact that an industry is unprofitable in a
developing country at present prices does not mean that it should not be
promoted, as the spectacular successes of Korea's (state-owned) steel
mili and the Japanese automobile industry testify.21
In the context of late development, whereby a country develops on
the basis of borrowed technologies, if a technology can be put to work,
the industry will become much more profitable than in the more
developed country where the technology originated. This is partly
because the borrower does not necessarily pay the full cost of
developing the technology but also because it can combine the
technology with cheaper (even when adjusting for its quality) labour.
Of course not ali producers who borrow technology become more
cost-effective than the lenders of the technology. This is so, first,
because technology is not a blueprint and it has a certain element that
cannot be codified, and therefore requires a (possibly infinitely long)
period of learning.22 Secondly, the firms in late-developing countries
do not have access to the social capital (for example skills of the
workforce, infrastructure), to which the firms in the developed
countries have ready access and for which they do not pay the full
costs (Abramovitz, 1986). Due to these reasons firms in late-
developing countries will need additional incentives to adopt new
technologies. And such additional incentives are exactly what the
states in many late-developing countries, from Germany down to
currently developing countries, have tried to provide through tariff
protection and other forms of state-crealed rents (for examplc
subsidised credits).
Indeed, as we have shown above, the industrial policy measures used
by the Korean state are not radically different in k.ind from those often
associated with economic failure in many other developing countries.
This is contrary to the widespread belief that the Korean state's role,
'apart from the promotion of shipbuilding and steel ... has been to
create a modern infrastructure, to provide a stable incentive system, and
to ensure that government bureaucracy will help rather than hinder
exports' (Balassa, 1988, p. S286), while in other dcveloping countries,
for example the Latín American countries, 'there are pervasive controls
of investment, prices, and imports and decisions are generally made on a
case by case basis, thereby creating uncertainty for business decisions'
Action - Korea
120 120
120
(Balassa, 1988, p. S287). That is, industrial policy in Korea has been no
different in kind from that in other countries in that its essence has been
to entice firms into new industries through state-created rents. However,
why does industrial policy work so well in Korea and not in many other
countries? This is because, as we pointed out earlier, there are dangers
associated with industrial promotion through state-created rents, which
Korea was able to avoid. The question then becomes: how was Korea
able to avoid, or at least minimise, these dangers?
One obvious danger of industrial policy is that it may, by opening up

opportunities to acquire wealth through unproductive activities such as
influence-peddling, divert entrepreneurial efforts away from productive
activities. This is the rent-seeking argument. However, if the waste
from rent-seeking activities has been a major obstacle to growth in
Turkey and India (for example Krueger, 1974; Mohammad and
Whalley, 1984), why has this apparently not been the case in Korea?23
As we pointed out earlier, rent-seeking costs are fundamentally
transaction costs expended in the process of seeking rents (which
involve activities like information collection, influence-peddling, and
bargaining), and have to be strictly differentiated from the rent itself,
which is a pure transfer. Therefore the mere existence of state-created
rents - and therefore the opportunity of rent-seeking - does not mean
that resources will actually be spent on rent-seeking. The realised
magnitude of rent-seeking costs in a society will depend on how state-
created rents can be obtained and through which process. For example,
as we have shown in earlier chapters, if firms can acquire rents through
bribery, and if there is little 'second-tier' rent-seeking, the actual
amount of resources spent on rent-seeking may not be large.
In this context it has often been suggested that less resources are spent
on influencing the state in Korea, because there is not much point in
spending resource to influence a 'hard' state (in the Myrdalian sense)
like that of Korea (for example Bardhan, 1984). One difficulty with this
view, as is well known to anyone familiar with Korean politics and
business, is that the country by no means lacks huge corruption scandals
on a regular basis. The head of one of the country's largest companies is
reported to have recently complained that '[t]he govemment has all the
power and you have to purchase approval ... [andas a result] we payas
much in extortion -· legal, semi-legal and illegal extortion - as we do in
legitirnate taxes' (FEER, 30 May 1991, p. 54). This shows that the
Korean state certainly is subject to influence, if Jess than other 'softer'
states. The explanation has to be more sophisticated.
120 TheIndustrial Policy in of
120 120
120 120
A part of the solution to this puzzle líes in the fact that the Korean state
is subject to influence, but mainly to influence from a small, exclusive
group of agents, the chaebols (literally - financia! clans). Although the
practice has produced sorne undesirable distributional consequences,
limiting the opportunity of rent-seeking to the chaebols seems to have
reduced rent-seeking costs in Korea in severa! ways. Firstly, when a
small number of people have exclusive access to rents, rent-seeking
activities will be less frequent and of smaller magnitude because others
may not join in the rent-seeking contests, knowing that they have little
chance of success in influencing the state (this is what Bhagwati, 1988,
calls the 'brother-in-law theorem'). Secondly, since the chaebols as a
group have exclusive access to the rent markets, they need to spend
few resources to acquire information about the nature of the present
opponent, because they are frequently confronted with the same
adversaries in different rent-seeking contests. Thirdly, the fact that the
chaebols are conglomerates, with stakes in multiple markets, also
reduces rent-seeking costs by the 'bundling of issues'. A bargaining
solution can be more easily devised if there are other related bargains
that allow more room for arranging side-payments (Schelling, 1960,
pp. 32-3). For exarnple, in the 1980 industrial reorganisation, Daewoo,
the third largest chaebol, remained in the passenger-car market as one
of the duopolists, but was forced to exit from the diesel-engine industry
and was forced to specialise in a cheaper variety of product in the
electronic-switching-system industry; Hyundai, the second largest
chaebol, remained in the passenger-car industry in return for forced
specialisation in the diese! engine and heavy electrical-machinery
industries (more on this later).
Even when the potential waste of rent-seeking is fully realised (and we

have repeatedly discussed why this may not be), we think that rent-
seeking is not the biggest danger of using state-created rents for
industrial development. As we mentioned earlier (see Chapter 1.4.2),
once a rent is granted it is highly likely that an entry barrier will be set
up around the rent market, which will discourage the potential entrants
from spending resources to dislodge the incumbent. A more serious
danger of state-created rent is that state intervention may protect or
even encourage inefficient producers or production methods, with long-
lasting efficiency consequences. In Korea access to state-created rents
is limited to the chaebols which, as conglomerates, are able to operate
equally well in almost any line of business. Because of this practice it
121 121
121 121
matters Jcss who gets the rent in Korea than in other developing
countries. In other words, state-created rents in Korea may generate a
certain amount of once-and-for-all rent-seeking costs, which are likely
to be small anyway (dueto the reasons given above), but generate few
long-lasting production inefficiencies.
In our opinion the most serious problem with industrial policy is that,
once implemented, state-created rents may be difficult to withdraw
due to political pressure from the recipients of such rents. The
existence of infant industries that refuse to 'grow up' in many
developing countries is a testimony to such a danger (see Bell et al.,
1984). As emphasised by economists such as Marx, Schumpeter and
Richardson, the beneficia! role of rents as a mcans to Jure (positive
rents) and force (negativc rents) firms into more productive activities,
hinges on the fact that no rent accruing to the innovator is permanent.
In a situation where rents are created by the state, these rents may
cease to be transitory if the state is unable to withdraw them when
necessary. 24
Korea is no different from other countries in that industrial policy has

crcated many inefficient firms. However what differentiates Korea from
other countries is that the Korean state has been willing and able to
withdraw support whenever performance (revealed through exporting
and fierce competition in the domestic rnarket) lagged (Khan, 1989;
Amsden, 1989). Such state discipline, when combined with industrial
upgrading (which involves creation of new and often bigger rents in
more productive industries), has acted as a powerful incentive for firrns
to enhance their technological capabilities. The imposition of such
discipline, of course, has not been a purely technocratic procedure
whereby impartía! bureaucrats teach non-performers a Jesson, nor has it
becn a smooth and concensual procedure, as things are usually assumed
to be in East Asia. It has, rather, been a process of continuous
bargaining and conflict between the state and the prívate sector, which,
as we shall see below, sometimes had to be solved by forceful measures
that are difficult to imagine in other countries.25
In 1969 the proliferation of inefficient firms after a massive investrnent

boom in the late 1960s (see Table 4.7) prompted the Korean state to set
up a task force, accountable only to the Blue House (the presidential
office), to deal with the problem. Between 1969 and 1972, the task force
122 122
122 122
forced dozens of inefficient finns (the exact number was never made
available to the public) into mergers, sales and liquidation - sometimes
sweetened by debt roll-overs by the Korea Development Bank. The
programme eventually ended with the notorious 8-3 Decree (named after
the date of its announcement, 3 August 1972), involving a total freeze on
ali curb market loans which were eating into the profits of many finns
suffering financia) distress, with a subsequent reduction in their interest
rates and/or in a debt-equity swap (for details, see S. H. Lee, 1985).
After the investment boom of the late 1970s (see Table 4.7), which
led to temporary excess capacity in sorne major industries, the Korean
state stepped in again with the Reorganisation of Heavy and Chemical
Industries Programme in 1980. Four existing companies in the power-
generating-equipment industry were merged into Korea Heavy Industries
and Construction Co. (KHIC), which was subsequently nationalised on
the ground that the state support needed to make KHIC profitable was too
great to be given to a single private finn.26 In the passenger-car industry,
one of the three existing producers (Kia) was forced to exit and specialise
in trucks and buses with a promise that it would be allowed back in again
when the demand condition improved - this actually occurred (a
conditional entry; see Chapter 3.2.2).27 One of the three companies in
the naval diese) engine industry (Daewoo) was forced to exit, and the
other two were forced to split the market into two segments and to
specialise (Hyundai in over-6000 hp and Ssangyong in under-6000 hp
engines). In the heavy-electrical-machinery industry, where there were
eight companies, three (Hyosung, Ssangyong and Kolon) were merged
into one (Hyosung) and allowed to produce only highly specialised and
expensive products. A subsidiary of Hyundai was asked to produce only
for its sister companies. Four other minor companies were forced to
produce less sophisticated and cheaper products. Each of the four
companies in the electronic-switching-system industry (Samsung, Gold
Star, OPC and Daewoo) was forced to specialise in a different product.
The two companies in the copper-smelting industry were merged by
forcing one to buy the other's equity, which was supported by equity
participation of KDB anda moratorium on bank-loan repayment.
Another round of state-led mergers and liquidations of inefficient
finns occurred between 1984 and 1988 (for details, see S. H. Lee et al.,
1989, pp. 60-2; Leipziger, 1988). This time, the shipping, overseas
construction and fertiliser industries, which were identified as declining
industries, fonned the focus of the programme. In 1984 three fertiliser
producers were liquidated and 63 shipping companies were merged into
123 123
123 123
17. In 1986 a major reorganisation of the overseas construction industry
was implemented, again involving mergers and liquidations. And
between 1986 and 1988, 82 inefficient firms (23 of them in shipping and
overseas construction) were forced int.o liquidation or merger.
What is notable in the conducting of industrial reorganisation pro-

grammes in Korea is that even the economically and politically
powerful conglomerates, the chaebols, were not immune to state
discipline as individual agents, although as a group they were certainly
privileged in their access to various state-created rents. To Korean
policymakers it matters little who runs a business so long as it is run
efficiently. If a particular chaebol runs a plant well, fine; otherwise the
ownership has to be transferred to another chaebol or even to the public
sector (for example the nationalisation of KHIC). The fact that the
chaebols as conglomerates are potentially able to move into any line of
business (on the basis of their activities in related lines) rnakes it
difficult for a chaebol to keep a particular industry as its fiefdom. Unless
it remains reasonably efficient, other chaebols can easily persuade the
state that they can do a better job and get the state support in the next
round of capacity expansion in that industry.
Therefore the chaebols have had a powerful incentive to remain
efficient, especially when the loss of state support can mean a sharp
downturn in business in a few years' time, given the state control of
credit and the high leverage of Korean firms. Many of the chaebols that
lost state favour (for political and/or efficiency reasons) have either gone
into oblivion or been disbanded and their remnants distributed to other
chaebols, as exemplified by the fact that only two of the 10 biggest
chaebols in 1966 were among the top 10 in 1974; only five of the 1974
top 10 were in the 1980 top 10; and only six of the 1980 top 10 were in
the 1985 top 10 (Paik et al., 1988, p. 352, table 35).
4.5 THE POLITICS OF INDUSTRIAL POLICY IN KOREA
The Korean state has played a central role in the country's economic
development through its cunning use of state-created rents as an
instrument for industrial development. Of course such a result has been
possible only because the Korean state has been a strong state, able to
discipline firms whenever necessary. The power of the Korean state has
frequently been underrated, especially by sorne neoclassical eco-
124 124
124 124
nomists, on the ground that the 'size' of the Korean state (defined in
tenns of public-sector expenditure) is relatively small (for example
Balassa, 1988, but see Sachs, 1987, for sorne evidence to the contrary).
However, what matters for the effectiveness or ineffectiveness of state
intervention is not where the boundary of thc state as a legal entity Jies,
but how far it can exercise its influence. Public-sector expenditure as it
is usually defined is a very poor measure of this. What, then, was the
basis of the power of the Korean state?
It has often been suggested that the Korean state is strong because the
country's historical development left a social structure with no power-
ful social classes to contest state power (Hamilton, 1983; Lim, 1985;
Cumings, 1987; Evans, 1987; Amsden, 1989, ch. 2). The landed class
was eliminated through land refonns around about the Korean War, and
the incipient (largely socialist) political organisations of the working
class and fanners were also crushed during the war and the subsequent
era of Cold War politics. Moreover it is argued that the country's
history of Confucian tradition produced a society where the state
commands the moral high ground and draws in the best talent (for
example Luedde-Neurath, 1988).28 Toe long tradition of centralisation
in Korean history seems to have been another factor serving to
Jegitimatise the power of the central bureaucracy.29
We think these historical factors are extremely important, and perhaps
are what differentiates the Korean developmental experience most from
that of, say, India or Latin America.P The weakness of the social
classes was certainly important in deciding the balance of power
between the state and socicty in Korea. The Confucian belief in the
state as a Jegitimate social institution (if not necessarily in particular
governments and individual political leaders), often lacking in other
developing countries, also seems to have been an important factor in
rnaking state intervention effective in Korea. Moreover there are
reasons to believe that the relatively high cultural and ideological
homogeneity of Korean society - although it is not as absolute as is
often argued (there have been rather intense regional conflicts) - has
also helped in the effective implementation of government policies, as
we have discussed earlier.
However we think that such historical factors are, in themselves, not
enough to bring about a strong state and effective state intervention. If
such conditions have been present since the end of the Korean War,
why was the Korean state so weak and incompetent in the l 950s? And,
125 125
125 125
if Confucianism is conducive to a strong and competent state, why was
the Kuomintang government before their defeat to the Chinese
Communist Party so weak and incompetent? Although a full discussion
of the subject is beyond the scope of this chapter, we suggest that the
emergence of a strong state in Korea should be understood in relation
to the political agenda of the military régime of General Park Chung
Hee, which fundamentally shaped the political economy of the country
for decades to come. The strong state was as rnuch, if not more, an
outcome of calculated political moves and institutional innovations as
that of historical conditions and culture.
The political ideas of the top political decision-makers of General

Park's military régime were fundamentally shaped under the shadow of
the Japanese variety of corporatism.J' In tenns of their economics, the
early top Korean política) decision-makers were no fans of the free
rnarket, although they had to pay constant lip service to free-enterprise
economics, given the critica) importance of US support for the política)
survival of the régime.32 In addition, whatever little economic know-
ledge the early Korean economic bureaucrats - who were mainly
lawyers by training - had was not of neoclassical economics but of the
economic theories of Friedrich List, Joseph Schumpeter and, especially,
Karl Marx, which dominated Japanese academia and policy-making
circles in the first half of the twentieth century (see Morris-Suzuki,
1989). The major themes of Korean economic policy-rnaking, for
example, the concern with social waste from excessive competition (a
distinctly Marxian notion), the emphasis on scale econornies, the
obsession with capital accumulation and the desire to develop heavy
and chemical industries, make more sense when we understand the
intellectual background of the Korean economic bureaucracy. Given
such a background, it is more than natural that the political-economic
agenda of the Park régime was summarised in the First FYP document
as 'guided capitalism' (Gyodo Jabon-Juii, where the state plays a
guardian role.
As soon as it carne to power the Park régime moved swiftly to prepare

sorne institutional grounds for its political-economic agenda. One of
the first moves was to nationalise ali the banks and thereby gain control
over the financia) flows in the economy. Subsequently new state-owned
banks (for example the Korean Exchange Bank, the Bank for Medium
and Small Firms, the Ex-Im Bank) were set up overa period of time,
126 126
126 126
resulting in full state control over investment loans.T' At the same time,
the Park régime imprisoned many prominent businessmen on the
charge of having accumulated wealth through illicit means (for
example by using political connections), but they were later released in
retum for their promise to 'serve the nation through enterprise', which
basically meant building new plants in state-designated industries (on
the so-called 'Illicit Wealth Accumulation' episode; see Jones and
Sakong, 1980, pp. 69-70, 281-2). With these two major political
blows, the business community suddenly became, morally, a criminal
on parole and subject to 'serving the nation through enterprise', and,
economically, a paper tiger with little power to rnake investment
decisions - the ultimate capitalist prerogative.
Another important institutional innovation made by the Park
régime was the centralisation of economic policy-making power in
the hands of a super-ministry, the Economic Planning Board (EPB),
headed by the deputy prime minister (see Whang, 1991, pp. 86-7).
The integration of both planning and budgeting authorities within the
EPB eliminated the conflict of interests between the planning and
industrial ministries (which are usually more interested in long-term
investments) and the finance ministry . (which is usually more
interested in short-term stability), if at the cost of concentrating
power within the govcrnment (which may be objected to 011 other
grounds). Elimination of such conflict made the implementation of
industrial policy in Korea more effective than in other industrial
policy-states such as Japan and France, where such conflict has been a
problem.J"
Even the much-vaunted cultural and ideological homogeneity of
Korean society was not purely historical bounty that the nation
accidentally stumbled on. The Park régime mobilised the nation with
the ideology of 'Renaissance of the Nation' through the building of
Jarip Gyongjé (independent economy). 35 Workers were described by
the state-controlled media and state-issued school textbooks as 'indus-
trial soldiers' fighting a patriotic war against poverty (although the
labour movement was brutally suppressed) and businessmen were
given medals for achieving export targets as if they were generals who
had won major battles. Farmers were mobilised into semi-compulsory
(unpaid) labour for rural infrastructural development a la Mao Tse
Tung, through the Sémaul (new village) movement (Michell, 1982,
pp. 205-8). Although not ali such ideological mobilisations were
successful (for example, the Sémaul movement was much resented) and
127 127
127 127
sorne rightly criticised them as 'militarising' the society (for example
Halliday, 1980), it is undeniable that they were important in promoting
ideological homogeneity.
The Korean state has continued to dominate economic and moral issues
throughout the country's developmental period. State control over
credit, which has been the most effective means of controlling private
firms, given their high leverage, has continued throughout.36 Although
sorne of the state-owned banks (the so-called 'city banks') were partially
privatised in 1982, the independence of these banks is almost nil
because of their lending to high-borrowing finns and their consequent
dependence on the central bank, which is under the full control of the
state. 37 In fact, following their privatisation, the share of policy loans -
loans with subsidised interest rates and/or priorities in credit rationing -
actually increased, from 56.0 pcr cent (1962-81) to 67.6 per cent
(1982-5), making it very difficult to argue that state control over the
banking sector has loosened (also see Cole and Park, 1983, p. 173). In
addition to their freedom to make loan decisions, the banks' freedom to
set interest rates has also been severely limited. Despite the legal
deregulation of interest rates on loans and long-term deposits in
December 1988, it was reported in 1991 that '[i]nterest rates are still
strictly controlled by guidance from the Bank of Korea and the Ministry
ofFinance, despite the legal deregulation' (FEER, 30 May, 1991, p. 52).
In addition to its control over bank loans, the Korean state has
maintained tight foreign-exchange controls. The buying and selling of
foreign exchange has been tightly regulated, and up until a few years
ago it was illegal (subject to prison sentence) to possess foreign
cxchange except for state-approved business purposes. The state's
control over foreign loans and foreign direct investrnent has been near-
absolute. Although foreign borrowing and, to a lesser degree, foreign
dircct investment have not been discouraged, the state has had the final
say in deciding whether a certain loan or foreign direct investrnent is
permitted, and on what terms.
Although far less important than its control over financia] resources,
the state's control over material resources through public enterprises
should not be ignored. The Korean state has owned various strategic
industries, including oil, coa) (partly), gas, fertilíser, steel and electri-
city. The fact that such crucial intermediate inputs as oil, coal, gas,
electricity, steel and fertiliser are supplied by public enterprises is
anothcr important factor contributing to the power of the state in Korea.
128 128
128 128
Of course the regular threat by the Korean state that it is going to use its
power to discipline those firms which do not comply with its policies
does not always materialise, partly because large business firms have a
strong influence on policy formation and implementation. Nevertheless
the threat is not an idle one and it is often realised, as exemplified by
the freezing of bank credit (on 8 May 1991) to 14 subsidiaries of eight
conglomerates that had not complied with state pressure to sell non-
business land (EIU, 1991, p. 20). A still more dramatic example of how
far the Korean state can go, if it chooses, was the Kukje-group incident
in 1985, when the state deliberately bankrupted the inefficiently-run
Kukje group, the then seventh largest conglomerate in the country, by
ordering its major lending bank not to honour its cheques, although it is
believed that the decision to Jet Kukje go under was in part motivated
by its lukewarm attitude to meeting the ruling party's financia)
demands (for details, see FEER, 21 April 1988, pp. 58-60).
CONCLUSION
The Korean experience shows what an intelligent industrial policy can

achieve. To its credit, the Korean state has recognised the limited ability
of the market to coordinate interdependent decisions. As expressed in
the policymakers' continued concern for excessive competition and the
resulting social waste, the possibility of coordination failure and its costs
have been taken most seriously when conducting industrial policy in
Korea. Again to its credit, the Korean state has not been bound by its
own free-enterprise rhetoric and it has been more than willing to
intervene in the market to prevent and remedy coordination failures.
That the Korean state has put a restraint on the operations of the
market does not mean that it did not believe in the power of the market.
If anything, it has taken the market more seriously than anyone else
has done as, for example, can be seen in its ali-out attempt to help local
firms to break into the world market. What differentiates the Korean
state's attitude toward the market from that ofmany other states is that
it has taken a dynamic view of the market whereby problems of tech-
nical change and learning, rather than allocative efficiency, have taken
centre stage. lt has fully recognised the inadequacy of the market
mechanism to bring about non-marginal change in a late-developing
context and it has deliberately created rents in areas where, given the
129 129
129 129
present state of local technological capability, the prospect of techno-
logical development was the greatest.
130 130
130 130
Of course such intervention by the Korean state would not have been
as successful as it has been were it not for the historical, political and
institutional conditions we have examined - for example the cultural
and ideological homogeneity of society, the Confucian tradition, an
élite bureaucracy and the state's control over the financia) flows in the
economy. However, as we have demonstrated, these conditions are not
entirely God-sent gifts which other countries cannot aspire to possess.
They have been brought to their present forms through an intense
process of political struggle, ideological campaigning, and institutional
innovation. Only when we recognise this fact will we be able to avoid
both the ahistoricism of sorne orthodox economists who recommend a
simplified and biased version of the Korean model as a panacea and
the scepticism of sorne heterodox economists who believe that other
countries can never learn from such a 'special' case as Korea.
Conclusion: The Market, the
State and Institutions
What is the best way to coordinate the activities of independent, but

interdependent, agents with divergent interests and dispersed know-
ledge? This has been one of the major concems of economists since
the birth of the discipline. Sorne early economists, notably Adam Smith
and Karl Marx, explored this question in relation to two problems. One
was the problem of coordination within economic enterprises - that is,
the study of the conscious management of the relationships between
individuals cooperating within the same enterprise. The other was the
problem of coordination between such enterprises - this was seen as the
study of spontaneous order of the market mechanism that regulates the
behaviour of and the relationships between enterprises without the need
for conscious design.
With the rise of modern economics, economists became almost
exclusively concerned with the second type of coordination problem -
spontaneous regulation through the market - and the first type of co-
ordination problem - conscious coordination within economic
organisations - was almost entirely dropped from the research agenda
of economics (Pagano, 1985). And the inquiry into the functioning of
the market continued to be made in a manner which largely ignored the
social nature of the problem. By implicitly assuming a complete and
costlessly enforced set of property and other rights underlying the
market, the study of the struggle between social groups over the
definition of the rights system was ignored (North, 1990b). Further-
more, by supposing the hypothetical system of instantaneous and
costless recontracting administered by the Walrasian auctioneer, even
the coordination problem within the given system of rights was
assumed to be solved at no real cost (Richardson, 1971).
Recently, however, economists have become more critical of the
conventional approach. As we mentioned earlier, the learning literature
and the labour-process literature are returning to the question of
coordination and the division of labour within the enterprise. New
institutional economics looks at not only market coordination but also
131
132 Conclusion
132 132
132 132
non-market coordination within and between enterprises, and also at
the deterrninants of the scope of individual enterprises. The govern-
ment-failure literature discusses the problems of the centralised forrn of
coordination between enterprises through state intervention. What is
becoming increasingly clear from these writings is that effective
coordination both within economic units and between such units can
only be achieved at a cost, and that different forrns of coordination have
different costs under different conditions.
By introducing the costs of achieving coordination, or transaction
costs in more popular terms, into the picture, we effectively break away
from the neoclassical tradition in which perfect coordination is
achieved without any cost, either by the Walrasian auctioneer (as in the
general-equilibrium version) or by the benevolent, omniscient and
omnipotent state (as in the welfare-economics version). In this new
frarnework, the market is but one of many possible economic institu-
tions, or coordination mechanisms, each of which incurs certain trans-
action costs of different types and magnitudes in resol ving the
coordination problem (Pagano, 1985, ch. 8). By adopting a theoretical
framework that recognises that all institutions are costly to run, we
implicitly reject the common view which assigns theoretical primacy to
the market and views other institutions as merely surrogates, which are
simply 'rneans of achieving the benefits of collective action in situ-
ations in which the price system fails' (Arrow, 1974, p. 33).
By regarding the market as just one of many alternative economic
institutions, we also come to realise that the efficient operation of the
market depends on many more institutional arrangements than those
which are usually recognised in the conventional literature.1 To begin
with, we need at least a mínimum degree of morality to have any
exchange which is intertemporal, as otherwise the costs of policing and
contract-enforcement will be prohibitive (McPherson, 1984). Also,
existing exchange relationships are definable only in relation to the
existing system of property and other rights (for example social-security
entitlements, working conditions, rights for clean air), and the en-
forcement of such rights necessarily requires certain institutions - the
police, the courts, and the prison system. And in this process the state
not only plays the role of the ultimate guarantor of property rights and
other legal forms but also extensively intervenes in order to establish
the market itself asan institution (Polanyi, 1957; Coase, 1988; see also
Chapter 1.2.3). Thus seen, the seemingly 'instítuüon-free' market
mechanism is sustainable only as a part of the intricate fabric of various
133 Conclusion
133 133
133 133
institutions. As Simon (1983) succinctly puts it, '[m]arket structures are

no substitute for the whole web of social interactions' (p. 78).2
Introducing transaction costs - or the costs of coordination - and

thereby emphasising that ali economic institutions are costly to run
allowed us to develop a theory of state intervention which overcomes
the limitations of sorne dominant theories of state intervention. By
recognising that state intervention carries certain costs, we were able to
avoid the naive view of state intervention as the ultimate solution to
the coordination problem that is held by welfare economists and the
proponents of central planning. At the same time, by recognising that
allocating resources through the market is also costly, we were able to
reject the biased view of the government-failure school that the costs of
state intervention almost invariably exceed the costs of market
transactions.
Our theory of state intervention also suggests that there are many
possible types of state intervention other than those implied in the
conventional analyses. In addition to the manipulation of the price
signals in the manner prescribed by welfare economics, the state can
take measures that are intended to reduce the costs of coordination, or
transaction costs, in the private sector. For this purpose, it may reduce
uncertainty in the economy through aggregate demand management,
institute and administer a well-defined property rights system, super-
sede certain private transactions, change the institutions and values of
society, or e ven simply draw attention to a focal point for coordination.
One interesting point that emerges from our discussion of the di verse
forms of state intervention is that state intervention need not be expens-
ive, as is implied by the conventional discussion which equates a state
that has a Iarger expenditure (as a proportion of nationa1 income) with a
more interventionist state. For example, it is often argued that the
Japanese state is not very interventionist on the ground that it has a Jow
ratio of government expenditure to GDP (for examp]e, World Bank,
1991, p. 40, box 2.2). However, as the discussions in Chapters 2 and 3
of this book show, the Japanese state has actively intervened in
industrial matters. This apparent puzzle that a 'smal1' state can a]so be
very interventionist is resolved when we realise that sorne types of state
intervention (for example providing the focal point) can improve
greatly the efficiency of the economy at little cost because they have
only very modest requirements for information processing and bargain-
ing. The point is that fostering credib]e commitments to the coordinated
134 Conclusion
134 134
134 134
outcome among the parties involved need not involve financia)

transfers, although they may help the process.
Combining our theory of state intervention with sorne recent develop-
ments in the literature on industrial organisation, technical change and
evolutionary economics, we set out to provide an analytical discussion
of industrial policy. In our discussion of industrial policy we argued,
first of ali, that in many modern industries where interdependence and
asset specificity are important, coordination failures are real possibilities
and can be costly, especially when combined with unforeseen demand
shocks. Our argument was that the state can coordinate individual
decisions ex ante to prevent coordination failures in such contexts - an
argument analogous to the role of the firm as a device of coordination
discussed by Coase (1937) and Williamson (1975). Toen we discussed
how, given the limited codifiability of human (especially technological)
knowledge, the state can intervene in industries in such a way that
would encourage experimentation and learning, especially in industries
at the earlier phase of the product cycle. Together with coordination of
investments in complimentary projects, such intervention would serve to
amplify the dynamism of the capitalist economy rather than diminish it,
as it is often argued - an argument similar to the Keynesian argument
about the socialisation of risk but with a new twist.
Obviously industrial policy is not a panacea. As many sceptics point
out, there are sorne serious difficulties involved in its effective opera-
tion. Informational constraints mean that industrial policy may not be
effectively designed and implemented, resulting in inefficiencies. lt
may also create perverse incentives for rent-seeking or turn itself into a
vehicle for the advancement of the interests of redistributive coalitions
or the preservation of conservative interests. Many also rightly point
out that industrial policy poses sorne serious political questions regard-
ing legitimacy and democratic control. We also recognised that the lack
of certain supporting institutions may make industrial policy inapplic-
able to certain countries which lack such institutions. While recognis-
ing all these problerns, we pointed out that sorne of these arguments are
based on biased assumptions, and we suggested sorne ways to alleviate
these problems.
The conducting of industrial policy in Korea provided sorne good

examples that clearly brought out the theoretical points made in the
earlier chapters. On the one hand the Korean state has used various
means of ex ante coordination (for ex.ample entry restrictions) to
135 Conclusion
135 135
135 135
minimise the static inefficiency from possible coordination failures of
the market. On the other hand it avoided the dynamic inefficiency that
may rcsult from the use of those means intended to reduce static
inefficiency of the market - the possibility that productivity growth
may slow down because there is not enough competitive pressure - by
imposing harsh discipline on the finns through its control over the
financia) flows in the economy. Thus seen, Korea provides a good
example of how active intervention by a strong developmental state in
industrial development can, under certain conditions, minimise static
incfficiency while maximising the dynamic efficiency of the economy.
The conclusion that emerges from our discussíon is that neither the
market, nor the state, nor any other economic institution is perfect as a
coordination mechanism. Each institution has its costs and benefits, and
is therefore bctter than others under certain conditions and worse under
other conditions. This means that different countries facing different
conditions can, and should, have different mixes of the rnarket, the state
and other institutions. And in fact, as we observed in the Introduction,
even economies that are usually Jumped together as 'capitalist' or
'market' economies have been based on substantially different institu-
tional mixes. In other words, there are many feasible ways of managing
a capitalist economy.
This point, which may seem obvious to the rcader by now, however
does not seem to be widely recognised by policy-makers in many refonn-
ing cconomies of the ex-socialist world and the developing world. Toe
reform packages applicd in these countries are modeled on Anglo-Saxon
capitalism, and they are often a crude caricature of it al that. The possibil-
ity that this may not be the only, or the most efficient, way of managing a
capitalist economy seems to be largely ignored. This is a disturbing state
of affairs, especially when within the Anglo-Saxon economies them-
selves there is growing disillusionment with their own models of econo-
mic management. Now their policy-makers are thinking seriously about
policy changes which recognise the positive role that the state and other
non-market institutions can play in the economy, as exemplified in the
renewed interest in industrial policy in these countries.
Of course this does not mean that the refonning countries would
necessarily succeed if they were to follow sorne other model, say that of
Scandinavia or East Asia. lf there are many different ways of combining
different economic institutions, this means that each country has to decide
on the exact mix between the market, the state and other institutions on
136 Conclusion
136 136
136 136
the basis of a careful consideration of its economic, political and social
conditions, rather than blindly follow one model or another. Needless to say,
finding an appropriate mix of the market, the state and other non- market
institutions is not easy. Such a task, especially in a world of ever- changing
technology, is formidable, as the sceptics point out. However, as we have
repeatedly pointed out with reference to countries such Japan, Korea and
France, it can in principie be met through a process of institutional
leaming and innovation.
Notes and References
Introduction
l. For more detailed discussions on the influence of Keynesian ideas on

postwar economic policy-rnaking, see essays in Hall (ed.), 1989.
2. See King, 1987, and Gamble, 1988, for analyses of new-right ideas.
3. For an evaluation of new-right economic policies in the UK, see essays
in Green (ed.), 1989.
4. Goldthorpe (ed.), 1984, Schott, 1984, Bruno and Sachs, 1985 and
Katzenstein, 1985, are earlier works in the literature. For more recent
theoretical developments, see Rowthom and Glyn, 1990; Rowthom,
1990; Pekkarinen et al. (eds), 1992. For left-wing critiques of the neo-
corporatist experience, see Panitch, 1981; Pontusson, 1987. Also see
Maier, 1987, for a discussion of the historical origins of corporatist ideas
in Western Europe.
5. LO (1963) provides theoretical backings for such policies from the trade
unions' point of view.
6. See Jacquemin (ed.), 1984; Johnson (ed.), 1984; Norton, 1986;
Thompson (ed.), 1989.
7. Deyo (ed.), 1987, provides an overview of the political economy of
Japan and the East Asían NICs. White (ed.), 1988, has useful essays on
the role of the state in the East Asian NICs. Magaziner and Hout, 1980,
Johnson, 1982; Dore, 1986; and Okimoto, 1989, provide detailed
accounts of the role of the state in postwar Japan. Amsden, 1985, and
Wade, 1990, do the same for Taiwan. For Korea, see Jones and Sakong,
1980; Amsden, 1989; and Chapter 4 of this book.
8. Coase, 1937, Coase, 1960, Richardson, 1960, and Simon, 1983, are
seminal works in the tradition. Williamson, 1975, has established itself
as a modern classic. For more recent theoretical developments, see
Pagano, 1985; Williamson, 1985; Matthews, 1986; Langlois (ed.), 1986;
Putterman (ed.), 1986; Aoki et al. (eds), 1990.
9. Schumpeter, 1987, is a classic work in this tradition. For sorne important
contemporary works, see Rosenberg, 1976; Nelson and Winter, 1982;
Fransman, 1986; Dosi et al. (eds), 1988; Johnson and Lundvall, 1989.
10. See Mueller, 1979; Schott, 1984; and Cullis and Jones, 1987, for a
literature survey. Also see Chapters 1 and 2 of this book.
1 Theories of State Intervention
l. Baumol, 1965, is a classic exposition of the literature. For textbook

expositions, see Musgrave and Musgrave, 1984, and Stiglitz, 1988. For
more recent theoretical developments, see Schotter, 1985; Cullis and
Jones, 1987.
137
138 Notes and
References to
to pp. 8-9
pp. 9-17 138
138 138
138
2. It has become customary to define a public good, following Samuelson

(1954), as a good with 'jointness in supply' or 'non-rivalness in
consumption' (that is, one person's consumption not reducing anothers'
consumption), Non-excludability is rcgarded as an additional, but not
necessary, feature of a public good (see for example Oakland, 1987, pp.
485-6; Laffont, 1988, pp. 33-5), or at best as one of the two (together
with non-rivalness) defining characteristics (see for example Stiglitz,
1988, pp. 119-22). However, along with Olson (1965), we think that
non-cxcludability should be the primary defining characteristic of a
public good, because a good with non-rivalness in consumption can be a
private good if there are means to exclude other individuals (that is, the
so-called club goods). This means that the public-good problem is
fundamentally a property-rights problem. Good historical examples in
this regard are English public parks, which used to be private goods (or,
more exactly, club goods for the upper classes) but later became public
goods (for every citizen). Such a change in the nature of the goods
occurred not because the parks suddenly acquired non-rivalness in
consumption but because the rising political power of the commoners
made it impossible to exclude them from access to the parks (a change
in the property rights structure). 1 thank Bob Rowthorn for suggesting
this example.
3. Here it is interesting to note sorne parallel between this argument and
sorne Marxist theories of the state that regard the provision of public
goods as one ofthe major functions ofthe state (Mandel, 1975; Hirsch,
1978; Fine and Harris, 1979). For example, Hirsch (1978) argues that
the capitalist state should and does intervene in areas 'where the
"principies of exclusion" cannot be guaranteed' (p. 92). According to
Mandel (1975), the capitalist state should provide the general technical
preconditions of production (for example means of transport and
communication) and the provision of the general social preconditions of
production (for example law and order, a currency system) (p. 476) - in
other words, public goods.
4. Mishan (1982, p. 154) suggests that, when narning the public good, the
term 'public' should be avoided because of the misleading connotation
of the term that any good supplied by the state may be called 'public',
and that it should be substituted by the term 'collective'.
5. Moreover, given that cooperation is more likely to emerge when the
relationship is of a long-term nature (Axelrod, 1984), the state can
promote cooperation for the provision of public goods by encouraging
the formation of long-term relationships among private-sector agents -
for example, the provision of legal backing for industry associations or
forum for labour-capital dialogue.
6. Toe logic behind this is as follows. Regardless of the competitiveness of
the market, firms will equate their marginal costs with their marginal
revenues in order to maximise their profits. In competitive markets, each
firm faces a horizontal demand curve, and therefore marginal revenue
coincides with average revenue (that is, price). In a non-competitive
market where firms face negatively-sloped demand curves, marginal
139 Notes and
References to
to pp. 8-9
pp. 9-17 139
139 139
139
revenue - whích should stíll be equated wíth marginal cost to maxímise

profit - will be less than average revenue (that is, príce). Consequently
finns in a non-competitíve markct will produce less at a higher príce
than they would íf the market were competí tí ve.
7. Morcover, there are non-efficiency reasons to regulate monopoly. For
example it is argued that monopoly has to be regulated because it leads
to thc concentratíon of economíc and politícal powcr in prívate hands,
whích are not easíly subject to democratíc control (Cullis and Jones,
1987, p. 182).
8. Antí-interventíonísts argue that, even when scale economy díctates a non-
competítí ve market structure, the state should regulate prívate finns
rather than set upa public enterpríse, because prívate monopoly is better
than public monopoly. Accordíng to thís argument, prívate monopoly is
subject to sorne fonn of market discipline (for exarnple takeover,
bankruptcy) and is therefore more responsíve to the changing needs of
the consumers, whereas publíc monopoly is unlikely to be subject to such
discipline. Although we do not have enough space to díscuss the meríts
and demeríts of publíc monopoly vis-a-vis prívate monopoly, it has to be
poínted out there are ways and means to discipline publíc enterpríses (for
a detaíled discussion of thís issue, see Chang and Singh, 1993).
9. Nclson and Wínter (1982) generate sorne computer-símulatíon results
where, under certaín plausible assumptíons concemíng firm behavíour, a
reasonably compelitíve market transfonns ítself ínto an oligopolistic onc
through a stochastíc process.
10. Toe concept of freedom employed by Friedman is that of negatíve
frcedom ('freedom from'), but it is not the only acceptable concept of
freedom. There is another widely-accepted concept of posítíve freedom
('freedom to'). On different notions of freedom, see Berlín, 1969.
11. The 'Iiberalism' of Peacock and Rowley is what is usually called
'libertaríanísm'. They argue, quite correctly, that thís posítíon should be
dífferentíated from Paretían arguments (for example market-failure
arguments), because the latter are ends-oriented whereas the former is
process-oriented (also see Rowley, 1983).
12. Hayek (1972) contends that non-market coordínation of the economy
('planníng' in hís termínology) is doomed to technícal faílure due to
ínsunnountable ínformation costs. This ímportant poínt need not detaín
us here sínce our maín concem is the politico-phílosophícal side of the
argument. For a discussion of the ínformational problem in central
planníng, see Lavoíe, 1985. For an exposítíon of the informatíon
problem in general, see Arrow, 1974. See also Chapters J.4.1, 2.1.1 and
3.4.1 of thís book.
13. Although Brennan and Buchanan (1985) reject the 'state-of-nature-as-a-
fictíon' type of criticism as irrelevanl, they are correct only when one
makes such crítícísm and at Lhe same time endorses contractarianism (as
Brennan and Buchanan themselves admit: sce p. 22), a posítion that is
not held in this chapter.
14. For exarnple Bobbío (1987) justifies the state's concern for social
ínequalíty on contractarian grounds (p. 136).
140 Notes and
Notes andReferences
Referencestotopp. 24-9
pp. 17-24 140
140 140
140 140
15. For Marxist theories of the state, see Milíband, 1969, and Jessop,1982.
16. According to neoclassical política) economy, what differentiates the
market from the polity is not the different behavioural principie of the
participants (always self-interested) but the different constraints that
they face (the individual bearing or otherwise of the whole effects of
one' s decision) (McCormick and Tollison, 1981, p. 5). Still emergent,
this school has many different names, such as economics of politics
(Buchanan et al., 1978), neoclassical política) economy (Colander [ed.],
1984), public-choice school (Mueller, 1979), new political economy
(Findlay, 1990), or positive política) economy (Alt and Shepsle [eds],
1990).
17. In sorne models of interest-group theory, interest groups - acting as
maximisers of utility (or wealth) - make demands to claim the largest
possible amount of state largesse, which the politicians who make up
the state distribute with a view to maximising their votes, and not to
maximising social welfare. The result often is the generation of policies
that are 'irrational' from the social point of view, for example 'political
business cycles'.
18. From the viewpoint of extreme methodological individualism this
argument is a non sequitur, because if these norms restrict the formation
of a certain interest group, the individuals who belong to that potential
interest group can organise themselvcs first to remove such restrictions.
However, in the real world, as Eggertsson (1990) argues, 'various
factors work against special interest groups and limit their effectiveness,
including laws against bribery, rules regulating the behaviour of lobbyists,
and public-spiritedness of civil servants and legislators' (pp. 66-7).
Moreover norms may have arisen due to the utility-maximising
behaviour of individuals but they acquire an autonomous Iife once they
come into being, because individuals who join the society by birth or
immigration after these norms have been established will internalisc
them through the socialisation processes.
19. Moreover, even when assuming that there are no unnecessary activities
initiated by the budget-maximising attempts of the bureaucrats, the state
activities may not be fully cost-efficient, Toe fact that the state bureau-
cracy is in a monopolistic position may mean that there is x-inefficiency
in their activities (see Peacock, 1979b, pp. 123-4). x-inefficiency, as
defined by Leibenstein (l 966), is generated when production units use
inputs in the right proportion, but on an unnecessarily large scale. Thus,
even when strict marginal pricing is conducted within the state
bureaucracy, it may operate with considerable slack due to the lack of
competitive pressures.
20. Marxists emphasise this problem when they discuss the nced for the
state in a class society to ensure the Iegitimacy of the system (see
Gramsci, 1988; O'Connor, 1973; Gough, 1979).
21. Even from a purely selfish point of view, there is far more scope for
cooperation in a stable organisation like the bureaucracy than in a
strictly atomised world. Firstly, since an organisation by definition
implies a long-lasting relationship between a given set of people, there is
141 Notes and
Notes andReferences
Referencestotopp. 24-9
pp. 17-24 141
141 141
141 141
much more possibility of evolutionary cooperation, that is, development

of cooperative behaviour through continued reciprocity (see Axelrod,
1984). Sccondly, the environment to which a self-seeker has to adapt
himself/herself is not just the material world, but also the human actors
around him/her (Simon, 1991). Toen adaptation will inevitably mean
collaboration with those other actors (this is an analogy to Dawkins'
genetic argument; see Dawkins, 1986, pp. 170-1). In an organisation, its
members will be treatcd more as a part of the environment than those
who people encounter on a more prccarious basis in the atomised world.
22. Toe informational problem also has a politico-philosophical dimension,
especially from the point of view of the individualism-contractarianism,
As Rowley (1983) argucs, the belief 'that reason is paramount and that
appropriate rule of social choice can be devised to encompass public
sector decision-rnaking in a manner fully compatible with the Paretian
emphasis upon the primacy of individual preferences' is misguided
bccause the pcrvasive uncertaintics in human life make it inevitable that
such as approach ends up with the imposition of the social decisíon-
makers' individual preference 'to fill the void that otherwise would
destroy the smooth calculus of social choice' (pp. 35-6).
23. In this sense, the self-seeking bureaucrats approach (see Chapter 1.3.3)
is also partly an informational argument. However we included it in the
political-cconomy literature bccause it emphasiscs the difference between
the objectives of the public, or the top politicians, and the bureaucrats.
An asymmetric information problcm would exist even if the bureaucrats
genuinely want to serve the public interest, bccause of the limited scope
for knowlcdge transmission (Hayek, 1949a, 1978).
24. It is interesting to note that this explanation of the dynamic rent-
seeking process in a competitive setting is similar to the Marxian
argument on surplus profit (Marx, 1981) and the Schumpeterian argu-
ment on entrepreneurial profit (Schumpeter, 1961, 1987). More on this
in later chapters of the book.
25. Buchanan argues that, unless no single set of winners can sccure benefits
on each and every occasion, a 'constitutional revolution' will bccome
more feasible with the increase in opportunities for rent-seeking
behaviour, since each group gains more than it loses in net from the
revolution. In the case of a single set of groups constantly securing
benefits, a 'nonconstitutional revolution' - or violent revolution -
becomes a possibility (Buchanan, 1980b, p. 366).
26. Of course this point does not hold if the bribed had to spend extra time to
change the state's dccision in the briber's favour. In this case, the bribery
should be regarded, at least from the analytical point of view, as an
implicit wage to the bribed rather than as pure bribery (apure transfer).
27. That bribery does not have direct efficiency consequences - that it is a
pure transfer - does not mean that it has no efficiency consequences. It
may have negative indirect efficiency consequences when it enables less
efficient producers to get monopoly rights (sec below). Moreover wide-
spread bribing can lcad to loss in legitimacy of the political-cconomic
system of the country (Krueger, 1974). The loss of legitimacy may create
142 Notes
Notes and
and References to pp. 38-45
29-38 142
142 142
142 142
uncooperative altitudes among the underprivileged and therefore more

monitoring and enforcement costs will have to be incurred. Bribery,
however, may have positive indirect efficiency consequences when it is
used as a 'signa!' for one's superior ability as a producer (as heavy
advertising may be used as a signa! for superior quality).
28. 1 thank John Toye for reminding me of this importan! point.
29. Toe proposal for constitutional revolution is reminiscent of the 'rigidities'
argument put forward by the new right, whose favourite recipe for a
restoration of dynamism in advanced capitalist economies is to undo
oligopolistic arrangements, enforce the principies of fair trade, break
union monopolies, break institutional constraints that prevent wages
from 'finding their real leve!' and make labour mobile and rnarket-
responsive again. See sorne essays in Balassa and Giersch (eds), 1986.
For a critique of such views, see Johnson and Lundvall, 1989.
30. Market failure does not necessarily mean that state intervention is the
solution. As new institutional economics argues, the history of capitalism
is full of institutional innovations to deal with the failure of the market to
coordinate individual activities. The rise of the factory system, joint-stock
companies, rnultidivisional firms, and vertical integration and conglom-
eration are but a few such innovations (see for example essays in
Langlois [ed.], 1986).
2 A New Institutionalist Theory of State Intervention

1. This view of rationality and decision-making has been employed not
only by the orthodox economists but also by the proponents of central
planning. On this point, see Lavoie, 1985.
2. On the role of expertise in decision-making, see Simon, 1983. On the
problem of knowledge transfer, see Hayek, 1949a, l949b, 1978, and
Chapter 3.3.1 of this book.
3. In 1957 the Polish Economic Council suggested that '[p]lanning is
improved not by introducing a multiplicity of indices, by making
extrernely detailed projects, and by formal balancing, but by securing a
more profound economic analysis and well-founded estimates of
economic development in those areas where precise economic calculation
is impossible' [emphasis added] (quoted from Brus, 1972, p. 154).
4. Public enterprises in an industry with public and prívate firms may also
play a valuable role in reducing the informational asymmetry between
the state and the firms in that industry, by providing the state with ñrst-
hand knowledge about the workings of the industry.
S. These conditions are assumed to be satisfied in empirical studies that ·
estímate the magnitude of resources involved in rent-seeking in sorne
developing countries. These studies assume that an amount of resources
equivalent to the rent is dissipated through the competitive proccss,
yielding estimates that put thc magnitude of resources involved at 7-45
per cent of GNP in countries like Turkey and India (for example Krueger,
1974; Mohammad & Whalley, 1984). However, as we discuss below,
these conditions are frequently not rnet, and thcrefore the quantitative
143 Notes
Notes and
and References to pp. 38-45
29-38 143
143 143
143 143
estimates of these studies should be accepted with great caution.

6. Farrell ( 1987) raises the same point from the perspective of information
economics when he says that 'if a central authority fails to succeed in
commiuing himself to ignore information, it can lead to excessive
incentives for subjects to try to influence his decisions; this is the idea of
"rent-seeking" behaviour, which often has important social costs'
[emphasis added] (p. 121).
7. Sorne qualifications, which emerged from a discussion with Terence Moll,
may be needcd here. First of ali, it may be necessary to invest resources to
qualify for the queue (you do not get a license to run a petrochemical plant
just because you applied earlier). In this case, people will 'waste' sorne
resources in order to qualify, although it is possible that in the process
they will create socially beneficia) by-products (see below), for example in
the form of enhanced managerial skílls (to prove that you are able to runa
modem plant well). Secondly, it may be possible to jump the queue by
paying the adrninistrator of the queue, although the payment to the
administrator should not be counted as a rent-seeking cost if it does not
make the administrator do any more work (see Chapter 1.4.2).
8. For example, ali firms in a protected industry may benefit from tariff
protection whether or not they contributed to lobbying efforts of the
industry association.
9. Strategic uncertainty (uncertainty dueto interdependence) is a crucial
element in a game-like situation. After ali, the use of randomised (or
mixed) strategies is a way to prevent other players from capitalising
on the certainty of one's future action by deliberately creating uncer-
tainty about one's future course of action. The existence of parametric
uncertainty (uncertainty due to nature) also affects the magnitude of
rent-seeking 'waste'. If there is uncertainty about the size of the rent
(for example the case of a monopoly right to an infant industry whose
growth prospect is uncertain), the outcome of rent-seeking is not clear
(Fisher, 1985), although we may suppose a strong bias against the
standard conclusión of total rent dissipation (Rogerson, 1982).
IO. Moreover, monopolistic or oligopolistic incumbency in a rcnt market
helps one to build up the ability to enter into other such markets. However,
as we discuss later in the Korean context, rather than multiplying rent-
seeking costs, such 'interlocked' rent-seeking may, under certain
conditions, work to decrease such costs (see Chapter 4.4.2).
11. However, education may generate sorne positive externalities, thus
partially offsetting the social loss from rent-seeking (a case of the
beneficia! by-stander effect), 1 thank Jay Min Lee for raising this point.
12. Although it may sound objectionable to say that rent-seeking may create
sorne values because the rent-seekers derive utility from the very act of
rent-seeking, this is only because we are already assuming that the rent-
seekers are sínister people. lf we imagine that the rent-seekers are those
poor tish farrners who want to acquire the right to restrict entry into their
stretch of river by a chemical factory or workers fighting for better
working conditions, the difficulty of accepting this argument may be
reduced.
144 Notes
Notes and
and References
References to pp.
pp. 47-52
45-7 144
144 144
144
13. For example, a society may encourage competition in education while

restricting competition in politics, while another does the opposite.
14. On the question of 'learning', see Rosenberg, 1976, 1982; Amsden and
Hikino (forthcoming); and Chapter 3 of this book.
15. See Braverman, 1974; Marglin, 1974; Bowles, 1985; and Pagano, 1985,
on the Marxian labour-process Iiterature. See essays in Akerlof and
Yellen (eds), 1986, for the efficiency-wage argument. See Alchian and
Demsetz, 1972, for the theory of team production.
16. There exists no agreed or exact definition of 'new institutional
economics', which still is emcrgcnt as a school. Transaction-cost
economics a la Oliver Williamson (for example Williamson, 1975,
1988) and property-rights economics a la Ronald Coase (for example
Coase, 1960; Alchian and Dernsetz, 1972; North, l 990b) are probably
the two most important strands within new institutional economics
(Putterman, 1986, p. 20, n. 27). However there also exist strong
influences from the Schurnpeterian, Austrian and behaviouralist (a la
Herbert Simon) traditions (Langlois, 1986a, pp. 1-2). Eggertsson (1990,
pp. 7-10) argues that transaction-cost economics - what he calls new
institutional economics - is different from property-rights economics -
what he calls neoinstitutionalist economics - in that it rejects the 'hard-
core' assumption of neoclassical economics about full rationality and
adopts the heterodox concept of bounded rationality. Although
Eggertsson's point is valid, the demarcation between the above two
groups is not as clear as he suggests, and the two groups share enough
commonalities to be put under the single heading of 'new institutional
economics'.
17. Needless to say, whether or not it can better allocate resources than the
market is another matter.
18. To date, there is no agreed definition of transaction costs, but we are in
general agreement with Matthews (1986), who defines transaction costs as
'the costs of arranging a contract ex ante and monitoring and enforcing it
ex post, as opposed to production costs, which are the costs of executing
the contract' (p. 906). At this point it should also be noted that it is not
easy to differentiate the production-cost elements and the transaction cost
elements in practice, even of a single activity, not to mention the activities
of the firms and of other organisations which partially perform economic
functions (for example the state, the family). For example it is difficult to
tell how much of a foreman's labour belongs to the domain of production
cost (for example his technical advice to the supervisee) and how much to
the domain of transaction cost (for example his monitoring activities). 1
thank Bob Rowthom for reminding me of this ímportant point.
19. Matthews (1986) is opposed to defining transaction costs as the costs of
interaction between people on the ground that there are production costs
that involve relations between people, particularly the costs of providing
personal services (p. 906). However the object of transactíon in the
provision of personal service is the service and not the human being who
is providing the service. Imagine a future society where ali personal
services are conducted by robots. Even in this society, there will stil! be
145 Notes
Notes and
and References
References to pp.
pp. 47-52
45-7 145
145 145
145
transaction costs due to the interaction between the recipient of the

services and the owner of the robots. lf this is the case, it may be useful to
define transaction costs as 'the costs of interaction between people'.
20. One intriguing point that emerges by reinterpreting rent-seeking costs
as variants of transaction costs is that someone who thinks rent-seeking
is wasteful is compelled to take the controversia) position that other
activities that incur transaction costs - informational, bargaining and
enforcement costs (for example stockbrokers, lawyers, judges,
policemen) - are also wasteful or unproductive (for discussions of
unproductive activities, see Wolff, 1987, ch. 2, and Boss, 1990). There
may be two possible responses to this point, neither of which are
convincing. ()) It may be argued that rent-seeking costs are artificially
created by state intervention, whereas other transaction costs are natural.
However influencing the state is only one of many ways of changing the
existing property rights structure and is not more or less natural than
other methods of changing it. (2) It may be argued that other transaction
costs are necessary evils whereas rent-seeking costs are unnecessary
evils because the former are more than offset by gains in efficiency -
which is achieved through the transfer of resources to the best (or at
least better) users - but the latter are not. However such difference exists
only because any efficiency gain from the state intervention which starts
off the rent-seeking process is assumed away (for an example, see
Buchanan, l 980b, p. 359)
21. Expanding on the above example of the pure coordination problem, this
will be the case when a society, in which people have been driving on
any side of the road they Iike, tries belatedly to decide on one side.
Given the higher possibility of accidents when people drive on both
sides, everyone will prefer there to be a designated side. However
different people will prefer different sides because say, they have cars
with steering wheels on different sides and/or may be reluctant to
change their driving habits.
22. This may be somewhat counterintuitive for those who are familiar with
the ncoclassical model of perfect competition, where the the existence of
many agcnts docs not lead to a coordination failure. However such a
result obtains only because sorne stringent assumptions are made in this
model to eliminate the interdependence between individual decisions
and therefore any need for coordination. Por a more detailed discussion
of this point, see Chapter 3.2.1.
23. In addition to the reduction of transaction costs involved in people's
atternpts to overcome the strategic uncertainty inherent in the
coordination problem, the state also has an important role to play in
reducing what Koopmans (1957) calls 'primary uncertainty', which arises
from random acts of nature and unpredictable changes in consumers'
preferences (p. 163). First of ali, the state can provide information whose
provisión is subject to scale economy, for example information regarding
the condition of the world market. It is well known that the information
on export markets collected by the Japanese and Korean government
trading agencies (JETRO and KOTRA) and diplomatic services played
146 Notes
Notes and
and References pp. 55-61
References totopp. 52-4 146
146 146
146 146
an important role in the exporting successes of these countríes. Secondly,

the state can provide infonnation with externalities, for exarnple technical
knowledge. Basic technical knowledge generated by non-profit-rnaking
institutions (for exarnple universities) will generate externalities in the
sense that those who utílise that information can improve their
productivity without paying the full cost for it. lf such a piece of
information is not readily available to the firms because of a lack of
communícation between them and the knowledge-generating institutions,
the state can act as a sort of information-clearinghouse, as is actually
practiced in many countríes.
24. Organising society will not only reduce transaction costs in the prívate
sector but also the transaction costs involved in state intervention, by
delegating the responsibility of monitoring and enforcernent to organisa-
tions such as industry associations and labour unions. Indeed the early
Marxíst belief in the increasing feasibility of central planning was also
partly based on the beliefthat the continued process of 'concentration and
centralisation of capital' would reduce the (transaction) costs involved in
state intervention (for a critical discussion of this view, see Tomlinson,
1982).
25. Here the tenn 'ideology' is used as a shorthand for a set of social norms
and values and not as meaning 'false belief" or 'pseudo-science', as in
sorne studies of ideology. See McLellan (1986) for different concepts of
ideology.
26. Many ideologies frame the decision-making process so as to encourage
people to identify themselves with people who are outside the usual
boundaries of an individual's limited personal and other channels of
identification (professional, ethnic and so on). For example, the
American ideology of industrial productivity and the Gennan and ltalian
fascist ideologies, which dismissed the liberal concept of trade-off
between growth and equity and emphasised the non-zero-sum nature of
distributive struggle, were accepted by different classes with conflicting
interests (Maier, 1987, chs 1-3).
27. Of course, as Hall (1987, p. 155) argues, this type of planning exercise
cannot be purely indicative and would probably have to involve a degree
of coercion in order to be effective, not least because most coordination
problems involve sorne conflicts of interest, as in our example of the
generalised coordination problem.
28. Commissioning pilot research projects, organising cooperative (among
firms) research projects, and providing 'signalling' subsidies for sunrise
industries are other examples of the state providing of a focal point (for
examples from various countries, see Porter, 1990, especially pp. 635-6,
639-40).
29. Our theory is also able to accommodate insights from the political
economy Iiterature by explicitly introducing the role of institutions and
ideas (or values) in the political process, although we did not fully
pursue this line of inquiry. For sorne works trying to incorporate the role
of institutions and ideas in explaíning the economic policy process, see
Tornlinson, 1980; Evans et al. (eds) 1985; and Hall (ed.), 1989.
147 Notes
Notes and
and References pp. 55-61
References totopp. 52-4 147
147 147
147 147

1. See NEDO, 1978; Stout, 1979; Cairncross et al., 1983.
2. See Singh, 1977; Blackaby (ed.), 1979; Rowthorn and Wells, 1987.
3. See Pinder (ed.), 1982; Jacquemin (ed.), 1984; Cox (ed.), 1986; Duchéne
and Shepherd (eds), 1987.
4. For sorne interesting reviews of the debate, see Norton, 1986;
Thompson, 1989.
5. See essays in Kirzkowski (ed.), 1984, and Krugman (ed.), 1988.
6. For example Dombusch et al., 1988, take a pro-rnanufacturing position
but recommend better macroeconomic managemenl rather than industrial
policy as the major solution lo the currenl industrial decline in the USA.
7. Of course this is not lo argue that a change in the employmenl structure
has no irnpact on our socioeconomic life. Thc growing importance of
service activities may have a significanl impact on people's life-styles,
on their relalionship with others, on their perceplion of the world, and so
on, especially by providing people with different work experiences from
those in the manufacturing sector (on the 'constitutive' nature of labour
processcs, see Bowles and Gintis, 1990). J thank Michael Landesmann
for raising this importan! point,
8. Lawrence (1984) argues against using industrial policy as a surrogate
regional or group-oriented policy on the grounds that 'particular
objectives - such as meeting national defense needs, redistributing
incomes, and promoting regional development - can ali be achieved by
more precise policies' (p. 115).
9. Our definition of industrial policy is based on a stylised version of
industrial policy conducted in 'industrial-policy states' such as Japan,
France and Korea. On Japanese industrial policy, see Magaziner and
Hout, 1980; Johnson, 1982; Dore, 1986; Okimoto, 1989; Best, 1990,
ch. 6. On Korean industrial policy, see Jones and Sakong, 1980; Luedde-
Neurath, 1986; Amsden,1989; and Chapter 4 of this book. On French
industrial policy, see Cohen, 1977; Hayward, 1986; Hall, 1987.
10. For example, in their study of Japanese industrial policy, Magaziner and
Hout ( 1980) document that MJTJ will often 'suggesl that a company
participate in an unappealing foreign investrnent projecl or delay a
capacity addition to accomplish a broader end' [emphasis added]
(p. 34).
11. This part of the definition helps us deal with those who downgrade the
importance of industrial policy, say in Japan, by saying that Japan would
have grown quickly and become rich anyway given factors like the
'catching-up effect' and high savings ratios (for example Krugman,
1984). From our perspective, however, the important point is not
whether Japan would have become rich or not without industrial policy
but whether the structure of the present Japanese economy is in line with
what the Japanese state intended when it implemented industrial policy
measures (by choosing what technology to deploy, where to channel the
savings and so on). And in this respecl, there is no doubt that Japanese
industrial policy has played a crucial role.
148 Notes and
References to
to pp. ó2-9
pp. 69-73 148
148 148
148 148
12. Moreover, as Williamson (1988) states, 'it does not suffice to demon-
strate that a condition of large numbers competition obtains at the outset.
Jt is also necessary to examine whether this continues or if, by reason of
transaction specific investments and incomplete contracting, a condition
of bilateral trading evo/ves thereafter' (p. 71 ).
13. What Hayek calls the 'competition as a state of affairs' view is still
dominant in the field of industrial economics (on different notions of
competition, sec Hayek, 1949b, 1978; McNulty, 1968; O'Driscoll, 1986).
For example, even a most updated industrial economics textbook, Tirole
( 1988), argues that replacement of one monopolist by another through a
patent race 'does not mean competition, as one monopolist replaces
another' (p. 396, n. 12). However Hayek would have argued that the
replacement was a result of 'the activity called competition',
14. Dobb (1925, ch. 23) expresses a similar concem. He describes the
capitalist economy as an 'economic anarchy' that gives fluidity to the
economy but at the cost of instability due to coordination failure. In
particular, he points out that miscalculations by competing firms might
not cancel out because expectations tend to move in the same direction.
15. lt should, however, be noted that Marx had another vision of socialist
society organised on the basis of a more democratic and less specialised
division of labour, which Pagano (1985) aptly calls 'anti-firm
communism' (p. 60).
16. Amadeo and Banuri (1991) argue that there is a correlation between the
liquidity of the assets owned by different groups (for example financiers,
industrialists, workers) and their degrees of support for unregulated
competition in the market.
17. Sorne recent developments in mainstream theory attempt lo incorporate
these observations through modelling 'wasteful R&D' and the like.
However this type of model is not generally extended to the theory of
competition in general.
18. The very diversity of coordination devices in a capitalisl economy is a
testimony to the diversity of coordination problems to be sol ved. And this
is one reason why we emphasise the particularistic nature of industrial
policy, since, to be successful, it has to be custorn-designed to fit the
nature of the coordination problem in volved in a particular instance.
19. Marx's concept of 'constant capital' (which was absent in the Ricardian
systern) and the Austrian concept of 'roundabout methods of production'
are two importan! ways of theorising such a characteristic.
20. Of course there are other options open to firms. One is to find a hitherto
unexploited market, say, through exporting. Another is to diversify into
other related industries, as shown in the examples of sorne Japanese firrns
in declining industries like textiles, brewing and food-processing that have
successfully diversified into biotechnology (see Okimoto, 1989, p. l 28).
21. However, devising a side-payrnents scheme is not easy because the
estímate of future costs and benefits from exit may differ among the
agents concemed. For example, in the branches of the Japanese textile
industry that are dominated by many small-scale firrns, who were hard-
pressed by imports from the NICs in the early 1980s, the govemment
149 Notes and
References to
to pp. ó2-9
pp. 69-73 149
149 149
149 149
took the view that exit compensation should be financed by the remaining
firms who would benefit from such exit, while the remaining firms argued
that the exiters' share of the total market was minute (and hence benefit
negligible) and that any room left in the market by the exit of home
producers was likely to be taken up by imports (see Dore, 1986, p. 236).
22. In this case, side-payments will mainly involve compensation for
workers who are laid off. For example, in the case of the Japanese
shipbuilding industry, additional unemployment benefits and special
placement services for workers, provided by the state, were important in
arranging a speedy capacity-scrapping arrangement (see Renshaw, 1986,
p. 145; Dore, 1986, p. 143).
23. In the case of the Japanese textile industry, sorne equipment was
mothballed in 1978, but it was eventually scrapped in 1981 as the
demand downtum proved to be permanent (Dore, 1986, pp. 235-6).
24. More generally, given the limited human capacity to process
information, the introduction of rigidity in behaviour through such long-
term binding contracts may well be essential to achieve rational
decisions (Simon, 1983; Heiner, 1983).
25. In this particular case, no exit was negotiated, and the cut was graduated
to the size of the firm, ranging from 40 per cent for the seven biggest
firms to 15 per cent for the 21 smallest (Dore, 1986, p. 145).
26. And this is why '[c]onsumers do not wish to contract for their future
purchases because they cannot foretell what their future needs and
opportunities will be; and produccrs do not generally wish to commit
themselves to forward purchases of inputs because they cannot predict
the productive possibilities that will be open to them' (Richardson, 1971,
p. 437).
27. In the case of the Japanese aluminium-smelting industry, one reason pul
forward for not cutting capacity to the leve) dictated by current relative
prices (mainly due to the oil price hike and the conscquent rise in
electricity prices) was the need to maintain a sizeable industry to support
an R&D capacity, which is an important precondition for regaining
intemational competitiveness if the cost situation improves in the future
(Dore, 1986, p. 143).
28. While arguing that the coordination failures of the market entails enorm-
ous waste and therefore needs to be replaced by less wasteful ex ante
coordination through central planning, Marx was also a precursor of the
Schumpeterian, and to sorne extent the Austrian, 'process' view of com-
petition, which emphasises the role of market competition in developing
the 'forces of production' (see, for example, Marx, 1981, pp. 373-4).
29. In contrast, in the ncoclassical frarnework every piece of information (or
knowledge) is seen as obtainable, albeit ata cost (for example search cost)
(for similar views, see Heiner, 1988, p. 148, and Pelikan, 1988, p. 385).
30. The importance, in the nineteenth century, of the migration of skilled
technicians in transmitting technical knowledge from one to another part
of the then industrialised world (that is, Europe and the USA)
documented by Rosenberg (1976, pp. 154-5) shows the difficulty of
codifying technical knowledge.
150 Notes
Notes and
and References
Referencestotopp.
pp. 77-86
73-6 150
150 150
150 150
3 J. For evolutionary arguments in economics, see Alchian, 1950, and

Nelson and Winter, 1982.
32. Toe same view of economic evolution líes in the French state's claim
that it was involved in rescue operations of declining sectors not 'to save
endangered spieces but to provide funds for their mutation' [emphasis
added] (Fortune, 9 April 1978; quoted in Hall, 1987, p. 190).
33. On the notion of cumulative causation, see Young, 1928; Kaldor, 1985;
Stigler, 1951.
34. Of course computer finns may decide on in-house production of semi-
conductors, but there is no guarantee that scale economies in computer
production and semiconductor production will be ofthe same magnitude.
lf the semiconductor industry is subject to a larger scale economy (as is
the case), .the in-house production option will be costly compared with
the option of production by independent semiconductor producers,
35. Porter ( 1990) reports that '[i]n the United States and often in Europe, the
process of reaching technical standards is frequently protracted as firrns
jockey for their individual positions. In Japan, MITl has frequcntly
applied significant pressure on firrns to set basic standards, pushing them
to move on to the next stage in the ínnovation cycle' (p. 653). Sce also
the examples of the Japanese computer and machine tools industries in
Dore, 1986, pp. 134-6.
36. It may not matter, for example, whether a country goes for
superconductivity or biotechnology, but it matters whether enough
complementary investrnents are made in either of these industries.
37. For a classic discussion of the problem of coordinating complementary
investments, see Richardson, 1960. Also related are concepts like
Hirschman's 'linkages' (Hirschman, 1958) and Dahmén's 'development
blocks' (Dahmén, 1988).
38. Magaziner and Hout (1980) argue that 'MITI's greatest strength appears
to be its understanding of the competitive stages through which an
industry moves and its ability to fashion appropriate policy' (p. 38).
They document that '[f]or businesses in the early, rapid growth phases
of development ... policy calls for protection from foreign ccmpetition,
concentration among producers, government support of the industry's
cash flow, and stimulation of new technology .... For businesses which
are aiready internationally competitive . . . government assistance
recedes [with the signiñcant exception of occasional officially
sanctioned recession and export cartels co-ordinated by the industry
associations] as it is no longer necessary. Finally, for businesses in
competitive decline, MITI becomes active again, this time trying to
bring about capacity reduction and rationalisation' (pp. 38-9).
39. Of course we should not forget the possibility of rejuvenation, whereby
a new series of exogenous technical changes turn a mature industry into
a young one again, although we may not go as far as Pierre Dreyfus, a
former French minister of industry, who argued that '[t]here are no
condemned sectors; there are only outmoded technologies' (quoted in
Hall, 1987, p. 210). l thank Nathan Rosenberg for reminding me of this
important point.
151 Notes
Notes and
and References
Referencestotopp.
pp. 77-86
73-6 151
151 151
151 151
40. A good example of the shift in policy emphasis according to product

cycle is the production cartel among six Japanese computer peripheral
equipment producers organised by MITI in the late 1970s. "The products
handled through the cartel included [mainly] standard peripherals whose
design had stabilised and where further innovation was remole'
(Magaziner and Hout, 1980, pp. 83-4).
41. On the role of diversity in the economic system, see Johnson and
Lundvall, 1989, pp. 103-4. On the role of genetic diversity in the
biological world, see Axelrod, 1984, p. 170.
42. See the 'long purse story' (pp. 337-9) and the case of high fixed-R&D
costs (p. 414) in Tirole (1988) for the implications of an imperfect
capital market for R&D activities,
43. Needless to say, dueto the limited human ability to process inforrnation,
the greater availability of inforrnation does not guarantee a better
decision (see Chapter 2.1.2).
44. Of course it has to be recognised that there exists 'second-rnover
disadvantage', because the first-movers would reap more rents from
innovation. 1 owe this point to Sandeep Kapur. See also Landes, 1990;
and Amsden and Hikino, forthcoming.
45. Interestingly sorne Marxist denunciation of the 'revisionist' strategies of
'co-opted' trade-union leaderships in favour of rank-and-file militancy (for
example Panitch, 1981) is based on the same view of inforrnation. As
Tomlinson ( 1982) argues, this position ultimately depends on 'a belief that
the experience of the ordinary employee, hislher experience of the
oppressive and exploitative relations of capitalist wage labour will
guarantee the appropriate socialist direction in the struggle. Toe experience
gives a privileged access to the appropriate means to change the capitalist
enterprise whereas the experience of the trade union official may lead him
or her to a different and wrong assessment' [emphasis original] (pp. 44-5).
46. For discussions of the different relationships between the state and
finance in different OECD countries and their implications for industrial
adjustment, see Zysman, 1983, and Cox (ed.), 1986. On the same issue in
the French context, see Hall, 1987. For the Korean case, see Chapter 4.
47. For a discussion of this problem in the French context, see Hall, 1987,
pp. 176-80.
48. And he continues: as a consequence, 'harrnony can only be obtained by
coercion or by a series of compro mises ... and purely strategic con-
siderations may tend more and more to override any considerations of
maximum welfare and efficiency'. It 'may also produce class struggle',
and, '[i]n conflicts of this kind, a considerable part of the economic
resources of a community may be consumed, either in their conduct or
in their prevention [emphasis added]' (p. 158).
49. It has to be pointed out, however, that parliamentary representation (the
representatíon of individuals as individuals) is not the only legitimate
form of representation. Representation along class lines (as in
Scandinavian social corporatism) or even issues (as in sorne American
lobbying organisations) are too pervasive to be dismissed as illegitimate
(see Maier, 1987, on different forms of representation).
152 Notes and References to pp. 87-109
i09-18 152
152 152
152 152
50. Interestingly, American authors such as Badaracco and Yoffie (1983),

Schultze (1983, pp. 9-10), and Lawrence (1984, pp. 112-15) usually
emphasise the absence of an élite bureaucracy and British authors such
as Hare (1985, pp. 112-13) ernphasise the hostility between the state
and the capitalists as the major obstacles to an effective industrial policy,
reflecting the institutional characteristics in their respective countries.
4 Industrial Policy in Action - The Case of Korea

1. The earlier view of Korea as a free-market and free-trade economy is
represented by Ranis and Fei, 1975, and Balassa, 1982. Iones and
Sakong, 1980, is a seminal work in the altemative explanation, ern-
phasising the role of the state. Luedde-Neurath, 1986, provides a detailed
empírica! study criticising the free-trade view of Korea. Amsden, 1989,
is, to date, the most comprehensive study in this altemative tradition.
2. It is interesting to note that recently one of the authors just quoted, by
emphasising the role of 'institutional/organisational changes orches-
tratcd by the govemments' of Korea and Taiwan, in effect contradicted
his earlier position (see Ranis, 1989).
3. Although there has been more trade liberalisation since Luedde-
Neurath's study, ali of the following 'invisible' import restrictions are
still being imposcd, if less widely and less rigorously.
4. For instance, the domestic price of importcd scotch whisky, whose tariff
was 100 per cent, was over nine times that of the c.i.f. price after the
imposition of various inland taxes, for example liquor tax, luxury con-
sumption tax and value added tax (Luedde-Neurath, 1986, p. 130).
5. In other words, '[b]y paying depositors low real interest rates and by
controlling capital outflows, the govemment implicitly taxed depositors,
then channelled the proceeds to favoured sectors for investment'
(Dombusch and Park, 1987, p. 418).
6. In 1980 Korea experienccd negative growth for the first time since the
industrialisation drive began in the early l 960s. For a different inter-
pretation of the early l 980s crisis, see Chang, 1987.
7. This is reminiscent of the situation in Japan during the l 950s and l 960s,
when 'only machinery which could not be made in [the country] was
allowed to be importcd' (Magaziner and Hout, 1980, p. 75).
8. The Korean shipbuilding industry raiscd its share of world shipbuilding
output from nil in 1973 (when the first modern shipyard started
production) to 4 per cent in 1980 and then to 21.6 per cent in 1986. This
made Korea the second largest shipbuilding nation in the world (next to
Japan), producing more ships (in gross tonnes) than Western Europeas a
whole, whose world-production share was 12.2 per cent in 1986 (NRI,
1988, pp. 162-4).
9. In Korean newspapers one always reads that 'the government has
allowcd', and not that 'firm A has decidcd on', a certain percentage price
rise of a product.
10. The most important policy documents include various Five-Year-Plan
(FYP) documents and the White Paper on the Economy (WP), produced
i09-18 153
153 153
153 153
annually by the Economic Planning Board (EPB), the planning rninistry.

The First FYP (1962-6) was issued in 1961 and revised in 1964. Sorne
of the later FYPs, for example, the Fourth and the Sixth, were also
rcvised during the plan period, but the changes were mainly ones of
macroeconomic forecasting rather than of policy.
l 1. For a critique of the dependency view on Korea, see Chang, 1990.
12. WP (1968) states that 'it is ncedless to say that the attainment or
otherwise of efficient scale of production is the most fundamental
determinant of productivity' (p. 174).
13. On the same practice in France, see Hall, 1987, p. 167.
14. On the issue of technology assimilation, see Rosenberg, 1982;
Abramovitz, l 986; Dore, 1989.
15. Of course this does not necessarily mean that it is impossible to develop
a national technological capability by inviting multinationals in. For
example Singapore has developed a sophisticated technological capa-
bility with a large multinational presence (see Evans, 1987, pp. 208-9).
Likcwise there has bcen a substantial technological diffusion through
subcontracting by multinationals in the Malaysian electronics industry in
Penang (see Rasiah, 1990). The importan! question then becomes: what
kind of policies are necessary to promote technological transfer by
multinationals?
16. In the country's earlier period of development, the Japanese state also
legislated that no repatriation of eamings or capital would be guaranteed
from foreign investment in scctors that were deemed to generate few
extemalities (for example marketing), and that, even in other industries,
repatriation would be guaranteed only when the investment 'contributes
to the development of the domestic industry' (Magaziner and Hout,
1980, p. 56).
17. The most importan! tool of industrial policy in Korea has been policy loans
- loans with subsidised interest rates and/or priorities in the (ubiquitous)
credit rationing. Policy loans accounted for 57.9 per cent of total bank
loans made between 1962 and 1985 (Y. S. Lee et al., 1987, p. 53, table III-
1 ). See 4.5 below for a more detailed discussion of the use of policy loans.
18. The concept of 'basic' industry is also employed in other industrial-
policy states such as France and Japan. In early French planning, coa),
steel, cernent, agricultura! machinery, electricity and transportation were
identified as 'basic' sectors (Cohen, l 977, p. 86). In Japan the office of
MITI that deals with the steel, chemicals and fertiliser industries is
called the 'Basic Industries Bureau' (Magaziner and Hout, 1980, p. 34).
l 9. This is in contras! with the Japanese practice whereby industries become
eligible for assistance according to the Structurally Depressed Industries
Law only if more than two-thirds of the producers petition MITI
(Magaziner and Hout, 1980, p. 40).
20. For details, see Kim, 1989, pp. 36-62; Paik et al., 1988, pp. 46-7; S. H.
Lee et al., 1989, pp. 64-73.
21 . Aftcr the three abovementioned unsuccessful attempts to raise money for
the stcel mili, the Korean state finally secured funds through war
repatriation from the Japanese government and loans from the Japanese
154
154 Ex-Im Bank (EPB, 1982, pp. 92-4). Ten years after it started operation, the
state-owned steel mili (POSCO) became the fourth largest and one of the
most cost-efficient steelmakers in the world (Amsden, 1989, pp. 298-9).
Toe success of the Japanese automobile industry is too obvious to discuss
further.
22. On the problem of 'learning', see Rosenberg, 1982; Abramovitz, 1986;
Dore, 1989; Amsden and Hikino (forthcoming).
23. On the basis of his theory of prescriptive state intervention, Bhagwati
(1988) suggests that 'prescriptive governments provide fewer induce-
ments for such unproductive activities, because the prescriptions leave
large areas open for initiatives' (p. 100). Nevertheless prescriptive
policy is no less prone to rent-seeking, because the logic of rent-seeking
is precisely that even if there are many areas open for private initiatives,
individuals will not engage in activities in such areas if their rates of
return are lower than in unproductive areas. For example it is well
known in Korea that obtaining subsidised credits (for example by being
a producer in a priority sector) is a very profitable business in itself,
because one can divert that money into the curb market where real
interest rates are much higher (see Table 4.6).
24. Toe introduction of limited lifetimes for rationalisation programmes due
to the implementation of IDL may be understood as an attempt to make
the state-created rents more 'transitory'.
25. For a discussion of the role of bargaining and conflicts in the process
of imposition of state discipline on the prívate sector, see Khan,
1989.
26. In addition to rolling over KHJC's debt (details not released), the state
tried to boost its activities by giving it monopoly rights to produce
certain power-generation components and certain heavy-construction
equipment (FEER, 2 June 1983, pp. 67-8).
27. The history of the passenger-car industry characterises how the Korean
state shapes up a targeted industry (the following account is based on
KDB, 1981, pp. 501-6). Toe Korean passenger car industry started in
1962 by assembling imported semi-knock-down (SKD) kits. In 1965
there was a state-led merger between the two existing passenger-car
makers (Sénara and Shinjin) into one (Shinjin). In 1968 there were two
more entries (Hyundai and Asia), and thereafter new entries were
banned until 1972, when a lorry producer, Kia, was allowed to enter. In
1974 the government announced the Long-tenn Plan for Promotion of
the Automobile Jndustry to develop local passenger-car models - at this
point producers were still assembling complete-knock-down (CKD) kits
from Toyota, Ford, GM and Fiat - and forced Asia, which had failed to
submit a plan for the development of its own models, to exit and sell its
capacities to Kia. From 1974 local models were developed (Kia's
'Brisa' in 1974, Hyundai's 'Pony' in 1975, GMK's 'Gemini' in 1977).
In the meantime Shinjin withdrew from GMK (its joint venture with GM
since 1972) due to a sales crisis of the newly introduced Chevrolet
models. Shinjin's stake was bought by the state-owned Korea
Development Bank (KDB) (hence a KDB-GM joint venture), which
Notes and References to pp. 122-5 155
ultimately sold it to Daewoo in 1978 (it has been a GM-Daewoo joint

venture since then until 1992). In the 1980 reorganisation of the sector
the original plan had been to force Kia to exit and to merge Hyundai and
Daewoo, but the plan was dropped because Korean policymakers would
not accept Daewoo's wish to abandon its local models in favour of the
'world car' models of GM. The market was maintained as a duopoly
between Hyundai and Daewoo until 1987, when Kia was allowed to re-
enter.
28. It has to be pointed out that Confucianism in itself is not necessarily
beneficia) to economic development. lts contempt for commercial and
industrial pursuits - the merchant and the craftsman occupied the two
lowest castes in the traditional Confucian social hierarchy - could have
acted as an obstacle to industrialisation. Actually, Confucian contempt
for commercial and industrial pursuits was often blamed for the
relatively poor Korean economic performance in the 1950s.
29. Korea has traditionally been even more centralised than other Confucian
countries. Toe Japanese feudal system was fairly decentralised until the
Meiji Restoration, and the Chinese system, because of the sheer size of
the country, hada strong tendency to dissolve into a decentralised one
except in the heights of a dynasty.
30. For a comparison of Korea with India, see Datta-Chaudhuri, 1990. For
comparisons of Korea with Latin America, see Fishlow, 1990; Shapiro
and Taylor, 1990; Hughes and Singh, 1991.
31. Park Chung Hee was a fierce nationalist strongly influenced by
corporatist ideas: his education in the Japanese military academy in
Manchuria, which he joined after a brief career as a school teacher, left
him deeply influenced by the Japanese version of corporatism. Such
influence is illustrated well by his naming his 1972 mid-career political
coup (to guarantee himself a lifetime presidency) the 'October
Restoration', after the Meiji Restoration. Park was also strongly
influenced by communism, which played an important role in the
Korean nationalist movement during Japanese colonial rule. His brother
was an influential local communist leader, and he himself was sentenced
to death (but earned an amnesty by publicly denouncing communism) as
one of the leaders of a communist mutiny within the Korean army in
1949. His close political aides shared similar ideas. On the debate
concerning the fascist nature of Japanese corporatism, see Halliday,
1975, pp. 133-40. On the influence of fascist ideology on the Japanese
élite during the interwar period, see Johnson, 1982, ch. 4.
32. With the growing sophistication of the economy, which puts more
informational demand on the state, there is a growing recognition among
Korean policymakers of the need for decentralisation in economic
policy-making. Hence, talk of the government's commitment to the
'private-led economy' in later policy documents (especially the Fifth
and Sixth-FYP documents). However such an emphasis, paradoxically,
testifies to the persisting dominance of the state - if the economy had
already been privately led, no comments of this sort would have been
necessary.
156
33.
156 Short-term working capital, given credit rationing, was mainly provided
either by small non-bank financia! institutions or by the curb market.
34. For example, in the French case the Planning Commission had to engage
in interdepartmental competition with the Ministry of Finance which had
'control over macroeconomic management, the How of funds and all
forms of public spending' (Hall, 1987, p. 172). See Johnson, 1982, p. 256,
for the conftict between MITI and the Ministry of Finance in Japan.
35. On the issue of ideological mobilisation in the Korean developmental
process, see You and Chang ( 1993).
36. The average equity participation ratio (as a proportion of total assets)
for Korean manufacturing firms was 22. 7 per cent during 1971-9 (Cha,
1983, table V-30). For 1986 it was 22.2 per cent, which is even lower
than for Japanese firms (28.3 per cent), which are known for their high
leverages. Compare this with the 47 per cent for US firms (1986) and the
46.8 per cent for Taiwanese firms ( 1985) (see Paik el al., 1988, p. 43,
table 1-6).
37. 'The govemment [still) appoints senior bank officials and major credit
allocation decisions have traditionally been cleared with government
authorities' (FEER, 21 April 1988, p. 58).
Conclusion: The Market, the State and Institutions
1. I benefited greatly from discussions with Bob Rowthom and Richard

Wright in clarifying this point.
2. In this respect the following passage by Coase is also illuminating:
'[T)he [neoclassical) analysis proceeds in terms of a comparison
between a state of laissez faire and sorne kind of ideal world. This
approach inevitably leads to a looseness of thought since the nature of
the altematives being compared is never clear. In a state of laissez faire,
is there a monetary, a legal ora political system and if so, what are they?
In an ideal world, would there be a monetary, a legal or a política)
system and if so, what would they be? The answers to all questions are
shrouded in mystery and every man is free to draw whatever conclusion
he líkes. Actually very líttle analysis is required to show that an ideal
world is better than a state of laissez faire, unless the definitions of a
state of laissez faire and an ideal world happens to be the same. But the
whole discussion is largely irrelevant for questions of economic policy.
Since whatever we may have in mind as our ideal world, it is clear that
we have not yet discovered how to gel to it from where we are. A better
approach would seem lo be to start our analysis witb a situation
approximating that whicb actually exists, lo examine the effects of a
proposed policy change and to attempl to decide whether the new
institution would be, in total, be/ter or worse than the original one. In
this way, conclusions for policy would have sorne relevance to the actual
situation' [emphasis added) (Coase, 1960, p. 43).

Ha Joon Chang The Political Economy of Industrial Policy Palgrave Macmillan 1993

Uploaded by

Copyright:

Available Formats

Ha Joon Chang The Political Economy of Industrial Policy Palgrave Macmillan 1993

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ha Joon Chang The Political Economy of Industrial Policy Palgrave Macmillan 1993

Uploaded by

Copyright:

Available Formats

The Political Economy

First edition 1994

ISBN 0-333-58862-2 hardcover

First publíshed in the United States of Ameríca 1994 by

All rights reserved. No reproduction, copy or transmission of

Printed in Great Britain by

1 Theories of State Intervention: A Review of the 7

2 A New Institutionalist Theory of State Intervention 33

2.2.1The vulnerability of the state 38

3 The Political Economy of Industrial Policy 55

4 Industrial Policy in Action - The Case of Korea 91

4.1 A brief overview of the Korean economic 92

Conclusion: The Market, the State and Institutions 131

Notes and References 137

the special symposium in World Development, 1994, no. 4; essays in

which nicely sums up the spirit of my argument. To develop this insight

Chang, H-J. (1995) 'Explaining "Flexible Rigidities" in East Asia' in T. Killick

This book is a revised version of my PhD dissertation at the Faculty of

Ashish Arora, Anuradha Basu, Pranab Bardhan, William Baumol,

Cambridge, England HA-JOON CHANG

1980s, partly prompted by the failure of the US and British

(often accused of creating rigidities in the labour market which obstruct

Another rnajor atternpt to find alternative patterns of econornic

just the result of macroeconomic mismanagement but of lagging

The industrial policy literature, in contrast with the social corporatism

ways, concemed about the development of the division of labour in the

STRUCTURE OF THE BOOK

system by influencing the state. The standard recommendation from the

1.1 EFFIClENCY: THEMARKET-FAILURELITERATURE

The most developed literature on state intervention is that of market fail-

1.1.1 Public Goods

In the economics literature, goods are usually classified into two

1.1.2 Non-Competitive Markets

The existence of scale economies and/or collusive behaviour can result

A powerful argument against state intervention to 'correct' non-

anti-trust intervention which does not eliminate 'distortions' from ali

One of the assumptions in textbook economics is that each individual

Sorne anti-interventionists try to dismiss the extemality argument for

the previous section we discussed the market-failure argument, the

1.2.1 Paternalism and Contractarianism

The above argument is often branded as paternalism by those who

rule. Otherwise, however fair the outcome of a politico-economic system

1.2.2 What should the State do?

On the basis of individualism-contractarianisrn, it is argued that any state

1.2.3 Sorne Reflections

Denouncing any moral judgment other than those based on narrowly

first time since Queen Elizabeth's reign were effectively supervised

Of course most contractarians are not so silly as to believe in the 'state

common assumption in economics, but nothing more than an assurnp-

1.3 INTENTION: THE POLITICAL ECONOMY LITERATURE The

political-economy viewpoints, both on the left (for example the

1.3.1 The Autonomous-State Approach

which interest groups feed their policy inputs, resulting in sorne

1.3.2 The Interest-Group Approach

Another group of arguments which question the intention of the state is

1.3.3 The Selí'-Seeking-Bureaucrats Approach

Another important critique of the 'benevolence' assumption in welfare

scribed as peculiar forms of self-interest - she tells the truth because

As the interest-group approach emphasises, groups in society engage in