Economic Development in Pacific Asia

Economic Development in Pacific Asia During the period 1965–1990 East Asia was the world’s fastest-growing region. Eco...

Author: Hafiz Akhand

44 downloads 1603 Views 703KB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Economic Development in Pacific Asia

During the period 1965–1990 East Asia was the world’s fastest-growing region. Economic Development in Pacific Asia provides illuminating, non-technical perspectives on key facets of the region’s economies. The text focuses on the eight countries which accounted for the majority of the economic growth: Japan, Hong Kong, Taiwan, Republic of Korea, Indonesia, Malaysia, Singapore and Thailand. Empirical evidence is used to provide a revealing, multidimensional statistical profile of the countries as well as the region as a whole. Rather than present a statistical history of each country, the text highlights the relative performance in terms of the variables which are studied within each chapter. Akhand and Gupta examine a range of popular topics including: the relative role of factor accumulation versus technology change; factor price distortions and employment; growth, poverty and income distribution; the Asian crisis; and corruption. In addition, topics rarely covered in the current economics literature are investigated, such as urbanization, the gender gap and the digital divide. This book provides an accessible and wide-ranging assessment of the existing evidence and current arguments on East Asian economic development. Economic Development in Pacific Asia highlights the lessons to be drawn and the policy implications they may lead to. This book will be a valuable addition to economists, policy makers and those interested in Asian economic affairs. Hafiz Akhand is Associate Professor of Economics at the University of Regina, Canada. His research interests are in the areas of foreign aid and the economics of growth. Kanhaya Gupta is Professor Emeritus of Economics at the University of Alberta, Canada. His research interests are in the areas of economic development, foreign aid and the economics of growth.

Economic Development in Pacific Asia

Hafiz Akhand and Kanhaya Gupta

First published 2006 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by Routledge 270 Madison Ave, New York, NY 10016 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.”

© 2006 Hafiz Akhand and Kanhaya Gupta All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data A catalog record for this book has been requested ISBN 0-203-09979-6 Master e-book ISBN

ISBN 0–415–28867–3 (hbk) ISBN 0–415–28868–1 (pbk)

To Cheryl and Latif Hafiz Akhand To the memory of my parents Kanhaya Gupta

Contents

List of figures List of tables Acknowledgements

ix xi xv

1 Introduction

1

2 Record of transformation

5

3 The East Asian Miracle: perspiration or inspiration?

22

Approaches to the estimation of total factor productivity 24 Productivity growth empirics: the tyranny of numbers 37 Productivity growth empirics: an assessment 55 Conclusion 57 4 Factor price distortions and economic activity

58

Factor price distortions, economic welfare and capital accumulation 58 Factor price distortions and growth 59 Estimates of factor price distortions 60 Conclusion 62 5 Growth, poverty and income distribution Growth and poverty reduction 63 Growth and income inequality 64 Poverty reduction, growth and redistributive policies 65 Conclusion 66

63

viii Contents 6 East Asian financial crisis

67

The nature and extent of the crisis 71 Assessing the empirical evidence 74 Conclusion 82 7 Corruption, growth and foreign direct investment

84

The nature and extent of corruption in East Asia 85 Determinants of corruption 88 Corruption and growth 93 Corruption and foreign direct investment 98 Conclusion 103 8 Urbanization

105

Indicators of urbanization 105 Conclusion 109 9 Gender gap

111

Mortality inequality 112 Natality inequality 114 Basic facility inequality 115 Professional inequality 115 Ownership inequality 118 Household inequality 119 10 Digital divide

122

11 Globalization and labour market outcomes

133

12 What we have learned and what remains to be done

138

Notes References Index

141 148 157

Figures

6.1 Exchange rates in crisis countries, January 1997–December 1999 6.2 Current account balance in crisis countries, 1990–2000 6.3 GDP growth rates in crisis countries, 1990–2000 6.4 Stock prices in crisis countries, January 1997–December 1999 6.5 Euro-money country risk ratings 7.1 Corruption in East Asian countries, 1984–2004 7.2 Corruption in the G-7 countries, 1990–2004

68 71 74 75 75 89 90

Tables

2.1 Average annual growth rates, 1965–1999 2.2 Structure of production: distribution of GDP, 1960 and 1999 2.3 The importance of trade and globalization, 1970 and 1999 2.4 Demographic transition 2.5 Trends in human development or some quality of life indicators 2.6 Population below the international poverty line 2.7 Commercial energy: use and efficiency 2.8 Selected indicators of human capital accumulation 2.9 Selected indicators of factor accumulation 2.10 Selected determinants of technological change 2.11 Selected indicators of domestic policies 3.1 Total factor productivity growth, 1966–1990/91 3.2 Sources of growth, 1960–1994 3.3 Total factor productivity growth, eight high-performing Asian economies, 1960–1989 3.4 Sources of growth in a model without human capital 3.5 Sources of growth in a model with human capital 3.6 Total factor productivity growth, 1960–1990 3.7 Productivity growth, 1960–1985 3.8 Growth accounting, 1978–1996 3.9 Total factor productivity growth 4.1 Distortions in the 1970s 5.1 Population below the international poverty line 5.2 Inequality in East Asia

6 8 9 11 12 13 15 16 17 19 20 39 41 43 44 46 48 50 52 54 60 64 64

xii Tables 5.3 Decomposition of poverty reduction into growth and redistribution components 6.1 Key economic indicators in East Asian economies 6.2 Short-term debt and reserves in crisis countries, 1994 and 1997 6.3 Real exchange rate, 1991–1997 6.4 External financing of the five crisis countries, 1994–1998 6.5 Annual output growth in East Asian countries, 1970–1998 6.6 Fundamental imbalances and Asian crisis (crisis and non-crisis countries) 6.7 Fundamental imbalances and Asian crisis (crisis countries) 6.8 New Probit regression results predicting the onset of financial crisis 7.1 A comparison of various indicators of corruption in East Asia 7.2 Corruption in East Asia, 1984–2004 7.3 Corruption in G-7 countries, 1984–2004 7.4 Determinants of perceived corruption 7.5 Corruption and growth, 1960–1992 7.6 Corruption and growth revisited, 1984–2003 7.7 Effects of corruption on foreign direct investment 7.8 Corruption and foreign direct investment revisited, 1984–2003 8.1 Indicators of urbanization 8.2 Traffic and congestion 8.3 Urban environment (a) housing and transport and (b) services 9.1 Male–female income 9.2 Gender inequality in economic activity 9.3 Mortality inequality 9.4 Natality inequality 9.5 Basic facility inequality: gender inequality in education 9.6 Professional inequality: gender empowerment measure 9.7 Ownership inequality: women’s political participation 9.8 Household inequality: gender work burden

66 69 70 71 73 74 77 78 79 87 87 88 93 94 97 101 103 106 107 108 112 113 114 114 116 117 118 120

Tables xiii 9.9 10.1 10.2 10.3 10.4 10.5 10.6 10.7 11.1 11.2

Venture capital spread across the world The information age (old technology) The information age (new technology) Cross-country correlations: explanatory and explained variables Cross-country correlations: explanatory variables OLS estimates SUR estimates Comparing the R 2 values Indicators of integration with the global economy Global financial flows

121 124 125 127 128 129 130 131 134 135

Acknowledgements

Special thanks to Cheryl Millar for carefully reading the manuscript and for her advice. Thanks as well to Charlene Hill for her help in preparing the manuscript and Ryan Wood for research assistance. In addition, Hafiz Akhand would like to thank Kanhaya Gupta for being the best co-author with whom one could hope to work.

1

Introduction

The so-called ‘East Asian Miracle’, the rapid growth of the Chinese economy over the past few decades and the recent spurt in the growth rate of the Indian economy, particularly in its information technology (IT) sector, have drawn considerable attention in the research agendas of academic scholars as well as other researchers and international organizations like the International Monetary Fund (IMF) and the World Bank. A good deal of the research work is widely scattered and, as one might expect, tends to be quite technical and at the same time cover a wide variety of issues. It is our strong belief that there is a wide constituency of readers, both economists, including students, and policy makers and others interested in Asian economic affairs, who would be interested in a rigorous though not necessarily technical, accessible and wide-ranging assessment of the existing evidence and arguments on some of the major issues under one roof. If any such undertaking is to be manageable, one must make choices about countries and the topics to be covered. As for the countries, our choice is relatively straightforward. The East Asian countries have demonstrated a robust economic performance over an extended period of time. For example, considering the period 1965–1990, the twenty-three East Asian countries grew faster than any other region of the world. But the most remarkable part of even this remarkable performance is the fact that only eight of the East Asian countries accounted for most of this growth. The eight countries are: Japan, the ‘Four Little Tigers’ – Hong Kong, the Republic of Korea (from here on, also South Korea), Singapore and Taiwan, China and the three newly industrialized countries (NICs) – Indonesia, Malaysia and Thailand. And although the Philippines is sometimes included for purposes of analysis, it lagged

2 Introduction behind considerably. Consequently, these eight or nine (including the Philippines) countries are the focus of this study. As for the topics to be covered, the decision was more complicated. Our choice, as reflected in the chapter scheme, essentially reflects three considerations. First, they reflect our preferences, developed over a significant period of teaching and research. Second, many of the topics are routinely included in courses in development economics, and those which are not, are beginning to be; for example, gender gap, digital divide and globalization. And, finally, we have tried to include topics which in our view, though not fashionable, are none the less important and are likely to be even more so in future – for example, urbanization and corruption, to name just a few. A brief outline of the rest of the chapters now follows, which should give a better idea about the contents of the book. In Chapter 2, using existing data sets, we try to provide a brief, but multidimensional, statistical profile of the countries of our sample as well as for the region. No attempt has been made to provide a statistical history of the eight or nine countries. Rather, the intention is to highlight relative performance in terms of the variables which are analysed in the rest of the chapters. Chapter 3 takes up one of the most contentious but interesting topics in the entire book. This has to do with the relative roles of factor accumulation versus technological change. The debate has been particularly furious about the ‘Four Little Tigers’ already identified; namely, Hong Kong, South Korea, Singapore and Taiwan, China. This chapter provides a detailed survey of the literature, both regarding methodological issues as well as the empirical evidence and attempts are made to draw appropriate lessons. Chapter 4 takes up the question regarding the role of factor price distortions. There has been much debate, no doubt inspired by neoclassical economics, that there exists significant factor price distortions in the developing countries which have serious consequences for resource allocation in general, but particularly for employment. Since these distortions are generally caused by public policies, the distortions have policy implications. After providing a brief and non-technical analytical background, the chapter tries to summarize the existing evidence on the extent of factor price distortions. An attempt is also made to summarize arguments and evidence on the causes of these distortions. Following this, a summary and assessment of the effects of these distortions are provided.

Introduction

3

Chapter 5 takes up the theme ‘Growth, Poverty and Income Distribution’. This topic has attracted particular attention because many of the countries of our sample have been able to achieve rapid growth while at the same time achieving reductions in income inequalities and poverty. This chapter provides a survey of this evidence and also examines possible danger signals for the future. Chapter 6 deals with an issue which continues to attract attention far beyond the East Asian region. This chapter tries to provide a brief but solid summary of the causes, implications and the reforms to deal with what is now popularly know as the ‘Asian crisis’. Few issues have caught as much attention or aroused as much passion as corruption. This topic is taken in Chapter 7. The problem with this topic is that it is too broad, it has too broad a scope – social, political and economic. Since there already is a vast literature on corruption, we very briefly summarize a rather specific section of this literature devoted to identifying the causes of corruption, and the effects of corruption on growth and foreign direct investment (FDI). In addition, we provide some of our own estimates of the effects of corruption on growth and FDI. Chapter 8 on ‘Urbanization’ tries to draw attention to an area which is rarely, if ever, included in books on economic development. This topic is generally included in geography, urban planning and demography. But our goal is rather limited. We aim to analyse the data on urbanization given in the World Development Indicators and provide a preliminary analysis of this information insofar as it bears on the issue of sustainability. Chapter 9 on the ‘Gender Gap’ is certainly an unorthodox one for a book on economic development, but not on Asia or East Asia. But our goal is quite limited, partly, as we will see, due to limitations of the data. However, this chapter offers some interesting ideas as to why we continue to have the ‘gender gap’ in these countries and what might be done to bridge the gap. Chapter 10 on the ‘Digital Divide’ may be viewed in the same spirit as the previous chapter on the gender gap. In spite of the explosion of interest in the information technology industry, this topic does not get included in books on economic development. We offer a brief survey of this topic. Globalization is a catchword these days and means so many things that it almost ends up meaning nothing. We try to define it very narrowly for

4 Introduction our purpose and then summarize its implications for the labour market in Chapter 11. In a book of this kind, there can be no standard type of conclusions. This is a book on ‘some major themes’ and no doubt one can argue that many, if not all, of the themes covered are interrelated, but that is not the aim of this book. Of course, we can argue that many of the topics are interrelated, but what we have done is to treat them individually, so readers can get a handle on separate issues and then draw their own conclusions. For this reason, Chapter 12 is not only brief, but also reflects our view of what the literature says, what lessons we can possibly draw, and what remains to be done.

2

Record of transformation

The basic aim of this chapter is to highlight some of the major changes that have taken place in the countries of our sample, which often earn the appellation of the ‘East Asia Miracle’. It should be noted that our aim is not to provide a detailed account of the economic history of the East Asian economies. The major rationale for including the statistical material that we have in this chapter is that these materials reflect our choice of the topics covered in the subsequent chapters. We provide brief descriptions of the tables included in this chapter along with their rationale and how they are connected to the chapters that follow. We start with Table 2.1 which details the aggregative aspects of growth in the region and their comparison with the rest of the world. The table reflects the aggregative performance as well as the sectoral performance. The improvement in individual welfare can also be read if we assume that improvements in household consumption expenditure reflect improvements in individual welfare. Given this interpretation and the rest of the data in this table, we can see that the countries of East Asia and the Pacific outperformed countries in every other group over the period considered. This is so whether judged in terms of the growth rate of the gross domestic product (GDP) or per capita GDP growth, or the growth rates of value added in the three major sectors or household consumption expenditure. Of course, there are variations in the individual countries included in East Asia and the Pacific region. A simple way to appreciate the dimensions of their outstanding performance is to look at their per capita GDP growth rate. Of the seven countries included here, the lowest estimate for the period 1965–1999 is for Japan, and even in that case the estimate is 3.4 per cent, a remarkable figure indeed. The average for the region was 5.6 per cent as against

0.1 2.4 −0.2

3.0 4.7 2.6 3.3

World

Source: Adapted from World Development Indicators on CD-ROM, 2001.

1.6

5.4 4.8 3.4 4.3 6.3 6.6 5.1 5.6 — 1.4

7.3 6.9 4.1 7.0 8.3 8.1 7.3 7.4 — 3.5

2.2

2.9 1.9

4.2

— 3.8 −0.2 2.9 −1.5 2.1 3.9 3.6 — 2.7

Agriculture

Total

Per capita

Value added

Gross domestic product

Hong Kong Indonesia Japan Malaysia Singapore South Korea Thailand East Asia and Pacific Europe and Central Asia Latin America and Caribbean Middle East and North Africa South Asia Sub-Saharan Africa

Country/area

Table 2.1 Average annual growth rates, 1965–1999 (%)

—

5.5 2.4

1.3

— 8.7 4.3 8.1 8.4 10.8 9.4 9.6 — 3.3

Industry

—

5.6 3.1

4.0

— 7.5 4.6 7.7 8.4 7.6 7.2 7.8 — 3.8

Services

3.4

4.4 2.8

—

7.4 7.0 4.0 5.9 6.6 7.8 6.1 6.7 — 3.5

Household final consumption expenditure

Record of transformation 7 a mere 1.6 per cent for the world as a whole. We can appreciate their performance further, if we consider the growth rates of three sectors. An interesting comparison is between South Asia and East Asia and the Pacific. Finally, the last column is very revealing. The growth rate of household final consumption expenditure was 6.7 per cent in East Asia and the Pacific while it was only 3.4 per cent for the world as a whole. Quite apart from the fact that the growth rate of East Asia and the Pacific was almost twice that of the world, what is even more striking is the absolute magnitude of the growth rate. It meant that while it took nearly twenty-one years for per capita consumption expenditure to double for the world, for East Asia and the Pacific it took only about eleven years, roughly a decade, or just about half the time. The evidence from Table 2.2 immediately points to the importance of the structure of production over the period covered. In this table we consider the evidence across the countries of the sample as well as over time. An outstanding feature of the table is the decline in the share of agriculture for each of the countries included in the table. By 1999, with the exception of Indonesia, the share of agriculture had declined to about 10 per cent or lower for all but Indonesia. The dominance of the share of the service sector is obvious for each of the countries. The extraordinary role played by this sector in Hong Kong, Japan, South Korea and Singapore is obvious, although its importance cannot be overemphasized for the remaining countries also. But the main point that emerges from this table is that the rapid growth rates detailed in Table 2.1 were accompanied by significant changes in the structure of production. Along with the rapid changes in the structure of production shown in Table 2.2, it has also been suggested that the growth of these economies was accompanied by or caused by rapid changes in the export performance of these economies. Put differently, it is often asserted that their growth was in no small measure due to their openness to trade and foreign investment. Table 2.3 shows the importance of trade. Columns 2 and 3 give the shares of trade in 1970 and 1999 while the next two columns give the shares of manufactures as a percentage of total exports for 1980 and 1998. By 1998 we can easily see the important role played both by trade as well as manufactures in total exports. Without saying anything about causality here, it is worthwhile asking whether the dynamic role often ascribed to the external sector had more to do with the manufacturing sector or with the industrial sector broadly defined. Regardless, what is clear is that for the countries under consideration, trade played a significant role.

4 54 13 37 4 40 41

0 19 2 11 0 5 10

34 14 42 18 18 19 18

1960

1960

1999

Industry

Agriculture

15 43 36 46 36 44 40

1999 25 8 33 9 12 12 11

1960

Manufacturing

Source: Adapted from World Development Indicators, 1979, and World Development Indicators on CD-ROM, 2001.

Hong Kong Indonesia Japan Malaysia Singapore South Korea Thailand

Country

Table 2.2 Structure of production: distribution of GDP, 1960 and 1999 (%)

6 25 24 32 26 32 32

1999

62 32 45 45 78 41 41

1960

Services

85 37 62 43 64 51 50

1999

28

181 28 20 79 232 37 34 24 — 20 — 12 47 52

261 62 19 218 — 77 102 70 77 34 57 30 60 66

96 — 95 19 47 90 25 45 — 20 6 54 12

80

95 45 94 79 86 91 71 82 57 49 17 78 —

1998

1980

1970

1999

Manufactures (as % of total exports)

Trade (as % of GDP)


World

Hong Kong Indonesia Japan Malaysia Singapore South Korea Thailand East Asia and Pacific Europe and Central Asia Latin America and Caribbean Middle East and North Africa South Asia Sub-Saharan Africa

Country/area

Table 2.3 The importance of trade and globalization, 1970 and 1999

10


Along with the transformation in economic sectors noted above, these economies also exhibited remarkable transformation in terms of demographic features. Some of these demographic factors are shown in Table 2.4. We concentrate on four variables: crude birth rate, crude death rate, total fertility rate and annual population growth rate. A comparison with the developed countries is revealing. Thus, with the exception of Japan, the crude birth rate in 1960 was almost twice that of the developed countries, but it had moved close to that of the developed countries by the end of 1994. Similar movement can be discerned in the case of the death rate. Of course, the total fertility rate reflects the movement of the two. By 1994 the total fertility rate in many of the countries in the sample had declined to even lower than that of the developed countries. Finally, if we consider the annual population growth rate over the period 1960–1994, it had declined to even lower than that of the industrial countries in Japan and South Korea, while it was close to 2 per cent for all except Malaysia. This significant decline in population growth is reflected in the sharp increase in the growth of per capita GDP noted in Table 2.1 earlier. This also indicates how economic and demographic factors must have interacted in these countries. Reference was made to the increase in welfare in Table 2.1. That was done in terms of the increase in the rate of household consumption expenditure. While that indicator is useful, we might also look at some other indicators which try to answer the same question. For this purpose, we consider a number of indicators of the quality of life. Needless to say, any selection of indicators for this purpose must remain arbitrary and the indicators considered in Table 2.5 are no exception. These indicators generally lie behind the Human Development Index (HDI) of the United Nations Development Programme (UNDP). There are many ways to read this table. For example, we can make an intercountry intertemporal comparison, or we can compare the countries of our sample with those of the other developing countries or the industrial countries. Thus, the comparison we make depends on the purpose at hand. For our case, we consider the performance of the countries in our sample with those of the industrial countries. Considering one indicator at a time, we can see that by 1994, life expectancy at birth had exceeded in Hong Kong and Singapore, in addition to Japan which had already surpassed the industrial countries earlier. The other countries had closed the gap significantly as well. Equally significant improvements can be noted in the case of infant mortality rates as well as adult literacy rates. Perhaps a comparison of the real GDP in per capita terms (at purchasing

35 47 18 39 38 41 46 20

11.8 23.5 15.0c 27.8 18.5 15.3 16.6 13.1

7 23 8c 9 8 13 17 10

5.8 8.0 6.0 5.0 4.8 6.3 6.1 10.0

1994b

1960a

1960a

1994b

Crude death rate

Crude birth rate

2.6 4.9 1.8 3.8 2.2 2.8 4.5

1977a

Notes a From World Development Indicators, 1979. b From Human Development Report, 1997. c Refers to 1977 and is from World Development Indicators, 1979.

1.3 2.8 1.8 3.5 1.8 1.6 1.8 1.7

1994b

Total fertility rate

Source: Adapted from World Development Indicators, 1979, and Human Development Report, 1997.

Hong Kong Indonesia Japan Malaysia Singapore South Korea Thailand Industrialized countries

Country/group

Table 2.4 Demographic transition

2.0 2.1 0.8 2.6 2.1 1.7 2.3 0.8

1960–94b

Annual population growth rate

66.2 41.2 71.6a 53.9 64.5 53.9 52.3 68.6 46.0

79.0 63.5 79.8 71.2 77.1 71.5 69.5 73.8 62.1

43.0 139.0 15.2a 72.0 36.0 85.0 103.0 39.0 149.0

5.0 53.0 5.5b 12.0 5.0 10.0 29.0 14.0 64.0

1994

1960

1960

1994

Infant mortality rate (per 000 live births)

Life expectancy at birth (years)

Notes a 1968 from World Tables, 1989–1990. b 1987 from World Tables, 1989–1990.

Source: Adapted from Human Development Report, 1997, and World Tables, 1989–1990.

Hong Kong Indonesia Japan Malaysia Singapore South Korea Thailand Industrialized countries All developing countries

Country/group

Table 2.5 Trends in human development or some quality of life indicators

— 83 83 — 98 94 — 64

60 — 88 79 — 43

1994

— 54

1970

Adult literacy rate (%)

915

1,783 2,409 690 985

2,323 490

1960

22,310 3,740 34,630 8,865 20,987 10,656 7,104 17,221 2,923

1994

Real GDP per capita (PPP $)

Record of transformation 13 Table 2.6 Population below the international poverty line (population below $2 a day) Country

Latest year

Population below poverty line (%)

Indonesia Malaysia Singapore South Korea Thailand

1999

55.3 n.a. n.a. >2.0 28.2

1993 1998

Source: Adapted from table 2.6, World Development Indicators on CD-ROM, 2001.

power parity dollar, PPP $) is most revealing. By 1994, it had exceeded in Hong Kong, Japan and Singapore and had risen significantly in the remaining countries. If we compare the figure for the developing countries in general, we find that the performance of these countries can only be described as being astounding. Turning to Table 2.6, we can examine another feature of the ‘welfare’ experienced by the populations of these countries. One might reasonably ask, what good is it to have a high growth rate if most of the population is mired in poverty and/or its income distribution is very lopsided? This table tries to shed some light on these kinds of questions. The table provides data on the percentage of the population below the poverty line, poverty being measured by $2 per day. In the three countries that we have the data for in Table 2.6, we can see that the disparity is very substantial, being less than 2 per cent for South Korea, to almost 28 per cent for Thailand and then almost twice that of Thailand for Indonesia. Of course one might argue that $2 per day is too high a standard to adopt, but by the same token, we may argue that it is too low as well. What is obvious, however, is that there is still widespread poverty in the East Asian economies as well as inter-country disparity. Next we consider their performance in terms of some of the inputs. Starting with Table 2.7, we first consider the use and the efficiency of commercial energy. An interesting way to look at these numbers will be to compare them over time with the performance of the ‘high’ income countries. But first, we can see that there was considerable variation in the performance of the different countries in East Asia. That is not surprising given the fact that the growth performance of these countries was not even either. A more revealing figure is the one given in column 4. For the period 1980–1998, the average annual growth rate

14


of commercial energy consumption per capita was 1 per cent for the ‘high’ income countries whereas it was 3 per cent for the East Asia and the Pacific countries. The performance was even more striking if we consider the individual countries, except Indonesia. To get some idea about the efficiency with which commercial energy was used, we can look at the last two columns. For the high-income countries, the figure was 4.6. It was equalled or exceeded by Hong Kong, Indonesia, Japan and Thailand. From the efficient use of commercial energy considerations in Table 2.7, we turn our attention to human capital accumulation in Table 2.8. In this table, we consider three components of the gross enrolment ratio, namely, primary, secondary and tertiary, each expressed as a percentage of the relevant age group. In this case again, we look at intertemporal as well as inter-country comparisons. Without entering into discussion about which is more important, we simply note that by 1997, the latest year for which we have the relevant data, the primary gross enrolment ratio for each of the countries in Table 2.11 had reached close to that of the high-income countries. For the secondary and tertiary gross enrolment ratios, the story is not quite as remarkable. With the exception of Japan and South Korea, the gross enrolment ratio for secondary school is nowhere near that of the high-income countries. In case of the tertiary ratio, only South Korea surpasses the high-income countries. While most of the countries in the sample have made significant strides, much remains to be done by these countries. Table 2.9 tells the story about their performance in terms of savings and investment. Without going into the controversial aspects of this topic, here we only consider the story as it emerges from the data in this table. Consider then the performance of these countries as far as their performance regarding gross domestic savings was concerned. The first point to observe is that Japan had reached the ‘saturation’ point in 1960 and could hardly maintain the level reached in 1960. A similar story is told about South Korea after 1989. But when we consider Malaysia and Singapore we find that such ‘saturation’ cannot be inherent in the nature of the economy concerned. In any case, what emerges from this table is that these countries were able to generate very high rates of domestic savings and, as we shall see in Chapter 3, this is one of the major sources of debate among those who consider their performance a ‘miracle’ and those think that it was the inevitable outcome of just ‘sweating it out’. But here we only note that, over the period 1960–1999, many of these countries were able to increase their domestic savings rates to

3,745

424 91 1,474 312 670 237 80

1965a 1,079 402 2,967 888 2,656 1,082 488 571 4,796

1980

Kg of oil equivalent

2,479 604 4,035 1,967 7,681 3,519 1,153 857 5,366

1998

Commercial energy consumption per capita

6.4 2.2 3.3 2.7 2.3 2.8 3.0 — 2.2

1980–98 4.5 2.9 2.3 5.1 7.8 8.3 6.4 3.0 1.0

1980

Average annual growth rate (%)

Notes a The data for 1965 are from World Development Report, 1986. b Refers to ‘industrial market economies’.

8.5 4.6 6.0 3.9 3.1 4.0 5.1 — 4.6

1998

GDP per unit of energy use (PPP $ per kg of oil equivalent)

Source: Adapted from World Development Report, 1986, and World Development Indicators on CD-ROM, 2001.

Hong Kong Indonesia Japan Malaysia Singapore South Korea Thailand East Asia and Pacific High incomeb

Country

Table 2.7 Commercial energy: use and efficiency

87 67 103 96 111 94 83 114

94 113 101 101 94 94 89 102

24 6 74 19 32 27 12 58

1960

1960

1997

Secondary

Primary

73 56 103 64 74 102 59 106

1997

Gross enrolment ratio (as % of relevant age group)

10 4 31 4 8 15 15 35

1980

Tertiary

Note a Refers to ‘industrial market economies’.

Source: Adapted from World Development Report, 1986, and World Development Indicators, 1979 and 2000.

Hong Kong Indonesia Japan Malaysia Singapore South Korea Thailand High incomea

Country

Table 2.8 Selected indicators of human capital accumulation

28 11 43 11 39 68 21 59

1997

1 8 34 27 3 25b 17

36 32 33 34 44 37 34

31 32 28 47 52 34 33

19 8 34 14 11 11 16

27 31 32 32 37 38 41

1990 25 24 26 22 33 27 21

1999

1960a

1999

1960a 1990

Gross domestic investment (as % of GDP)

Gross domestic saving (as % of GDP)

56 56 64 51 55 54 53

1960a

74 64 69 62 51 70 68

1999

% of working age population (15–64 years)

Notes a The data for 1960/1968 are from the World Development Indicators, 1979. b Refers to 1968.


Hong Kong Indonesia Japan Malaysia Singapore South Korea Thailand

Country

Table 2.9 Selected indicators of factor accumulation

18


rather extraordinary levels and that raises a very important question, how? Gross domestic investment as a percentage of GDP is shown in columns 4 to 6. We could carry out a detailed analysis but that would not serve any purpose for now, so we merely note that the behaviour of investment simply mirrors that of gross domestic savings. The role of labour force growth is reflected in the next column which gives the percentage of population of working age. With the exception of Singapore, the ratio increased for each of the other countries. This increased percentage of population of working age provided another source of potential growth. Technological change is often asserted as one of the sources of economic growth, so that it is useful to look at how these countries performed in terms of some of the determinants of technological change. We consider four determinants: personal computers per 1,000 people; scientists and engineers in research and development (R&D) per million people; R&D expenditure as a percentage of gross national income (GNI); and patent applications filed in 1997. Table 2.10 reports the data on these determinants. For the first three indicators, we compare the countries of our sample with the United States. In 1998, in terms of computers per 1,000 people, none of the countries came close to the United States, but the variation among the countries was quite high. Thus, while Indonesia had only 8.2 computers per 1,000 people, the corresponding number for Hong Kong, Japan and Singapore was 254.2, 237.2 and 458.4, respectively. Even for South Korea it was about 157. For the next determinant, scientists and engineers in R&D per million people, only Japan, South Korea and Singapore stand out, although Japan exceeded the United States. One of the more important determinants of technological change is considered to be the expenditure on R&D as a percentage of GNI. Over the period 1987–1997, both Japan and South Korea exceeded the United States. And Singapore was a respectable 1.13. But the other countries were further behind. It is interesting to note that the average for the high-income countries was 2.36 while for the East Asia and Pacific countries it was 1.32, just a little over half of that. In other words, these countries have a long way to go to catch up with the high-income countries. Finally, we consider the patent applications filed to 1997. Drawing the distinction between those filed by residents and non-residents, we can see that except for Japan and South Korea, this activity is dominated by the non-residents, indicating lack of substantial domestic research activity. When compared with the United States, we find that the United States almost overwhelms the

— 182 4,909 93 2,318 2,193 103 3,676 492 3,166

1987–97b

1998b 254.2 8.2 237.2 58.6 458.4 156.8 21.6 458.6 14.1 311.2

Scientists and engineers in R&D (per million people)

Personal computers (per 000 people)

— 0.07 2.80 0.24 1.13 2.82 0.13 2.63 1.32 2.36

1987–97a

R&D expenditure (as % of GNI)

Notes a World Development Indicators, 2001. b World Development Report, 2000/2001, table 19. c Refers to ‘industrial market economies’.

Source: Adapted from World Development Indicators, 2001, and World Development Report, 2000/2001, table 19.

Hong Kong Indonesia Japan Malaysia Singapore South Korea Thailand United States East Asia and Pacific High incomec

Country/region

Table 2.10 Selected determinants of technological change

26 — 351,487 179 8,188 92,798 238 125,808 106,342 648,093

Residents

2,359 4,517 66,487 6,272 29,467 37,184 5,205 110,884 184,288 2,137,327

Non-residents

Patent applications in 1997

— 0.3 1.3 1.1 1.0 0.2 0.4 2.2

1960

— 0.7 5.9 1.4 1.2 2.3 1.9 6.0

1998

— 0.7 5.9 1.4 1.2 2.3 1.9 6.0

1990–98

Public expenditure on health (as % of GDP)a

Notes a Public expenditure on health in 1960 is given as a percentage of GNP. b The observation is for 1980.

Source: Adapted from various issues of Human Development Report.

Hong Kong Indonesia Japan Malaysia Singapore South Korea Thailand Industrialized countries

Country/region

Table 2.11 Selected indicators of domestic policies

2.9 1.4 3.6 4.9 3.0 3.7 4.8 5.4

2.4b 2.5 4.0 2.9 2.8 2.0 2.3 3.7

— 5.8 1.0 1.9 0.4 6.0 2.6 6.3

1960

1997

1960

— 2.3 1.0 2.2 5.7 3.4 2.3 2.4

1997

Military expenditure (as % of GNP)

Public education expenditure (as % of GNP)

Record of transformation 21 other countries. Furthermore, patent applications filed by residents and non-residents appear to be equally important in the case of the United States. It has been often claimed that the rapid growth of the East Asian economies was in no small measure due to the domestic policies adopted by these countries. While this is not the place to discuss that topic, it is useful to consider some statistical dimensions of some of the policies adopted in these countries. Table 2.11 provides information on three indicators of domestic policies for different periods of time. We consider three types of expenditure as indicative of national priorities and therefore significant national policies. These are: public expenditure on health, education and the military. In terms of public expenditure on health, in 1960, compared to the industrial countries, where it was 2.2 per cent, it varied from 0.2 per cent for South Korea to 1.3 per cent for Japan. But by 1998 it had risen to 5.9 per cent for Japan and 2.3 per cent for South Korea. But for the other countries the figure was very low. It should be noticed that the figure had risen to 6 per cent for the United States. With respect to public education expenditure, with the exception of Indonesia, the percentage of GNP devoted to education increased for the other countries. By 1997, while none of the countries had caught up with the high-income countries, Thailand and Malaysia had made great strides, reaching almost 5 per cent. Finally we look at military expenditure. At least a priori, we can argue that military expenditure, being unproductive, is not conducive to increasing growth. Thus, comparing with the industrial countries, which spent 2.4 per cent of their GNP on the military, except Japan, Singapore and South Korea far exceeded this number, and the other three countries spent almost as much in terms of the percentage of GNP. We thus find that the countries pursued policies which were both conducive to and also detrimental to growth.

3

The East Asian Miracle Perspiration or inspiration?

The previous chapter documents an impressive record of high and sustained economic growth in East Asia.1 Between 1965 and 1999, real gross domestic product (GDP) per capita in these East Asian economies has grown (on average) at an annual rate of 3.4 per cent or higher. This is more than double the world growth rate of real GDP per capita (see Table 2.1). The East Asian economies also outperformed the industrial economies during the post-war period. Consider, for example, the case of Singapore. Between 1966 and 1990 the economy of Singapore grew at an impressive rate of 8.5 per cent per annum. This is three times the rate at which the United States economy grew during the same period. During 1966–1990 per capita income in Singapore grew at the rate of 6.6 per cent per annum, which roughly amounts to a doubling of per capita income in Singapore every ten years. An achievement such as this is considered by many to be an economic miracle. The extraordinary performance of the East Asian economies has led to a large and varied literature devoted to uncovering the factors underlying such sustained and impressive growth. Two opposing views on the sources of the East Asian Miracle emerge from this literature, namely the accumulation view and the assimilation view. According to the accumulation view, the level and growth of productivity are roughly the same across countries. Consequently, cross-country differences in output levels and growth rates are due primarily to differences in physical and human capital. Proponents of the accumulation view therefore argue that the basis for East Asian growth is the maintenance of high rates of physical and human capital accumulation over a number of decades.2

The East Asian Miracle 23 However, not everyone agrees with the accumulation view. Romer (1993), a proponent of the assimilation view, for example, argues that ‘idea gaps’ are much more important than ‘object gaps’ in explaining cross-country differences in output levels and growth rates.3 According to the assimilation view, the East Asian success resulted from East Asian economies’ ability to adopt technologies, developed in more advanced economies, at a low cost. Unlike the accumulation view, the assimilation view therefore emphasizes the role of assimilation of new technology. The East Asian economies’ exceptional ability to master foreign technology makes their growth experience unique. The dispute regarding sources of growth in East Asia can be conveniently characterized in terms of an aggregate production function Y = F (A, X) where Y is output, A is total factor productivity (TFP) and X is a composite factor input encompassing physical and human capital. In terms of the production function, the disagreement is over the relative importance of A and X in explaining the East Asian performance record. The assimilation view (from here on, A view) assigns a greater role to A. By contrast, the accumulation view (from here on, X view) insists that X is more important. Establishing which view is correct is an essential step towards extracting lessons from the East Asian growth experience because the positive and normative implications of the A view can be dramatically different from those of the X view. In this chapter we re-examine the debate on the relative importance of productivity growth versus physical and human capital accumulation in explaining the exceptionally high rate of output growth in East Asia. The chapter begins with a brief review of the methodological issues, namely, the notion of TFP, techniques used to estimate TFP, and potential problems with the estimation methods most widely used in the literature. The chapter also presents a wide range of TFP growth estimates for the eight East Asian economies. A large number of empirical studies reviewed in this chapter find little or no evidence of rapid TFP growth in the East Asian economies.4 According to these studies, the extraordinary rate of output growth in the East Asian economies is due primarily to an equally impressive rate of factor accumulation. These studies therefore provide strong evidence in favour of the accumulation view. Findings such as these have led many economists to conclude that the industrial revolution in East Asia can be explained largely in terms of transition dynamics in

24

The East Asian Miracle

a neoclassical growth model.5 This conclusion suggests that improving living standards requires investment in physical and human capital, paid for in large part through foregone current consumption. By contrast, other empirical studies reviewed in this chapter suggest that TFP growth accounts for most of the output growth in the East Asian economies.6 These findings – consistent with the assimilation view – have led many to conclude that rapid economic growth in East Asia can be explained largely by these countries’ ability to successfully catch up with technology.7 This view suggests that a poor country aspiring to emulate industrialized nations need not incur as great an opportunity cost. Findings such as these therefore imply a much more optimistic growth prospect for less developed economies. The discrepancy in TFP growth estimates reported in the empirical literature is the root cause of disagreement among economists about the relative importance of factor accumulation and productivity growth in explaining the extraordinary rate of output growth in East Asia. Consequently, no consensus has been reached on the lessons to be learned from the East Asian Miracle. As we will see in this chapter, this discrepancy in TFP growth estimates arises primarily due to the differing estimation methodologies used in the empirical studies. This chapter will therefore provide a critical evaluation of the empirical literature in light of rationales underlying the methodologies used in estimating TFP growth. Concluding remarks follow.

Approaches to the estimation of total factor productivity One of the stylized facts that has emerged from the accumulation of empirical work on TFP growth at the economy-wide level has been that roughly one-third to one-half of output growth can be attributed to TFP growth – often described as the rate of technological progress. It is, however, well understood that TFP growth has to be interpreted broadly to include changes in health and education levels, allocation and x-efficiency, as well as factors affecting the motivation of workers. As Abramovitz (1956) put it more succinctly, TFP growth is really a ‘measure of our ignorance’, precisely because it is a residual.8 This ignorance covers many components, some wanted (such as the effects of technical and organizational innovation), others unwanted (such as measurement error, omitted variables, aggregation bias and model misspecification).

The East Asian Miracle 25 Nadiri (1970) argues that if factor inputs are measured correctly and the function governing their interactions is specified correctly then the residual TFP growth should be zero. Consequently, the main focus of the empirical effort has been towards better measurement of factor inputs and more precise estimation of the production function itself. In this section we provide a brief review of these efforts. The main emphasis, however, is on various methods of estimating TFP growth commonly used in the productivity growth empirics literature.9 Total factor productivity The standard model of economic growth seeks to explain the long-term trend in the potential output of an economy by dividing it into two components: (1) the component which can be explained by the growth in factor inputs used in production; and (2) the component which can be explained by improvements in the efficiency with which inputs are used in the production process. The second component is called TFP growth.10 TFP extends the concept of single factor productivity, such as labour productivity (that is, output per unit of labour), to more than one factor. In other words, the idea is to measure productivity taking into account all factors of production employed in the production process. Thus, if there is a single aggregate output Q and more than one factor input to be taken into account, then the measure of TFP (denoted A), in its simplest form, is given by: Qt i=1 ωit Xit

At =

(3.1)

where Xi s are factor inputs used in producing output, ωi s are appropriate weights and the subscript t is time period. It follows from equation (3.1) that TFP is merely a ratio of output to a weighted sum of inputs, i=1 ωi Xi . Simply put, TFP is output per joint unit of all the factor inputs used. Index number approach There are two ways to determine the weights in equation (3.1) above: arithmetic weighting due to Abramovitz (1956) and Kendrick (1961), and geometric weighting due to Solow (1957) and Jorgenson and

26


Griliches (1967). Both weighting methods are deeply rooted in the literature and lead to similar measures of TFP growth. The geometric indices, however, are more widely used in the literature. We discuss the two approaches in turn. The arithmetic weighting approach is based on the linear homogeneous production function:

Qt = At

ρ δi Xit

i=1

−1/ρ

Xit

(3.2)

i=1

Equation (3.2) reflects the assumption that parameters of the production technology, δi and ρ, across time or firms or countries are the same except the efficiency parameter A. The TFP measure given by equation (3.1) combined with equation (3.2) implies that: At =

Qt ρ −1/ρ i=1 δi Xit i=1 Xit

(3.3)

The productivity growth that occurs over two points in time, say t − 1 and t, or the efficiency differences between two firms or countries can be represented by the fractional change in A or as the ratio of the TFP measures for the two observations minus unity. Thus, using t − 1 as the reference time period (or firm or country) and denoting the price of input i to be wi , TFP growth according to the arithmetic weighting approach is: Qt /Qt−1 dAt −1 = At−1 i=1 wi Xit / i=1 wi Xit−1

(3.4)

That is, the TFP growth of observation t is the TFP of observation t relative to observation t − 1. By contrast, the geometric weighting approach is based on the Cobb– Douglas production function: Qt = At

Xitαi

(3.5)

i=1

The technology parameter αi is the elasticity of output with respect to input Xi .11 The production function, given by equation (3.5), is assumed to exhibit constant returns to scale. That is, the technology parameters,

The East Asian Miracle 27 αi s, sum to unity. The definition of TFP given by equation (3.1) combined with equation (3.5) leads to the TFP measure: Qt αi i=1 Xit

At =

(3.6)

Here At is referred to as exogenous, disembodied and Hicks-neutral technical progress, and is measured as a shift in an aggregate production function. Thus, the notion of TFP can be reinterpreted as an index of all those factors other than Xi s which are not explicitly accounted for but which contribute to the production of output. The literature points out that these factors are: management capabilities, organizational competence, research and development (R&D), the intersectoral transfer of resources, increasing returns to scale, embodied technical progress and technology diffusion. At is therefore a measure of these elements. Logarithmic differentiation of equation (3.6) with respect to time yields the TFP growth measure: aˆ t = qˆt −

αi xît

(3.7)

i=1

From here on, in this chapter, a circumflex or ‘hat’ over a variable will denote a proportional growth rate. Thus, aˆ t ≡ dAt /At , qˆt ≡ dQt /Qt and xît ≡ dXit /Xit . Equation (3.7) says that TFP growth is output growth less the weighted sum of input growths. Consequently, the TFP growth measure given by equation (3.7) is often referred to as the Solow residual, for it is the residual growth rate of output not explained by the growth in factor inputs. The key finding of Solow (1957) is that aˆ t is, in theory, equal to the growth rate of the Hicksian efficiency parameter. The TFP growth measures characterized by equations (3.4) and (3.7) are true index numbers in that they can be directly computed from prices and quantities. The simplicity of this index number approach led to its ubiquitous use in early productivity growth literature.12 Primal approach The standard primal approach to growth accounting begins with the neoclassical production function: Qt = At F (X1t , X2t , . . . , Xnt )

(3.8)

28


where A is the level of technology and Xi s are the factor inputs such as physical capital, human capital, labour and so on. Equation (3.8) indicates that the technology factor A appears in a Hicks-neutral way – an assumption frequently made in the literature. The factor inputs are often disaggregated among types or qualities as in Jorgenson and Griliches (1967). The growth accounting literature has shown that the growth rate of output can be decomposed in components associated with factor accumulation and technological progress. Logarithmic differentiation of equation (3.8) with respect to time yields such a decomposition, and is given by the equation: qˆt = aˆ t +

ηit xît

(3.9)

i=1

where ηit ≡ FXi Xit /Qt is the elasticity of output with respect to factor input Xi with FXi being the (social) marginal product of Xi . TFP growth, measured by the rate of technological progress here, can be calculated from equation (3.9) as a residual: aˆ t = qˆt −

ηit xît

(3.10)

i=1

Equation (3.10), however, is difficult to implement because it requires knowledge of the social marginal products FXi . In practice, therefore, the computation of TFP typically assumes that the marginal social products can be measured by observed prices. In particular, if the factors are paid their social marginal products, so that FXi = wi with wi being the factor price, then the primal estimate of TFP growth implied by equation (3.10) is given by: aˆ t = qˆt −

sit xît

(3.11)

i=1

where sit ≡ wit Xit /Qt is the factor Xi s share in output. If all of the income associated with Qt is claimed by the factor inputs, then Qt = i=1 wit Xit which implies i=1 sit = 1. The equality of output to total factor income is consistent with equality between the factor prices and social marginal products if the production function (3.8) exhibits constant returns to scale in factor inputs – that is, if

The East Asian Miracle 29 Qt = i=1 FXi Xit . Using the condition i=1 sit = 1 and designating factor X1 as labour, equation (3.11) can be written in intensive form to yield an alternative measure of TFP growth:

aˆ t = qˆ˜ t −

sit xˆ˜ it

(3.12)

i=2

where q˜ ≡ Q/X1 and x˜i ≡ Xi /X1 are quantities per unit of labour. The estimated TFP growth rate, given by equation (3.11) or (3.12), is computed at each date by using time-series data on output and factor input growth rates, qˆ and xî s, as well as factor income shares, si s. In practice, researchers report an average of the computed aˆ values for a designated time period. In subsequent works, several refinements have been made to the TFP growth measure given by equation (3.11). One of these refinements is the measurement of inputs and the disaggregation of input composition to allow for changing input quality.13 Jorgenson and Griliches (1967) and Jorgenson et al. (1987) demonstrate the importance of disaggregating the inputs by quality class. For example, Xi in equation (3.8) can be viewed as a vector of quantities of capital of various kinds, say short-lived and long-lived capital. For a given required rate of return on capital, the rental price of type or quality j is higher if it has a higher depreciation rate – say, due to a higher rate of physical deterioration or economic obsolescence. Consequently, a shift from long-lived capital (for example, buildings) to short-lived capital (for example, machinery) would account for part of the output growth. Failure to allow for this change in capital quality tends to overstate TFP growth computed according to equation (3.11). Because data are not continuous over time but come in discrete time units, Jorgenson and Griliches (1967) propose a further refinement to the TFP growth measure given by equation (3.11), which is a Divisia index.14 In particular, they introduce a discrete-time approximation to the Divisia index derived from the Tronqvist index. The Tronqvist approximation to the TFP measure given by equation (3.11) is obtained by replacing the continuous-time income shares sit s in (3.11) with the average between-period shares. Thus, in the Tronqvist approximation, input Xi s discrete-time income share between periods t and t − 1 is s¯i ≡ (sit + sit−1 )/2. As well, continuoustime growth rates are replaced by their discrete-time counterparts. The rate of growth of input Xi between periods t and t − 1, for

30


example, is ln (Xit /Xit−1 ). Furthermore, to allow for input quality changes, the growth rate of (aggregate) inputs is expressed as a weighted 15 average of the growth rate of the sub-inputs. Thus, ln (Xit /Xit−1 ) = ¯ ij ln Xij t /Xij t−1 where σ¯ ij ≡ (σij t + σij t−1 )/2 with σij being j =1 σ the share of sub-input Xij in total payments to its aggregate input Xi . These refinements to the TFP growth measure given by equation (3.11) lead to the transcendental logarithmic (translog) index of TFP growth:16

Qt = ln Qt−1

TFPt, t−1

−

i

Xit s¯i ln Xit−1

(3.13)

where TFPt, t−1 is the translog index of TFP growth between periods t and t − 1. The index provides a measure of the increase in output attributable to the time-related shift in the production function. Dual approach The dual approach to growth accounting was first introduced by Jorgenson and Griliches (1967), and in recent years it has been applied to the study of productivity growth in East Asia by Hsieh (1999, 2002). We have seen above that the primal approach to growth accounting computes TFP growth rates from growth rates of factor quantities. The dual approach, by contrast, computes TFP growth rates from growth rates of factor prices. The dual approach begins with the basic national income accounting identity that output equals factor incomes: Qt =

wit Xit

(3.14)

i=1

where, as indicated earlier, wi (Xi ) is the unit price (quantity) of the ith factor.17 Differentiating (3.14) with respect to time t and dividing by Qt , one obtains: qˆt =

sit (wˆ it + xît )

(3.15)

i=1

where, as before, si s are the factor income shares. By placing the terms involving the growth rates of factor quantities on the left-hand side of

The East Asian Miracle 31 equation (3.15), one obtains two alternative (but equivalent) measures of TFP growth: aˆ t = qˆt − =

sit xît

(3.16)

i=1

sit wˆ it

(3.17)

i=1

Equation (3.16) is just equation (3.11) introduced earlier. It then follows that the primal estimate of TFP growth, given by equations (3.11) and (3.16), equals the share-weighted growth in factor prices – that is, the right-hand side of equation (3.17). The dual measure of TFP growth is therefore obtained as a shareweighted growth in factor prices. The notion underlying the dual measure of TFP growth is that rising factor prices (for factors of a given quality) can be sustained only if output is increasing for given inputs. Therefore the appropriately weighted average of the growth in factor prices measures the extent of TFP growth. The dual estimate of TFP growth, given by equation (3.17), is computed at each date using time-series data on growth in factor prices and factor income shares. Note, however, that the dual estimate of the TFP growth rate uses the same factor income shares as the primal estimate (see equations 3.16 and 3.17). In practice, an average of the computed TFP growth rates is reported. It is worth noting that the equality of primal and dual estimates of TFP growth, implied by equations (3.16) and (3.17), only requires that equation (3.14) hold. No assumptions about the relations of factor prices to social marginal products or about the form of the production are therefore necessary. The two TFP growth measures will differ only if the national accounts data are inconsistent with the data on factor prices. Econometric approach The index number approach is appealing in its simplicity; it also provides a first-order approximation to productivity measurement. The index number approach, however, is based on restrictive assumptions such as the absence of non-constant returns to scale and perfect competition in both input and output markets. It also requires that firms optimize. Another shortcoming of this approach is that by its very construction it cannot distinguish between a shift in the production function (the pure

32


technical change effect) and movement along the function due to scale and relative price changes. Such a distinction is crucial to understanding the forces that shape the process of productivity advancement. In this sub-section we describe an alternative approach called the econometric approach to productivity measurement. The econometric approach is a flexible technique, used to identify the sources of productivity growth. The estimation technique begins by explicitly specifying the underlying cost or production structure. The estimated parameters of the underlying cost or production model are then used to derive an index of TFP growth. The econometric approach has the flexibility to incorporate pertinent features of the market and industry structure as well as technological features that affect firms’ and industries’ productivities. Using the econometric approach it is therefore possible to identify the sources of productivity growth such as shifts in the production function, increases in the scale of operation, differential utilization of factor inputs, innovations by undertaking R&D, and changes in market and industry structure. Cross-country regression analysis As noted earlier, a direct implementation of equation (3.9) is difficult because the output elasticities in equation (3.9) are not directly observable. If, however, factor markets are perfectly competitive then the output elasticities are equal to the income shares of factor inputs. In practice, therefore, the computation of TFP growth typically proceeds under the assumption that the output elasticity coefficients in equation (3.9) equal income shares. Unfortunately, reliable measures of factor income shares are not available for most developing countries. An alternative approach would therefore be to estimate a cross-country growth regression in the form of equation (3.9) – that is, to regress the growth rate of output on the growth rates of factor inputs.18 The estimated coefficients on the input growth rates then measure output elasticities, while the intercept measures the TFP growth rate. The main attraction of this approach is that it dispenses with the assumption that the social marginal products of factor inputs coincide with the observable factor prices. Instead, one could estimate, as do Young (1994) and Marti (1996), a cross-country regression of the form: qˆ˜ k = α +

βi xˆ˜ ik + k

(3.18)

The East Asian Miracle 33 where qˆ˜ is the growth rate of output per worker, xˆ˜ i is the growth rate of factor input i per worker, is the TFP growth rate and the subscript k refers to country k. Equation (3.18) is an empirical implementation of equation (3.7) above. Thus, this alternative approach exploits the idea that TFP growth can be interpreted as the Solow residual, the residual growth rate of output not explained by the growth in factor inputs. Meta-production function approach Most TFP growth measurement exercises assume constant returns to scale, neutral technological progress, and profit maximization with competitive output and input markets. However, these assumptions are arbitrary and the resulting TFP growth estimates are sensitive to these assumptions. Young (1992, 1995), for example, assumes constant returns to scale. By contrast, Denison (1967) assumes that returns to scale are 1.1 – that is, increasing returns to scale. It is, however, well known that a high degree of assumed returns to scale, in general, results in a low estimate of TFP growth rate. Similarly, TFP growth is independent of the quantities of inputs if technological progress is neutral. If, however, technological progress is biased then the TFP growth estimate will depend on the input quantities. Kim and Lau (1994, 1995) argue that the assumptions frequently made in the traditional TFP growth measurement literature are inconsistent with the data. In particular, they find that the data reject the assumptions that returns to scale are unity, technological progress is neutral and profit is maximized. Kim and Lau (1994, 1995) therefore develop and implement an econometric approach called the meta-production function approach which dispenses with these assumptions.19 Instead, this framework directly tests the conventional assumptions for TFP growth measurement. A meta-production function, introduced by Hayami and Ruttan (1970, 1985), is defined as the common underlying production function representing the input–output relationship of a given industry across all countries. In terms of efficiency-equivalent quantities of outputs and inputs, the meta-production function is given by: Qkt = F (X1kt , . . . , Xnkt ),

∀ k = 1, . . . , N

(3.19)

where Qkt and Xikt s are the efficiency-equivalent quantities of output and inputs respectively of country k in period t. Equation (3.19)

34


reflects one of the basic assumptions of the meta-production function approach: that all the N countries have the same underlying production function F (·). Countries, however, may operate on different parts of F (·). The efficiency-equivalent quantities of output and inputs of each country are not directly observable. Qkt s and Xikt s are, however, linked to the measured quantities of outputs Qkt s and inputs Xikt s through time-varying, country-specific (as well as output and input-specific) augmentation factors AQk (t)s and Aik (t)s respectively. Thus: Qkt = AQk (t)Qkt

(3.20)

Xikt = Aik (t)Xikt

(3.21)

Combining equations (3.20) and (3.21) with equation (3.19), one obtains the aggregate production function in terms of measured quantities of output and inputs: Qkt = F [A1k (t)X1kt , . . . , Ank (t)Xnkt ]/AQk (t) = F (X1kt , . . . , Xnkt , t),

∀ k = 1, . . . , N

(3.22)

It follows from equation (3.22) that the inverse of the outputaugmentation factor AQk (t) can be interpreted as the level of output efficiency in country k at time t. In an empirical implementation of the meta-production function approach one needs to specify the aggregate production function (3.19) as well as the augmentation factors AQk (t)s and Aik (t)s. Kim and Lau (1994, 1995) specify the aggregate production function to be the translog production function which, in terms of efficiency-equivalent quantities of output and inputs, takes the form: ln Qkt = ln Q0 +

i=1

ai ln Xikt +

1 Bij ln Xikt ln Xj kt 2 i=1 j =1

(3.23) In their empirical implementation, Kim and Lau (1994, 1995) also assume that the augmentation factors AQk (t)s and Aik (t)s have a

The East Asian Miracle 35 constant geometric form with respect to time. Consequently: Qkt = AQk (1 + γQk )t Qkt

(3.24)

Xikt = Aik (1 + γik ) Xikt

(3.25)

t

Here the constants AQk (γQk )s and Aik (γik )s are the augmentation level (rate) parameters for output and inputs respectively. Equations (3.23) through (3.25) yield the translog production function in terms of measured quantities of output and inputs: ln Qkt = ln Q0 + δk0 +

δik ln Xikt

i=1

1 + Bij ln Xikt ln Xj kt 2 i=1 j =1 + φk0 + φk1 t + ik ln Xikt t,

∀ k = 1, . . . , N

i=1

(3.26) Thus, equation (3.26) is the translog functional form of equation (3.22). We note further that equation (3.26) contains two types of parameters: country-specific parameters and parameters which are not countryspecific. The country-specific parameters such as δs, φs and s depend on the technology parameters as and Bs as well as on the input and output augmentation parameters As and γ s which may differ across countries. By contrast, the parameters Bs are not country-specific. Equation (3.26) therefore implies that Bs must be identical across countries. Thus, the parameters Bs provide a basis for testing the hypothesis that there is a single aggregate meta-production function for all countries. Equation (3.26) is the most general specification possible under the following maintained assumptions: there is a common underlying aggregate meta-production function across all countries, the functional form of the meta-production function is translog, and the input and output augmentation factors have a constant geometric form with respect to time. Restrictive forms under different hypotheses on the returns to scale, the nature of technical progress and the structure of technology can therefore be obtained as special cases of equation (3.26). The empirical aggregate meta-production function (3.26) is estimated from pooled cross-country time-series data. The estimated production

36


function can then be used to compute TFP growth rates. Boskin and Lau (1990) suggest the following procedure. Let ln Fˆ (·) be the estimated translog production function. Then the average annual rate of TFP growth in country k between the intial period 0 and terminal period T is: TFPk0 = [ln Fˆ (X1k0 , . . . , Xnk0 , T ) − ln Fˆ (X1k0 , . . . , Xnk0 , 0)]/T where Xik0 s are the quantities of inputs in the initial period. An alternative estimate of the average annual rate of TFP growth is given by: TFPkT = [ln Fˆ (X1kT , . . . , XnkT , T ) − ln Fˆ (X1kT , . . . , XnkT , 0)]/T where XikT s are the quantities of inputs in the terminal period. The TFP growth estimates TFPk0 and TFPkT will not, in general, coincide unless technological progress is neutral. Boskin and Lau (1990) therefore suggest that the average annual rate of TFP growth be computed as an arithmetic average of the two estimates, TFPk0 and TFPkT . Thus, the average annual rate of TFP growth in country k between periods 0 and T is TFP = (TFPk0 + TFPkT )/2. The econometric approach implemented in Kim and Lau (1994, 1995) is based on the econometric estimation of an aggregate production function called the meta-production function. The main advantage of this approach is that it does not depend on the assumptions of constant returns to scale, neutral technological progress and profit maximization with competitive output and input markets. Instead, these assumptions are directly tested. Consequently, the resulting TFP growth estimates are likely to be more accurate. Error correction model Nehru and Dhareshwar (1994) point out one serious drawback of the conventional regression approach to estimating TFP growth. They note that TFP growth estimates within this framework are based on estimated production functions expressed in terms of growth rates, which measure only the short-run responses of output to changes in inputs. The production function, however, is an expression that models the longrun relationship between the level of output and the level of inputs. Nehru and Dhareshwar (1994) suggest that the potential dangers of spurious regression results between non-stationary variables prevented

The East Asian Miracle 37 researchers from exploiting the long-run relationship between output and inputs. Consequently, a considerable amount of information contained in data on levels of output and inputs which is relevant for productivity growth analysis remains unused. According to Nehru and Dhareshwar (1994), this potentially valuable information on the long-run relationship between levels of output and inputs can be gainfully exploited in an error correction model. The primary advantage of the error correction model is that it provides a way of separating the long-run relationship between the levels of output and inputs from the short-run responses of output to changes in inputs. Nehru and Dhareshwar (1994) suggest that one estimate the following log-linear version of the Cobb–Douglas production function: ln Qkt = ln A(0) +

k

φk Dk +

k

λk t +

αi ln Xikt + kt

i

(3.27) and apply the Engle–Granger test for cointegration using the equation: kt = µ0 + µ1 kt−1 + µ2 kt−1

(3.28)

Equations (3.27) and (3.28) are estimated using pooled cross-country time-series data. The estimation procedure consists of three steps. First, equation (3.27) is estimated using the method of ordinary least squares (OLS). Second, the variance of the predicted errors ˆkt s are then used to weight the variables in equation (3.27) to correct for heteroskedasticity. Equation (3.27) with the weighted variables is then re-estimated using OLS. Finally, the predicted errors from the re-estimated equation (3.27) are used for the Engle–Granger test in equation (3.28). If the hypothesis that µ2 = 0 can be rejected at a reasonable level of significance, then one may conclude that the errors from the homoskedastic cointegrating regression are stationary and the cointegrating vectors describe a nonspurious relationship. The estimation procedure results in an estimate of the TFP growth rate for country k, and is given by φk + λk .

Productivity growth empirics: the tyranny of numbers In this section we summarize some of the most influential empirical works on TFP growth in East Asia.20 We begin with empirical

38


studies which support the accumulation view. These studies find little or no evidence supporting the hypothesis that East Asia’s miraculous growth has been associated with rapid productivity growth or closing the knowledge or technology gaps. Instead, these empirical works suggest that the basis of East Asian growth has been the maintenance of high rates of factor accumulation over a number of decades. Tables 3.1–3.6 summarize the findings of a select number of such empirical studies. As mentioned previously, however, not everyone agrees with the accumulation view of the East Asian Miracle. For example, researchers such as Klenow and Rodriguez-Clare (1997), Sarel (1997) and Hsieh (1999, 2002) show that the rate of TFP growth in East Asian economies has not been as small as claimed by the proponents of the accumulation view. We therefore also briefly discuss these empirical works. Tables 3.7–3.9 summarize the main findings of these studies. The accumulation view Young (1992) uses the primal approach to growth accounting to estimate TFP growth for Hong Kong and Singapore. He finds practically no TFP growth for Singapore and less than spectacular TFP growth in Hong Kong.21 He also finds that during the 1960s through 1980s Singapore experienced an enormous decline in the implied rate of return on capital. Hong Kong, on the other hand, experienced only a modest decrease in the implied rate of return on capital. Young (1995) uses the same methodology as in his 1992 study to extend the analysis to the Republic of Korea and Taiwan. His findings, similar to those of Young (1992), are summarized in Table 3.1. In particular, Young (1995) finds that the average annual growth rate of TFP in Hong Kong over the period 1966–1991 was 2.3 per cent. This accounts for about 31 per cent of the overall output growth in Hong Kong (see Table 3.1). He further finds that during 1966–1990, TFP in the Republic of Korea grew at an average rate of 1.7 per cent per annum. Thus, TFP growth accounted for only about 16 per cent of the overall output growth in the Republic of Korea. For Singapore, Young (1995) finds that TFP growth during 1966–1990 was only 0.2 per cent per annum. It then follows from Table 3.1 that TFP growth accounted for only a meagre 2 per cent of output growth in Singapore during 1966–1990. Finally, according to Young (1995), TFP growth in Taiwan during the same period was 2.6 per cent, accounting for about 28 per cent of the overall output growth.

The East Asian Miracle 39 Table 3.1 Total factor productivity growth, 1966–1990/91 Country

Labour share

Growth rate (% points per year) of: Output

Hong Konga

0.63

7.3

Korea,b Republic of

0.70

10.3

Singapore

0.51

8.7

Taiwan,b China

0.74

9.4

Weighted capital

Weighted labour

Total factor productivity

8.0 (41) 13.7 (40) 11.5 (65) 12.3 (34)

3.2 (28) 6.4 (44) 5.7 (33) 4.9 (38)

2.3 (31) 1.7 (16) 0.2 (2) 2.6 (28)

Sources: Adapted from tables V–VIII, Young (1995). Numbers in parentheses are authors’ calculations based on estimates reported in Young (1995), and indicate contributions of capital, labour and total factor productivity growth as percentage of output growth. Notes Weighted inputs are translog indices of factor input growth, with labour services measured by hours of work; see Young (1995) for further details. a The reported labour share and the average annual growth rates for Hong Kong are for the 1966–1991 period. b Economy-wide estimates, excluding agriculture.

Young (1994) conducts a cross-country regression analysis of growth and uses the residuals to determine the importance of TFP growth in 118 countries over the period 1970–1985. His conclusions are similar to ones reached on the basis of the primal approach to growth accounting analysis in Young (1992, 1995). In particular, Young (1994) finds that while TFP growth in Hong Kong was relatively high, it was practically zero in Singapore. He also finds that TFP in East Asian economies had not been higher than TFP in many other parts of the world.22 Based on the above findings, Young (1995) argues that the centralized economy of Singapore has always emphasized movement up the technological ladder and forced its citizens to save too much in order to finance technological advancements. New technologies were adopted much too soon without realizing the benefits of learning-by-doing at each stage of technological progress. This premature adoption of new technologies in Singapore actually resulted in increasing costs of production. According to Young (1995), the main source of growth in Singapore has been factor accumulation. He notes that even though Singapore started its development process much later than Hong Kong, it traversed many of the same industries in a much more compressed time frame, which (Young contends) accounts for the lack of productivity growth in

40


Singapore. The implication of this finding is that the growth prospects of other developing countries trying to emulate Singapore’s success are rather gloomy. Similar conclusions are echoed in Krugman (1994). Like Young (1992, 1995), Collins and Bosworth (1996) also use the primal approach to growth accounting, but for a large set of countries. They adopt the neoclassical analysis of economics growth which assumes a stable underlying relationship between output Q, the inputs physical capital K and labour L, and technology A: Qt = At F (Kt , Ht , Lt )

(3.29)

where Ht is an index of labour quality and hence Ht Lt is the skilladjusted measure of the labour input. Assuming that the shares of income paid to the factor inputs can be used to measure their relative importance in the production process, the above equation implies an index of TFP growth aˆ t : aˆ t = qˆt − sK kˆt − sL ˆt

(3.30)

where sK and sL are capital’s and labour’s share, respectively, in output and ˆ is the rate of growth of skill-adjusted labour input.23 Equation (3.30) implies that TFP growth is the rate of growth output qˆt less the share-weighted growth of the factor inputs kˆt and ˆt . Unlike Young (1995), Collins and Bosworth (1996) assume fixed weights across both time and countries with sK = 0.35 and sL = 0.65. TFP growth estimates for the newly industrialized countries (NICs) in East Asia as well as for other industrialized countries (including the United States) and selected regions are reported in Table 3.2. According to Collins and Bosworth (1996), TFP growth for Singapore averages 1.5 per cent per year and accounts for about 30 per cent of output growth during 1960–1994. For more recent periods the role of TFP growth is found to be more pronounced. Between 1984 and 1994, for example, Collins and Bosworth (1996) find that TFP growth averages 3.1 per cent per year and accounts for about 52 per cent of the output growth in Singapore.24 Although the TFP growth estimates reported in Collins and Bosworth (1996) are slightly higher than those of Young (1992, 1995), the results seem to suggest that TFP growth in East Asia was not particularly high relative to TFP growth in other industrialized countries and regions. Like the other accumulationists, Collins and Bosworth (1996) therefore also conclude that factor accumulation was more important for the post-war growth in East Asia.

The East Asian Miracle 41 Table 3.2 Sources of growth (percentage points per year), 1960–1994 Country/region

East Asian NICs: Indonesia

Growth of outputa

3.4

Korea, Republic of

5.7

Malaysia

3.8

Singapore

5.4

Thailand

5.0

Taiwan, China

5.8

Other countries/regions: United States

1.1

East Asia

4.2

Latin America

1.5

Other industrialized countries

2.9

Contribution of: Physical capitala

Educationa


2.1 (62) 3.3 (58) 2.3 (61) 3.4 (63) 2.7 (54) 3.1 (53)

0.5 (15) 0.8 (14) 0.5 (13) 0.4 (7) 0.4 (8) 0.6 (10)

0.8 (23) 1.5 (28) 0.9 (26) 1.5 (30) 1.8 (38) 2.0 (37)

0.4 (36) 2.5 (60) 0.9 (60) 1.5 (52)

0.4 (36) 0.6 (14) 0.4 (27) 0.4 (14)

0.3 (28) 1.1 (26) 0.2 (13) 1.1 (34)

Source: Adapted from tables 6 and 7, Collins and Bosworth (1996). Numbers in parentheses are authors’ calculations based on estimates reported in Collins and Bosworth (1996), and indicate contributions of physical capital, education and total factor productivity growth as percentage of output growth. Notes A capital’s production share of 0.35 and a labour’s production share of 0.65, respectively, are used to compute the contributions of physical capital per worker and education per worker to the growth rate of output per worker. a Growth of output per worker and contributions of physical capital per worker and education per worker.

The World Bank study The East Asian Miracle (1993) begins, as do Collins and Bosworth (1996), with the aggregate production function of the form: Qt = At F (Kt , Et , Lt )

(3.31)

were Et is a measure of human capital endowments. The aggregate production function (3.31) implies the following measure of TFP growth,

42


denoted aˆ t : aˆ t = qˆt − sK kˆt − sE eˆt

(3.32)

where sK and sE are the elasticities of output with respect to physical and human capital. Thus, equation (3.32) implies that TFP growth is a residual of the rate of growth of output per worker qˆt − ˆt after deducting the contributions of human and physical capital accumulation. The output elasticities in equation (3.32) are not directly observable. Under the assumption of competitive factor markets, however, the output elasticities are equal to the income shares of factor inputs. Consequently, most empirical applications of equation (3.32) estimate the output elasticity coefficients with income shares.25 However, reliable measures of factor income shares are not available for most developing countries. The World Bank study therefore estimates the output elasticities sK and sE directly using a simple, cross-country production function. To be specific, the World Bank study regresses annual log output growth on log physical and human capital growth, and log labour growth during the period 1960–1990, subject to the constraint that the regression coefficients sum to unity. To allow estimates of individual rates of TFP changes for each of the eightyseven countries in the sample, country-specific dummy variables are also included in the regression. The estimated TFP growth rates for the eight high-performing Asian economies (HPAEs) – Hong Kong, Indonesia, Japan, the Republic of Korea, Malaysia, Singapore, Taiwan and Thailand – are reported in Table 3.3. Estimates based on the output elasticities implied by the estimated production function for the full sample of eighty-seven countries are given in column 3. The TFP growth estimates implied by the estimated production function representing only the high-income-economy input–output relationships are reported in column 4 of Table 3.3. It follows from column 3, Table 3.3, that the role of TFP growth in the high-performing East Asian economies has not been as small as claimed by Young (1992, 1994, 1995). This apparent contradiction, the 1993 World Bank study argues, results from the unusually low estimate of the elasticity of output with respect to capital from the whole cross-country sample. The World Bank study notes that the sample of countries considered includes a sub-set of developing countries that has positive net investment and human capital growth rates but negative output growth rates. In effect, the marginal products of

The East Asian Miracle 43 Table 3.3 Total factor productivity growth, eight high-performing Asian economies, 1960–1989 Economy

Growth rate (% points per year) Real GDP per worker

Hong Kong

5.6925

Indonesia

3.5591

Japan

5.0145

Korea, Republic of

5.8066

Malaysia

3.4886

Singapore

5.2147

Taiwan, China

5.6806

Thailand

4.0262

Total factor productivity Full samplea

High-income economiesb

3.6470 (64) 1.2543 (35) 3.4776 (69) 3.1021 (53) 1.0755 (31) 1.1911 (23) 3.7604 (66) 2.4960 (62)

2.4113 (42) −0.7953 (−22) 1.4274 (28) 0.2355 (4) −1.3369 (−38) −3.0112 (−58) 1.2829 (23) 0.5466 (14)

Source: Adapted from World Bank (1993). The real GDP growth per worker and contributions of the total factor productivity growth as percentage of the real GDP growth (numbers in parentheses) are authors’ calculations based on data from the Penn World Tables Mark 5.6; see Summers and Heston (1991). Notes The eight high-performing Asian economies are Hong Kong, Indonesia, Japan, the Republic of Korea, Malaysia, Singapore, Taiwan and Thailand. Absent figures for Latin America and sub-Saharan Africa reflect the fact that these figures were not calculated. a Estimates based on the output elasticities implied by the estimated production function for the full sample of eighty-seven economies. b Total factor productivity growth estimates implied by the estimated production function reflecting only the high-income-economy input–output relationships.

both physical and human capital were negative in these economies, reducing the estimated elasticities of output with respect to physical and human capital in the production function based on the full sample. When the production function is re-estimated on the basis of only highincome-economy input–output relationships, the estimated elasticity of output with respect to capital rises to more conventional levels.26 The TFP growth estimates, based on the revised estimate of the elasticity of output with respect to capital, are broadly consistent with the findings of Young.

44


Table 3.4 Sources of growth in a model without human capital Country

East Asian NICs: Hong Kong Korea, Republic of Singapore Taiwan, China G-5 countries: France Japan Germany United Kingdom United States

Time period

GDP a

Relative contributionb of: Capital

Labour

Technical progress

1966–90 1960–90 1964–90 1953–90

7.8 8.6 8.9 8.7

74 80 68 85

26 20 32 15

0 0 0 0

1957–90 1957–90 1960–90 1957–90 1948–90

3.7 6.7 3.2 2.5 3.1

37 56 40 39 26

−1 5 −7 4 26

65 39 66 57 49

Source: Adapted from tables 3.1 and 7.2, Kim and Lau (1994). Notes a Average annual growth rate of real GDP (percentage points). b Relative contributions of inputs as percentage of output growth.

Kim and Lau (1994) employ the meta-production function approach to estimate TFP growth rates, during the post-war period, for the four East Asian newly industrialized countries (NICs) – Hong Kong, the Republic of Korea, Singapore and Taiwan – and the Group of Five (G-5) industrialized countries – France, Japan, Germany, the United Kingdom and the United States. The approach amounts to estimating a regression, pooling time-series data for a cross-section of countries. Kim and Lau (1994) pool data for the four East Asian NICs and the G-5 countries from the late 1940s to 1990. As shown in Table 3.4, the precise period covered varies from country to country and is generally shorter for the East Asian NICs than for the G-5 countries. Kim and Lau (1994) implement the estimation procedure by fitting a translog production function with a time trend, a proxy for technical efficiency, to the time-series data pooled from the nine countries. To test whether technical progress is of the augmentation type, they also include augmentation factors in the aggregate meta-production function. Finally, in order to test the hypotheses of profit maximization and competitive markets, Kim and Lau (1994) also estimate the equation for the share of labour cost in the value of output. The main findings of Kim and Lau (1994), summarized in Table 3.4, are enumerated below: 1

The four East Asian NICs and the G-5 countries share the same aggregate meta-production function as well as factor (level)

The East Asian Miracle 45

2

3

4

augmentation parameters. In particular, Kim and Lau (1994) find that technical progress in the nine countries is capital-augmenting. Thus, according to Kim and Lau (1994), technical progress in these countries can be represented as purely capital-augmenting. The standard assumptions which underly growth accounting are rejected by the data. The assumptions include: homogeneity of production function in factor inputs, constant returns to scale, neutral technological progress and profit maximization. The data reject the hypothesis of zero technical progress for the G-5 countries, but not for the four East Asian NICs. This is apparent in Table 3.4, which shows that the contribution of technical progress to post-war real GDP growth in the East Asian NICs is zero. By contrast, technical progress accounts for about 40–70 per cent of post-war real GDP growth in the G-5 countries. Finally, the level of technology (relative to that in the United States) in the East Asian NICs has been declining since the 1950s. In the 1950s the level of technology in the NICs was 25 per cent of the technology level in the United States. By 1990 the level of technology in the NICs had declined to only 20 per cent that of the United States. Kim and Lau (1994) therefore argue that there is no evidence of convergence in technology.

Kim and Lau (1995) extend their 1994 study to examine the importance of human capital growth. In particular, they (re)estimate an aggregate meta-production function with three inputs – physical capital (K), human capital (H ) and labour (L), using annual data drawn from the four East Asian NICs and the G-5 countries. The quantity of human capital is measured in terms of the total number of years of education per person of the working age population. Results based on the re-estimated aggregated meta-production function are reported in Table 3.5. These results reinforce their previous findings. As in their 1994 study, Kim and Lau (1995) find that the aggregate production functions of all nine countries are identical in terms of the efficiency-equivalent units of output and factor inputs. Furthermore, the aggregate meta-production function of the nine countries is of the form: Yit = F (Ai (t)Kitλ Hit1−λ , Lit ) where λ is a constant common to all nine countries. The above aggregate meta-production function implies that technical progress

46


Table 3.5 Sources of growth in a model with human capital Country

East Asian NICs: Hong Kong Korea, Republic of Singapore Taiwan, China G-5 countries: France Japan Germany United Kingdom United States

GDPa

Relative contributionb of: Capital

Human capital

Labour

Technical progress

7.8 8.6 8.9 8.7

66 67 63 75

11 14 13 11

22 19 25 14

0 0 0 0

3.7 6.7 3.2 2.5 3.1

33 48 36 35 23

4 3 5 6 4

−1 6 −7 4 30

63 43 66 55 43

Source: Adapted from table 3.1, Kim and Lau (1994), and table 5, Kim and Lau (1995). Notes Time periods are the same as in Table 3.4. a Average annual growth rate of real GDP (percentage points). b Relative contributions of inputs as percentage of output growth.

in all nine countries can be simultaneously represented as purely capital-augmenting and purely human-capital-augmenting. Based on the estimated (modified) aggregate meta-production function, Kim and Lau (1995) also carry out a growth accounting exercise. They find that technical progress continues to be the most important source of economic growth in the G-5 countries (Japan excepted); physical capital is the second most important source of growth (except for the United States). By contrast, the most important source of economic growth in the East Asian NICs is physical capital accumulation, accounting for more than 60 per cent of their real GDP growth. Growth in labour and human capital accounts for the residual output growth in the four NICs. Kim and Lau (1995) also find that human capital is relatively more important as a source of growth for the NICs. Kim and Lau (1994) provide the following justifications for their finding that exogenous technical progress is not an important source of growth in the East Asian NICs: 1

Kim and Lau (1994, 1995) use gross capital stock as a measure of physical capital. To the extent that physical capital depreciation is significant, the measured capital stock will overstate the true capital


2 3

4

5

6

7

8

stock. Consequently, the estimated capital augmentation rate may be biased downwards. Until recently, (R&D) investment in the East Asian NICs has been very low. Industries in the NICs employed mature technologies and imported capital goods at prices reflecting amortized R&D as well as other development costs. Capital goods installed in the NICs are likely to be of the ‘off the shelf’ variety and the prospect of indigenous improvements is limited. It is quite possible that the existing technical progress is embodied primarily in capital goods used in high-technology industries. As a result, the NICs may not have benefited from technical progress to the same extent as the industrialized countries. It is possible that the software component of investments – that is, managerial methods, institutional environments and supporting infrastructure – lags behind the hardware component. If so, then the full productivity potential of capital goods cannot be realized. The NICs’ poor natural resource endowment and the lack of a labour force with advanced scientific knowledge may have offset potential gains from technical progress. Measured GDP may not reflect the contributions of all the inputs to output.

Nehru and Dhareshwar (1994) present new estimates of TFP growth for ninety-three industrial and developing countries at the economywide level for the period 1960–1990. These estimates are based on a production function which does not incorporate human capital as a factor of production. In order to emphasize the role of human capital, Nehru and Dhareshwar (1994) also provide estimates of TFP growth which incorporate human capital as a factor of production for eightythree countries for the period 1960–1987. Table 3.6 reproduces these TFP growth estimates for six East Asian NICs and five OECD countries. According to Nehru and Dhareshwar (1994), the key innovations consist of incorporating human capital as a factor of production in the ‘old’ growth model, using new data on factors of production, and applying cross-country error correction models for estimating the parameters of long-run input–output relations. Nehru and Dhareshwar (1994) observe that conventional cross-country estimates of production functions measure only the short-run responses of output to changes in factor inputs.

48


Table 3.6 Total factor productivity growth, 1960–1990 Country

East Asian NICs: Indonesia Korea, Republic of Malaysia Singapore Thailand Taiwan, China OECD countries: France Japan Germany United Kingdom United States

Total factor productivity (average annual % changes) First-difference model

Error correction model

Without human capital (1960–90)

With human capital (1960–87)

Without human capital (1960–90)

With human capital (1960–87)

1.05 2.43 1.11 1.22 1.73 2.21

0.55 2.21 0.96 0.61 2.21 —

0.19 0.71 −0.18 −0.61 0.09 0.41

0.12 0.55 0.09 −0.77 0.75 —

1.05 1.82 0.69 0.76 1.07

1.64 2.68 1.42 1.07 1.74

0.08 0.03 0.00 0.06 0.56

0.60 0.72 0.79 0.15 1.08

Source: Adapted from Nehru and Dhareshwar (1994), annex 1. Estimates of total factor productivity growth rates for Hong Kong and Taiwan (model with human capital) are not reported in Nehru and Dhareshwar (1994).

But the production function is an expression that models a long-run relationship between the level of output and the level of inputs. This long-run relationship has not been explored in previous empirical work on TFP growth because of the danger of spurious regression results between non-stationary variables. Consequently, analysts have not been able to exploit a considerable amount of information contained in data on levels of output and inputs relevant to productivity growth analysis. Nehru and Dhareshwar (1994) argue that this potentially valuable information on the long-run relationship between levels of output and inputs can be gainfully used in productivity growth analysis if an error correction model is employed. Nehru and Dhareshwar (1994) find that human capital accumulation is far more important in explaining growth than is indicated in other studies because human capital plays a dual role in the development process: as a standard factor of production to be accumulated and as a source of learning and entrepreneurship.27 Contrary to the other productivity growth studies, Nehru and Dhareshwar (1994) also find that TFP growth

The East Asian Miracle 49 in high-income countries has been comparable to the faster growing low-income and middle-income countries.28 Furthermore, they find that the fastest growing developing economies have based their growth more on the speed with which they have accumulated physical and human capital than on high TFP growth. Young (1992, 1994, 1995) arrives at much the same conclusion, as do the 1993 World Bank study, Kim and Lau (1994, 1995), Collins and Bosworth (1996) and Nadiri and Son (1998). Findings such as these led Krugman (1994) to conclude that ‘perspiration’ and not ‘inspiration’ powered the East Asian Miracle. Clearly, this has important implications for the maintenance of the East Asian Miracle (as well as for other developing countries aspiring to join the league of wealthy nations), because growth driven by perspiration is not sustainable if capital is subject to diminishing marginal returns, whereas costless gains in productivity may continue as long as inspiration (or imitation) permits. An assimilationist reinterpretation The minority role of TFP growth which is documented in the empirical growth literature reviewed above, and apparent in Tables 3.1– 3.6, led researchers such as Klenow and Rodriguez-Clare (1997), Sarel (1997), Hsieh (1999, 2002) and Bernanke and Gürkaynak (2001) to re-examine the evidence supporting the accumulation view of the East Asian Miracle. These researchers find that the role of TFP growth in East Asian economies has not been as small as claimed by their predecessors. We now turn to these works. Like Young (1992, 1995), Klenow and Rodriguez-Clare (1997) also employ the primal approach to growth accounting to estimate TFP growth for ninety-eight industrialized and developing countries for the period 1960–1985. However, unlike Young (1992, 1995), they adopt the Mankiw–Romer–Weil growth equation (see Mankiw et al., 1992): α/(1−α−β) β/(1−α−β) ht

qt = At kt

(3.33)

where q is output per worker, A is productivity, k is physical capital per unit of output, h is human capital per unit of output, α is the elasticity of output with respect to physical capital and β is the elasticity of output with respect to human capital. Equation (3.33) implies the following

50


expression for TFP: aˆ t = qˆt −

α 1−α−β

kˆt −

β 1−α−β

hˆ t

(3.34)

Following Mankiw et al. (1992), Klenow and Rodriguez-Clare (1997) set α = 0.30 and β = 0.28. To obtain TFP growth estimates, they then apply data on output per worker, the physical capital-to-output ratio, and the human capital-to-output ratio to equation (3.34). Table 3.7 reports these estimates for the eight East Asian economies. The estimation results imply that growth in human capital per worker has a modest role in explaining growth. More important, the results show that TFP growth accounts for most of the growth of output per worker in East Asian economies. This finding extends to the sample of ninety-eight countries considered in Klenow and Rodriguez-Clare Table 3.7 Productivity growth, 1960–1985 Country

Labour share

Growth rate (% points per year) of: q

k

h

A

0.53 (7) 1.88 (35) 2.01 (27) 2.32 (31) 1.30 (25) 2.40 (34) 1.76 (24) 0.87 (17)

1.08 (13) 0.95 (16) 0.51 (6) 1.77 (22) 1.21 (22) 0.16 (2) 1.52 (19) 0.62 (11)

4.39 (80) 1.91 (49) 3.53 (67) 2.54 (47) 2.00 (53) 3.29 (64) 3.03 (57) 2.66 (72)

Hong Kong

0.70

5.49

Indonesia

0.70

3.89

Japan

0.70

5.30

Korea, Republic of

0.70

5.37

Malaysia

0.70

3.74

Singapore

0.70

5.11

Taiwan, China

0.70

5.30

Thailand

0.70

3.70

Sources: Adapted from Klenow and Rodriguez-Clare (1997), data appendix. Numbers in parentheses are authors’ calculations based on estimates reported in Klenow and Rodriguez-Clare (1997), and indicate contributions of physical capital, human capital and productivity growth as percentage of output growth. Notes q output per worker, k physical capital-to-output ratio, h human capital-to-output ratio, A productivity.

The East Asian Miracle 51 (1997): they find that 90 per cent of cross-country differences in output per worker is attributable to differences in productivity growth. Needless to say, these findings are quite contrary to the accumulation view of the East Asian Miracle. Earlier we reviewed several studies which conclude that TFP growth rates in East Asian economies are not nearly as spectacular as the output growth rates in these economies. Sarel (1997) argues that the findings of such studies are driven primarily by their heavy reliance on national accounts data for measures of various variables and, in particular, for the computation of the factor shares. Like some of the previous studies, Sarel (1997) also uses the primal approach to growth accounting. That is, he computes TFP growth according to the equation: aˆ t = qˆt − sK kˆt − sL ˆt

(3.35)

where, as before, sK and sL are the factor income shares. Unlike some of the previous studies, however, Sarel (1997) uses internationally comparable data from the Summers–Heston database. More important, he employs an alternative method of estimating sK and sL which incorporates the view that technological factor shares are determined by the industrial structure of the economy as well as its level of development. This alternative method yields estimates of factor shares which are not necessarily equal to income shares as measured in the national accounts. Using this new methodology of estimating factor shares and consistent cross-country methods of estimating growth rates of output, capital and labour, Sarel (1997) computes TFP growth rates for five Asian economies and the United States. The estimation results for the period 1978–1996 as well as for the sub-period 1991–1996 are reproduced in Table 3.8. The results for the period 1978–1996 show an impressive growth rate for TFP in Malaysia (2.00 per cent per year), Singapore (2.23 per cent per year) and Thailand (2.03 per cent per year), and a relatively strong rate for Indonesia (1.16 per cent per year). By contrast, the estimated rate of TFP growth over the same period for the United States was only 0.29 per cent per year. TFP growth rates over the sub-period 1991–1996 for these economies were higher (Malaysia excepted) than their 1978–1996 rates. Furthermore, Sarel (1997) finds that the proportion of per capita output

4.54 5.09 5.24 1.07

Malaysia

Singapore

Thailand

United States

1.26

6.51

4.91

5.35

5.11

8.97 (62) 6.86 (47) 6.45 (42) 7.32 (41) 1.63 (44)

6.98 (44) 8.25 (50) 5.60 (40) 11.13 (50) 1.19 (27)

0.93 (13) 0.58 (9) 1.06 (14) 1.51 (21) 0.43 (28)

1978–96 0.96 (13) 0.97 (12) 0.77 (10) 1.37 (15) 0.43 (25)

1991–96

Effective labour per person

1.16 (25) 2.00 (44) 2.23 (44) 2.03 (39) 0.29 (27)

1978–96

2.20 (43) 2.00 (37) 2.46 (50) 2.25 (35) 0.61 (49)

1991–96


Notes Numbers in parentheses are contributions of physical capital per person, effective labour per person and total factor productivity growth as percentage of growth of output per person.

Source: Adapted from table 2, Sarel (1997).

4.74

1991–96

1978–96

1978–96

1991–96

Capital per person

Output per person

Average annual growth (%) of:

Indonesia

Country

Table 3.8 Growth accounting, 1978–1996

The East Asian Miracle 53 growth attributable to TFP growth is not systematically different in East Asian economies and the United States. Hsieh (1999, 2002) employs the dual approach to growth accounting to re-estimate TFP growth for the East Asian countries investigated in Young (1995). The dual measure of TFP growth, denoted aˆ t , is obtained as a share-weighted growth in factor prices and is given by the equation: aˆ t = sK rˆt − sL wˆ t

(3.36)

where r is the real rental price of capital, w is the real wage rate, and as before, sK and sL are the factor income shares. Hsieh (1999, 2002) demonstrates that equation (3.36) can be readily derived from the basic national income accounting identity that national output is equal to factor incomes.29 Therefore, the equality between the primal and dual measures of TFP growth is an implication of the condition that output equals factor incomes. The two measures of TFP growth will differ only if the national accounts are inconsistent with the data on factor prices. Using the aggregate factor shares estimates provided in Young (1995) in equation (3.36), Hsieh (1999, 2002) produces a new set of TFP growth estimates for the four East Asian countries. These results are reproduced in Table 3.9. It follows from Table 3.9 that, while the dual estimates of TFP growth for Hong Kong and the Republic of Korea are similar to the primal estimates, they exceed the primal estimates by roughly 1 per cent per year for Taiwan and by more than 2 per cent per year for Singapore. According to Hsieh (1999, 2002), the main reason for the large discrepancy for Singapore is that, despite the high rate of capital accumulation documented in the national accounts of Singapore, the return to capital has remained constant in Singapore. Furthermore, changes in the risk premium, financial market controls, taxes on capital and public investment subsidies do not explain the absence of any trend in the rental rate of capital in Singapore. These results therefore suggest that technology may have played a much larger role in the economic transformation of Singapore than the primal estimates would have one believe. It is apparent from the empirical studies we have reviewed thus far that the empirical work on TFP growth has been a tyranny of numbers. From ‘back of an envelope’ calculations (for example, see Young, 1994) to the most detailed studies (for example, see Young, 1995, and Hsieh, 2002), the purpose of this cumulative empirical work on TFP growth has been to provide TFP growth estimates supporting the authors’ views on

54


Table 3.9 Total factor productivity growth Real interest rate

Time period

Labour share

Growth rate (% points per year) of: Rental price of capital

Hong Kong: Prime lending rate Call money rate Earnings–price ratio Korea, Republic of: Kerb-market loana Deposit rate Discount rate Singapore: Return on equity Average lending rate Earnings–price ratio Taiwan, China: Kerb-market loana Secured loan rate Treasury bill rateb

Wages

Total factor productivity Dual

Primal

1966–91 1966–91 1973–91

0.63 0.63 0.62

−1.13 −1.53 0.96

4.05 4.05 4.14

2.12 1.98 2.92

2.30 2.30 2.18

1966–90 1966–90 1966–90

0.70 0.70 0.70

−3.95 −3.41 −4.91

4.38 4.38 4.38

1.91 2.07 1.62

1.70 1.70 1.70

1971–90 1968–90 1973–90

0.51 0.51 0.51

−0.20 1.64 −0.50

3.17 2.67 3.63

1.52 2.16 1.61

−0.69 −0.22 −0.66

1966–90 1966–90 1975–90

0.74 0.74 0.75

−0.36 −2.01 −2.07

5.26 5.26 5.79

3.79 3.36 3.79

2.60 2.60 2.70

Sources: Adapted from tables 1–4, Hsieh (2002). The primal estimates of total factor productivity and aggregate factor shares are from Young (1995). Notes The dual estimates of total factor productivity growth rates are the (factor shares) weighted growth rate of quality-adjusted real wages and rental price of capital. The return on equity (in Singapore) and earnings–price ratio (in Hong Kong and Singapore) are used as real interest rates. All other measures of return to capital are used as nominal interest rates. The implied real interest rates are obtained by subtracting the ex post inflation rates from the nominal interest rates. Hsieh (2002) provides further details. a Kerb-market loan rate. b Three-month treasury bill rate.

the sources of growth in East Asia. Consequently, the war of numbers has resulted in a wide range of TFP growth estimates which are very sensitive to the specific assumptions of each study, the methodology and data used, and the time period covered in each study. Often, these studies lead to contradictory results. It appears that by reworking the data one can arrive at almost any plausible conclusion.30 This highlights the general fragility of inferences drawn in the productivity growth empirics literature devoted to explaining the East Asian Miracle. Caution should therefore be used when interpreting the conclusions of this literature.


Productivity growth empirics: an assessment In this section we provide a critical assessment of the TFP growth estimates presented above and the methods used in producing these estimates. We highlight different theoretical and empirical problems encountered by the estimation methods used and their implications for the meaning and interpretation of the measures of productivity growth. We also evaluate the existing productivity growth empirics literature in light of its ability to enhance our understanding of the East Asian growth experience and to help resolve the controversy regarding TFP growth in the East Asian economies.31

Primal estimates of TFP growth Typical growth accounting exercises employ growth accounting equations such as (3.13) where the observed factor shares sit s from the national accounts are used to estimate the output elasticities with respect to factor inputs. It is apparent from equation (3.13) that sit s play an important role in the computation of TFP growth rates. Pack (2001) raises several issues with regards to the measurement and interpretation of the factor shares sit s. Some growth accounting studies assume that an identical international production function – with identical parameters – exists for all countries, and that firms in each country can move along this common international production function. There are, however, significant problems with this presumption. For example, as noted in Pack (2001), if the Republic of Korea was on the same production function as the United States in the 1980s, given the observed differences in capital–labour ratios (19,349 and 60,057 respectively) and factor shares (0.52 and 0.24 respectively) in the 1980s in the two countries, the implied elasticity of substitution is 0.4, which is lower than most econometric estimates. That is, under the assumption that all countries are on the same production function, the data imply a low elasticity of substitution σ . This is problematic for growth accounting because a low value of σ combined with labouraugmenting technological change could maintain constant capital shares within a country. This, however, would imply that the capital share itself is affected by technical progress. As noted above, the factor shares sit s within a country may themselves be endogenous, reflecting technical or structural change. Growth accounting exercises such as Young (1992, 1995) and Bosworth and

56


Collins (1996) assume Hicks-neutral technological change. Kim and Lau (1994, 1995), however, find that this hypothesis is not supported by the data. Nelson and Pack (1999) show that sit s are indeed endogenous if technological progress is Harrod-neutral – that is, labour-augmenting. If technical change is, in fact, labour-augmenting, then the capital shares used in conventional growth accounting exercises would be smaller, leading to a higher estimate of TFP growth. Dual estimates of TFP growth In order to circumvent the possibility that physical quantities of inputs are incorrectly measured, particularly capital stock, Hseih (1999, 2002) estimates TFP growth for four East Asian countries using the dual approach to growth accounting. This, however, presumes that factor prices are measured more accurately than factor quantities. As demonstrated in equations (3.16) and (3.17) above, dual and primal estimates of TFP growth should coincide. This, however, is not the case. Hseih (1999, 2002) presents several estimates of TFP growth, based on different methods of computing the user cost of capital (Table 3.9). It is evident from these estimates that, in each case, the dual estimate differs from the primal estimate, and in most cases the dual exceeds the primal. Furthermore, the dual estimate for Singapore is considerably higher than the estimate reported in Young (1995). Although Hseih’s results are subject to caveats because he assumes that factor shares are exogenous, the exercise demonstrates the wide range of estimates of economy-wide TFP growth rates calculated from growth accounting. Production function estimates of TFP growth The main advantage of the econometric approach is that it dispenses with the assumption that the factor social marginal products coincide with the observable factor prices. The disadvantages of the regression approach, as Barro (1999) points out, are several. In typical growth regressions, input growth rates are treated as exogenous. Input growth rates, however, cannot usually be regarded as exogenous with respect to changes in the rate of technological progress, in particular. If this were the case, the factor growth rates within this framework would receive credit for correlated variations in unobservable technological change. This would result in an underestimate of the TFP growth rate.

The East Asian Miracle 57 Furthermore, if the input growth rates are measured with errors, then the OLS estimates of the coefficients of the input growth rates would provide inconsistent estimates of output elasticities with respect to factor inputs. This errors in measurement problem is likely to be acute for the growth rate of capital input, where the measured capital is unlikely to correspond well to the stock of capital currently utilized in production. As Barro (1999) points out, this problem often leads to low estimates of the contribution of capital accumulation to output growth when highfrequency data are used.

Conclusion The empirical results reviewed in this chapter present a wide range of TFP growth estimates for East Asian countries. The observed differences in TFP growth estimates are due to the methodology and data used, the time periods covered in each study and so on. Despite the apparent lack of consensus on the TFP growth estimates for the East Asian economies, two main conclusions can be drawn from the existing TFP growth empirics. First, a large part of the observed output growth in the East Asian economies over the last three decades is due to factor accumulation. Second, a significant portion of the observed output growth in these countries still remains unexplained. In view of the results presented in this chapter, it is safe to conclude that the debate on the sources of the East Asian growth has yet to be settled. There is clearly some scope for agreement between the assimilation and the accumulation views. The insights from microanalysis akin to Wade (1990) can play an important complementary role in our understanding of the nature of the East Asian Miracle.

4

Factor price distortions and economic activity

In the tradition of neoclassical economics, it is no surprise that there has been so much discussion about the role of factor and commodity prices in resource allocation and the choice of techniques in the developing countries. However, our aim is not to go through the theoretical debate on these issues. Rather we are concerned with the assessment of the empirical literature in this area and even here, we are particularly interested with the role of factor price distortions on employment, though we make a brief excursion into some other implications of factor prices distortions too. The major difficulty in assessing the evidence is that there is not much quantitative evidence to assess. For our purpose, the most relevant and extensive evidence is found in Gupta (1989) which deals with the role of factor price distortions in the developing countries. Unfortunately, even this work does not cover all the countries included in our sample of the East Asian economies. In addition to this evidence, we consider two other types of works: those dealing with the implications of factor price distortions for economic growth and the other with the implications for welfare and capital accumulation. While there is much empirical work on the role of factor price distortions on growth, we consider only three works in some detail. Our reasons for this choice are that these works not only summarize the earlier empirical works, but also represent important analytical contributions. These works are Aggarwala (1983), Jones (1994) and Thomas and Wang (1996).

Factor price distortions, economic welfare and capital accumulation Before assessing the evidence on the impact on employment, we consider some of the literature which deals with the empirical aspects of

Factor price distortions 59 factor price distortions for economic welfare and capital accumulation. We concentrate on the work by Kwon and Paik (1995). This follows the earlier works by Dougherty and Selowsky (1973) on Colombia and Floystad (1975) on Norway, which estimated the static welfare cost of factor price distortions. Their approach was based on a partial equilibrium framework and concluded that the cost was second order of magnitude and therefore not of much significance. However, de Melo (1977), which used a computable general equilibrium (CGE) model, concluded that the effect amounted to a 13.3 per cent reduction in GDP, an obviously far greater cost. Kwon and Paik (1995) revisited de Melo’s study, but only for South Korea. They also used a CGE model, but made significant extensions to the de Melo study. For example, unlike de Melo they allowed for distortions not only in the price of labour but also in the price of capital as well as distinguishing between autonomous differentials and distortions. Using a variety of experiments, they arrive at conclusions which are more along the lines of Dougherty and Selowsky (1973) and Floystad (1975) than de Melo (1977); namely, that the elimination of factor price distortions did not contribute significantly to an increase in GDP or welfare. Kwon and Paik also conducted one further experiment and that was the implication of eliminating factor price distortion on capital accumulation and the concentration of capital stock. They concluded that ‘distortions may lead to more rapid capital formation and higher concentration of capital stock’ (p. 664). While the reader is invited to examine their paper in detail and also read Dougherty and Selowsky for earlier evidence on this issue, our conclusion for now is that, going by the details of the findings on South Korea, we need to be cautious about the potential gains from the elimination of factor price distortions.

Factor price distortions and growth Turning to the studies on the impact of factor price distortions on growth, we concentrate on three studies: Aggarwala (1983), Jones (1994) and Thomas and Wang (1996). Aggarwala was probably the first study to examine the issue in detail. As one might expect, his was an experimental study and as such suffered from many limitations, the most important being that much of the evidence on distortions was qualitative in nature, so that no numerical estimates of the cost of factor price distortions were possible. An example of such qualitative evidence is to be found in Table 4.1. This table gives qualitative evidence on the extent of factor price distortions in some of the countries of our sample. It is clear that

60

Factor price distortions Table 4.1 Distortions in the 1970s Country

Exchange rate

Capital

Labour

Inflation

Thailand South Korea Malaysia Philippines Indonesia

L L L L M

L M M M M

L L M L L

L M L L M

Source: Adapted from Aggarwala (1983), figure 1. Note L low distortion, M medium distortion.

these data cannot be used to assess the quantitative cost of employment foregone as a result of factor price distortions. None the less, from his study, Aggarwala concluded that such distortions did cause a reduction in growth. Jones (1994) also examined the same question, namely, did factor price distortions reduce the growth rate? His approach was in the spirit of Barro (1991). Using some indicators of factor price distortions, he estimated its significance in a Barro-type growth equation. Of course, his study is subject to all the limitations inherent in the Barro type of approach. But, that aside, his major finding is that the relative price of machinery, among the prices of investment, is of the highest consequence. To the extent that taxes and subsidies can influence such prices, his findings have policy implications. Thomas and Wang essentially follow in the footsteps of Barro and Jones. Their paper distinguishes between government intervention and distortion and then calculates composite indices of distortion. These composite indices are regressed on total factor productivity and per capita GDP growth rates. They conclude that ‘part of the East Asia success is attributable to its superior policy framework’ (p. 278).

Estimates of factor price distortions Before offering some summary assessment of the above works, we turn our attention to a detailed review of Gupta (1989). For our purpose, this work provides the most relevant evidence. This is because in order to evaluate the cost of factor price distortions we need two types of information. They are: quantitative estimates of factor price distortions and, second, some method of measuring the cost of these distortions, in

Factor price distortions 61 terms of, say, employment foregone. Gupta (1989) provides both types of information. In order to estimate the extent of factor price distortions, Gupta points out that the most appropriate procedure would be to estimate the shadow prices of the factors concerned, but such a task is too big with such a large sample. He instead adopts a partial equilibrium approach which consists of using the marginal productivity condition under perfectly competitive markets. Using this condition, he estimates the shadow wage rate. The shadow price of capital is estimated by exploiting the definition of the user cost of capital for alternative definitions of the real interest rates. Once these shadow prices are estimated, the distortions are measured as deviations of actual prices from the shadow prices. The deviations thus provide the quantitative estimates of factor price distortions. Gupta (1989) offers estimates of factor price distortions for only two of the countries included in our sample of East Asia. They are South Korea and Singapore. He provided estimates for the wage rates and the user cost of capital. The unit of observation is three-digit industries as well as for all industries and three groupings of industries, namely consumer goods, intermediate goods and capital goods. Gupta’s estimates clearly show that at different levels of aggregation there was significant percentage deviation of actual from the shadow factor prices, where factors refer to labour and capital. It is interesting to note that the degree of distortion was more marked for South Korea than for Singapore. But it should be kept in mind that these estimates are quite dated as of now and whether there is still any factor price distortion in these countries is anybody’s guess. We now turn to the estimates of the cost of the factor price distortions mentioned above for South Korea and Singapore. As mentioned above, Gupta’s (1989) work provides extensive estimates of the short-run and the long-run evidence on the elasticities of employment with respect to wage rates and the rental cost of capital. Gupta then argues that the cost of the distortions is given by the product of the relevant elasticity and the factor price distortion. The loss in employment provided the cost of factor price distortion. Unfortunately, Gupta provides rather scant evidence on the cost of factor price distortions. For the long-run cost of wage distortions he provided some estimates only for eleven three-digit industries and only for South Korea. With the exception of a few three-digit industries of the remaining industries, the cost of wage distortion was really small. For loss due to distortion in the cost of capital Gupta (1989) provided estimates, again, only for South Korea.

62

Factor price distortions

For the manufacturing sector as whole, South Korea suffered a loss of 3.65 per cent in terms of employment foregone.

Conclusion What general conclusions can be drawn from the above survey? First a caveat. Our empirical evidence is highly selective and therefore the question arises as to its representativeness. Keeping that in mind, it seems to us that factor price distortions ought not to be encouraged, but the extent to which they are harmful is still an open issue.

5

Growth, poverty and income distribution

We have already seen the growth performance of the countries in our sample. We now turn to their experience with respect to poverty reduction and reduction in income inequality. There has been much debate about the relationship between poverty and income inequality reduction and growth. Therefore it would be interesting to see what patterns emerge about these variables for the countries of East Asia. What do we do is consider first the experience with respect to poverty reduction. In this respect, we look at both intertemporal experience for a given country and inter-country experience for a specific period of time. Of course, it may be interesting to carry out a multi-dimensional comparison, but that may not always be possible.

Growth and poverty reduction In order to enable as much comparability as possible, both over time and across countries, we use international poverty line measures. These measures are two in nature: $1 a day and $2 a day. We do not have data over time. So, in this case, only inter-country comparisons are possible. Even from the limited data in Table 5.1, we can draw a number of interesting inferences. For one thing, confining ourselves to the recent period, it is clear that, by a somewhat more adequate poverty measure, there are still countries in the East Asian region which have large sections of the population which are poor. On the other hand, Indonesia and Thailand show that, by the $1 a day measure, the population below the international poverty line is very low. This is something to be applauded.

64

Growth, poverty and incomes

Table 5.1 Population below the international poverty line (%) Country Hong Kong Indonesia Japan Malaysia Philippines Singapore South Korea Thailand

Survey year 1999

1993 1998

Population below $1 a day

Population below $2 a day

— 7.7 — — — — >2.0 >2.0

— 55.3 — — — — >2.0 28.2

Source: Adapted from table 2.6, World Development Indicators on CD-ROM, 2001.

Table 5.2 Inequality in East Asia Country, period

Hong Kong, 1971–91 Singapore, 1973–89 Taiwan, 1985–95 South Korea, 1970–88 Malaysia, 1973–89 Thailand, 1975–92 Indonesia, 1970–95 Philippines, 1985–94

Measured variable

Income per household Income per household Income per household Income per household Income per capita Expenditure per capita Expenditure per capita Expenditure per capita

Gini coefficient First year

Last year

40.9 41.0 29.0 33.3 50.1 36.4 34.9 41.0

45.0 39.0 31.7 33.6 45.9 46.2 34.2 42.9

Source: Adapted from World Bank (1997a). Note Inequality is measured in averaged Gini coefficients.

Growth and income inequality Having looked at absolute poverty, it would be interesting to look at their experience with respect to income distribution. Has overall income distribution worsened? Have these countries become more egalitarian as their overall growth rates have gone up? Some interesting light is shed on these questions by Table 5.2. The most important implication of this table is that high growth rate and improved income distribution do not always go together.

Growth, poverty and incomes 65

Poverty reduction, growth and redistributive policies But more important, the above data on poverty and income distribution and the previous evidence on the growth experience make one wonder whether there is any relationship between poverty reduction, growth and redistributive policies. A priori, there is no reason to believe that high growth rates should necessarily lead to either a reduction in the number of poor people (regardless of how the poverty line is defined) or an improvement in the distribution of income. In fact, one could as plausibly argue that the benefits of high and rapid growth could as easily be concentrated in the upper tail of income distribution. To shed some light on the relationship between changes in poverty, redistributive policies and growth, we draw upon work by Datt and Ravallion (1992). They show that the change in poverty can be decomposed into that due to economic growth and the other due to distributional changes. Thus, we can write: Pt − Pt−1 = [P (Z/µt , Lt−1 ) − P (Z/µt−1 , Lt−1 )] + [P (Z/µt−1 , Lt ) − P (Z/µt−1 , Lt−1 )] + t

(5.1)

where P is poverty, Z is the poverty line, µ is mean expenditure, L is the Lorenz curve and is the residual. The first term, P (Z/µt , Lt−1 ) − P (Z/µt−1 , Lt−1 ), on the right-hand side of equation (5.1), represents the growth component, while the second, P (Z/µt−1 , Lt ) − P (Z/µt−1 , Lt−1 ), represents the redistribution component. The components can be measured first by assigning explicit functional forms and by assigning specific time periods to t and t − 1, and then proceeding as follows. The growth component can be estimated by calculating the change in poverty by fixing the Lorenz curve at the initial date t − 1 and letting mean expenditure grow. The redistributive effect can be estimated by varying the Lorenz curve and fixing the mean expenditure at the initial date t − 1. Ahuja et al. (1997) apply this decomposition methodology to four countries of our sample. The results are reproduced in Table 5.3, which makes it clear that, over the period covered, poverty declined in all four countries and the decline was primarily due to growth. In fact, in some cases, the redistribution policies contributed to increasing poverty and

66

Growth, poverty and incomes

Table 5.3 Decomposition of poverty reduction into growth and redistribution components Country

Period

Decline in head-count indexa

Growth component

Redistribution component

Residual

Malaysia Thailand

1973–89 1975–86 1986–92 1970–78 1978–84 1984–95 1985–88 1988–91 1991–94

19.1 −1.9 10.0 3.8 26.7 23.6 5.0 −1.2 1.7

16.4 6.1 10.0 7.6 18.5 22.4 5.2 2.9 0.8

3.9 −11.0 −1.5 −2.7 3.4 −3.1 −0.3 −4.1 1.0

−1.2 3.0 1.5 −1.1 4.8 4.3 0.1 0.0 −0.1

Indonesia Philippines

Source: Adapted from Ahuja et al. (1992), table 5.1. Notes Declines are denoted by positive entries, while negative entries denote contributions to increasing poverty. a Percentage points.

net outcome was a decline in the poverty index because the growth effect swamped the redistribution effect.

Conclusion These results once again raise the issue dealt with in Chapter 3, namely, if growth is the major reducer of poverty, then what caused high and sustained growth in the countries of our sample, and further, since growth is only a necessary and not a sufficient condition for reducing poverty, what sort of mechanisms operate between growth, income distribution and poverty? Policy issues come into play here. While there is much work in these areas and much speculative work, as one might expect, a good deal of the work bearing on these issues is not only speculative and amounts to ex post theorizing but in fact covers virtually all of economics. It would serve no useful purpose summarizing any specific piece or pieces of work. Instead we would merely urge the reader to keep in mind the diversity of experience noted above and then search for answers as they may seem relevant. The material presented in Chapter 3 provides a fairly detailed approach to how the reader may go about this task.

6

East Asian financial crisis

In July 1997 the currency and financial crises erupted unexpectedly in Thailand and gradually spread to most of the Association of Southeast Asian Nations (ASEAN) economies.1 Figure 6.1 documents the behaviour of the exchange rate (Asian currency in terms of the US dollar) in Indonesia, South Korea, Malaysia, the Philippines and Thailand from January 1997 to December 1999. In July 1997 the Thai baht came under attack. Within a few months, Indonesia, South Korea, Malaysia and the Philippines succumbed to the currency crisis. Figure 6.1 shows that in July 1998 the Thai baht, the Korean won, the Malaysian ringgit and the Philippine peso were reduced to 60 per cent of their January 1997 levels, and the Indonesian rupiah declined to 15 per cent. Many observers believe that liberalization of capital account, significantly increased capital flows, poor prudential supervision of banking systems and unproductive investments, all under a fixed exchange rate regime, together increased the vulnerability of East Asian economies dramatically in the first half of the 1990s. How well do these factors distinguish the crisis-affected economies in East Asia from other economies in the region that did not succumb to the financial and currency crisis of 1997? Table 6.1 highlights the differences between the five crisis economies as well as three non-crisis economies (China, Japan, and Taiwan) in East Asia on key economic variables.2 Inflexible exchange rates alone did not cause the crisis. Table 6.1 shows that China maintained an exchange rate that was pegged against the US dollar. Inadequate prudential regulation alone did not cause the crisis either. Table 6.1 indicates that prudential regulation in Malaysia and the Philippines was arguably better than that in Japan. In spite of this, both Malaysia and the Philippines

68

East Asian financial crisis 1.1 IDN KOR MYS PHL THA

Exchange rate: January 1997 = 1

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 1997

1998

1999

2000

Year

Figure 6.1 Exchange rates in crisis countries, January 1997–December 1999 (US$). Source: Figures derived from International Financial Statistics, available on CD-ROM.

succumbed to the Asian financial crisis, but Japan did not. Did the capital account liberalization alone cause the crisis? The answer is not clear either. Both China and Taiwan maintained restrictions on capital movements. But then so did South Korea and yet it was affected by the currency and financial crisis. Perhaps capital account liberalization combined with inflexible exchange rates and inadequate prudential regulation led to the financial crisis in 1997–1998. One economic variable which clearly separates the crisis countries from the non-crisis countries is the ratio of short-term debt to reserves. With the exception of Malaysia and the Philippines, the crisis countries had a very high short-term debtto-reserves ratio relative to those of non-crisis countries.3 Nevertheless, it is not clear from Tables 6.1–6.3 and Figure 6.2 exactly what caused the Asian financial crisis. Consequently, despite the fact that over seven years have passed since the outbreak of the Asian crisis, there is no consensus yet on what caused the crisis. However, two explanations seem to have gained currency: ‘fundamental distortions’ and ‘self-fulfilling prophecies’. According to the fundamentalist view, popularized by Corsetti et al. (1999), fundamental imbalances in the crisis-affected countries caused the Asian crisis.4 By contrast, the proponents of the self-fulfilling prophecies view,

CP MF MF MF PG PG FF MF

Indonesia South Korea Malaysia Philippines Thailand China Japan Taiwan

Y Y N N Y N N N

Significant accumulation of short-term foreign debt in relation to foreign exchange reserves P N Y Y Y N N N

Significant real exchange rate appreciation in 1996

Y P Y Y Y N Y P

Capital account liberalization

Notes CP crawling peg, FF free floating, MF managed float, PG peg, P partial, M moderate, L large, N no, Y yes.

Source: Adapted from Noble and Ravenhill (2000), table 1.1.

Exchange rate

Country

Table 6.1 Key economic indicators in East Asian economies

Y Y P P Y Y Y P

Inadequate prudential regulation

M M M M L N N N

Current account deficit in 1996

18,822 35,204 8,203 2,646 27,151

Short-term debta

June 1994

10,915 21,685 32,608 6,527 27,375

Reservesa

Notes a Million dollars. b Ratio.

Source: Adapted from table 3, Radelet and Sachs (1998a).

Indonesia South Korea Malaysia Philippines Thailand

Country

1.724 1.623 0.252 0.405 0.992

Debt-toreservesb

Table 6.2 Short-term debt and reserves in crisis countries, 1994 and 1997

34,661 70,612 16,268 8,293 45,567

Short-term debt a

June 1997

20,336 34,070 26,588 9,781 31,361

Reservesa

1.704 2.073 0.612 0.848 1.453

Debt-toreservesb

East Asian financial crisis 71 Table 6.3 Real exchange rate (year end), 1991–1997 Country

1991

1992

1993

1994

1995

1996

1997

Indonesia South Korea Malaysia Philippines Thailand Taiwan

99.6 91.5 96.9 103.1 99.0 95.7

100.8 87.7 109.7 107.1 99.7 95.7

103.8 85.2 111.0 97.4 101.9 91.4

101.0 84.7 107.1 111.7 98.3 92.6

100.5 87.7 106.9 109.6 101.7 90.4

105.4 87.2 112.1 116.4 107.6 89.6

62.4 58.6 84.9 90.9 72.4 89.2

Source: Adapted from table 17, Corsetti et al. (1999).

20

Current account (% of GDP)

15

IDN KOR MYS PHL THA

10

5

0

–5

–10 1990

1991

1992

1993

1994

1995 Year

1996

1997

1998

1999

2000

Figure 6.2 Current account balance in crisis countries, 1990–2000. Source: Figures derived from World Development Indicators, available on CD-ROM.

with Radelet and Sachs (1998a, b) being the leading ones, believe that the roots of the crisis lay in the inherent instabilities in international financial markets.5 In this chapter we briefly review the existing empirical evidence on the two explanations of the Asian financial crisis as well as present some new empirical evidence.

The nature and extent of the crisis The currency and financial crisis began in Thailand following a sudden reversal of net capital flows to the five crisis countries: Indonesia,

72


South Korea, Malaysia, the Philippines and Thailand. Table 6.4 documents the magnitude and suddenness of the financial reversal. Private net inflows to the five crisis countries increased from $40.5 billion in 1994 to $93.0 billion in 1996. This long period of inflow, however, unexpectedly reversed in 1997 with an estimated net outflow of $12.1 billion. This amounts to a reversal of capital flows of around $105.1 billion (from an inflow of $93.0 billion in 1996 to an outflow of $12.1 billion in 1997) and represents approximately 11 per cent of the pre-crisis dollar GDP of the five crisis countries. This remarkable and unexpected swing of capital flows triggered sharp drops both in the stock market and in the value of the baht against the US dollar. By early 1998 the crisis spilled over to Indonesia, South Korea, Malaysia and the Philippines. Figures 6.1 and 6.3, respectively, show the evolution of the value of the domestic currency in terms of dollars for the five crisis economies and of the stock prices from January 1997 to December 1998. With the exception of Indonesia, the short-term debt-to-reserve ratio in all the crisis countries increased over the 1994–1997 period. In particular, Table 6.2 shows that the debt-to-reserve ratio in three of the five crisis countries (Indonesia, South Korea and Thailand) exceeds unity. Another aspect of the Asian crisis is that, with the exception of South Korea, all the currencies that crashed in 1997 and early 1998 had experienced a real appreciation. This is apparent in Table 6.3. Between 1991 and 1996 the real exchange rate had appreciated by 6 per cent in Indonesia, 16 per cent in Malaysia, 13 per cent in the Philippines and 9 per cent in Thailand. Finally, the patterns of macroeconomic adjustment in the crisisaffected countries are worth noting. Table 6.5 shows that with the exception of the Philippines, GDP growth rates in the crisis countries were in excess of 7 and 8 per cent during 1970–1996 and 1990–1996, respectively. GDP growth, however, collapsed in 1998. Figure 6.3 shows that this setback was short-lived. GDP growth in the five crisis countries recovered strongly in 1999 and 2000. Table 6.4 indicates that portfolio investment flows into these countries also reduced drastically in 1997 but quickly recovered in 1998. Note, however, that foreign direct investment was much more stable than portfolio investments. In fact, Table 6.4 shows that foreign direct investment in the five crisis countries increased over the 1994–1998 period. Figure 6.4 shows that the stock markets in the crisis countries collapsed in 1998,

47.4 40.5 12.2 4.7 7.6

External financing (net) Private inflows (net) Equity investment Direct investment Portfolio investment

80.9 77.4 15.5 4.9 10.6

1995 92.8 93.0 19.1 7.0 12.1

1996

15.2 −12.1 −4.5 7.2 −11.6

1997

Notes Figures are sums over data for the five crisis countries: Indonesia, South Korea, Malaysia, the Philippines and Thailand.

Source: Adapted from table 1, Radelet and Sachs (1998a).

1994

Item

Table 6.4 External financing of the five crisis countries, 1994–1998 (US$ billion)

15.2 −9.4 7.9 9.8 −1.9

1998

74


Table 6.5 Annual output growth in East Asian countries, 1970–1998 (%) Country

1970–96

1970–89

1990–96

1997

1998

Indonesia South Korea Malaysia Philippines Thailand

7.3 7.9 7.5 3.6 7.7

7.1 8.0 6.8 3.9 7.4

8.0 7.6 9.5 2.8 8.6

4.7 5.0 7.3 5.2 −1.4

−13.1 −6.7 −7.4 −0.6 −10.5

Source: Authors’ calculations based on the data obtained from World Development Indicators, available on CD-ROM. 15 IDN KOR MYS PHL THA

GDP per capita growth rate

10

5

0

–5

–10

–15 1990

1991

1992

1993

1994

1995 1996 Year

1997

1998

1999

2000

Figure 6.3 GDP growth rates in crisis countries, 1990–2000. Source: Figures derived from World Development Indicators, available on CD-ROM.

but recovered sharply in 1999. In sum, the five crisis economies suffered a deep, sharp shock without lasting effects. Figure 6.5 shows that the 1997 crisis was largely unanticipated, although that many observers saw some danger in late 1996 – for example, the overvaluation of the baht. No one, however, anticipated the kind of financial meltdown that actually took place in East Asia.

Assessing the empirical evidence This section not only summarizes the existing empirical evidence on the two explanations of the financial crisis, but also, and more important,

2

IDN KOR MYS PHL THA

Stock price index: January 1997 =1

1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1997

1998

1999

2000

Year

Figure 6.4 Stock prices in crisis countries, January 1997–December 1999. Source: Figures derived from International Financial Statistics, available on CD-ROM.

80 IDN KOR MYS PHL

70

Country risk ratings

60

THA HKG JPN SGP

50 40 30 20 10 0 199301

199401

199501 199601 Year and month

Figure 6.5 Euro-money country risk ratings. Source: Figures derived from Radelet and Sachs (1998b), table 6.

199701

199801

76


provides new estimates about the controversy. In the discussion, we try to show why, in our view, the empirical estimates remain inconclusive. First, we consider the Corsetti et al. view that the Asian financial crisis reflects structural and policy distortions in the countries of the region.6 This view, Corsetti et al. (1999) suggest, is founded on their analysis of the empirical evidence on the five crisis countries – Indonesia, Malaysia, the Philippines, South Korea and Thailand – and four non-crisis countries: China, Hong Kong, Singapore and Taiwan. Their assessment of the fundamental imbalances in these East Asian countries identified a list of structural factors such as current account deficits and foreign indebtedness, growth and inflation rates, budget deficits, real exchange rates, excessive bank lending, credit growth and financial fragility, capital flows and political instability which may have caused the crisis. One of the weaknesses in the Corsetti et al. analysis of the Asian crisis is that they never formally confront their hypothesis with the data. Perhaps this is because there is virtually no theoretical literature on the fundamentalist school. We use the data provided in Corsetti et al. to estimate cross-country regressions, implied by the Corsetti et al. analysis, and formally investigate whether the fundamental imbalances were the root cause of the Asian crisis. Tables 6.6 and 6.7 report the estimation results. Corsetti et al.’s analysis does not specify a regression model. Instead, they provide a list of potential indicators of macroeconomic imbalances which may have caused the Asian crisis. To minimize the potential effect of this specification uncertainty on our findings, we estimate five different regression models. In each case, we regress a one-year forward percentage change (from period t to t + 1) in the nominal exchange rate on a sub-set of macroeconomic fundamentals listed in Corsetti et al. In our model, an increase represents a depreciation. Table 6.6 reports panel regression results. The panel consists of the nine crisis and non-crisis countries in East Asia, investigated over 1990–1997. Two sets of results emerge. First we find that factors such as short-term foreign debt as a percentage of total reserves, total foreign debt as a percentage of total reserves, current account balance, capital flows, fiscal balance, real exchange rate change, GDP growth and inflation rates are not significantly correlated with expected currency depreciation. In model 4, short-term foreign debt is significant at the 5 per cent level, but the regression coefficient appears with the wrong sign. Table 6.6 shows that the growth in bank lending to the private sector is significantly positively correlated with the expected exchange rate depreciation (see models 2, 3 and 5). Thus, the high growth in bank

East Asian financial crisis 77 Table 6.6 Fundamental imbalances and Asian crisis (crisis and non-crisis countries) Independent variable

Model 1

Short-term debt/reserves Total foreign debt/reserves GDP growth rate Inflation rate Bank lending growth

2

3

Current account surplus/GDP Real exchange rate change Interest rate spread

5 0.24 (1.53) −0.06 (−1.70) −0.65 (−0.42) 1.91 (1.59) 1.21 (2.06)a −4.28 (−1.56) −0.97 (−1.19)

0.12 (0.93)

0.14 (1.07)

−0.00 (−1.97)b

−0.82 (−0.49) 0.42 (0.58) 1.17 (1.58)

−0.50 (−0.31) 1.32 (1.14) 1.03 (1.74)b

−1.28 (−1.69)

−0.91 (−1.25)

−0.02 (−0.01) 1.76 (1.39) 1.02 (1.75)b −3.16 (−1.12) −0.72 (−0.93)

−0.75 (−0.51) 1.30 (1.36) 0.92 (1.60) 1.15 (1.20)

−0.67 (−0.90) 5.20 (2.09)a 1.66 (2.15)a

−0.62 (−0.84) 5.50 (2.21)a 1.61 (2.07)a

−0.09 (−0.12) −0.99 (−1.62) 5.68 (3.34)a 1.33 (2.20)a

0.43 42

0.43 42

0.43 56

0.18 (1.62)

Fiscal balance Capital inflow/GDP

4

Political instability

4.55 (2.10)a 2.25 (3.83)a

R¯ 2 No. of observations

0.42 42

−0.48 (−0.66) 7.04 (2.92)a 0.85 (1.18) 0.44 42

Source: Authors’ regressions. Data on all independent variables, except the interest rate spread, are from Corsetti et al. (1999). The interest rate spread measures the difference between the lending rate in a sample country and that in Japan. Data on lending rates are from the IMF, International Financial Statistics, available on CD-ROM. Notes The dependent variable is a one-year forward percentage change (from period t to t + 1) in nominal exchange rate. An increase represents a depreciation. The panel consists of the nine crisis and non-crisis countries in East Asia, investigated over 1990–1997. All regressions include a constant term. t-ratios, based on heteroskedasticity-consistent standard errors, are in parentheses. R¯ 2 is the adjusted R 2 . a Significance at 5%. b Significance at 10%.

lending may have played some role in the onset of the Asian crisis. This finding, however, is not robust. The regression coefficient is insignificant in models 1 and 4. Furthermore, the regression results reported in Table 6.7 show that growth in bank lending is never statistically significant. Second, we find that the interest rate spread and political

78


Table 6.7 Fundamental imbalances and Asian crisis (crisis countries) Independent variable

Model 1

Short-term debt/reserves

2

3

4

5

0.16 (1.24)

0.11 (0.59)

0.11 (0.60)

0.18 (1.09)

−1.48 (−0.73) 1.01 (0.36) 0.94 (1.19)

−0.90 (−0.38) 0.72 (0.30) 0.90 (1.29)

−0.73 (−0.95)

−0.63 (−0.71)

−0.37 (−0.14) 0.17 (0.06) 0.90 (1.29) −2.41 (−0.68) −0.64 (−0.74)

−0.93 (−0.37) 2.24 (0.70) 0.82 (1.26) −1.57 (−0.40)

Total foreign debt/reserves GDP growth rate Inflation rate Bank lending growth Fiscal balance Capital inflow/GDP Current account surplus/GDP Real exchange rate change Interest rate spread Political instability

5.48 (2.56)a 2.88 (3.06)a

R¯ 2 No. of observations

0.43 35

−0.47 −0.47 (−0.46) (−0.46) 5.72 6.05 (2.35)a (2.44)a 2.49 2.31 (2.91)a (2.42)a 0.42 35

0.40 35

0.20 (1.07) −0.07 (−1.91)b −0.49 (−0.21) 0.72 (0.28) 1.05 (1.53) −4.15 (−1.19) −1.15 (−1.33)

−2.67 (−1.38) −0.56 −0.30 (−0.58) (−0.30) 5.35 8.32 (2.26)a (3.38)a 2.67 1.35 (2.67)a (1.61) 0.41 35

0.42 35

Source: See Table 6.6. Notes See Table 6.6. The panel consists of the five crisis countries in East Asia, investigated over 1990–1997.

instability are significantly positively correlated with the onset of the Asian crisis. Table 6.7 reports the estimation results based on a panel of the five crisis countries over the period 1990–1997. The object of this exercise is to see whether our main findings reported above are robust. Our results show that the interest rate spread and political instability are the only two variables which significantly correlate with the onset of Asian crisis.7 Our regression results show no clear evidence in favour of the view that fundamental imbalances were the root cause of the Asian currency and financial crisis in 1997. Instead, our results suggest that the fact that banks in Japan could raise deposit funds at a low cost and lend them at a substantially higher rate in crisis-affected countries was

East Asian financial crisis 79 Table 6.8 New Probit regression results predicting the onset of financial crisis Independent variable

Model 1

Short-term foreign debt/reserves

2 0.31 (1.86)b

3 0.49 (1.60)

4 0.52 (1.62)

Total foreign debt/reserves

−0.00 (−0.28) 0.03 (0.70)

−0.70 (−1.93)a 0.01 (0.21) 0.06 (1.01)

Freedom from corruption Private credit buildup Capital inflow/GDP

0.01 (1.02) −1.87 (−3.17)a

−0.04 (−1.84)b 0.02 (1.25) −0.16 (−0.10)

−0.20 (−1.61) −0.04 (−1.86)b 0.02 (1.24) −0.77 (−0.44)

−0.04 (−1.84)b 0.02 (1.23) −0.69 (−0.04)

0.10 74

0.46 48

0.53 54

0.47 48

Current account surplus/GDP Real exchange rate change Interest rate spread Constant Pseudo R 2 No. of observations

−0.57 (−1.60) 0.01 (0.15)

0.54 (1.02) −0.02 (−0.11) −0.72 (−1.80)b 0.01 (0.20) 0.06 (0.97)

Source: Authors’ regressions. Data on short-term foreign debt, total debt, reserves, private credit, GDP, current account balance, and capital inflow are from the World Bank, World Development Indicators, available on CD-ROM; on lending rates from the IMF, International Financial Statistics, also available on CD-ROM; on real exchange rate from the J. P. Morgan RER series; and on corruption from Political Risk Services, a private firm that annually publishes the International Risk Guide. Notes As in Radelet and Sachs (1998a), the dependent variable is a dummy which takes value of 1 for a given country in year t if that country experiences a sharp reversal of capital flow: from an inflow of capital in year t − 1 to an outflow of capital in year t. The panel consists of the twenty-two countries investigated over 1994–1997 by Radelet and Sachs plus China and Hong Kong. Following Radelet and Sachs (1998a) we exclude a country from the panel in subsequent years after a country has experienced a crisis. All independent variables included in regressions are the same as in Radelet and Sachs (1998a), except the interest rate spread which measures the difference between the lending rates in the twenty-four countries in our sample and the lending rate in Japan. Asymptotic t ratios are in parentheses. a Significance at 5%. b Significance at 10%.

largely responsible for the crisis. This finding is reinforced by the Probit regression results reported in Radelet and Sachs (1998a) as well as in Table 6.8. Our results further suggest that political instability and political uncertainty in Asia also played a significant role in the onset of the Asian crisis. Thus, our findings provide support in favour of the view that a deterioration in expectations about the political and financial

80


environment can contribute to a balance of payments and exchange rate crisis.8 Corsetti et al. (1999) detail the extent of the political instability and political uncertainty in Asia. Radelet and Sachs provide a formal test for the self-fulfilling expectations hypothesis. First, they find that a higher ratio of short-term foreign debt to reserves is strongly positively correlated with the onset of a crisis. Furthermore, Radelet and Sachs suggest that this finding is robust because the estimated coefficient of the ratio of short-term foreign debt to reserves is significant at the 5 per cent level in each of the four different specifications of the model they estimate. Second, they find that private credit build-up is also significantly positively correlated with the onset of a crisis. Finally, they find that the total debt to reserves ratio, capital inflow to GDP ratio, current account surplus to GDP ratio, real exchange rate change and the level of corruption are not significantly correlated with the onset of financial crisis. Radelet and Sachs suggest that these results provide evidence in favour of the hypothesis that the East Asian financial crisis was primarily caused by the inherent instabilities in international capital markets. Radelet and Sachs (1998a) note that the variables used in their regression models are subject to measurement error. Consequently, their estimation results may be biased. There are two additional reasons why one should treat the Radelet and Sachs findings with caution. Barry Bosworth observes that banks could raise deposit funds with a 1 per cent deposit rate in Japan and lend at 10 per cent to South Korea, for example.9 He suggests that this huge interest rate spread may have been a significant inducement to the expansion of capital inflows into the crisis-affected countries in East Asia. The estimation results reported in Radelet and Sachs (1998a) ignore the role which the incredibly low interest rates in Japan may have played in the East Asian crisis. Christopher Sims noted that a shift in the maturity of debt before the crisis was endogenous and therefore it is inappropriate to include, as Radelet and Sachs (1998a) do, the amount of short-term debt as an exogenous variable in regressions.10 The estimation results reported therefore are likely to be biased. We re-estimate the Radelet and Sachs regressions to see whether the low interest rate in Japan was a contributing factor. Table 6.8 reports our estimation results.11 Data on short-term foreign debt, total debt, reserves, private credit, GDP, current account balance and capital inflow are from the World Bank World Development Indicators, available on CD-ROM; on lending rates from the IMF, International Financial Statistics, also

East Asian financial crisis 81 available on CD-ROM; on real exchange rate from the J. P. Morgan RER series; and on corruption from Political Risk Services Inc., a private firm that annually publishes the International Risk Guide. All right-hand-side variables included in the four different versions of the Probit regression are the same as in Radelet and Sachs (1998a) with the exception of the new independent variable, interest rate spread, which measures the difference between the lending rates in the twenty-four countries in our sample and the lending rate in Japan.12 A comparison of our regression results reported in Table 6.8 with those in Radelet and Sachs (1998a), table 12 ( p. 45), indicates that the Radelet and Sachs results are reversed when the interest rate spread is included in their Probit regressions. In particular, we find that neither the ratio of short-term foreign debt to reserves nor the private credit build-up is significantly correlated with the incidence of currency crisis.13 We also find, as do Radelet and Sachs, variables such as total debt to reserves ratio, capital inflow to GDP ratio and current account surplus to GDP ratio are not significantly associated with the onset of financial crisis. More important, we find that the real exchange rate change, a measure of exchange rate overvaluation, is strongly associated with the onset of a crisis. The estimated regression coefficient is negative, as expected, and significant at the 5 per cent level in a specification that includes the real exchange rate change. This is contrary to the results reported in Radelet and Sachs. They themselves, however, were surprised by their result that real exchange rates were not a significant factor in the financial crisis.14 Our findings suggest that this may have been caused by the exclusion of the interest rate spread from the regressions. Many studies of crisis episodes emphasize that a significant real exchange rate appreciation may result in a loss of competitiveness and a structural worsening of the trade balance, hence making a current account position unsustainable.15 Thus, real exchange rate appreciation could signal a crisis. Our results therefore provide evidence supporting this view. Finally, contrary to Radelet and Sachs (1998a), we find that after controlling for the level of short-term debt, bank credit and other variables, the level of corruption is significantly associated with financial crises. Table 6.8 shows that in each specification, the estimated coefficient on the level of corruption is of the correct sign and is significant in two out of three specifications. Our finding, however, is consistent with those of Wei (2001). As noted in Wei (2001), corruption increases the likelihood of a currency crisis by altering the composition of the capital inflows into a corrupt country. A high level of corruption reduces FDI

82


and increases international bank credits.16 Using data on bilateral bank loans and bilateral FDI during 1994–1996, Wei finds that more corrupt countries tend to have a lower ratio of FDI to bank credit. To sum up, it is apparent from Table 6.8 that our findings are strikingly different from those of Radelet and Sachs (1998a). This discrepancy arises primarily from the fact that Radelet and Sachs exclude the interest rate spread, an important explanatory variable, from the Probit regressions they estimate. The empirical evidence they provide in favour of the self-fulfilling hypothesis is therefore unconvincing. Jeanne (1997) presents a model of currency crisis in which selffulfilling speculation can arise following a bifurcation in the fundamentals. Thus, in such a model, a currency crisis may arise either from a deterioration of the fundamentals or from a sudden jump in the devaluation probability for a given level of the fundamentals. In the later case, the currency crisis is caused by the self-fulfilling speculation. Jeanne estimates the model in the case of the 1991–1993 French franc crisis and finds that while the different episodes of the crisis were associated with bad fundamentals, self-fulfilling speculation made these episodes considerably worse. Test results reported above show that inferences drawn with regards to the fundamentalist hypothesis and the self-fulfilling expectations hypothesis are not robust. This is because the tests implemented above are weak. In our view the test procedure developed in Jeanne is a superior one in that it nests both the fundamentalist and self-fulfilling expectations hypothesis. However, we do not implement the Jeanne test. But in our view it would be worthwhile implementing the test procedure developed by Jeanne for each of the crisis-affected countries in East Asia to see whether self-fulfilling speculation is the principal cause of the Asian crisis.

Conclusion The debate on what caused the Asian financial and currency crises has had important implications for policy prescription. According to the fundamentalists, the root causes of the crises lay in misguided economic policies that the governments in East Asian economies pursued for many years. Close government ties with business elites gave rise to regulation based on relations rather than rule. This led to moral hazard problems that encouraged business to engage in unduly risky ventures. Thus, the proponents of the fundamentalist view of the Asian crisis prescribe

East Asian financial crisis 83 that the system of relational banking and close ties between the state and private business must be dissolved to eliminate the moral hazard problems. The fundamentalists further argue that the problems of the crisis countries arose from the fact that the financial systems in these countries are not liberal enough. Further liberalization of financial systems is therefore a necessary step towards avoiding future financial and currency crises. By contrast, the advocates of the self-fulfilling prophecies view argue that the crisis arose because of premature liberalization of capital markets and speculative activities. The policy implications are international surveillance of the activities of hedge funds and measures to penalize short-term international capital flows. In spite of all the debate and controversy over the root causes of the crises, we cannot say for sure what caused them and how best to avoid such crises in future. The self-fulfilling prophecies view of the Asian crisis, however, is favoured for two reasons. First, ex ante fundamentals in the crisis-affected countries were not too far off when compared with the recent history of the crisis economies or other similar economies. In particular, the crisis-affected countries had pursued sensible macroeconomic policies and outward-oriented development policies for some time. Second, the pattern of widespread contagion observed during the Asian crisis lends weight to the self-fulfilling prophecies view. We note, however, that although the self-fulfilling prophecies view appeared to be strengthened by the Asian crisis, the interaction of the financial markets and governments is still not well understood. Consequently, the self-fulfilling prophecies view falls far short of offering reforms that would dampen the wide swings in confidence and capital flows that characterized the 1990s, and prevent future financial crises. While there is little consensus on what caused the crisis, many agree that the financial crisis in East Asia has been short-lived and the economic slowdown has been temporary. The speed with which the affected countries recovered from the economic slowdown is a clear indication that these countries have matured and are on their way to join the industrialized nations.

7

Corruption, growth and foreign direct investment

It is widely accepted that corruption has large economic costs. Mauro (1995) and Kaufmann et al. (1999), for example, find strong cross-country evidence that higher levels of corruption are associated with lower growth and lower levels of per capita income.1 Corruption also adversely affects foreign direct investment.2 According to the World Development Report 2002, corruption is a force that undermines well-functioning markets in three ways: as a tax, as a barrier to entry, and by upsetting the legitimacy of the state and its ability to provide institutions that support markets. Corruption, as a tax, distorts the choice between activities and lowers the returns to public and private investments. Unfortunately, corruption is much worse than a conventional tax because the revenues generated through corruption do not result in public spending on socially useful activities. Moreover, as noted in Shleifer and Vishny (1993), since corruption is illicit, there is much greater uncertainty over this form of ‘taxation’ than conventional forms, making the corruption tax even more costly. Wei (1997, 2000a) study the impact of corruption on FDI, and find that an increase in corruption comparable to the difference between Singapore (which has a very low level of perceived corruption) and Mexico (which has an average level of perceived corruption) would have the same negative effect on FDI as a fifty percentage point increase in marginal tax rates on foreign investment income. Fisman and Svensson (2000) study the effect of corruption on manufacturing firm growth in Uganda. They find that a 1 per cent increase in bribes paid by a firm led to a 3 per cent reduction in firm growth, while a 1 per cent increase in taxation reduced firm growth by only about 1 per cent. Finally, evidence from transition economies suggests that firms would be willing to pay significantly higher formal taxes if corruption were eradicated.

Corruption

85

Corruption also undermines competitive forces by creating barriers to entry into markets. A robustly competitive environment requires the continuous entry of new firms. But when potential new firms must pay bribes at every turn in order to register and begin operations, many will decide simply not to enter a market, making the market less competitive. Based on the evidence from transition economies, Tanzi and Davoodi (2001) find that the competition deterrent effect of corruption is important in transition economies and that small firms and new entrants were significantly more likely to report corruption as an obstacle to business. Mauro (1997) and Gupta et al. (2001) find that corruption is associated with lower public spending on health and education, which in turn limits opportunities for poor people to invest in their human capital and to participate in markets. Across countries, greater corruption is also associated with lower overall tax revenues. Finally, corruption subverts the legitimacy of the state itself and weakens the capacity of the state to provide institutions that support markets. A particularly destructive form of corruption, known as ‘state capture’ enables firms to overturn the entire political process to ensure that policies and regulations favourable to their business interests are implemented. This phenomenon of state capture has been studied most systematically in the transition economies of Eastern Europe and the former Soviet Union. However, the close and questionable links between businesses and governments are not unique to this region, nor are they unique to democratic systems. For example, Fisman (2001) finds that one-quarter of the value of the firms in Indonesia with close links to the Suharto regime was directly attributable to their political connections. In this chapter we re-examine the economic costs as well as consequences of corruption that are documented in the literature. Given the high costs of corruption, researchers and policy makers have increasingly focused on identifying the causes of corruption; we therefore also provide a brief survey of this literature. As well, we provide some new evidence on the effects of corruption on growth and foreign direct investment.

The nature and extent of corruption in East Asia Classical theoretical works such as Nye (1967), Rose-Ackerman (1975, 1978) and Shleifer and Vishny (1993) develop a rich notion of corruption. For example, Rose-Ackerman (1978) emphasizes the

86

Corruption

distinction between bribery to erect or to change the rules or laws to favour the payers and bribery to deviate from honest implementation of the existing rules or laws. Shleifer and Vishny (1993), by contrast, distinguish between organized (or efficient) corruption3 and disorganized (or inefficient) corruption.4 The survey-based corruption indices frequently used in the existing empirical literature do not capture the conceptual richness developed in the works of Nye, Rose-Ackerman and Shleifer and Vishny. Nevertheless, following the existing literature, we use survey-based corruption indices to analyse the nature and extent of corruption in the economies of East Asia. There are several measures of corruption currently available. The PRS index is one such index.5 This index measures corruption within the political system that threatens foreign investments by distorting the economic and financial environment, reducing the efficiency of government and business by enabling people to assume positions of power through patronage rather than ability, and introducing inherent instability into the political process. The corruption index ranges from 0 to 6 – 0 being the most corrupt and 6 being the least corrupt. The Transparency International (TI) is another corruption index used in the empirical literature. This survey-based index reflects the degree of corruption perceived by business people, risk analysts and the general public. The TI index ranges between 0 (highly corrupt) and 10 (least corrupt). The Business International (BI) index is a subjective measure of corruption across countries devised by the ‘in-house experts’ at Business International, a consulting firm. The BI index reflects the degree to which business transactions involve corruption or questionable payments. This index was first used in Mauro (1995); the index in this paper ranges from 1 (most corrupt) to 10 (least corrupt). Treisman (2000) uses an adjusted BI index (1 indicates least corrupt and 10 indicates most corrupt), in order to align rankings with the PRS and TI indices. Table 7.1 compares various measures of corruption for the eight East Asian countries. It follows from this table that, despite the differences in methods used to compile the indices, the three indices are highly correlated, suggesting similar rankings vis-à-vis the extent of corruption in the eight East Asian countries. The sample correlation coefficient between the PRS and TI indices is 0.83 and between the PRS and BI indices is −0.74. We use the PRS series because it is available for a longer period of time (1984–2004). Tables 7.2 and 7.3 summarize the nature and extent of corruption in the East Asian countries and the G-7 countries, respectively,

Corruption

87

Table 7.1 A comparison of various indicators of corruption in East Asia Country

PRS index 2000

TI index 2000

PRS index 1984

BI index 1980–83

Hong Kong Indonesia South Korea Malaysia Philippines Singapore Taiwan Thailand

3.3 1.9 3.0 3.0 2.4 4.0 2.1 2.3

7.7 1.7 4.0 4.8 2.8 9.1 5.5 3.2

5.0 1.0 2.8 5.0 0.0 6.0 4.0 3.0

2.0 8.5 4.3 4.0 5.5 0.0 3.3 8.5

Source: PRS index is obtained from the Political Risk Services Group, 1979–2004; TI index is from Lambsdorff (2000); and BI index is from table A.3, Treisman (2000). Notes The PRS index is a measure of corruption within the political system that is a threat to foreign investment by distorting the economic and financial environment, reducing the efficiency of government and business by: enabling people to assume positions of power through patronage rather than ability, and introducing inherent instability into the political process. The corruption index ranges between 0 (highly corrupt) and 6 (least corrupt). The TI index relates to perceptions of the degree of corruption as seen by business people, risk analysts and the general public. The TI index ranges between 0 (highly corrupt) and 10 (least corrupt). The BI index reflects the degree to which business transactions involve corruption or questionable payments. The index ranges between 0 (least corrupt) and 10 (most corrupt). The BI index reported in Treisman (2000) is the adjusted BI index, compiled by Paolo Mauro. The BI index used in Mauro (1995) ranges between 1 (most corrupt) and 10 (least corrupt). The sample correlation coefficient between the PRS and BI indices is −0.74 and that between the PRS and TI indices is 0.83.

Table 7.2 Corruption in East Asia, 1984–2004 Country

Mean corruption index

Maximum corruption index

Minimum corruption index

Hong Kong Indonesia South Korea Malaysia Philippines Singapore Taiwan Thailand

4.4 1.4 3.2 3.7 2.2 4.5 3.5 2.6

5.0 3.0 5.0 5.0 4.0 6.0 4.0 3.0

2.3 0.0 2.0 2.5 0.0 4.0 2.1 1.5

Source: Authors’ calculations based on data provided by the Political Risk Services Group, 1979– 2004.

from 1984 to 2004. Over this period, the average index of corruption in the eight East Asian countries is only 3.2. By contrast, the average corruption index in the G-7 countries for the same sample period is 4.7 (see Tables 7.2 and 7.3). Thus, the average level of corruption in the

88

Corruption Table 7.3 Corruption in G-7 countries, 1984–2004 Country

Mean corruption index

Maximum corruption index

Minimum corruption index

Canada France Germany Italy Japan United Kingdom United States

5.8 4.5 5.0 3.5 4.3 5.1 4.7

6.0 6.0 6.0 4.0 5.0 6.0 5.4

4.3 3.0 4.0 2.5 2.1 4.5 3.9

Source: Authors’ calculations based on data provided by the Political Risk Services Group, 1979–2004.

East Asian countries is much higher than the average level of corruption in the G-7 countries. Figures 7.1 and 7.2 show the trend of corruption for the eight countries in East Asia as well as for the G-7 countries during the period 1984–2004. They indicate that corruption remains a serious problem in most countries in East Asia, even following the economic crisis of 1997. Many believe that corruption played a crucial role in the currency crisis that started in Thailand with the collapse of the Thai baht.6 Some also feel that corruption was a factor in the overthrow of the Suharto government in Indonesia, the bankruptcies of the chaebols (family-owned large conglomerates) in Korea and social unrest in Malaysia. However, although the economic crisis led to major reforms of the financial systems, the crisis did not apparently serve as a wake-up call to the (perceived) problem of corruption. As Figure 7.1 indicates, the problem of corruption has not improved significantly following the Asian crisis. In most cases it is worse. Among the eight East Asian countries, Hong Kong and Singapore are the only countries where the perceived level of corruption is lower than the average level of corruption in East Asia.

Determinants of corruption The World Development Report, 1997 identified three primary causes of corruption: a distorted policy environment, which creates greater opportunities for public officials to manipulate rules for their own benefit; a weak judiciary that is unable to provide a credible threat of punishment when official misconduct is discovered; and poor civil service management and low public-sector pay. Subsequent research

Corruption

89

6 CHN HKG IDN MYS PHL SGP KOR TWN THA

5

Corruption index

4

3

2

1

0 1980

1985

1990

1995

2000

2005

Year

Figure 7.1 Corruption in East Asian countries, 1984–2004. Source: Figures derived from Political Risk Services, 1979–2004.

has highlighted additional factors contributing to corruption. These are discussed below. Economic factors A growing number of studies find that countries more open to international trade have lower corruption.7 This may reflect a combination of factors which are at work. First, greater openness induces more

90

Corruption 6 CAN FRA DEU ITA JPN GBR USA

5.5

5

Corruption index

4.5

4

3.5

3

2.5

2 1990

1992

1994

1996 1998 Year

2000

2002

2004

Figure 7.2 Corruption in the G-7 countries, 1990–2004. Source: Figures derived from Political Risk Services, 1979–2004.

competition. A more competitive environment lowers economic rents and hence reduces opportunities for corruption. Second, greater openness improves information flows. This, in turn, helps expose official wrongdoing and also create constituencies in support of anti-corruption activities among trading partners abroad. Third, countries that are naturally more disposed to trade because of favourable geographic

Corruption

91

characteristics will invest greater resources in developing institutions that make trade more attractive. Finally, Gatti (2001) finds that as countries dismantle formal tariff barriers to trade, opportunities for corruption decrease. A more complex regulatory environment breeds corruption. For example, Djankov and Murrell (2002) find that countries with more elaborate procedures for registering new businesses have higher levels of corruption. This finding reflects the fact that complex regulations increase opportunities for corruption. The Djankov and Murrell finding may also suggest that corrupt bureaucrats will favour the proliferation of rules and regulations that in turn create further opportunities for corruption.8 In any event, the more complex the rules, the greater the likelihood that officials will have discretion in how the rules are applied, creating opportunities for corruption. Cross-country evidence shows that, controlling for a variety of other factors, corruption is significantly greater in countries where inflation is high and variable: information about prices is difficult to obtain, creating greater opportunities for corruption in public procurement. Another factor which may also affect the level of corruption is publicsector pay. Empirical evidence on the relationship between public-sector pay and the level of corruption is, however, less clear-cut. There is plenty of informal evidence that the low wages available to civil servants in many developing countries drive them to take bribes in order to supplement their incomes. While at least one study has found systematic cross-county evidence that higher corruption is associated with lower wages in a sample of twenty eight countries, other studies covering more countries fail to establish the same association.9 Furthermore, many of these studies do not distinguish between countries where petty corruption (which is more likely to be influenced by salaries) and grand corruption (which is less likely to be influenced by salaries) are important. Careful country-specific studies are beginning to provide additional evidence about the relative importance of wages and other factors in corruption. For example, Di Tella and Shargrodsky (2000) study procurement contracts in public hospitals in Buenos Aires, Argentina, and find that a 10 per cent increase in the salary of procurement officers was associated with a 1.2 per cent reduction in prices paid for hospital supplies. However, this relationship between pay and performance emerged only after a crackdown on corruption had been in effect for a period of six months. The campaign against corruption also had significant effects on procurement prices, initially lowering them by an average of 18 per cent.

92

Corruption

This effect, however, weakened over time. It is worth noting that the crackdown on corruption achieved significant results without the threat of penalty for wrongdoing. Sociopolitical factors Treisman (2000) surveys various theories of the causes of corruption. He identifies a number of economic and sociopolitical factors which determine a country’s level of corruption.10 Using three annual indices of perceived corruption (for 1996, 1997 and 1998) prepared by Transparency International, Treisman tests various theories of the causes of corruption.11 To that end, he estimates multiple regressions with variables capturing a broad range of theoretically plausible determinants. Treisman’s main findings are summarized in Table 7.4.12 First, Treisman finds, as do La Porta et al. (1999), that countries with a Protestant tradition and those with more developed economies have higher-quality governments. In addition, according to the estimation results reported in Table 7.4, both factors are significantly and robustly associated with a lower level of perceived corruption. As well, the estimation results suggest that the level of perceived corruption is robustly negatively correlated with a country’s colonial past. In particular, the evidence indicates that countries with a history of British rule were robustly rated as less corrupt. Third, federal states were more ‘corrupt’ than unitary ones. Treisman suggests that the competition between autonomous levels of government to extract bribes may lead to ‘overgrazing of the commons’. Finally, Treisman further finds that, while the current degree of democracy is not a significant determinant of the perceived level of corruption, a long period of exposure to democracy leads to a lower level of perceived corruption. The empirical findings we surveyed provide insight into why fighting corruption in many countries has proved so difficult. Treisman (2000) finds that the past is as important as – or more important than – current policy in determining the level of corruption. Democratization has to be radical and long-lived and trade liberalization has to be extensive in order to reduce corruption significantly. One slightly more hopeful finding, however, emerges from the empirical work on the determinants of corruption. Even though corruption hinders growth, countries can at times grow their way out of corruption. The evidence suggests that if other factors lead to vigorous economic development, corruption is likely to decrease.

Corruption

93

Table 7.4 Determinants of perceived corruption Variable

Model 1

Common law system

0.65 (1.32) Former British colony −1.99a (1.01) Never a colony −2.63c (0.86) Per cent Protestant −0.04c (0.01) Ethnolinguistic division 0.00 (0.01) Fuel, metal & mineral exports 0.02b (0.01) Log GDP per capita

2

3

4

5

0.64 (0.71) −1.56b (0.59) −0.53 (0.43) −0.03c (0.01) −0.01 (0.01) 0.01 (0.01) −4.54c (0.55)

0.39 (0.60) −1.25b (0.52) −0.11 (0.37) −0.03c (0.01) −0.01 (0.01) 0.01 (0.01) −4.25c (0.61) 0.51a (0.30) −1.02b (0.48)

0.43 (0.61) −1.20b (0.54) −0.12 (0.39) −0.03c (0.01) −0.01 (0.01) 0.01 (0.01) −4.16c (0.67) 0.48 (0.30) −1.12b (0.52)

0.45 (0.65) −1.04 (0.79) −0.07 (0.34) −0.02b (0.01) −0.01 (0.01) −0.00 (0.01) −4.84c (1.00) 0.56b (0.26) −1.04 (0.63)

64 0.89

64 0.89

62 0.89

36 0.93

Federal Democracy No. of observations R2

64 0.58

Source: Adapted from table 2, Treisman (2000). Notes The dependent variable is the TI index for the year 1998. White heteroskedasticity corrected standard errors in parentheses. Treisman includes several other conditioning variables in the regression models 4 and 5. The estimated regression coefficients of these additional conditioning variables are statistically insignificant and therefore are not reproduced in this table. All regression models include a constant term which is significant in all the regressions at the 1% level. Treisman (2000) provides further details. a p value less than 0.10. b p value less than 0.05. c p value less than 0.01.

Corruption and growth In this section we provide a brief review of the existing empirical works on the effects of corruption on growth. In addition, we provide some new evidence on the empirical relationship between corruption and growth. The empirical evidence reviewed, as well as the new evidence presented below, suggests that the debate over the effects of corruption on growth is far from over.

94

Corruption

Empirical evidence: a review There is a growing empirical literature devoted to quantifying the effect of corruption on growth. Bardhan (1997) provides an extensive survey of this literature. However, these studies have not yet provided any empirically conclusive evidence on the quantitative relationship between the level of corruption and the rate of economic growth. Mauro (1995) conducted the first empirical study of the relationship between corruption and economic growth. Within a standard growth regression, he embedded the BI index – a subjective measure of corruption across countries – and found that countries perceived to be more corrupt tend to grow more slowly. Méndez and Sepúlvea (2001) provide cross-country evidence in favour of the ‘efficient grease’ hypothesis which states that firms that pay more bribes face less red tape and have better access to public funds and bank loans (hence reducing their cost of capital). Following Mauro (1995), Méndez and Sepúlvea also augmented the standard growth regression by including a measure of corruption and corruption squared. However, instead of using the BI index, they use the PRS index. The main findings of Méndez and Sepúlvea are reproduced in Table 7.5. Contrary to Mauro (1995), Méndez and Sepúlvea find evidence in favour of the bribery-as-grease hypothesis. In particular, they find corruption

Table 7.5 Corruption and growth, 1960–1992 Variable

Model 1

Model 2

Corruption

0.00260 (2.2390)

0.01210 (3.8870) −0.00080 (−3.2080) 0.00003 (0.0060)

Corruption squared Asia

0.00340 (0.5010)

No. of observations R2

35 0.809

35 0.868

Source: Adapted from table 1, Méndez and Sepúlvea (2001). Notes All results obtained with ordinary least squares regression. Méndez and Sepúlvea include several other conditioning variables in their regression models. The estimated regression coefficients of these additional conditioning variables are suppressed in this table. See Méndez and Sepúlvea (2001) for further details. Numbers in parentheses are t ratios.

Corruption

95

to be growth-enhancing at low levels of incidence and growth-reducing at high levels of incidence. This evidence, Méndez and Sepúlvea note, implies there is a positive level of corruption that maximizes long-run growth. Méndez and Sepúlvea note that their findings can be explained in a simple capital accumulation model where bribery allows investors to avoid regulations imposed by the government and where the bribe is decided via a bargaining process. In a model of this genre, when the number of corrupt officials is low, paying a bribe allows the investor to bypass regulations at a relatively lower cost. As the practice of bribery becomes widespread, and corrupt officials gain bargaining power, bribes become expensive and frequent enough to impose a net burden on the investor. A potential problem with studies based on cross-country regressions, Wei (2000c) points out, is that many factors in a country are correlated, so it may be difficult to disentangle the effects attributable solely to corruption. For example, corruption may be highly correlated with the poor quality of public servants, a factor that may retard growth whether corruption exists or not. According to Wei (2000c), one way to overcome this problem is to examine evidence at the firm level. Kaufmann and Wei (1999) use three worldwide firm surveys to re-examine the ‘efficient grease’ hypothesis. To guide their analysis, Kaufmann and Wei construct a simple model where the bureaucrat who takes the bribe also sets the red tape.13 Kaufmann and Wei distinguish between nominal bureaucratic harassment (for example, statutory tax and tariff rates, or the waiting time for a permit without bribery) and effective harassment (for example, actual tax and tariff rates, or queuing time after the firm pays a bribe).14 The Kaufmann and Wei model predicts that, in equilibrium, the bureaucrat would impose just enough nominal harassment to obtain the maximum bribe without inducing the firm to leave the country. Consequently, nominal harassment and bribes are correlated positively across firms. Furthermore, the model suggests that for firms with weak outside options (and hence more willingness to tolerate higher bribe demands), nominal harassment can be sufficiently high that in equilibrium across firms, effective harassment and bribes are correlated positively. This is contrary to the efficient grease hypothesis. Using firm-level data obtained from three sources,15 Kaufmann and Wei (1999) find that across firms the measures of effective red tape are correlated positively with bribery. They observe that the positive

96

Corruption

correlation could be attributable to response biases that are correlated positively across the survey questions: measures of harassment and bribery used in Kaufmann and Wei (1999) are all based on subjective survey responses. Consequently, Kaufmann and Wei construct a measure of possible perception bias at the individual firm level based on the firm’s rating of the provision of public goods and services that are arguably identical across firms – that is, a rating of overall infrastructure, power supply and mail delivery. When such measures of perception bias are included in regressions of effective harassment on bribery, Kaufmann and Wei find that the coefficients on bribery generally become smaller. The estimated coefficients on bribery, however, remain positive and statistically significant. Using a unique survey of firms in Uganda that includes direct information on the monetary value of bribes, Svensson (1999) demonstrates that bribery tends to rise with firms’ profitability and to decline with reversibility of investment. These findings are consistent with the predictions of the Kaufmann and Wei model as well as the empirical findings of Kaufmann and Wei (1999). These results do not imply that, in a generally corrupt environment, an individual firm can necessarily do better by paying fewer bribes. Instead, they suggest that measures which increase all firms’ ability to resist bribe demands may not only reduce bribes but also reduce red tape. New evidence of the effects of corruption on growth Existing studies of the empirical relationship between growth and corruption commonly rely on the framework of single cross-country regressions. In this framework, however, it is econometrically difficult to allow for country-specific factors. Recent empirical growth literature finds that country-specific factors are important in explaining crosscountry growth experiences, and the omission of such factors from growth regressions can result in an omitted variable bias. The panel data approach, first implemented by Islam (1995) in growth empirics, can overcome this omitted variable bias. In this sub-section we therefore re-examine the empirical relationship between corruption and growth in a panel framework. We implement the panel data approach by dividing the study period 1984–2003 into five data points for each country: each data point spans a four-year period. Consequently, all variables (with the exception of GDP per capita) are averages over the periods 1984–1987, 1988–1991, and so on.

Corruption

97

Data on GDP per capita growth, GDP per capita, population growth, shares of investment and government consumption spending in GDP, secondary school enrolment, inflation rate and openness (measured by trade, in percentage of GDP) are obtained from the World Development Indicators, 2004. The index of corruption used ranges from 0 to 6 – 0 being the most corrupt and 6 being the least corrupt – and is provided by the Political Risk Services Group, 1979–2004. Table 7.6 reports the estimation results obtained by implementing a pooled regression (OLS). Our main finding is that the perceived level of corruption is not significantly negatively correlated with growth. The evidence suggests that after accounting for factors such as GDP per capita, investment share in GDP, government consumption spending share in GDP, human capital, and political instability, there is no residual growth which can be explained by the level of perceived Table 7.6 Corruption and growth revisited, 1984–2003 Variable

Model 1

Model 2

Corruption

−0.11568 (−0.8484) 0.03912 (0.0642) 0.13670 (1.4470) −0.00009 (−2.8270) 0.15088 (6.1520) 0.01567 (2.1250) −0.08782 (−3.2970) −0.00100 (−5.7080) 0.00338 (1.1400) −0.07866 (−4.9800)

−0.24576 (−0.6107) 0.01965 (0.3680) 0.03159 (0.0520) 0.13424 (1.4100) −0.00009 (−2.8240) 0.15135 (6.2100) 0.01546 (2.1020) −0.08730 (−3.2770) −0.00102 (−5.4620) 0.00347 (1.1670) −0.07900 (−4.9700)

512 0.509

512 0.568

Corruption squared East Asia Population growth Initial GDP per capita Investment Secondary school enrolment Government consumption spending Inflation Openness Political instability No. of observations R2

Notes All results obtained with ordinary least squares regression. Numbers in parentheses are t ratios, based on heteroskedasticity-consistent standard errors. All regressions include a constant term whose estimate is not reported.

98

Corruption

corruption. Furthermore, we find no evidence that over the sample period 1984–2003 the growth experience of the East Asian countries was adversely affected by the higher than average level of corruption observed in these countries.

Corruption and foreign direct investment In this section, we assess the existing empirical evidence as well as provide new evidence on three questions regarding the effect of corruption on foreign direct investment (FDI).16 First, does an increased level of perceived corruption deter inward FDI? Second, does East Asia constitute a special group of host countries in that corruption has no or a reduced effect on inward FDI in East Asia? Third, is there any deterrent effect of the United States Foreign Corrupt Practices Act (FCPA)? Our assessment of the existing empirical evidence as well as the new evidence we provide suggest that the empirical relationship between corruption and FDI is fragile. Inference about the effect of an increased level of corruption on FDI is sensitive to the specification of the regression model estimated and/or the data used. Empirical evidence: a review The first question concerning the effect of corruption on FDI is partly motivated by the claim that China has rampant corruption.17 Yet, of developing countries, China has been the largest recipient of international investment. Even its FDI flow-to-GDP ratio has been the highest among developing countries. Several empirical studies have failed to find a significant negative relationship between corruption and inward FDI. For example, in a study of US firms’ foreign investment, Wheeler and Mody (1992) do not find a significant correlation between the size of FDI and a host country’s risk factor, a composite measure that includes perception of corruption as one of the components. Based on their findings, the authors conclude that the importance of the risk factor should be discounted, although it would not be impossible to assign it some small weight as a decision factor. Similarly, using total inward FDI, Hines (1995) does not find a negative correlation between total inward FDI and the level of perceived corruption in host countries. Hines observes that while the equations fit poorly, it is noteworthy that local corruption has an insignificant effect on the post-1977 growth of FDI.

Corruption

99

In spite of these findings, many observers believe that corruption reduces inward FDI. Why is the empirical evidence so hard to come by? Wei (1997) provides an answer. He observes that Wheeler and Mody (1992) mix the corruption measure together with twelve other indicators to form one regressor called RISK. These other indicators include ‘attitude of opposition groups towards FDI’, ‘government support for private business activity’ and ‘overall living environment for expatriates’, which may not be overwhelmingly correlated with government corruption, may not be precisely measured or may not be as important for FDI as one might think. Consequently, the noise-to-signal ratio for the composite measure (RISK) may be too high to show up significantly in the regressions. Hines (1995) uses total inward FDI from the International Monetary Fund’s International Financial Statistics database which may also be too noisy. Wei (1997) notes that this may have contributed to the Hines finding that inward FDI is not significantly negatively correlated with the level of corruption. The second question concerns East Asia as a host region. As is apparent in Tables 7.2 and 7.3, many East Asian countries have a high level of corruption. Despite this high level of perceived corruption in East Asia, the region has been the largest recipient of FDI.18 One plausible explanation for this East Asia paradox is that corruption has been part of the culture in these economies for a long time. It is likely that countervailing institutions have been developed to circumvent corruption so that its negative effect on FDI is minimized. An alternative view is that East Asia is not unique. The large volume of FDI into East Asia is attracted by the region’s higher than average growth rate. Within the East Asian region, a less corrupt country still receives larger FDI than more corrupt countries. That is, if the countries in the region were less corrupt, they would have received even more FDI than what they already have. Wei (1997) is the first to provide systematic evidence on the East Asia paradox. He finds no evidence that international investors are less sensitive to corruption in East Asia. Using a different data set, we also find that international investors do not treat East Asia differently (see the next sub-section). Our findings, however, differ from that of Wei in that we do not find any evidence of a statistically significant effect of corruption on FDI. The third question with regards to the effect of corruption on FDI arises from the concern of the government of the United States that the Foreign Corrupt Practices Act (FCPA) of 1977 may have undermined

100

Corruption

American firms’ competitiveness in overseas markets relative to firms from Europe, Japan and elsewhere.19 Using country dummies as a measure of corruption, Beck et al. (1991) find a statistically significant effect of corruption on American export competitiveness. The effect, however, is quantitatively small. Controlling for the GDP growth rate of the host country, Hines (1995) finds that corruption adversely affects the growth of US-controlled FDI during 1977–1982, their capital/labour ratio, the incidence of joint ventures, and aircraft exports. He claims this as evidence that the FCPA has undermined American firms’ competitiveness relative to other countries. Wei (1997) notes several reasons why Hines’s claim may require further investigation. First, corruption may reduce FDI from non-American investors to the extent that they feel morally obligated to avoid bribery. Second, American firms may be just as clever at finding covert substitutes for bribery payments as other investors. Third, the degree of corruption in host countries tends to be highly correlated with many other dimensions of government quality, such as the extent of bureaucracy and red tape, or the quality of the legal system. These features are likely to affect non-American investors as well. To attribute the finding that FDI from the United States is negatively correlated with corruption to the FCPA, one needs to control for the response of all FDI to corruption. Wei (1997) re-examines the effects of corruption on inward FDI using a broader panel of bilateral FDI data with a more comprehensive list of control variables. He finds that corruption in a host country indeed depresses inward FDI in a way that is statistically significant and quantitatively large. Table 7.7, adapted from tables 4, 6 and 9 in Wei (1997), summarizes the main findings of Wei (1997). Table 7.7 shows that both an increase in the tax rate on multinational firms and corruption level in the host country reduce inward FDI. In particular, Table 7.7 implies that an increase in the corruption level from that of Singapore (a BI rating of zero) to that of Mexico (a BI rating of 6.75) is equivalent to raising the tax rate by twenty-one to twenty-four percentage points. Wei finds no evidence that international investors are less sensitive to corruption in East Asia. This is apparent in column 2 in Table 7.7. Wei notes that though the size of the annual inflow is large, over 60 per cent of the annual inflow in each of the last ten years came from overseas Chinese, notably those in Hong Kong. By contrast, FDI from the ten largest source countries in the world accounts for a relatively small

Corruption

101

Table 7.7 Effects of corruption on foreign direct investment Variable Tax rate Corruption

1

2 −3.68a (0.70) −0.09a (0.03)

East Asia East Asia × Corruption China Singapore Hong Kong United States × Corruption

3 −4.28a (0.90) −0.16a (0.04) 0.99a (0.40) 0.04 (0.07) −0.90a (0.27) −0.89a (0.43) −1.22a (0.37)

United States × East Asia United States × Corruption × East Asia Japan × Corruption Japan × East Asia Japan × Corruption × East Asia Includes other conditioning variables Includes source dummies No. of observations Log likelihood function

Yes Yes 450 1433.69

Yes Yes 450 1440.19

−3.73a (0.71) −0.10a (0.03)

−0.14 (0.13) 0.35 (1.22) 0.06 (0.16) 0.12a (0.06) −0.68 (0.57) −0.07 (0.08) Yes 450 1435.70

Source: Adapted from tables 4, 6 and 9, Wei (1997). Notes Reported estimates are based on modified Tobit estimation. Numbers in parentheses are the Eicker-White standard errors. a Significant at the 5% significance level. All regressions have additional conditional variables and source dummies that are not reported here. For further details see Wei (1997).

portion of total FDI going into China. In view of the estimation results reported in Table 7.7, China is in fact an underachiever as a host for FDI from the major source countries. This finding is consistent with the view that investors from the major source countries prefer to go to less corrupt countries. This finding also suggests that overseas Chinese are apparently less sensitive to corruption, possibly because they are better able to use personal connections to substitute for the rule of law.

102

Corruption

Finally, Table 7.7 shows that American investors are averse to host country corruption but not necessarily more so than other investors, in spite of the unique Foreign Corrupt Practices Act. Results reported in Table 7.7 also provide some weak evidence that Japanese investors are less sensitive to corruption, possibly reflecting the way business transactions are conducted in Japan. New evidence of the effects of corruption on FDI Wei (1997) is the only work which provides a systematic quantitative analysis of the effects of corruption on FDI. Wei, however, studies the flows of FDI over a short period of time (from 1990 to 1991). Furthermore, in estimating the empirical relationship between FDI and corruption, he uses the BI index of perceived corruption for the period 1980–1983.20 However, researchers have begun to use time series of indices of perceived corruption prepared by business risk analysts and polling organizations, based on survey responses of businessmen and local residents.21 It is therefore reasonable to ask whether the main findings of Wei (1997) would survive if panel data spanning a longer period of time were used. We therefore re-examine the empirical relationship between FDI and corruption over a much longer period, from 1984 to 2003. Unlike Wei, however, we use four-year averages of the net inflow of FDI as a percentage of GDP as the dependent variable in our regression model. The data on FDI are obtained from the World Development Indicators, 2004. Table 7.8 reports our findings. Even though the regression model we estimate is not as rich as the one which Wei (1997) estimates, using richer panel data we are able to replicate the main results reported in Wei. In particular, we find that both the corporate tax rate and political instability are significantly negatively related to FDI. This is consistent with the results reported in Wei. As well, we find that in the sample of countries we have studied, corruption discourages FDI. Contrary to Wei, however, we find that this adverse effect of corruption on FDI is statistically insignificant. Finally, like Wei, we find no evidence that FDI flows are any more sensitive to corruption in East Asia than in other countries or regions. We conclude this section by noting that the estimation results reported in Table 7.8 are based on a regression model which excludes a set of explanatory variables which might be significantly correlated with FDI. Table 7.7, for example, shows that FDI is significantly positively

Corruption

103

Table 7.8 Corruption and foreign direct investment revisited, 1984–2003 Variable

Model 1

Model 2

Tax rate

−2.6395 (−3.0210) −0.0653 (−0.3737) −1.1852 (−1.1140)

Political instability

−6.0425 (−5.1783)

−2.5228 (−3.0190) −0.0716 (−0.4015) −1.1813 (−1.1160) −2.5228 (−0.5253) −3.0115 (−3.2696)

No. of observations R2

187 0.377

187 0.378

Corruption East Asia East Asia × Corruption

Notes All results obtained with ordinary least squares regression. Numbers in parentheses are t ratios, based on heteroskedasticity-consistent standard errors. All regressions include a constant term whose estimate is not reported.

related to log(GDP) and log( population) while significantly negatively correlated with log(distance). Consequently, our estimation results may suffer from model misspecification bias, and therefore should be regarded merely as suggestive and preliminary. Nevertheless, our finding that corruption has no statistically significant effect on FDI is not new. Previous studies such as Wheeler and Mody (1992) and Hines (1995) also fail to find a significantly negative effect of corruption on FDI.

Conclusion The empirical evidence on the determinants of corruption and its effects on growth and foreign direct investment we have reviewed in this chapter do not provide conclusive answers to the questions posed at the beginning of the chapter. There are several plausible explanations for the lack of compelling evidence on the effects of corruption. First, as noted earlier, the measures of corruption used in the existing empirical literature do not capture the conceptual richness of the classical theoretical work on corruption. Instead, the survey-based corruption indices used in this chapter as well as in the existing empirical works refer primarily to the administration of rules or laws pertinent to foreign firms, and are

104

Corruption

probably weighted by efficiency level as perceived by those who were surveyed. Second, as noted earlier in this chapter, corruption can have many detrimental effects in an economy: corruption may retard growth, possibly as a result of reduced domestic investment (Mauro, 1995; Knack and Keefer, 1995); corruption can discourage foreign direct inward investment in a host country by imposing undue costs on the source countries (Wei, 1997, 2000a); corruption may raise the probability of a currency crisis by altering the composition of a country’s capital inflows (Wei, 2001); corruption often contributes to unfair income or wealth distribution; and corruption can breed political instability. All these important aspects of corruption may interact with each other. None of the existing empirical works on corruption investigates this joint interaction. Finally, corruption has been part of the culture in East Asian economies for a long time. Countervailing institutions may have been developed to circumvent corruption so that its negative effect on growth and foreign direct investment is minimized.

8

Urbanization

The literature on this topic is vast,1 but mostly it relates to demography, urban planning and geography, which is not the main concern here. Rather, the purpose here is much more limited and focused. In particular, we concentrate on the data relating to urbanization given in the World Development Indicators and provide a preliminary analysis of that information as far as it bears on the issue of sustainability. Sustainability is a concept which, at best, is amorphous and means different things to different people. We confine ourselves to issues which were raised in the initial proposal. We start with an examination of the trend towards urbanization in the countries of our sample. Then we look at what the existing evidence has to say about the sustainability of this trend, say, in terms of adequate and affordable housing, access to safe drinking water, proper sanitation, electricity and means of communication like telephones. In the process we will suggest our meaning of sustainability.

Indicators of urbanization Table 8.1 details some of the indicators of urbanization. Consider columns 1 and 2 first. These are, probably, the best single summary measures of urbanization of a country. Leaving aside Hong Kong and Singapore, which are city states and therefore very close to being fully urbanized, the trend towards urbanization in the rest of the countries of the sample is clear. In 1980 the five East Asian countries (Indonesia, South Korea, Malaysia, the Philippines and Thailand) represent fairly low levels of urban population, but the trend towards increased urbanization over the next twenty years is clear, with varying degrees of progress. South Korea’s progress is most remarkable, overtaking Japan.

106

Urbanization

Table 8.1 Indicators of urbanization (percentage of total population) Country

Urban population 1980

Hong Kong Indonesia Japan South Korea Malaysia Philippines Singapore Thailand

92 22 76 57 42 38 100 17

Population in urban agglomerations of more than 1 million

Population in largest city

1999 100 40 79 81 57 58 100 21

1980

2000

2015

91 8 34 40 7 14 100 10

100 10 38 47 6 16 100 12

100 12 39 45 6 17 100 15

1980

2000

100 18 25 38 16 33 100 59

100 13 26 26 10 25 100 56

Source: Adapted from World Development Indicators, 2001, table 3.10.

On the lower end is Thailand, with very slow increase in the proportion of urban population. The other two indicators of urbanization, population in urban agglomerations of more than one million as a percentage of total population and population in the largest city as percentage of the urban population, also shed interesting light on the nature and extent of urbanization in the eight countries being considered. But what is clear from this table is that the trend towards urbanization is irreversible and its implications, whether good or bad, cannot be avoided. Tables 8.2, 8.3a and 8.3b shed light on different implications of urbanization. Table 8.2 can be looked at in a variety of ways. For example, we can consider it as reflecting vehicular intensity and its accompanying implications for road congestion. Using some indicator of this congestion as being optimal, we can then talk about whether the levels of congestion are sustainable or not. The problem in this kind of interpretation is that the levels of congestion also reflect levels of development, so that low levels of congestion may also reflect low levels of development. Thus consider Thailand. While one might consider low levels of congestion as being good, yet these low levels may simply reflect low levels of development, certainly not a desirable attribute. Therefore, to the extent that increased traffic and increased road congestion accompany increased urbanization, they might simply mean increased urban affluence, rather than urban degradation. Somewhat similar conflicting implications emerge from Tables 8.3a and 8.3b. These tables give selected information about a few of the urban centres in some of the countries of our sample. In all but one case

Urbanization 107 Table 8.2 Traffic and congestion Country

Hong Kong Indonesia Japan South Korea Malaysia Philippines Singapore Thailand

Motor vehiclesa

Motor vehiclesb

Passenger carsa

Twowheelersa

1990

1999

1990

1999

1990

1999

1990

1999

66 16 469 79 124 10 146 46

78 25 560 238 300 31 164 106

253 10 52 60 26 4 142 36

276 14 61 120 69 11 170 97

42 7 283 48 101 7 101 14

58 14 395 167 170 10 119 28

4 34 146 32 167 6 45 86

5 62 115 59 224 14 41 174

Source: Adapted from table 3.12, World Development Indicators, 2001. Notes a Per thousand people. b Per kilometre of road.

the data are given for one year only. With a single data point we cannot draw many conclusions. But for what it is worth, the table still does shed interesting light on the sustainability issue from a variety of angles. For comparison’s sake, we use Hull, a city in Canada, a developed country. We can reasonably argue that, as far as access to services, for example, is concerned, there is no reason why the developing countries should settle for anything less than what is found in the developed countries. Keeping this standard of achievement in mind, consider the first statistical column, which gives the percentage of people with secure tenure. In Hull the relevant figure is 100 per cent. For the six cities included in the tables, two cities, namely Semarang in Indonesia and Bangkok in Thailand, still lie way behind. Property security is well below 100 per cent and, assuming that the percentage of people with secure tenure is lower than in Semarang and Bangkok, we can reasonably be sure that the East Asian countries have a long way to go in terms of this indicator. The housing sector is an important sector from multiple points of view. It both affects and is affected by the behaviour of the rest of the economy. A detailed analysis of this sector is not warranted here, but the single indicator in Table 8.3a, the ‘House price to income ratio’, is a very useful one. As a gross indicator, it reflects affordability status, except that to the extent that the statistic is a ratio and both the price and the household incomes are influenced by a myriad of factors, both structural and non-structural, it is difficult to say a priori what specific

Table 8.3a Urban environment – housing and transport Country

Canada Indonesia South Korea Malaysia Philippines Thailand

City

Hull Jakarta Semarang Surabaya Hanam Pusan Seoul Penung Cebu Bangkok Chiang Mai

Secure tenurea 1998 100.0 95.5 80.2 97.6

95.0 77.2 96.5

House price to income ratio 1990

1998

3.46

14.6 — 3.4 3.7 4.0 5.7 7.2 13.3 8.8 8.8

4.15

Work trips by public transport 1998 16 — — 18 — 39 71 55

Travel time to work (minutes) 1998

35 42 60 40 35 60 30

28 5

Sources: All figures, except the house price to income ratio for 1990, are adapted from World Development Indicators, 2001, table 3.11. The 1990 figures are from McGee and Robinson (1995), p. 119. Notes a Percentage of people with secure tenure.

Table 8.3b Urban environment – services Country

Canada Indonesia South Korea Malaysia Philippines Thailand

City

Hull Jakarta Semarang Surabaya Hanam Pusan Seoul Penung Cebu Bangkok Chiang Mai

% of households with access to services Potable water 1998

Sewerage connection 1998

Electricity 1998

Telephone 1998

100 50 34 41 81 98 100 99 41 99 95

100 65

100 99 85 89 100 100 100 100 80 100 100

100

56 68 69 99 92 100 60

71 100 100 100 98 25 60 75

Sources: All figures, except the house price to income ratio for 1990, are adapted from World Development Indicators, 2001, table 3.11. The 1990 figures are from McGee and Robinson (1995), p. 119.

Urbanization 109 factors could account for its inter-temporal behaviour. We have data for two time periods for two cities: for 1990 and 1998 for Jakarta (Indonesia) and Bangkok (Thailand). What is striking about both is the magnitude of the change. Thus, the ratio more than quadrupled for Jakarta, while it more than doubled for Bangkok. But why such a wide difference? Concentrating on the year 1998 only, the ratio varies from a low of 3.4 for Surabaya (Indonesia) to a high of 14.6 for Jakarta (Indonesia). What is one to make of such differentials? Being a ratio, it could reflect just about any kind of behaviour of price and income. Unfortunately, no comparable data are available for other cities to enable a comparative study. The only thing one can say is that the house price-to-income ratio varies widely, and whether it reflects any well defined economic reality would require a detailed analysis of the housing markets in different cities. The next two columns, work trips by public transport (in per cent) and travel time to work (minutes), reflect yet other aspects of urbanization. Taking Seoul (South Korea) as the example of most urbanized city among the cities listed, we can see that the other cities have a long way to go. The last four columns are very revealing in at least two ways (Table 8.3b). First, they include Hull – a city from a highly urbanized country – and, second, the four columns refer to services which are commonly found in most urban societies: water, sewerage, electricity and telephone. We could, of course, add other means of communication, but that topic is dealt with elsewhere. Since the percentage for Hull is 100, we take that number as the desirable upper bound. Taking access to potable water, the situation is not optimal even in all the cities covered here, although a great deal of progress has been made in countries other than Indonesia and the Philippines. But then, reflecting back to the growth performance of ASEAN, we could see that these two countries do not stand out as star performers. The column representing sewerage connection, not surprisingly, shows somewhat lower levels of access than potable water. Access to electricity suggests quite high levels, while telephone access has reached 100 per cent in South Korea and nearly 100 per cent in Penang (Malaysia).

Conclusion Of course, from these few cities we cannot conclude that access to the four services is widespread across all the urban areas. But the question

110

Urbanization

of ‘sustainability’ needs to be answered. On the one hand, we can say that the levels of access achieved so far are not necessarily the ones which can be considered optimal and in the few cases where they have achieved such levels, whether those levels can be continued remains to be seen. Much would depend on public attitudes. If, like the developed countries, where universal access to such services is taken for granted as part of ‘normal’ life, then surely such access will become sustainable.

9

Gender gap

Gender gap is not a topic which is commonly discussed in books on economic development. The reasons for this exclusion are varied, but they need not concern us here. Instead, we use the available aggregate data to shed some light on the nature and the extent of gender inequality or gender gap. Perhaps the single most revealing indicator of gender gap is the female–male per capita income gap. This statistic summarizes the interplay of many underlying factors. Some of the pertinent data are given in Table 9.1. The data are for a single year, 1997, but still we can make a number of observations from these limited data. First and foremost is the fact that the male per capita income exceeds the female per capita income (all measured in PPP $) for each of the countries listed in the table. Furthermore, with the exception of the Philippines and Thailand, the ratio of male to female income per capita income is more than two to one. And even for the Philippines and Thailand, the difference is more than 60 per cent. This income gap is accompanied by other attributes of gender inequality in economic activity. A useful way to examine this issue is by considering some of the structural features of an economy. We consider three: female economic activity rate; structure of employment by economic activity; and the ratio of contributing family workers. The data on these three factors are given in Table 9.2. But we concentrate on the third feature. Column 3 of Table 9.2 gives the female participation rate as a percentage of the male participation rate. If we consider Norway, which was ranked No. 1 in 1999 with an HDI of 84, as the norm for the developed countries, we can see that only Thailand equalled Norway. Every other country in the tale was significantly behind, including Japan and Singapore. This lower ratio, as we know, is reflected in the statistics

112

Gender gap Table 9.1 Male–female income Country

Japan Singapore Hong Kong South Korea Malaysia Thailand Philippines Indonesia

Real GDP per capita (PPP $) in 1997 Female

Male

14,625 18,947 15,180 8,388 5,115 5,000 2,510 2,359

33,893 37,833 32,688 18,708 11,081 8,382 4,513 4,626

Source: Adapted from Human Development Report, 1999, pp. 138–140.

in Table 9.1. The data on the structure of employment by economic activity once again complements the data in Table 9.1. Since generally the wage rates tend to be lower in the service sector, the dominance of females in this sector ultimately reflects lower per capita incomes for females in Table 9.1. With these broad aspects of gender inequality before us, we now consider some more specific forms. It is, of course, true that we probably cannot define all types of gender inequalities. But a good start can be made by adopting the classification proposed by Sen (2001). However, to elucidate this classification, Sen refers to evidence from South Asia, in particular from India. But given the composition of our sample of countries, this will not be very useful for our purpose. Instead, we use data relating to the countries of our sample, which can be regarded as a direct or an indirect indicator of a particular type of gender inequality.

Mortality inequality Two interesting aspects of the demographic profile of females and males are death rates and life expectancy. The data on these variables are given in Table 9.3. A consistent feature of this table is that the death rate for males exceeds that for females, while, not unexpectedly, life expectancy for females is higher than for males, across all the countries included in the sample. This seems like a remarkable feature of the East Asian economies. This is not the place to speculate on the gender bias in favour of females in these respects, but none the less it would seem to be the case that some underlying actors contributing to a lower mortality rate and higher life expectancy for females must be operating in these countries.

51.1 49.1 50.2 53.0 47.8 72.9 49.4 55.0 58.9

106 103 105 110 111 97 107 115 114

67 63 64 69 60 84 61 67 84

6 (·) (·) 13 14 51 28 42 2

Female

Ratec

Ratea

Indexb

Agriculture

Notes a Rate in %. b Index (1958 = 1000). c Rate as percentage of male economic activity rate.

5 (·) (·) 10 19 49 48 41 7

Male 24 15 25 21 30 17 13 16 10

Female

Industry

39 31 34 38 36 22 19 21 35

Male

Employment by economic activity, 1994–97

Female economic activity rate (age 15 and above), 1999


Japan Hong Kong Singapore South Korea Malaysia Thailand Philippines Indonesia Norway

Country

Table 9.2 Gender inequality in economic activity

69 85 75 66 56 32 59 42 87

Female

Services

55 69 66 52 45 28 33 39 59

Male 82 — 75 88 — 66 — — 67

Female

18 — 25 12 — 34 — — 33

Male

Contributing family workers (as % of total), 1994–97

114

Gender gap Table 9.3 Mortality inequality Country

Hong Kong Japan South Korea Malaysia Philippines Singapore Thailand Indonesia

Death rate by sex

Age in years (at birth)

Year

Male

Female

Year

Male

Female

1997 1997 1995 1997 1991 1998 1990

5.5 8.1 6.1 5.1 5.6 5.5 5.2

4.2 6.5 4.7 3.8 3.8 4.4 3.7

1997 1997 1995 1997 1991 1998 1995–2000 1995–2000

76.8 77.2 69.5 69.6 63.1 75.2 65.8 63.3

82.2 83.8 77.4 74.5 66.7 79.3 72.0 67.0

Source: Adapted from Demographic Yearbook, 1998.

Table 9.4 Natality inequality: live births, all ages, by sex Country

Year

Male

Female

Hong Kong Japan South Korea Malaysia Philippines Singapore Thailand

1997 1997 1997 1997 1993 1998 1994

30,872 610,905 346,762 223,244 875,540 22,612 494,485

28,362 580,760 319,871 208,748 805,356 21,225 465,763

Source: Adapted from Demographic Yearbook, 1998.

Natality inequality This refers to a preference for boys over girls by parents and may be responsible for the use of gender-identifying devices in developing countries. While we do not have data on ‘missing women’ as offered by Sen for the South Asian economies, the bit of evidence in Table 9.4 on live births by all ages and by sex provides some indication of this phenomenon. Table 9.4 shows that, for the year under consideration, the number of male live births exceeded the number of female live births for each of the seven countries. Of course, these data need not necessarily reflect a preference for boys over girls. And yet we must wonder at the consistency of the evidence. In particular, we need to ask whether it is the outcome of a natural phenomenon or the result of deliberate choice by parents. If the latter, then it certainly would suggest that natality inequality exists across all the countries in our sample and its social, economic and demographic implications must be examined.

Gender gap 115

Basic facility inequality Sen’s next form of gender inequality is described as ‘basic facility inequality’. Once again, we have no concrete indicator of this form of inequality. We consider adult literacy, youth literacy, and net primary enrolment, as rough indicators of this form of gender inequality. Table 9.5 gives data on female rates as well as female rates as a percentage of male rates for a specific year or time period. Considering the female rate as a percentage of the male rate, we notice some inequality against females in adult literacy, though almost none in youth literacy or net primary enrolment, wherever we have the data. Thus, at least in terms of ‘basic facility inequality’, as judged here, the countries of our sample would seem to have done well.

Professional inequality Once again, it is difficult to measure this form of inequality in any concrete way. Consequently, we use some of the data provided by the Human Development Report, 2001 on gender empowerment. The logic here is that as the proportion of females who are trained professionally increases, so would their representation in professional and political bodies, and thus reduce professional inequality. The relevant data are given in Table 9.6. Looking at the value of the overall ‘gender empowerment measure’ (GEM) and comparing with Norway, which in 2001 had the highest HDI, we can see that Norway had a value of 0.836 while the corresponding value for the East Asian economies was substantially lower. It is interesting to see that, even for Japan, the value was 0.520. This low value of the gender empowerment measure is reflected in the data on professional inequality measures. Thus, in terms of seats held by women in parliament, the percentage does not exceed even 15 and in many of the countries it does not even exceed 10. The inadequacy of this number or the extent of professional inequality is most vividly demonstrated when we notice that in Norway the corresponding figure was 36.4 per cent. The situation for some of the countries improves somewhat when we consider female legislators, senior managers and managers (as percentage of the total). However, the very low percentage for Japan, South Korea and Singapore stands out once again. Unlike the decision-making power exercised by females as indicated by the data in columns 2 and 3 of Table 9.6, it is interesting to notice that the percentage of female professional and technical workers is substantially

89.7 88.0 96.2 82.8 93.5 94.9 81.3

93.0 92.0 97.0 91.0 96.0 100.0 89.0

99.8 99.8 99.8 97.4 98.3 98.7 96.8

Ratec

Ratea

Rateb

Youth literacy (female), 1999

Adult literacy (female), 1999

Notes a Rate in % (age fifteen and above). b Rate as percentage of adult male literacy rate. c Rate as percentage of adult female literacy rate. d Rate as percentage of male youth literacy rate. e Ratio in %. f Ratio as percentage of male net primary enrolment rate.



Country

Table 9.5 Basic facility inequality: gender inequality in education

101.0 100.0 100.0 100.0 99.0 100.0 99.0

Rated

91.0 — 93.0 — — — 93.0 100.0

Ratioe

103.0 — 101.0 — — — 97.0 100.0

Ratiof

Net primary enrolment (female), 1995–97

0.520 — 0.509 0.358 0.503 — 0.470 — 0.836


10.8 — 6.5 5.9 14.5 — 11.8 8.0 36.4

Seats in parliament held by women (as % of total) 9.0 22.0 21.0 5.0 21.0 22.0 33.0 — 31.0

Female legislators, senior officials and managers (as % of total)

Note a Ratio of estimated female to male earned income.


Value

Country

Table 9.6 Professional inequality: gender empowerment measure (GEM)

44.0 38.0 42.0 31.0 44.0 55.0 63.0 — 58.0

Female professional and technical workers (as % of total)

0.43 — 0.49 0.45 0.46 — 0.55 — 0.63

Income ratioa

118

Gender gap

higher. This might indicate that males find it more agreeable to accept females as colleagues when the former do not find the latter as threatening. Finally, the overall implications of ‘professional inequality’ are well-reflected in the last column which shows the income differentials. It can be seen that except for the Philippines, the ratio of estimated female to male earned income was below 50 per cent, which was also considerably lower than 63 per cent for Norway.

Ownership inequality The main idea in this form of inequality is that the laws of the land and/or the institutional arrangements create inequality in the ownership of property. A rigorous defence of this type of inequality would require a study of the legal framework of a country, its inheritance laws, etc. But we take an indirect approach as above. We argue that the greater the participation of women in the government, in particular in decisionmaking positions and in market activities as contrasted with non-market activities, the greater the power they would be able to exercise in a way which would reduce ownership inequality by leading to more equitable property ownership laws. Some suggestive evidence on this line is offered in Table 9.7. This table shows the proportion of women with senior decision-making powers in the government. Once again, column 2 and even 3 are most revealing. At the ministerial level in 1998, in Japan and Singapore, this proportion was zero. It is not being suggested here that the only way Table 9.7 Ownership inequality: women’s political participation (%) Country


Women in government, 1998 At all levels

At ministerial level

At sub-ministerial level

2.2 — 6.7 — 13.7 6.3 17.0 1.6 22.2

0.0 — 0.0 — 15.6 4.0 9.5 3.4 20.0

2.8 — 8.3 — 12.9 6.8 18.8 1.3 22.9


Gender gap 119 to protect and enhance the rights of any group is to have representation from that group in the ministry. None the less, it is the case that with such complete dominance by males of the legislation make-up body in these countries, one should not be surprised if the inheritance and other property laws are not skewed towards males. The final four columns are also illustrative of a similar phenomenon. We notice that females dominate non-market activities, suggesting thereby that they have a lesser voice in decision making, which may again suggest greater ownership inequality.

Household inequality Even assuming the absence of the above forms of gender inequality, Sen suggests that even then there might be intra-household gender inequalities. This means that domestic tasks may be institutionalized in such a way that they would act unfavourably for females. Some suggestive evidence is offered in Table 9.8. While the number of countries referred to is only three, covering different time periods and, except for South Korea, covering either rural or urban areas, yet the broad picture that emerges is similar. Females end up sharing the domestic burden of work more than males. This inequity is more pronounced in rural areas than in urban areas. While the above types of inequalities can shed important light on the impediments faced by females in their march towards equality with males, they leave out one of the more important ones in today’s environment. And this has to do with possible bias against women-led businesses to access venture capital. In the ‘New Economy’, new ventures require high capital needs but are also often characterized by high risk. Consequently, they find it hard to raise equity capital and the needs are often met by venture capital. But, it has been suggested by Greene et al. (2001) that there might be gender bias in this access to venture capital. In other words, women-led businesses are often bypassed by venture capital firms. Since the ‘New Economy’ firms or high growth potential firms are often the means by which new accumulation of wealth often takes place, this vehicle is thus denied to females. Unfortunately, there is very little evidence to help us to sort out this source of inequality. But still consider the evidence presented in Table 9.9. We can notice a number of interesting points. First, most of the venture capital is made available in the United States. Looking at the supply of venture capital in 2000, the United States was by far the most dominant country.

1992 1975–77 1990

Indonesia (urban) Philippines (rural) South Korea (national)

398 546 488

Female

Note a Ratio (in %) of female to male work time per day.

366 452 480

Male

Work time (minutes per day)

Burden of work

Source: Adapted from Human Development Report, 2000, p. 263.

Year

Country

Table 9.8 Household inequality: gender work burden

109 121 102

Ratioa

35 29 34

Female

Market

Activities

86 84 56

Male

65 71 66

Female

Non-market

14 16 44

Male

Gender gap 121 Table 9.9 Venture capital spread across the world: current investment (US$ million) Country

1995

2000

United States United Kingdom Japan Germany France Hong Kong Singapore Sweden Israel India Finland China South Korea Philippines South Africa

4,566 19 21 13 8 245 5 — 8 3 — — 1 2 —

103,170 2,937 1,665 1,211 1,124 769 651 560 474 342 217 84 65 9 3

Source: Adapted from table 2.4, Human Development Report, 2001.

But at the same time, it is beginning to emerge even in developing and small countries too; for example China, India, Israel and Singapore. But Greene et al. (2001) have suggested, as mentioned above, the possibility of a gender bias in the supply of venture capital in the United States. It is possible to speculate that if such a bias exists in the United States, then the existence of a similar bias cannot be ruled out in other countries as well, in particular the East Asian economies. But, of course, concrete evidence on this question will be far more useful.

10 Digital divide

This topic, like many others in this book, is both controversial and confusing. It is not our intention to survey the entire field. Rather we focus on a narrow issue, namely the determinants of the major components of the information technology (IT) industry. These determinants, if properly identified, can shed some light on how the IT industry is likely to expand in the Asian countries of our sample and the policy implications of the major determinants. One of the most unfortunate aspects of this topic for our purpose is the lack of suitable data for analytical purposes. This is most evident if we consider the kind of data which are available for some of the European countries and the United States. Consider, for example, the data summarized in Norris (2001). Norris provides data on ‘trends in Internet use and computer use in Europe and America, 1996–1999’, as well as the ‘social profile of Internet users in Europe, 1996–1999’. The data identify various determinants of Internet users. It includes age, income, category, age finished education, gender and occupational status. If such data were available for a large number of countries and/or for a long period of time, we could carry out quite a vigorous statistical analysis. Unfortunately, such data are not available for the Asian countries even for a very short period of time. Consequently, we are not in a position to conduct the kind of analysis which is possible for the United States and some of the European countries. However, all is not lost and we can pursue the approach followed by the Asian Development Bank (ADB).1 This approach exploits aggregate data, as we do, although our study differs from the Quibria et al. (2002) study (from here on, the ADB study) in many respects. But before we discuss our results and summarize the findings of the ADB study, it is useful to look at how the countries of our sample

Digital divide 123 compare with developed countries in terms of access to new and old information and communication technologies. This information is given in Table 10.1 for old technologies and in Table 10.2 for new. As we might expect, there is great deal of divergence and disparity among the various countries with regard to old technologies as well as new technologies. As more concrete examples, we can look at access to personal computers. Using Canada as an illustrative example of a developed country we can see that countries of our sample include those that significantly exceed the Canadian figure and those that fall far behind. The same holds true of new technology (for example, Internet users), although in this case, those falling behind are more important. A very revealing statistic is found in the last column which gives information and communications technology expenditure as a percentage of GDP for 1999. Here the differences are quite glaring. Hong Kong, Japan and Singapore are reasonably close to Canada, while Indonesia, South Korea, Malaysia, the Philippines and Thailand are far behind. This could mean many things, but merely speculating would serve no useful purpose. The subsequent analysis of the determinants of these components is likely to shed more significant light. The ADB study used six dependent variables: cellular phone, fax machine, Internet use, personal computer, main telephone line and television. Each of these variables was regressed on population, income and Asia as a dummy variable. In the second set of regressions, these same dependent variables were regressed on income, primary, secondary and tertiary education. Clearly both sets of equations can be considered as being misspecified, assuming that all independent variables are relevant variables. But leaving that aside, the ADB study reported that while income was significant in all cases, population and education variables on the Asian dummy were either statistically insignificant or had the wrong sign in most cases. The most distinctive feature of the two tables is the significance of per capita income. Population is significant for cellular phone and television, while the Asian dummy is insignificant in all cases. The results with respect to education are also mixed. While primary education plays no role, secondary education affects the use of telephones only and tertiary education determines the use of the Internet and personal computer. Given the ADB study and the lack of other works, and disaggregated data, we decided to experiment with aggregate models. As for estimation procedures, we decided to use not only the OLS as in the ADB study, but also SUR, because we thought that the seemingly unrelated regressions

792.0 24.0 578.0 393.0 158.0 79.0 360.0 63.0 159.0

Hong Kong Indonesia Japan South Korea Malaysia Philippines Singapore Thailand Canada

678.0 157.0 960.0 1,033.0 419.0 159.0 682.0 253.0 1,047.0

Radiosb

434.0 143.0 719.0 361.0 174.0 110.0 308.0 289.0 715.0

Sets

Televisionb

Notes a Per 1,000 people in 1996. b Per 1,000 people in 1999. c Cable subscribers per 1,000 people in 1999.


Daily newspapersa

Country

Table 10.1 The information age (old technology)

68.0 — 125.4 150.1 5.2 9.4 53.2 2.4 273.2

Cablec 58.0 0.9 127.0 — 8.1 — 25.8 2.5 35.8

Fax machinesb

297.6 9.1 286.9 181.8 68.7 16.9 436.6 22.7 360.8

Personal computersb

182.92 1.15 269.25 100.65 27.55 2.21 385.73 8.84 590.37

Hostsa

Internet

2,430.0 900.0 27,060.0 10,860.0 1,500.0 500.0 950.0 800.0 11,000.0

Usersb

18.0 9.0 41.0 12.0 1.0 31.0 15.0 25.0 12.0

Servicec

Notes a Per 1,000 people in July 2000. b Thousands of users in 1999. c Service provider charge (monthly) in 1998. d Telephone call charge (monthly) in 1998. e Secure servers in 2000. f Information and communications technology expenditure (percentage of GDP) in 1999.

5.0 6.0 14.0 14.0 8.0 0.0 5.0 33.0 0.0

Telephoned

Monthly access charge


Hong Kong Indonesia Japan South Korea Malaysia Philippines Singapore Thailand Canada

Country

Table 10.2 The information age (new technology)

475.0 54.0 4,139.0 313.0 128.0 59.0 483.0 103.0 4,530.0

Secure serverse

8.31 1.39 7.06 4.42 5.20 2.71 7.67 2.13 8.52

ICT expendituref

126

Digital divide

model was a particularly suitable one for this case. Our choice of the functional forms and the specific variables to be included was wide and problematic. Since we did not want to go on a data mining campaign, we decided to appeal to the literature for deciding on the specification. First we decided to experiment with respect to the income variable. We proceeded on the assumption that, on a priori grounds, non-linearity in income made sense, so we carried out estimation with income squared. However, this variable turned out to be either insignificant or with a wrong sign. This suggested to us that in the countries of our sample the ITC industry had a long way to go before it would reach saturation level. Next, we decided that given the current state of the literature we would like to test whether income, income distribution, tertiary education, scientists and engineering students (as a percentage of tertiary students), urbanization, gender and the special characteristics of the countries of our sample (as represented by the Asian dummy) played a significant role in the adoption and spread of the ITC industry. We did some further experimentation with respect to the indicators of urbanization and gender. As Table 10.3 shows, we tried two alternatives for urbanization and three for gender. The alternative definitions are given in Table 10.3. Before presenting and discussing our regression results, first we present cross-country correlations which will give some idea about correlations between the dependent variables and the explanatory variables as well as give some idea about possible multicollinearity between the explanatory variables. These cross-country correlations are reported in Tables 10.3 and 10.4. Looking at the cross-correlations in Table 10.3 we can notice that per capita income is significantly correlated with each of the dependent variables. The Gini coefficient, representing income distribution, has the correct sign and is also significant in four of the five cases. It is worth noting that the only case where the Gini coefficient has the wrong sign is also the case where the correlation coefficient is insignificant. Since the main objective of this chapter is to understand the main determinants of the digital divide in the countries of our sample, it is interesting to look at the OLS and the SUR estimates given in Tables 10.5 and 10.6. Since the SUR estimates are more efficient, we concentrate on them, although a comparison of the results in Tables 10.5 and 10.6 clearly shows that the two sets of estimates are not very different. The most significant aspect of Table 10.6 is the significance of per capita GDP as a determinant of the five components of the ITC industry.

0.834 (0.000) 0.897 (0.000) 0.906 (0.000) 0.928 (0.000) 0.724 (0.000)

GDPP

TERT −0.081 (0.440) −0.110 (0.291) −0.214 (0.039) −0.249 (0.016) −0.383 (0.000)

GINI −0.506 (0.000) −0.511 (0.000) 0.184 (0.076) −0.671 (0.000) −0.635 (0.000)

Explanatory variable

0.031 (0.768) 0.078 (0.455) 0.159 (0.127) 0.181 (0.082) 0.055 (0.601)

SEST 0.099 (0.344) 0.010 (0.924) 0.049 (0.639) −0.085 (0.417) −0.074 (0.478)

ASIA 0.506 (0.000) 0.580 (0.000) 0.501 (0.000) 0.654 (0.000) 0.423 (0.000)

UPO1 0.374 (0.000) 0.352 (0.000) 0.358 (0.000) 0.302 (0.003) 0.324 (0.001)

UPO2 0.207 (0.045) 0.196 (0.058) 0.184 (0.076) 0.300 (0.003) 0.248 (0.016)

GEN1

0.308 (0.003) 0.378 (0.000) 0.310 (0.002) 0.407 (0.000) 0.378 (0.000)

GEN2

0.290 (0.005) 0.446 (0.000) 0.359 (0.000) 0.498 (0.000) 0.360 (0.000)

GEN3

Notes TELP telephone lines per capita. CELP cellular mobile telephones per capita. TEVP television receivers per capita. INTP internet users per capita. PECP personal computers per capita. UPO1 urban population (% of total population). UPO2 population in urban agglomerations of more than 1 million (% of total population in 2000). TERT education per tertiary student (% of GNI per capita in 1997). SEST scientists and engineering students (% of total tertiary students during 1987–1997). GDPP GDP per capita in 1999, current international dollar (PPPs). ASIA Asian dummy (1 for the eight East Asian countries and 0 otherwise). GEN1 girls to boys ratio (tertiary level enrolment). GEN2 research and development scientists and engineers (number of females). GEN3 research and development technicians (number of females). GINI Gini index. Numbers in parentheses are the p values.

TEVP

TELP

CELP

PECP

INTP

Explained variable

Table 10.3 Cross-country correlations: explanatory and explained variables

Note See Table 10.3.

GEN3

GEN2

GEN1

UPO2

UPO1

ASIA

SEST

TERT

GDPP GINI

Explanatory variable — −0.584 (0.000) −0.237 (0.021) 0.127 (0.224) −0.052 (0.616) 0.580 (0.000) 0.304 (0.003) 0.241 (0.019) 0.425 (0.000) 0.484 (0.000)

GDPP — — — 0.141 (0.175) 0.081 (0.440) 0.093 (0.374) −0.245 (0.018) 0.101 (0.334) −0.199 (0.055) −0.321 (0.002) −0.306 (0.003)

GINI


— — — — — −0.174 (0.094) −0.187 (0.071) −0.405 (0.000) −0.297 (0.004) −0.173 (0.095) −0.157 (0.130) −0.110 (0.291)

TERT — — — — — — — −0.441 (0.000) 0.318 (0.002) 0.196 (0.058) 0.004 (0.974) −0.066 (0.526) 0.103 (0.321)

SEST

Table 10.4 Cross-country correlations: explanatory variables

— — — — — — — — — −0.238 (0.021) −0.033 (0.751) −0.207 (0.045) 0.298 (0.004) 0.052 (0.622)

ASIA — — — — — — — — — — — 0.467 (0.000) 0.187 (0.071) 0.216 (0.036) 0.199 (0.055)

UPO1 — — — — — — — — — — — — — −0.011 (0.915) 0.205 (0.048) 0.080 (0.445)

UPO2 — — — — — — — — — — — — — — — −0.130 (0.211) −0.015 (0.887)

GEN1

— — — — — — — — — — — — — — — — — 0.651 (0.000)

GEN2

— — — — — — — — — — — — — — — — — — —

GEN3

0.00001 (0.000) 0.00002 (0.000) 0.00002 (0.000) 0.00002 (0.000) 0.00001 (0.000)

GDPP

UPOP 0.00095 (0.270) 0.00094 (0.348) 0.00160 (0.223) 0.00127 (0.085) 0.00286 (0.024)

GINI −0.00020 (0.803) 0.00023 (0.767) −0.00209 (0.153) −0.00419 (0.000) −0.00661 (0.000)


0.00009 (0.103) 0.00009 (0.120) 0.00013 (0.125) 0.00005 (0.324) −0.00019 (0.015)

TERT −0.00077 (0.345) −0.00075 (0.287) 0.00337 (0.038) 0.00139 (0.175) −0.00178 (0.219)

SEST 0.00689 (0.814) −0.00486 (0.862) −0.01410 (0.742) 0.06508 (0.062) 0.02679 (0.554)

GEND

0.03181 (0.267) 0.01231 (0.582) 0.10998 (0.011) 0.01579 (0.605) −0.02562 (0.512)

ASIA

0.526

0.512

0.346

0.379

0.395

R2

44

45

45

44

45

N

Notes See Table 10.3 for variable definitions. Numbers in parentheses are the p values which are based on the heteroskedasticity consistent standard errors of the coefficient estimates. R 2 adjusted R 2 . N number of observations.

TEVP

TELP

CELP

PECP

INTP

Explained variable

Table 10.5 OLS estimates

0.00001 (0.000) 0.00002 (0.000) 0.00002 (0.000) 0.00002 (0.000) 0.00001 (0.000)

GDPP

UPOP 0.00081 (0.276) 0.00091 (0.213) 0.00192 (0.068) 0.00112 (0.165) 0.00280 (0.071)

GINI −0.00024 (0.819) 0.00023 (0.826) −0.00306 (0.041) −0.00442 (0.000) −0.00676 (0.002)


0.00010 (0.209) 0.00009 (0.230) 0.00015 (0.170) 0.00005 (0.571) −0.00019 (0.261)

TERT −0.00015 (0.883) −0.00062 (0.536) 0.00335 (0.020) 0.00195 (0.077) −0.00176 (0.407)

SEST

0.00906 (0.786) −0.00503 (0.878) 0.00410 (0.931) 0.06716 (0.064) 0.03140 (0.652)

GEND

0.04046 (0.161) 0.01419 (0.620) 0.10608 (0.010) 0.02149 (0.496) −0.02575 (0.671)

ASIA

Notes See Table 10.3 for variable definitions. Numbers in parentheses are the p values. System R 2 is 0.981 and the number of observations is forty-four.

TEVP

TELP

CELP

PECP

INTP

Explained variable

Table 10.6 SUR estimates

Digital divide 131 The GINI coefficient is significant for only three of the components, namely CELP, TELP and TEVP. We will have more to say about these results later on. Urban population is not significant in any of the equations at the 5 per cent level, although it is at about the 7 per cent level in the equations for CELP and TEVP. TERT is not significant in any of the equations. SEST is significant only in the case of CELP. Gender is not significant in any of the components. Finally, the Asian dummy is significant in the equation for CELP. It would seem that, apart from per capita income, the other variables do not add much to explaining the extent of the digital divide in the countries of the sample. In order to check whether variables other than the per capita income made a significant contribution to the determination of the five components of ITC, we conducted an experiment whose results are given in Table 10.7. This table gives the R 2 values of two regressions: the ‘Full regression’ and the ‘Residual regression’. The ‘Full regression’ refers to the case where each of the dependent variables is regressed on all the independent variables listed in Tables 10.5 and 10.6, while the ‘Residual regression’ shows R 2 which are derived via a two-stage procedure. In step 1 the residuals are derived by regressing each of the dependent variables on GDPP, and in step 2 these residuals are regressed on all the independent variables, except the DGPP. The R 2 s for the ‘Residual regression’ refer to the R 2 s of this second step. It seems from Table 10.7 that variables other than the Table 10.7 Comparing the R 2 values Explained variable INTP PECP CELP TELP TEVP

R 2 values Full regression

Residual regression

0.394 0.379 0.346 0.512 0.526

0.024 0.087 0.115 0.042 0.072

Notes See Table 10.3 for variable definitions. The reported R 2 values are the adjusted R 2 s. ‘Full regression’ refers to the case where each of the dependent variables is regressed on all the independent variables listed in Tables 10.5 and 10.6. ‘Residual regression’ shows R 2 values which are derived via a two-stage procedure. In step 1 the residuals are derived by regressing each of the dependent variables on GDPP and in step 2 these residuals are regressed on all the independent variables, except the GDPP. The R 2 s for ‘Residual regression’ refer to the R 2 s of this second step.

132

Digital divide

GDPP do not explain much variation in the five components of the ITC industry. What do these results mean? Perhaps the main lesson is that the economic health of a country, as measured by the per capita income, is the main indicator of its overall economic well-being, and countries whose economic growth is significant – for that eventually determines per capita income – are also likely to bridge the digital divide more rapidly. This does not mean that appropriate policies as related to the ITC industry are not important or relevant. But it does mean that policies which advance economic growth are of paramount importance. It means that it would be misguided to emphasize policies for enhancing ITC per se rather than as a complement to increasing economic growth.

11 Globalization and labour market outcomes

This chapter has two main objectives: the first is to briefly outline the nature and the extent of globalization in the countries of our sample and the second is to discuss the labour market outcomes for the same set of countries. Precisely, what is globalization? Globalization has no precise meaning, but loosely speaking, it indicates shrinking of national borders from a variety of perspectives. For example, it may indicate integration of labour, capital, knowledge and other kinds of national markets. The fuzziness or the imprecision of the definition can be seen from the fact that these days a nation can have instant access to the most up-to-date information about a variety of topics without at the same time having a highly educated labour force, vast resources of capital or first-rate centres for R&D. Therefore, to keep our discussion manageable, we confine ourselves to a few indicators of integration with the international or the global economy. These indicators are summarized in Tables 11.1 and 11.2. Table 11.1 provides the extent of linkages for eight countries of our sample in terms of the physical flows and the Table 11.2 provides information about financial flows. Considering trade in goods, perhaps the oldest and the most common form of indicator of integration of national markets, it shows that practically all the countries of our sample are quite open economies. Put differently, these countries would appear to be highly integrated with the rest of the world, even if we were to confine ourselves to trade in goods only. Needless to say, even in this sub-sample, some countries are better integrated than the others. Another interesting indicator, but again of the physical variety, is the gross private capital flows/gross foreign direct investment. Both are expressed as percentages of GDP measured

165.4 12.1 20.8 59.3 9.5 264.5 35.7 23.9

239.2 12.3 23.2 802.0 24.5 275.1 35.9 29.4

739.3 63.4 43.5 211.4 77.1 854.6 108.3 123.6

1,075.2 81.3 52.3 333.7 191.1 739.5 136.3 171.2

1999

1989

1989

1999

% of goods GDP

% of PPP GDP

Trade in goods


Hong Kong Indonesia Japan Malaysia Philippines Singapore South Korea Thailand

Country

Table 11.1 Indicators of integration with the global economy

— 0.4 11.1 4.2 1.2 56.8 3.3 3.8

1989

174.7 2.4 30.9 7.1 5.3 54.2 9.1 4.9

1999

Gross private capital flows (% of PPP GDP)

— 0.2 2.0 2.1 0.3 10.6 0.5 1.0

1989

35.3 0.9 1.2 0.8 0.3 13.3 2.1 1.8

1999

Gross FDI (% of PPP GDP)

— 3,235 — 770 639 — 1,038 4,399

— −8,416 — 3,247 4,915 — 6,409 2,471

— 1,093 1,777 2,333 530 5,575 788 2,444

— −2,745 12,308 1,553 573 6,984 9,333 6,213

1999

1990

1990

1999

Foreign direct investment

Net private capital flows


Hong Kong Indonesia Japan Malaysia Philippines Singapore South Korea Thailand

Country

Table 11.2 Global financial flows ($ million)

26 — −1,239 395 — 151 −87

—

1990

Bonds

— −1,458 — 747 3,895 — −1,414 −1,538

1999

Portfolio investment flows

— 312 — 293 0 — 518 449

1990

Equity

— 1,273 — 522 422 — 12,426 2,527

1999 — 1,804 — −617 −286 — −418 1,593

1990

— −5,486 — 426 25 — −13,935 −4,911

1999

Bank and traderelated lending

136

Globalization

in purchasing power parity (PPP) terms. These columns tell the same sort of story as the trade in goods figures. An interesting issue is: what are the implications of globalization for the countries of our sample? This is, as one might suspect, a rather broad question; in fact a very broad question. Does globalization cause higher growth rates? Lower income inequality? Reduce wage inequality? Reduce the unemployment rates of unskilled workers? Reduce or increase gender gaps? And so on. Put simply, this is such a contentious issue that we cannot expect to deal with it in a neat and clean way. Therefore, we confine our survey and discussion only to the labour market outcomes. And even in this case, we consider only the issue which has attracted considerable discussion. This deals with the possible effects of globalization on the wage rates of skilled and unskilled workers and their rates of employment and, of course, whether the inequality of the two rates increases or decreases as a consequence of globalization. Unfortunately, the evidence available is not extensive, particularly for the developing countries. Most of the theoretical and particularly the empirical literature has been confined to the developed countries and has been ably discussed and brought together by Wood (1994). This poses a real problem for us. Normally, in empirical work, either we have time-series studies or cross-section or simulation or pure cross-country studies. Each one of these studies has advantages and disadvantages. Whether we use one type or the other depends on the purpose at hand and as well on what is available. For our present purposes, the ideal approach would be timeseries studies for each of the countries of our sample. This, unfortunately, is not the case. As pointed out by Wood (1997), we have evidence only for the four tigers and that too for the 1960s and 1970s. For the other countries, the little evidence that we have is based on two unpublished studies by Robbins (1994a,b), which are not available to us, so we must rely on the summaries provided by Wood (1997). Use of unpublished and unavailable works and dependence on others’ summaries has numerous hazards. But keeping these reservations in mind, we briefly restate their findings. The most commonly cited evidence relates to the four little tigers, namely Hong Kong, South Korea, Singapore and Taiwan. This evidence supports the Heckscher–Ohlin prediction that globalization (openness to trade) leads to increased demand for unskilled labour relative to the demand for skilled labour, thus reducing the wage gap between the two. More precisely, the gap narrowed during the 1960s in South Korea and Taiwan and the 1970s in Singapore. In Hong Kong,

Globalization 137 on the other hand, in the 1950s the wage gap widened, although Wood points out that this was probably caused by a simultaneous increase in the relative supply of unskilled labour. Apart from these four cases, there is some evidence for Malaysia (Robbins, 1994a) and the Philippines (Robbins, 1994b), as reported by Wood (1997). According to Robbins (1994a), as reported by Wood (1997), wage differentials by level of education shrank in Malaysia from 1973 to 1989 and then continued this pattern in the early 1990s. Wood’s discussion does not make it clear whether this result could be attributed to openness to trade. The evidence about the Philippines is even less clear. Robbins examined the movements of wages by education for the period 1978–1988. But there was no clear trend by skill differential over this period. Wood (1997) then goes on to examine the evidence available for Latin America, which contradicts the outcome for East Asia. This muddies the waters. Does it mean that the predictions of the Heckscher–Ohlin theorem are not valid? Does it mean there are problems with the empirical work? Or does it mean that there are other problems? And, of course, could it mean that globalization is not the reason either for the shrinkage or the expansion of the wage inequalities? Wood (1997) cites his work as well as that of Rodrik (1997) as possible answers.

12 What we have learned and what remains to be done

This is not a traditional book, so no traditional conclusions are possible. Besides, no useful purpose would be served by mere repetition of the points already made. However, we believe that some obvious points do emerge and it would be worthwhile to highlight them. The survey provided in Chapter 6 about the Asian crisis brings out very sharply the fact that the most interesting feature of the crisis was the speed with which the economies under consideration bounced back. This highlights two facts: first, that the progress of the economies concerned was resilient and durable and second, that the crisis had been greatly exaggerated. In retrospect, the crisis did not indicate a long-term malaise. Nevertheless, the key question which awaits further investigation is how to avoid future financial crises. Many argue that the answer lies in stronger capital controls and a reform of the international financial architecture. When the discussion of the crisis is put together with Chapter 3 on the relative roles of technological change and factor accumulation, we are obliged to look at the debate of Chapter 3 in a somewhat different light. If the crisis was short-lived, then surely we must wonder if the nature of the underlying causes of the rapid growth of the eight or nine East Asian economies was not much more complex in nature rather than the simple ‘Perspiration versus Inspiration’. While this is not the place to speculate on the underlying causes, the fact does remain that the crisis and the subject matter of Chapter 3 cannot be separated. If there are lessons to be learnt from these countries, then they must be joint ones. In short, the Asian crisis shows the strength of these economies rather than weaknesses. Another point that emerges from a number of chapters has to do with the data. It is clear that if we are to analyse many of the issues in detail,

What we have learned

139

we need new data. That was clear, for example, in Chapters 8–11. Unless new and improved data are available, we are not likely to get very far. This point was most glaringly brought out in the discussion about access to venture capital and its implications for the gender gap. Similarly, on the topic of urbanization, before we can get anywhere we need more data, as pointed out in that chapter. The chapter on corruption pointed out the limitations of what can be done at the present and what needs to be done. Once again, most of that depends on the limitations of the data. As an example, it would be interesting to examine the role of corruption in determining public and private investment separately. Then, of course, there are different measures of corruption. For now, we have only survey data, but that is not necessarily the only type of data we can or should use. While factor markets are becoming freer, that does by no means mean that factor price distortions no longer exist. We saw in Chapter 4 that the information is rather old and confined to only a few countries, so it would be useful to enlarge the sample size and bring the information more up to date. Our survey has barely scratched the surface of ‘growth, poverty and income distribution’. This topic is complicated and vast in scope. Readers may find that they would wish to expand their understanding in different directions. In Chapters 7 and 10 we have provided some original evidence on a number of questions. For example, with respect to the effects of corruption on growth and FDI, we investigate whether the growth experience of the East Asian countries has been adversely affected by the higher than average level of corruption observed in these countries. Is East Asia a special group of host countries in that corruption has little or no effect on inward FDI in East Asia, possibly because corruption has been part of the culture or a way of life for a long time? With respect to the IT industry, we try to identify the determinants of the major components of the IT industry. Unfortunately, like our predecessors, we are unable to provide any conclusive evidence on any of these questions. In fact, our findings raise more questions than the answers they provide. Nevertheless, our results do suggest the usefulness of such exercises and we believe it would be useful not only to extend the work that we have reported, but to undertake more such work. We conclude by noting the obvious. The principal lessons to be drawn from the extraordinarily rapid and prolonged growth experience of the

140

What we have learned

East Asian countries await further research. The Asian crisis has been short-lived and the economic slowdown of the East Asian economies has been temporary. Despite the temporary setback, the progress of these economies has been resilient and durable. It is therefore evident that the East Asian countries have matured and are well on their way to join the industrialized nations.

Notes

3 The East Asian Miracle: perspiration or inspiration? 1 We define East Asia, as a region, to include the eight high-performing Asian economies: Hong Kong, Indonesia, Japan, the Republic of Korea, Malaysia, Singapore, Taiwan and Thailand. We confine our discussion to these eight East Asian economies because they have been the focus of much debate about the sources of growth in the region. See, for example, the World Bank study The East Asian Miracle (World Bank, 1993) and Krugman (1994). 2 Among others, Young (1992, 1994, 1995), Kim and Lau (1994), Krugman (1994) and Collins and Bosworth (1996) subscribe to the accumulation view of the East Asian growth experience. 3 Others who also subscribe to the assimilation view of the East Asian Miracle are Dahlman and Westphal (1981), Dahlman et al. (1987), Hobday (1994a,b,c, 1995), Pack (1992), Pack and Page (1994a,b) and Nelson and Pack (1999). 4 See, for example, Young (1992, 1994, 1995, 1998), Kim and Lau (1994) and Collins and Bosworth (1996). 5 See, for example, Krugman (1994) and Mankiw (1995). 6 See, for example, Klenow and Rodriguez-Clare (1997), Sarel (1997), Hsieh (1999, 2002) and Bernanke and Gürkaynak (2001). 7 See, for example, Pack (1992), Romer (1993) and Nelson and Pack (1999). 8 See equation (3.7) for a formal definition of TFP growth. 9 See Diewert (1992), Barro (1999), Hulten (2001) and Nadiri and Prucha (2001) for an extensive review of the estimation methods discussed below. 10 The US Bureau of Labor Statistics refers to TFP as ‘multifactor productivity’. Another term coined by Robert Solow (see Solow, 1957) frequently used in the productivity growth literature is the ‘Solow residual’. All three terms are synonymous in the TFP growth literature. 11 At cost-minimizing levels of inputs, αi also represents the input expenditure share for the factor input Xi .

142

Notes

12 See Hulten (2001) for a survey of this literature as well as relevant references. 13 Another refinement of the TFP growth measure given by equation (3.11) is reflected in the view that sectoral reallocation of resources is a key factor in productivity growth. A part of the growth process consists of transferring resources from low to high productivity sectors, particularly from agriculture to industry, where capital–labour ratios are higher, and in consequence, result in higher marginal products of labour. Failure to allow for such changes in sectoral composition will result in biased estimates of TFP growth. See, for example, Denision (1967), Chen (1977), Pack (1992), Nelson and Pack (1999) and Chanda and Dalgaard (2003) for empirical estimates. 14 The Divisia index is a weighted sum of growth rates, where the weights are the components’ share in total revenue. 15 This is assuming that aggregate inputs Xi s are constant returns to scale transcendental logarithmic Xij s. In particular, it is assu

indices of sub-inputs 1 med that Xi = exp j =1 θij ln Xij + 2 j =1 p=1 ijp ln Xij ln Xip , subject to the restrictions: j =1 θij = 1, and p=1 ijp = 0. 16 Young (1995) demonstrates that this translog index of TFP growth follows from the translog value-added production function – independently developed by Christensen et al. (1971, 1973), Griliches and Ringstad (1971) and Sargan (1971): 1 φi ln Xit + ij ln Xit ln Xj t Qt = exp φ0 + 2 i=1

i=1 j =1

where ln Xnt ≡ t denotes time, and where, under the assumption of constant returns to scale, the parameters function satisfy the n−1 of the production n−1 restrictions: n−1 i=1 φi = 1 and j =1 ij = i=1 in = 0. Diewert (1976) shows that the Tronqvist approximation to the Divisia index proposed by Jorgenson and Griliches (1967) is an exact index number if the production function (3.8) has the above translog form. In other words, the Tronqvist index is not an approximation at all. Instead, it is exact under the right conditions. Furthermore, because the translog production function can be regarded as a good second-order approximation to other production functions, the discrete-time Tronqvist index is a sensible choice even when the production function does not have the translog form. 17 For expositional ease, the equation (3.14) does not allow for changing input quality. The identity, however, can be easily extended to allow consideration of more finely differentiated inputs. 18 The World Bank (1993), for example, implements this approach. 19 Here we provide only a brief introduction to the estimation procedure. Boskin and Lau (1990) provide further details. See Pack (2001) for a geometric interpretation of the meta-production function framework.

Notes 143 20 As noted earlier, the productivity growth empirics literature devoted to exploring reasons for the exemplary growth performance of the East Asian economies is large and growing. The object of this chapter is not to provide an exhaustive survey of this vast literature. Instead, the chapter highlights only those empirical works which are considered to form the basis of the East Asian TFP growth debate. 21 In a growth accounting exercise, Tsao (1985) first revealed there is little evidence of TFP growth in Singapore. It is, however, Young (1992) who provides the more renowned results that ignited the controversy regarding TFP growth in the East Asian economies. 22 Marti (1996) presents a cross-country regression similar to that of Young (1994), but estimates using a more recent version of the Summers and Heston data set and for a slightly smaller number of countries (104 instead of 118 in Young, 1994). The results, however, are quite different from those reported in Young (1994). For example, according to Marti (1996), TFP growth in Singapore was 1.49 per cent during 1970–1985. Young (1994), by contrast, reports that the estimated rate of TFP growth in Singapore was only 0.1 per cent during the same period. 23 Recall that a circumflex or ‘hat’ over a variable denotes a proportional growth rate. 24 See table 6 in Collins and Bosworth (1996). 25 See, for example, Young (1992, 1995) and Collins and Bosworth (1996), discussed above. 26 See the World Bank study The East Asian Miracle (1993), table A1.2. 27 See, for example, Kim and Lau (1995). 28 See table 3.6 above. Annex 1 in Nehru and Dhareshwar (1994) provides further evidence. 29 The dual measure can also be derived from the dual cost function of any production function. See, for example, Good et al. (1997). 30 See, for example, the exchange between Young (1995) and Marti (1996) as well as the exchange between Young (1998) and Hsieh (2002). 31 See Pack (2001) and Filipe (1999, 2000) for further details.

6 East Asian financial crisis 1 In 1997 many of the fast-growing economies in South East Asia experienced currency and financial crises. The problems began in Thailand and gradually spilled over to Indonesia, Malaysia, South Korea and the Philippines. Other countries in the region such as China, Taiwan, Singapore and Hong Kong also felt the pressure. However, these economies managed to avoid the financial and currency crises of 1997. During this period financial markets were so volatile that the Asian currency and financial crisis spilled over into other emerging markets such as Russia, South Africa and Brazil.

144

Notes

2 Table 6.1 in essence summarizes the empirical evidence reported in Tables 6.2 and 6.3 and Figure 6.4. 3 Many observers believe that the level of short-term debt relative to foreign exchange reserves is one reason why the crisis was much worse in Indonesia, South Korea and Thailand than in Malaysia and the Philippines. 4 Others who subscribe to this view are Krause (1998) and Goldstein (1998). 5 The self-fulfilling prophecies view of the Asian financial crisis is founded on the theory of a bank panic or run. Banks exist because they transform maturities – that is, they borrow short and lend long – to provide liquidity services to their depositors. This mismatch of banks’ assets and liabilities makes them susceptible to a bank panic or run. A bank run occurs when depositors en masse rush to the bank and demand immediate withdrawals of bank deposits, fearing that the rush of others to withdraw will leave them with nothing. A bank panic occurs when a number of banks experience bank runs. The model of a bank panic or run is expounded by a number of authors, including Bryant (1980), Diamond and Dybvig (1983) and Freeman (1988). We note that this is just one of two main approaches to explaining herd behaviour in financial markets. The second approach explores the role of asymmetric information among creditors as a cause of financial market instability. Stiglitz and Weiss (1981), Mishkin (1991) and Banerjee (1992), among others, employ this approach to explain herd behaviour. Here we briefly sketch a bank panic or run model based on Champ and Freeman (2001). There are n three-period-lived small depositors. Each depositor receives y units of goods when young. In period zero, an individual depositor is not sure whether they will have to withdraw funds in period one or in period two to meet their idiosyncratic consumption needs. This uncertainty over the timing of consumption needs, however, is resolved in period one. The bank receives deposits in the amount of y unit of goods from each individual in period zero. In return, the bank offers a rate of return 1 if deposits are withdrawn in period one and a rate of return x > 1 if withdrawals are made in period two. Assuming that the bank can generally forecast the overall demand for withdrawals in period one, it can finance the promised returns under the following investment scheme. Let α be the fraction of individuals who will need to consume in period one. Then the overall demand for withdrawals in period one is αny. The bank leaves this amount in storage, which yields a rate of return 1 per period. It then lends the remaining funds (1 − α)ny for a long-term project yielding a rate of return x in period two. The only problem, however, arises in the unlikely event when depositors panic, not because they all need to consume in period one, but because the other depositors are also panicking. If more than αn depositors demand

Notes 145

6 7

8 9 10 11

12

13

14

15 16

immediate withdrawals of funds in period one, the bank must call in the loan on long-term investment. The investment project must then be terminated and sold for its salvage value v < 1 in period one. In the event that the salvage value of the investment project is insufficient to cover the unexpected withdrawals in period one, the bank will be forced into liquidation. The model therefore predicts a bank panic or run as one rational equilibrium. In this framework, the bank run occurs not when depositors fear that the bank has made a poor investment decision, but when individual depositors fear that other depositors are withdrawing funds from the bank, and thereby driving the bank into illiquidity and eventual liquidation. Inter-bank lending, temporary suspensions of withdrawals and government deposit insurance may prevent bank panic or runs. According to the proponents of the self-fulfilling prophecies view, these conditions did not exist in the crisis-affected economies of East Asia, the consequence being the Asian financial crisis. See Corsetti et al. (1999). We note that total foreign debt is significant at the 5 per cent level. This result, however, is sensitive to whether non-crisis countries are included in the sample. See Table 6.6. See, for example, Corsetti et al. (1999). See Bosworth (1998). See Sims (1998). We tried to address the endogeniety issue noted earlier by instrumenting short-term foreign debt to reserve ratio with its lagged value. The regression results, however, are not significantly different from the case when the short-term foreign debt to reserve ratio is used in regressions. We therefore do not report these estimation results. The two additional non-crisis countries which we include in our sample are China and Hong Kong. All variables are constructed in the same manner as in Radelet and Sachs (1998a). A lower corruption index represents greater corruption. We note that the ratio of short-term foreign debt to reserves appears in regression model 1 with a statistically significant positive coefficient. The regression coefficient, however, becomes insignificant when the real exchange rate change is included in the regression (see models 2–4). Needless to say, in each case the coefficient appears with the expected positive sign. ‘Somewhat surprisingly, our measure of real exchange rate overvaluation does not seem to be associated with financial crisis’ (Radelet and Sachs, 1998a, p. 48). See, for example, Corsetti et al. (1999). See Wei (2001) for further details.

146

Notes

7 Corruption, growth and foreign direct investment 1 Corruption can be beneficial when individuals’ day-to-day economic decisions are complicated by cumbersome and excessive government regulations. De Soto (1989) provides a descriptive analysis of this phenomenon. In fact, several studies have found corruption to be growth-enhancing as long as ‘it greases the wheels of the economy’ and allows more and faster economic transactions. Méndez and Sepúlvea (2001) provide cross-country evidence in favour of the ‘bribery as grease’ hypothesis. 2 See, for example, Wei (1997, 2000a) and Habib and Zurawicki (2002). 3 Corruption is considered organized (or efficient) when the payers of bribes can get things done after paying a relatively well defined bribe. 4 According to Shleifer and Vishny (1993), corruption is unorganized or inefficient when there is still a high degree of uncertainty even after paying the bribe. 5 The PRS index – a measure of the perceived level of corruption – is developed by Political Risk Services Inc., a private firm that annually publishes the International Risk Guide. 6 See, for example, Wei (2001). 7 See, for example, Treisman (2000) and Wei (2000b). 8 See Kaufmann and Wei (1999) and Svensson (1999). 9 VanRijckeghem and Weder (1997) find a negative relationship between wages and perceived corruption level. Rauch and Evans (2002) and Treisman (2000), however, fail to find evidence of a negative effect of wages on corruption. 10 In the preceding sub-section we have discussed some of the economic determinants of the level of corruption. In this sub-section we discuss the sociopolitical determinants of corruption. 11 As a robustness check, Treisman also examines the BI index, an index of perceived corruption compiled by Business International, for the early 1980s. 12 Adopted from Treisman (2000), table 2. 13 The existing literature (see, for example, Bardhan, 1997, and Shleifer and Vishny, 1993) suggest that the bribery-as-grease hypothesis depends on the assumption that red tape is set exogenously. 14 See Kaufmann and Wei (1999) for further details. 15 Kaufmann and Wei (1999) provide further details. 16 These questions were identified in Wei (1997). See also Wei (2000a). 17 Wei (1997) quotes a story from the Wall Street Journal which illustrates this point. The Chinese consume a huge quantity of foreign-made cigarettes: one in every three cigarettes is smoked in China, yet duty is not paid on 90 per cent of the imports (see ‘Smugglers Stoke B.A.T.’s Cigarette Sales in China’, The Wall Street Journal, 18 December 1996). The British American

Notes 147

18 19

20

21

Tobacco company is the largest supplier of foreign cigarettes in China. In 1995, the company sold 400 million cigarettes that were duty-paid, 3 billion in duty-free shops, 4 billion in special economic zones (many of which were transported illegally to other parts of China), and 38 billion to retailers who smuggled the cigarettes directly into China. Conversations with Hong Kong business people indicate there is a well-developed fee-for-service business in Hong Kong to smuggle goods through Chinese customs. There are at least four different ways to circumvent Chinese tariffs, most of which involve bribing Chinese customs officials. A business consultant who works for a major consulting firm in Hong Kong, owned by the United States, indicated that 90 per cent of foreign wine in the Chinese market is also smuggled into the country. The United Nations’ commentary on its publication, World Investment Directory 1992, vol. 1, p. 14, testifies to this fact. The FCPA was passed following the Watergate hearings of the early 1970s, when many American firms were discovered to be paying large bribes to foreign officials. To date, no other country has passed similar legislation. Thus, the United States is the only source country in the world that penalizes its multinationals or their officers with fines or jail terms for bribing foreign government officials. Wei (1997) also reports estimation results obtained using the TI index of perceived corruption. He does not, however, explain his results. Nonetheless, the estimation results obtained by using the TI index are similar. See, for example, Méndez and Sepúlvea (2001).

8 Urbanization 1 See McGee and Robinson (1995) and reference cited therein. 10 Digital divide 1 See, for example, Quibria et al. (2002).

References

Abramovitz, M. (1956). Resources and output trends in the United States since 1870. American Economic Review 46(2), 5–23. Aggarwala, R. (1983). Price distortions and growth in developing countries. World Bank Staff Working Papers No. 575. Washington DC: World Bank. Ahuja, V., B. Bidani, F. Ferreira and M. Walton (1997). Everyone’s Miracle: Revisiting Poverty and Inequality in East Asia. Washington DC: World Bank. Banerjee, A. V. (1992). A simple model of herd behaviour. Quarterly Journal of Economics 107(3), 797–817. Bardhan, P. (1997). Corruption and development: a review of issues. Journal of Economic Literature 35, 1320–1346. Barro, R. (1991). Economic growth in a cross-section of countries. Quarterly Journal of Economics, 407–443. Barro, R. J. (1999). Notes on growth accounting. Journal of Economic Growth 4(2), 119–137. Beck, P. J., M. W. Maher and A. E. Tschoegl (1991). The impact of the Foreign Corrupt Practices Act on US exports. Managerial and Decision Economics 12(4), 295–303. Bernanke, B. S. and R. S. Gürkaynak (2001). Is growth exogenous? Taking Mankiw, Romer, and Weil seriously. In B. S. Bernanke and J. J. Rotemberg (eds), NBER Macroeconomics Annual 1997, pp. 11–57. Cambridge MA: MIT Press. Boskin, M. and L. Lau (1990). Post-war economic growth of the Group-of-five countries: a new analysis. Technical Paper 217, Conference for Economic Policy Research. Stanford CA: Stanford University. Bosworth, B. P. (1998). Comment on the East Asian financial crisis: diagnosis, remedies, prospects. Brookings Papers on Economic Activity (1), 80–84. Bryant, J. (1980). A model of reserves, bank runs, and deposit insurance. Journal of Banking and Finance 4, 335–344. Champ, B. and S. Freeman (2001). Modeling Monetary Economies (2nd edn). Cambridge: Cambridge University Press.

References 149 Chanda, A. and C.-J. Dalgaard (2003). Dual economies and international total factor productivity differences. Working Paper. Raleigh NC: North Carolina State University. Chen, E. K. (1977). Factor inputs, total factor productivity, and economic growth: the Asian case. The Developing Economies 15(1), 121–143. Christensen, L., D. Cummings and D. Jorgenson (1981). Relative productivity levels, 1947–1973: an international comparison. European Economic Review 16(1), 61–95. Christensen, L. R., D. W. Jorgenson and L. J. Lau (1971). Conjugate duality and the transcendental logarithmic production function. Econometrica 39, 255–256. Christensen, L. R., D. W. Jorgenson and L. J. Lau (1973). Transcendental logarithmic production frontiers. Review of Economics and Statistics 55, 28–45. Collins, S. M. and B. P. Bosworth (1996). Economic growth in East Asia: accumulation versus assimilation. Brookings Papers on Economic Activity (2), 135–191. Corsetti, G., P. Pesenti and N. Roubini (1999). What caused the Asian currency and financial crisis? Japan and the World Economy 11, 305–373. Dahlman, C. and L. E. Westphal (1981). The meaning of technological mastery in relation of transfer of technology. Annals of the American Academy of Political and Social Science 458, 12–26. Dahlman, C., B. Ross-Larson and L. E. Westphal (1987). Managing technological development: lessons from the newly industrialized countries. World Development 15(6), 759–775. Datt, G. and M. Ravallion (1992). Growth and redistribution components of changes in poverty measures: a decomposition with applications to Brazil and India in the 1980s. Journal of Development Economics 38(2), 275–295. de Melo, J. A. P. (1977). Distortions in the factor market: some general equilibrium estimates. Review of Economics and Statistics 59(4), 398–405. Denison, E. F. (1967). Why Growth Rates Differ: Postwar Experiences in Nine Western Countries. Washington DC: Brookings Institution. De Soto, H. (1989). The Other Path: the Invisible Revolution in the Third World. New York: Harper and Row. Diamond, D. W. and P. H. Dybvig (1983). Bank runs, deposit insurance, and liquidity. Journal of Political Economy 91, 401–419. Diewert, E. M. (1976). Exact and superlative index numbers. Journal of Econometrics 4, 115–145. Diewert, E. M. (1992). The measurement of productivity. Bulletin of Economic Research 44, 163–198. Di Tella, R. and E. Shargrodsky (2000). The role of wages and auditing during a crackdown on corruption in the city of Buenos Aires. Boston MA: Harvard Business School. Processed.

150

References

Djankov, S. and P. Murrell (2002). Enterprise restructuring in transition: a quantitative survey. Journal of Economic Literature 40(3), 739–792. Dougherty, C. and M. Selowsky (1973). Measuring the effects of misallocation of labour. Review of Economics and Statistics 55(3), 386–390. Filipe, J. (1999). Total factor productivity growth in East Asia: a critical survey. Journal of Development Studies 35(4), 1–41. Filipe, J. (2000). On the myth and mystery of Singapore’s zero TFP. Asian Economic Journal 14(2), 187–209. Fisman, R. (2001). Estimating the value of political connections. American Economic Review 91(4), 1095–1012. Fisman, R. and J. Svensson (2000). Are corruption and taxation really harmful to growh? Firm-level evidence. Washington DC: World Bank and New York: Columbia University. Processed. Floystad, G. (1975). Distortions in the factor market: an empirical investigation. Review of Economics and Statistics 57(2), 200–213. Freeman, S. (1988). Banking as the provision of liquidity. Journal of Business 61, 45–64. Gatti, R. (2001). Corruption and trade tariffs or a case for uniform tariffs. Policy Research Department Working Paper No. 2216. Washington DC: World Bank. Goldstein, M. (1998). The Asian Financial Crisis: Causes, Cures and Systemic Implications. Washington DC: Institute for International Economics. Good, D. M., I. Nadiri and R. Sickles (1997). Index number and factor demand approaches to the estimation of productivity. In M. H. Pesaran and P. Schmidt (eds) Handbook of Applied Econometrics II, Microeconometrics. Oxford: Blackwell. Greene, P. G., C. G. Brush, M. M. Hard and P. Saparito (2001). Patterns of venture capital funding: is gender a factor? Venture Capital 3(1), 68–83. Griliches, Z. and V. Ringstad (1971). Economies of Scale and the Form of the Production Function. Amsterdam: North-Holland. Gupta, K. L. (1989). Industrialization and Employment in Developing Countries. London: Routledge. Gupta, S., H. Davoodi and E. Tiongson (2001). Corruption and the provision of health care and education services. In A. Jain (ed.), The Political Economy of Corruption. London: Routledge. Habib, M. and L. Zurawicki (2002). Corruption and foreign direct investment. Journal of International Business Studies 33(2), 291–307. Hayami, Y. and V. Ruttan (1970). Agricultural productivity differences among countries. American Economic Review 60, 895–911. Hayami, Y. and V. Ruttan (1985). Agricultural Development: an International Perspective. Baltimore MD: Johns Hopkins University Press. Hines, J. (1995). Forbidden payment: foreign bribery and American business after 1977. NBER Working Paper No. 6266.

References 151 Hobday, M. (1994a). Export-led technology development in the four dragons: the case of electronics. Development and Change 25, 333–361. Hobday, M. (1994b). Technological learning in Singapore: a test case of leapfrogging. Journal of Development Studies 30(3), 831–858. Hobday, M. (1994c). Innovation in East Asia: diversity and development. In M. Dodgson and R. Rothwell (eds), The Handbook of Industrial Innovation. Cheltenham: Edward Elgar. Hobday, M. (1995). Innovation in East Asia: The Challenge to Japan. Cheltenham: Edward Elgar. Hsieh, C.-T. (1999). Productivity growth and factor prices in East Asia. American Economic Review 89(2), 133–138. Hsieh, C.-T. (2002). What explains the industrial revolution in East Asia? Evidence from the factor markets. American Economic Review 92(3), 502–526. Hulten, C. R. (2001). Total factor productivity: a short biography. In C. R. Hulten, E. R. Dean and M. J. Harper (eds), New Developments in Productivity Analysis, pp. 1–47. Chicago IL: University of Chicago Press. Islam, N. (2002). Growth empirics: a panel data approach. Quarterly Journal of Economics 110(4), 1127–1170. Jeanne, O. (1997). Are currency crises self-fulfilling? A test. Journal of International Economics 43, 263–286. Jones, C. I. (1994). Economic growth and the relative price of capital. Journal of Monetary Economics 34(3), 359–382. Jorgenson, D. W. and Z. Griliches (1967). The explanation of productivity change. Review of Economic Studies 34(3), 349–383. Jorgenson, D. W., F. M. Gollop and B. M. Framueni (1987). Productivity and US Economic Growth. Cambridge MA: Harvard University Press. Kaufmann, D. and S.-J. Wei (1999). Does grease money speed up the wheels of commerce? Policy Research Department Working Paper No. 2254. Washington DC: World Bank. Kaufmann, D., A. Kraay and P. Zoido-Lobatón (1999). Governance matters. Policy Research Working Paper No. 2196. Washington DC: World Bank. Kendrick, J. (1961). Productivity Trends in the United States. New York: National Bureau of Economic Research. Kim, J.-I. and L. J. Lau (1994). The sources of economic growth of the East Asian newly industrialized countries. Journal of the Japanese and International Economies 8(3), 235–271. Kim, J.-I. and L. J. Lau (1995). The role of human capital in the economic growth of the East Asian newly industrialized countries. Asia-Pacific Economic Review 1(3), 3–22. Klenow, P. J. and A. Rodriguez-Clare (1997). The neoclassical revival in growth economics: has it gone too far? In B. S. Bernanke and J. J. Rotemberg (eds), NBER Macroeconomics Annual 1997. Cambridge MA: MIT Press.

152

References

Knack, S. and P. Keefer (1995). Institutions and economic performance: cross-country test using alternative institutional measures. Economics and Politics 7(3), 207–227. Krause, L. B. (1998). The Economics and Politics of the Asian Financial Crisis of 1997–98. New York: Council on Foreign Relations. Krugman, P. (1994). The myths of Asia’s miracle. Foreign Affairs 73(6), 62–78. Kwon, J. K. and H. Paik (1995). Factor price distortions, resource allocation, and growth: a computable general equilibrium analysis. Review of Economics and Statistics 77(4), 664–676. Lambsdorff, J. G. (2000). The precision and regional comparison of perceived levels of corruption: interpreting the results. Berlin: Transparency International and Göttingen: Göttingen University. La Porta, R., F. L. de Silanes, A. Shleifer and R. Vishny (1999). The quality of government. Journal of Law, Economics and Organization 15(1), 222–279. McGee, T. and I. Robinson (eds) (1995). The Mega-urban Regions of Southeast Asia. Vancouver WA: UBC Press. Mankiw, G. N. (1995). The growth of nations. Brookings Papers on Economic Activity (1), 275–326. Mankiw, G., D. Romer and D. Weil (1992). A contribution to the empirics of economic growth. Quarterly Journal of Economics 107(2), 407–437. Marti, C. (1996). Is there an East Asia miracle? Union Bank of Switzerland Economic Research Working Paper. Zurich: Union Bank of Switzerland. Mauro, P. (1995). Corruption and economic growth. Quarterly Journal of Economics 110(3), 682–712. Mauro, P. (1997). The effects of corruption on growth, investment, and government expenditure: a cross-country analysis. In K. A. Elliott (ed.), Corruption and the Global Economy. Washington DC: Institute for International Economics. Méndez, F. and F. Sepúlvea (2001). Is corruption harmful for growth? Evidence from a cross-section of countries. Working Paper. East Lansing MI: Michigan State University. Mishkin, F. S. (1991). Asymmetric information and financial crises: a historical perspective. In G. Hubbard (ed.), Financial Markets and Financial Crises. Chicago IL: University of Chicago Press. Nadiri, I. M. (1970). Some approaches to the theory and measurement of total factor productivity: a survey. Journal of Economic Literature 8(4), 1137–1177. Nadiri, I. M. and I. R. Prucha (2001). Dynamic factor demand models and productivity analysis. In C. R. Hulten, E. R. Dean and M. J. Harper (eds), New Developments in Productivity Analysis. Chicago IL: University of Chicago Press.

References 153 Nadiri, I. and W. Son (1998). Sources of growth in East Asian economies. New York: New York University. Nehru, V. and A. Dhareshwar (1994). New estimates of total factor productivity growth for developing and industrial countries. Policy Research Working Paper. Washington DC: World Bank. Nelson, R. and H. Pack (1999). The Asian miracle and modern growth theory. Economic Journal 109, 416–436. Noble, G. W. and J. Ravenhill (2000). Causes and consequences of the Asian financial crisis. In G. W. Noble and J. Ravenhill (eds), The Asian Financial Crisis and the Architecture of Global Finance. Cambridge: Cambridge University Press. Norris, P. (ed.) (2001). Digital Divide: Civic Engagement, Information Poverty and the Internet Worldwide. Cambridge: Cambridge University Press. Nye, J. (1967). Corruption and political development: a cost–benefit analysis. American Political Science Review 61(2), 417–427. Pack, H. (1992). Technology gaps between industrial and developing countries: are there dividends for latecomers? In L. H. Summers and S. Shah (eds), Proceedings of the World Bank Annual Conference on Development Economics. Washington DC: World Bank. Pack, H. (2001). Technological change and growth in East Asia: macro versus micro perspectives. In J. E. Stiglitz and S. Yusuf (eds), Rethinking the East Asian Miracle, 2001. New York: Oxford University Press. Pack, H. and J. M. Page (1994a). Accumulation, exports and growth in the high performing Asian economies. Carnegie-Rochester Conference Series on Public Policy 40, 199–236. Pack, H. and J. M. Page (1994b). Reply to Alwyn Young. Carnegie-Rochester Conference Series on Public Policy 40, 251–257. Quibria, M., S. N. Ahmed, T. Tschang and M.-L. Reyes-Masearaquit (2002). Digital divide: determinants and policies with special reference to Asia. ERD Working Paper Series No. 27. Manila: Asian Development Bank. Radelet, S. and J. Sachs (1998a). The East Asian financial crisis: diagnosis, remedies, prospects. Brookings Papers on Economic Activity (1), 1–90. Radelet, S. and J. Sachs (1998b). The onset of the East Asian financial crisis. NBER Working Paper No. 6680. Rauch, J. and P. Evans (2002). Bureaucratic structure and bureaucratic performance in less-developed countries. Journal of Public Economics 40(3), 739–792. Robbins, D. (1994a). Malaysian wage structure and its causes. Cambridge MA: Harvard Institute for International Development. Processed. Robbins, D. (1994b). Philippine wage and employment structure, 1973–93. Cambridge MA: Harvard Institute for International Development. Processed. Rodrik, D. (1997). Has Globalization Gone Too Far? Washington DC: Institute for International Economics.

154

References

Romer, P. M. (1993). Idea gaps and object gaps in economic development. Journal of Monetary Economics 32(3), 543–573. Rose-Ackerman, S. (1975). The economics of corruption. Journal of Public Economics 4, 187–203. Rose-Ackerman, S. (ed.) (1978). Corruption: A Study in Political Economy. New York: Academic Press. Sarel, M. (1997). Growth and productivity in ASEAN countries. IMF Working Paper. Washington DC: IMF. Sargan, D. J. (1971). Production functions. In P. Layard, J. Sargan, M. Ager and D. Jones (eds), Qualified Manpower and Economic Performance. London: Penguin Press. Sen, A. K. (2001). The many faces of gender inequality. New Republic, 35–40. Shleifer, A. and R. Vishny (1993). Corruption. Quarterly Journal of Economics 108(3), 599–617. Sims, C. (1998). Comment on the East Asian financial crisis: diagnosis, remedies, prospects. Brookings Papers on Economic Activity (1), 85. Solow, R. M. (1957). Technical change and the aggregate production function. Review of Economics and Statistics 39(3), 312–320. Stiglitz, J. E. and A. Weiss (1981). Credit rationing in markets with imperfect information. American Economic Review 71(3), 393–410. Summers, R. and A. Heston (1991). The Penn World Table (Mark 5): an expanded set of international comparisons, 1950–1988. Quarterly Journal of Economics 106(2), 327–368. Svensson, J. (1999). Who must pay bribes and how much? Evidence from a cross-section of firms. Policy Research Department Working Paper No. 2486. Washington DC: World Bank. Tanzi, V. and H. Davoodi (2001). Corruption, growth and public finance. In A. Jain (ed.), The Political Economy of Corruption. London: Routledge. Thomas, V. and Y. Wang (1996). Distortions, interventions, and productivity growth: is East Asia different? Economic Development and Cultural Change, 265–288. Treisman, D. (2000). The causes of corruption: a cross-national study. Journal of Public Economics 76, 399–457. Tsao, Y. (1985). Growth without productivity: Singapore manufacturing in the 1970s. Journal of Development Economics 19(1), 25–39. UN (1992). World Investment Directory, Vol. 1. New York: United Nations. UN (1998). Demographic Yearbook. New York: United Nations. UNDP (1997). Human Development Report. New York: Oxford University Press. UNDP (1999). Human Development Report. New York: Oxford University Press. UNDP (2000). Human Development Report. New York: Oxford University Press.

References 155 UNDP (2001). Human Development Report. New York: Oxford University Press. VanRijckeghem, C. and B. Weder (1997). Corruption and the rate of temptation: do wages in the civil service cause corruption? International Monetary Fund Working Paper No. 97/73. Washington DC: IMF. Wade, R. (1990). Governing the Market: Economic Theory and the Role of Government in East Asian Industrialization. Princeton NJ: Princeton University Press. Wei, S.-J. (1997). How taxing is corruption on international investors? NBER Working Paper No. 6030. Wei, S.-J. (2000a). How taxing is corruption on international investors? Review of Economics and Statistics 82(1), 1–11. Wei, S.-J. (2000b). Natural openness and good government. NBER Working Paper No. 7765. Wei, S.-J. (2000c). Local corruption and the global economy. NBER Reporter. Wei, S.-J. (2001). Negative alchemy? Corruption, composition of capital flows, and currency crisis. NBER Working Paper No. 8187. Wheeler, D. and A. Mody (1992). International investment location decisions: the case of US firms. Journal of International Economics 33, 57–76. Wood, A. (1994). North–South Trade, Employment and Inequality: Changing Fortunes in a Skill-driven World. Oxford: Clarendon Press. Wood, A. (1997). Openness and wage inequality in developing countries: the Latin American challenge to East Asian conventional wisdom. World Bank Economic Review 11, 33–57. World Bank (1979). World Development Indicators. Washington DC: World Bank. World Bank (1986). World Development Report. New York: Oxford University Press. World Bank (1993). The East Asian Miracle: Economic Growth and Public Policy. New York: Oxford University Press. World Bank (1997a). Income inequality in China. Report 16685-CHA. Washington DC: World Bank. World Bank (1997b). World Development Report. New York: Oxford University Press. World Bank (2000a). World Development Indicators. Washington DC: World Bank. World Bank (2000b). World Development Report. New York: Oxford University Press. World Bank (2001a). World Development Indicators on CD-ROM. Washington DC: World Bank. World Bank (2001b). World Development Report. New York: Oxford University Press.

156

References

World Bank (2002). World Development Report. New York: Oxford University Press. World Bank (2004). World Development Indicators on line. Washington DC: World Bank. Young, A. (1992). A tale of two cities: factor accumulation and technical change in Hong Kong and Singapore. In O. J. Blanchard and S. Fischer (eds), NBER Macroeconomics Annual 1992. Cambridge MA: MIT Press. Young, A. (1994). Lessons from the East Asian NICs: a contrarian view. European Economic Review 38(3–4), 964–973. Young, A. (1995). The tyranny of numbers: confronting the statiscial realities of the East Asian growth experience. Quarterly Journal of Economics 110(3), 641–680. Young, A. (1998). Alternative estimates of productivity growth in the NICs: a comment on the findings of Chang-tai Hsieh. NBER Working Paper No. 6657.

Index

Abramovitz, M. 24–25 accumulation view of East Asian Miracle 22–23, 38–49, 51 adult literacy 10, 12, 115 Aggarwala, R. 58–60 aggregative aspects of growth 5 agriculture 7 Ahuja, V. 65 American firms’ competitiveness, effect of corruption on 100 anti-corruption activities among trading partners abroad 90 Argentina, study of procurement contracts in public hospitals 91 arithmetic weighting approach 26 ASEAN see Association of Southeast Asian Nations Asian crisis see Asian financial crisis Asian Development Bank (ADB) 122; study 122–123 Asian financial crisis 3; assessing empirical evidence 74–82; contagion, widespread 83; currency crisis 67, 143; fundamental imbalances 77–78; nature and extent of 71–74; self-fulfilling prophecies view 83; short-lived 83, 138, 140 assimilation view of East Asian Miracle 22–24 Association of Southeast Asian Nations (ASEAN) 67, 109

baht see Thailand Bangkok 107–109 bank/banking: bilateral loans 82; deposit funds 80; lending 76–78; panic or run 144–145; relational 83 bankruptcies of chaebols (family-owned large conglomerates) in Korea 88 Bardhan, P. 94, 146 Barro, R. J. 56–57, 60, 141 Barro-type growth equation 60 basic facility inequality 115–116 basic national income accounting identity 30 Beck, P. J. 100 Bernanke, B. S. 49, 141 BI see Business International (BI) index birth rate, crude 10–11 Boskin, M. 142 Bosworth, B. P. 40–41, 44, 49, 55, 141, 143 Brazil 143 bribery 86, 95; of foreign government officials 147; as grease, hypothesis 94, 146; monetary value of 96 British rule, history of 92–93 Bryant, J. 144 budget deficits 76

158

Index

Business International (BI) index of perceived corruption 86–87, 94, 102 Canada 108, 123–125 capital: account, liberalization 68; accumulation 53, 57, 59; augmenting 46; controls, stronger 138; inflows to GDP ratio 76–78, 80–81; markets, premature liberalization of 83; net flows, reversal of 71; rental cost of 61; restrictions on movements 68 Cebu (Philippines) 108 cellular mobile telephones 123; per capita 127–131 Central Asia: annual growth rates 6; trade and globalization 9 Champ, B. 144 Chanda, A. 142 Chen, E. K. 142 Chiang Mai in Thailand 108 China 39, 68–69, 76, 121, 143, 145; economy 1; FDI flow-to-GDP ratio 98; as host for FDI 101; as non-crisis economy 67–68, 76–77, 145; rampant corruption 98 Chinese, overseas 100 Christensen, L. R. 142 civil service management, poor 88 Cobb–Douglas production function 26, 37 Collins, S. M. 40–41, 49, 56, 141, 143 Colombia 59 colonial past 92–93 commercial energy: consumption per capita 14; use and efficiency 13, 15 common law system 93 competitiveness 81 computable general equilibrium (CGE) model 59 conventional regression approach to estimating TFP growth, error correction model 36–37

corruption 3, 97, 103; campaign against 91; countervailing institutions 99, 104; determinants of 88–93; detrimental effects in economy 104; effects on growth and FDI 84, 98–103, 139; in G-7 countries 88, 90; grand 91; level of 80–81; limitations of data 139; measures of 86; new evidence of effects on growth 96–98; organized 86; and per capita income 84; perceived determinants of 93; petty 91; in public procurement 91; subversion of legitimacy of state 85; as tax 84; trend of 88; see also Business International (BI) index; Political Risk Services (PRS) Group, index; Transparency International (TI), annual index corruption in East Asia 88–89; comparison of various indicators of 87; indices, survey-based 86, 103, 146; nature and extent of 85–88 Corsetti, G. 68, 71, 76–77, 80, 145 crawling peg 69 credit: build-up, private 80; growth 76 crisis countries: external financing 73; fundamental imbalances and Asian crisis 77–78; macroeconomic adjustment patterns 72 currency: crisis 67, 82, 104; depreciation 76; see also Asian financial crisis current account: balance in crisis countries 71; deficits 76; real exchange rate change 80; surplus to GDP ratio 77–78, 81 Dahlman, C. 141 Dalgaard, C.-J. 142 Datt, G. 65 Davoodi, H. 85 death rates 112; crude 10–11

Index 159 de Melo, J. A. P. 59 democracy 93 democratization 92 Demographic Yearbook 114 demography/demographic: transition 10–11; urban planning 105 Denison, E. F. 33, 142 De Soto, H. 146 developing countries, choice of techniques in 58 Dhareshwar, A. 36–37, 47, 143 Diamond, D. W. 144 Diewert, E. M. 141–142 digital divide 3, 122; OLS estimates 129; SUR estimates 130 disorganized (or inefficient) corruption 86 Di Tella, R. 91 Divisia index 29, 142 Djankov, S. 91 domestic burden of work 119 domestic policies 21; selected indicators of 20–21 Dougherty, C. 59 dual approach to growth accounting 30; measure of TFP 31, 53, 56, 143 Dybvig, P. H. 144 East Asia 97; assimilationist reinterpretation of miracle 49–55; definition 141; financial crisis 67; as host region 99; and Pacific 6, 9, 15, 19; paradox 99; post-war growth 40; sources of growth 41 East Asian economies: fundamental imbalances in 76; high-performing 141; misguided economic policies governments in 82; real gross domestic product (GDP) per capita 22; temporary economic slowdown of 140 East Asian NICs see newly industrialized countries Eastern Europe, transition economies of 85

econometric approach 31–37; advantages 32; cross-country regression analysis 32–33 economy/economic conditions: corruption 89; factors 89–92; growth 132; indicators in East Asian economies, key 69; rents 90; sectors, transformation in 10; welfare 59 education 123; expenditure, public 20; per tertiary student 127–131 efficiency: equivalent units of output and factor inputs 45; improvements in 25 ‘efficient grease’ hypothesis 94–95 elasticity: of employment 61; of output 42 electricity 109 employment 58, 61–62 endogeniety issue 145 Engle–Granger test for cointegration 37 ethnolinguistic division 93 Euro-money country risk ratings 75 Europe 100; annual growth rates 6; trade and globalization 9 Evans, P. 146 exchange rate 67; in crises countries 68; fixed 67; free floating 69; overvaluation 81 export performance 7 factor: and commodity prices, role of resource allocation 58; incomes 30; markets 139 factor accumulation 28, 39–40; high rates of 38; selected indicators of 17; versus technological change 2 factor price distortions 2; composite indexes of 60; cost of 60; and economic activity 58; economic welfare and capital accumulation 58; estimates of 60–62; fundamental 68; and growth 59–60; quantitative estimates of 60; static welfare cost of 59

160

Index

family workers, ratio of contributing 111 FCPA see Foreign Corrupt Practices Act in United States FDI see foreign direct investment federal states and corruption 92 female per capita income see women fertility rate, total 10–11 Filipe, J. 143 finance/financial conditions: crisis avoidance 138; external 73; fragility 76; markets 143–144; predicting onset of crises 79 Finland 121 fiscal balance 77–78 Fisman, R. 84–85 Floystad, G. 59 Foreign Corrupt Practices Act in United States 98–99, 100, 116 foreign debt 145; to reserves ratio 145 foreign direct investment (FDI) 3, 72, 84; bank credit ratio 82; bilateral 82; inward effects of corruption on 100; large volume into East Asia 99 foreign exchange reserves 144; see also exchange rate foreign indebtedness 76 Four Little Tigers 1–2, 136 France 44, 48, 121; 1991–1993 franc crisis 82 Freeman, S. 144 fundamentalist hypothesis 82 G-5 (Group of Five) industrialized countries 44–46 G-7 countries 86; corruption in 88 Gatti, R. 92 GDP growth rates 72, 78; in crisis countries 74 GDP per capita 5, 10, 127–131; initial 97; log 93; real 12 gender 114, 126; bias in access to venture capital 119, 121; empowerment 115; gap 111, 136; inequality in economic

activity 111–113; intra-household inequalities 119 gender empowerment measure (GEM) 115 geometric weighting approach 26 Germany 44, 48, 121 global financial flows 135 globalization 3; implications of 136; and labour market outcomes 133; nature and extent of 133 Goldstein, M. 144 Good, D. M. 143 government: consumption spending 97; deposit insurance 145; quality 100; ties with business elites 82 Greene, P. G. 119, 121 Griliches, Z. 26, 28–30, 142 gross domestic investment 17; as percentage of GDP 18 gross domestic saving 17 gross enrolment ratios 14 gross national income (GNI) 18–19, 127; R&D expenditure as percentage 18 growth: accounting exercise 46; average annual rates 6; cross-country regression analysis 39; and income inequality 64; and inflation rates 76; of per capita GDP 10; and poverty reduction 63–64; rate of household final consumption expenditure 7; standard primal approach to accounting 27, 49, 55 Gupta, K. L. 58, 60–61 Gupta, S. 85 Gürkaynak, R. S. 49, 141 Habib, M. 146 Hanam in South Korea 108 Hayami, Y. 33 HDI see Human Development Index health, public expenditure on 20–21 Heckscher–Ohlin theorem 137; prediction on globalization 136 hedge funds 83

Index 161 herd behaviour 144; in financial markets 144 Heston, A. 43 Hicksian efficiency parameter 27 Hicks-neutral technological change 56; technical progress 27 high growth potential firms 119 high income countries 15–16, 18–19, 21; input–output relationships 43 high-performing Asian economies (HPAEs) 42; total factor productivity growth 43 Hines, J. 98–100, 103 Hobday, M. 141 Hong Kong 1, 53, 88, 135–136, 141, 143, 145; basic facility inequality 116; corruption in 87, 101; demographic transition 11; domestic policy indicators 20; as East Asian NIC 44–45; energy consumption 14; energy use and efficiency 15; factor accumulation indicators 17; FDI 100–101; GDP in per capita terms 13; gender inequality 113; global financial flows 135; human capital accumulation 16; human development 12; inequality in 64; information age 123–125; integration with global economy 134–136; life expectancy 10; male–female income 112; mortality inequality 114; natality inequality 114; as non-crisis country 76; as one of ‘Four little Tigers’ 2, 136; ownership inequality 118; population below international poverty line 64; professional inequality 117; in service sector 7; sources of growth 44–45; structure of production 8; technological change 18–19; total factor productivity growth 38–39, 42–43, 48, 50, 54; trade and globalization 9; traffic 107;

urbanization, indicators of 105–106; venture capital 121 household inequality 119–121; gender work burden 120 house price-to-income ratio 109 housing sector 107 Hsieh, C.-T. 30, 38, 49, 53, 56, 141 Hull in Canada 108–109 Hulten, C. R. 141, 142 human capital 45–47, 49; accumulation 14, 16, 48; augmenting 46; dual role in development process 48; endowments 41; selected indicators of 16; sources of growth in a model with 46; sources of growth in a model without 44 Human Development Index (HDI) 10, 111, 115 human development or quality of life indicators, trends in 12 Human Development Report 12, 20, 112–113, 115–118, 120–121 idea gaps 23 illiquidity 145 IMF see International Monetary Fund income: distribution 126; inequalities 3 index number approach 25–31; primal approach 27–30 India 121; economy 1 Indonesia 1; annual growth rate 6; annual output growth 74; basic facility inequality 116; corruption in 87–88, 101; currency crisis 67, 71–72, 76, 143; demographic transition 11; domestic policy indicators 20; as East Asian NIC 44–45; economic indicators 69; energy consumption 14; energy use and efficiency 15; external financing 73; factor accumulation indicators 17; factor price distortion 60; FDI 100–101; GDP in per capita terms 13; gender

162

Index

Indonesia (Continued) inequality 113; global financial flows 135; growth accounting 52; household inequality 120; as HPAE 42, 141; human capital accumulation 16; human development 12; inequality in 64; information age 123–125; integration with global economy 134–136; Jakarta 109; life expectancy 10; male–female income 112; mortality inequality 114; natality inequality 114; ownership inequality 118; population below international poverty line 63–64; poverty reduction, decomposition of 66; professional inequality 117; public education expenditure 21; real exchange rate 71–72; rupiah 67; in service sector 7; share of agriculture 7; short-term debt and reserves in 70, 144; sources of growth 41, 44–45; structure of production 8; Suharto 85, 102; Surabaya 109; technological change 18–19; total factor productivity growth 38–39, 42–43, 48, 50–51, 54; trade and globalization 9; traffic 107; urbanization, indicators of 105–108; venture capital 121 industrialized countries 11–12, 20, 40; sources of growth 41 industrial market economies 15–16, 19 inequality: in East Asia 64; in ownership of property 118 infant mortality rates 10, 12 inflation 91, 97; rate 77–78 information: age, new and old technology 124–125; asymmetric 144; flows 90 information and communication technologies: access to new and old 123; expenditure as a percentage of GDP 123

information technology (IT) industry: determinants of major components of 122; major components of 139 institutional environments 47 integration with global economy, indicators of 133–134 inter-bank lending 145 interest rate spread 77–78, 80–82 international capital: flows, short-term 83; markets, inherent instabilities in 80 international financial architecture, reform of 138 international financial markets, inherent instabilities in 71 international investors 100 International Monetary Fund (IMF) 1; International Financial Statistics 68, 75, 77, 80; International Financial Statistics database 99 international poverty line measures 63 International Risk Guide 81, 146 internet use 122–123; in Europe and America, trends in 122; per capita 127–131 investment 97 Islam 96 Israel 121 Jakarta (Indonesia) 108–109 Japan 1; annual growth rates 6; basic facility inequality 116; birth rate 10; corruption in 87–88, 101; demographic transition 11; domestic policy indicators 20; economic indicators 69; energy consumption 14; energy use and efficiency 15; factor accumulation indicators 17; FDI 100–101; as G-5 44–45; GDP in per capita terms 13; gender inequality 111, 113; global financial flows 135; as HPAE 42, 141; human capital accumulation 16; human development 12; inequality in 64;

Index 163 information age 123–125; integration with global economy 134–136; investors sensitive to corruption 102; lending rate in 80–81; life expectancy 10; male–female income 112; mortality inequality 114; natality inequality 114; as non-crisis country 67, 76; as OECD country 48; ownership inequality 118; population below international poverty line 13, 64; professional inequality 115–117; ‘saturation’ point 14; in service sector 7; sources of growth 44–45; structure of production 8; technological change 18–20; total factor productivity growth 38–39, 42–43, 48, 50, 54; trade and globalization 9; traffic 107; urbanization, indicators of 105–106; venture capital 121 Jeanne, O. 82; model of currency crisis 82 Jones, C. I. 58–60 Jorgenson, D. W. 25, 28–30, 142 judiciary, weak 88 Kaufmann, D. 84, 95–96, 146 Kaufmann and Wei model 95–96 Keefer, P. 104 Kendrick, J. 25 Kim, J.-I. 33–34, 44, 46, 49, 56, 141, 143 Kim and Lau on TFP growth measurements: econometric approach implemented in 33–36; findings of 44–47 Klenow, P. J. 38, 49–50, 141 Knack, S. 104 knowledge or technology gaps, closing 38 Korea, Republic of see South Korea Krause, L. B. 144 Krugman, P. 40, 49, 141 Kwon, J. K. 59

labour 45; with advanced scientific knowledge 47; augmenting 56; and human capital, growth in 46; market outcomes 133; productivity 25; unskilled 136 Lambsdorff, J. G. 87 La Porta, R. 92 Latin America 43, 137; and Caribbean 6, 9; sources of growth 41 Lau, L. 36, 142 Lau, L. J. 33–34, 44, 46, 49, 56, 141, 143 learning and entrepreneurship, source of 48 liberalization of financial systems 83 life expectancy 12, 112; at birth 10 McGee, T. 108, 147 Malaysia 1; annual growth rate 6; annual output growth 74; basic facility inequality 116; corruption in 87–88, 101; currency crisis 67, 71–72, 76, 143; demographic transition 10–11; domestic policy indicators 20; domestic savings 14; as East Asian NIC 1, 44–45; economic indicators 69; energy use and efficiency 15; external financing 73; factor accumulation indicators 17; factor price distortion 60; GDP in per capita terms 13; gender inequality 113; global financial flows 135; growth accounting 52; as HPAE 42–43, 141; human capital accumulation 16; human development 12; inequality in 64; information age 123–125; integration with global economy 134–136; life expectancy 10; male–female income 112; mortality inequality 114; natality inequality 114; ownership inequality 118; population below international poverty line 63–64; poverty reduction, decomposition of 66;

164

Index

Malaysia (Continued) professional inequality 117; public education expenditure 21; real exchange rate 71–72; ringgit 57; short-term debt and reserves in 68, 70, 143; social unrest in 88; sources of growth 41, 44–45; structure of production 8; technological change 18–19; total factor productivity growth 38–39, 42–43, 48, 50–51, 54; trade and globalization 9; traffic 107; urbanization, indicators of 105–108; venture capital 121 male–female income 112 male per capita income 111 managed float 69 managerial methods 47 Mankiw, G. 49–50, 141 Mankiw–Romer–Weil growth equation 49 Marti, C. 32, 143 Mauro, P. 84–87, 94, 104 Méndez, F. 94–95, 146–147 Méndez and Sepúlvea capital accumulation model 95 meta-production function 33–34, 44–46, 142 Mexico 84, 100 Middle East 6, 9 military expenditure 20–21 Mishkin, F. S. 144 Mody, A. 98, 103 moral hazard problems 82–83 Morgan, J. P. 81 Morgan RER series 81 mortality inequality 112–114 multifactor productivity 141 Murrell, P. 91 Nadiri, I. M. 25, 49, 141 natality inequality 114 Nehru, V. 36–37, 47, 143 Nelson, R. 56, 141–142 neoclassical growth model 24 neoclassical production function 27 ‘New Economy’ firms 119

newly industrialized countries (NICs) 1, 44, 45, 48; poor natural resource endowment 47; sources of growth 41 new technologies, adoption of 39 NICs see newly industrialized countries Noble, G. W. 69 non-American investors 100 non-constant returns to scale 31 non-crisis countries 145 Norris, P. 122 Norway 111, 113, 115, 117–118 Nye, J. 85–86 object gaps 23 OECD countries 47–48 open economies 97, 133, 137 output growth in East Asian Countries, annual 74 ownership inequality 118–119 Pack, H. 55–56, 141–142 Page, J. M. 141 Paik, H. 59 patent applications 18, 21 peg 69; see also exchange rate Penang (Malaysia) 108–109 Penn World Tables 43 per capita: GDP growth 5; income 123, 132 perfect competition 31 personal computers 18, 123; per capita 127–131 Philippines 1–2, 76, 115, 118; annual output growth 74; basic facility inequality 116; Cebu 108; corruption in 87–88, 101; currency crisis 67–68, 71–72, 76, 143; economic indicators 69; external financing 73; factor price distortion 60; gender inequality 113; global financial flows 135; household inequality (rural) 120; inequality in 64; information age 123–125; integration with global economy 134–136; male–female

Index 165 income 111–112; mortality inequality 114; natality inequality 114; ownership inequality 118; peso 67; population below international poverty line 63–64; poverty reduction, decomposition of 66; professional inequality 117; real exchange rate 71–72; short-term debt and reserves in 68, 70, 143; urbanization, indicators of 105–108; venture capital 121; wages by education 137 physical capital 23, 43, 49; accumulation 46 policy environment, distorted 88 political instability in Asia 76–79, 97, 103 Political Risk Services (PRS) Group 81, 87, 89, 90, 97, 146; index of perceived corruption 86–87, 94, 146 population 123; below international poverty line 13, 64; decline in growth 10; growth 97 portfolio investment flows 72 poverty 3; absolute 64; in East Asian economies 13; index 65; line 13; Lorenz curve 65; reduction, growth and redistributive policies 65–66 primal approach to growth accounting see growth private capital flows/foreign direct investment, gross 133 production: estimates of TFP growth 56; function 31, 48; importance of structure of 7; structure of gross domestic product, distribution of 8 productivity growth 26, 50; assessment of empirics 55–57; empirics 37–54; study of 30 professional inequality 115–118; gender empowerment measure (GEM) 117 property: ownership laws 118; security 107 Protestant tradition 92–93

PRS index see Political Risk Services Group Prucha, I. R. 141 prudential regulation, inadequate 67 public education expenditure 21 public sector pay 91; low 88 purchasing power parity (PPP) 136; power parity dollar 13 Pusan, South Korea 108 quality of legal system 100 Quibria, M. 122, 147 R&D expenditure as percentage of gross national income (GNI) 18 Radelet, S. 70–72, 75, 79–82, 145 Rauch, J. 146 Ravallion, M. 65 Ravenhill, J. 69 real exchange rate 71, 76, 81; change 77–78, 81 red tape 95, 100 Ringstad, V. 142 road congestion 106 Robbins, D. 136–137 Robinson, I. 108, 147 Rodriguez-Clare, A. 38, 49–50, 141 Rodrik, D. 137 Romer, P. M. 23, 141 Rose-Ackerman, S. 85–86 Russia 143 Ruttan, V. 33 Sachs, J. 70–72, 75, 79–82, 145 Sarel, M. 38, 49, 51–52, 141; TFP growth rates for five Asian economies and United States 51 Sargan, D. J. 142 saturation 14 savings and investment, performance in terms of 14 school enrolment: primary 115–116; secondary 97 scientists and engineers: in research and development 18; students 126–131

166

Index

self-fulfilling expectations hypothesis 68, 80, 82 Selowsky, M. 59 Semarang (Indonesia) 107–108 Sen, A. K. 112, 114–115, 119; classification proposed by 112 Seoul 108–109 Sepúlvea, F. 94–95, 146–147; see also Méndez and Sepúlvea capital accumulation model service sector, share of 7 sewerage 109 Shargrodsky, E. 91 Shleifer, A. 84–86, 146 short-term debt ratio to reserves 68, 72, 76–78, 80; in crisis countries 70 Sims, C. 80, 145 Singapore 1, 61, 84; annual growth rate 6; basic facility inequality 116; BI rating 88; corruption in 84, 87–88, 101; demographic transition 11; domestic policy indicators 20; as East Asian NIC 44–46; energy consumption 14; energy use and efficiency 15; factor accumulation indicators 17; factor price distortion 61; FDI 100–101; GDP in per capita terms 13; gender inequality 113; global financial flows 135; GNP 21; growth accounting 52; as HPAE 141; human capital accumulation 16; human development 12; inequality in 64; information age 123–125; integration with global economy 134–136; life expectancy 10; male–female income 111–112; mortality inequality 114; natality inequality 114; as non-crisis country 76; as one of ‘Four little Tigers’ 2, 136; ownership inequality 118; per capita income, rapid growth 22; population below international poverty line 64; professional inequality 117; in service

sector 7; source of growth in 39, 41, 44–46; structure of production 8; technological change 18–19; total factor productivity growth 38–40, 42–43, 48, 50–51, 54, 56, 142–143; trade and globalization 9; traffic 107; urbanization, indicators of 105–106; venture capital 121 skilled labour 136 social marginal products 28; of factor inputs 32 software component of investments 47 Solow, R. M. 25, 27, 141 Solow residual 33, 141 Son, W. 49 South Africa 121, 143 South Asia 6, 9; and East Asia and the Pacific, comparison 7 South Korea: ADB study 123; annual growth rate 6; annual output growth 74; basic facility inequality 116; CGE model 59; chaebols 88; corruption in 87–88, 101; currency crisis 67, 71–72, 76, 143; demographic transition 10–11; distortions in 1970s 60; domestic policy indicators 20–21; as East Asian NIC 44–46; economic indicators 69; energy use and efficiency 14–15; external financing 73; factor accumulation indicators 17; factor price distortion 61; gender inequality 113–115; global financial flows 135; household inequality 119–120; as HPAE 42, 141; human capital accumulation 14, 16; human development 12; inequality in 64; information age 123–125; integration with global economy 134–136; lending rate 80; male–female income 111–112; mortality inequality 114; natality inequality 114; as one of ‘Four little Tigers’ 2, 136; ownership

Index 167 inequality 118; population below international poverty line 13, 64; professional inequality 117; Pusan 108; real exchange rate 71–72; Seoul 108–109; in service sector 7; short-term debt and reserves in 68, 70, 144; source of growth in 39, 41, 44–46; structure of production 8; technological change 18–19; total factor productivity growth 38–40, 42–43, 48, 50, 53–54, 56, 142–143; trade and globalization 9; urbanization, indicators of 105–108; venture capital 121; won 67 Soviet Union, former 85 speculative activities 83 standard model of economic growth 25 state capture 85 Stiglitz, J. E. 144 stock markets in crisis countries 72 stock prices 75 sub-Saharan Africa 6, 9, 43 Suharto government in Indonesia 85, 88 Summers, R. 43 Summers–Heston database 51, 143 Surabaya (Indonesia) 108–109 sustainability 106, 110 Svensson, J. 84, 96, 146 Sweden 121 Taiwan, China 1; as East Asian NIC 44–45; economic indicators 69–70; as HPAE 42–43, 141; inequality in 64; as non-crisis country 67, 76, 143; as one of ‘Four little Tigers’ 2, 136; real exchange rate 70; restrictions on capital movement 68; sources of growth 41, 44–46; total factor productivity growth 38–39, 42–43, 48, 50, 54 Tanzi, V. 85 tariff barriers to trade 91

tax rate 103 technological change 18; selected determinants of 19 technological progress 24, 28, 46; Harrod-neutral 56 telephone 109; lines per capita 127–131 television 123; receivers per capita 127–131 TFP (total factor productivity) 23, 25; approaches to estimation of 24–37; in East Asia 37, 57; estimates 24; growth 39, 48; for Hong Kong 38; for newly industrialized countries (NICs) 40; for Singapore 38, 142–143; transcendental logarithmic (translog) index of growth 30; Tronqvist approximation 29 Thailand 1; annual growth rate 6; annual output growth 74; baht 67, 88, 72, 74; Bangkok 107–109; basic facility inequality 116; Chiang Mai 108; corruption in 87–88, 101; currency crisis 67, 71–72, 76, 143; demographic transition 10–11; distortions in 1970s 60; domestic policy indicators 20–21; as East Asian NIC 44–46; economic indicators 69; energy use and efficiency 14–15; external financing 73; factor accumulation indicators 17; gender inequality 113–115; global financial flows 135; growth accounting 52; as HPAE 42, 141; human capital accumulation 14, 16; human development 12; inequality in 64; information age 123–125; integration with global economy 134–136; male–female income 111–112; mortality inequality 114; natality inequality 114; ownership inequality 118; population below international poverty line 13, 64; poverty reduction 63, 66; professional

168

Index

Thailand (Continued) inequality 117; real exchange rate 71–72; short-term debt and reserves in 68, 70, 144; sources of growth in 39, 41, 44–46; structure of production 8; technological change 18–19; total factor productivity growth 38–40, 42–43, 48, 50, 53–54, 56, 142–143; trade and globalization 9; traffic 107; urbanization, indicators of 105–108 Thomas, V. 58–60 TI index see Transparency International Tiongson, E. 85 total debt to reserves ratio 80–81; see also short-term debt ratio to reserves total foreign debt as percentage of total reserves 76, 78; GDP growth rate 77; see also short-term debt ratio to reserves trade: balance 81; and globalization, importance of 9; in goods 133; liberalization 92 traffic and congestion 107 Transparency International (TI) 86; annual index of perceived corruption 87, 92 Treisman, D. 86–87, 92–93, 146 Tronqvist approximation: to Divisia index 142; to TFP measure 29 Tronqvist index 29, 142 Tsao, Y. 143 Uganda 84; survey of firms in 96 unemployment rates of unskilled workers 136 United Kingdom 44, 48, 121 United Nations Development Programme (UNDP) 10 United States 18–19, 44, 46, 48, 100, 122; Bureau of Labor Statistics 141; economy 22; firms’ foreign investment, study of 98; Foreign Corrupt Practices

Act 98–99, 102, 146; growth accounting 52; sources of growth 41; supply of venture capital 119 urbanization/urban conditions 3, 105, 126; affluence 106; agglomerations 106; degradation 106; environment services 108; indicators of 105–109; population 127–131 US see United States VanRijckeghem, C. 146 vehicular intensity 106 venture capital 119; access to women-led businesses in United States 119; current investment spread across world 121 Vishny, R. 84–86, 146 Wade, R. 57 wage: differentials by level of education 137; effect on corruption 146; gap 137; inequalities 136–137; rates 61, 112 Wang, Y. 58–60 water 109 Watergate 147 Weder, B. 146 Wei, S.-J. 81, 84, 95–96, 99–102, 104, 146 Weiss, A. 144 welfare: experienced by populations 13; increase in 10 Westphal, L. E. 141 Wheeler, D. 98–99, 103 women: in business, access to venture capital 119; economic activity rate 111; female per capita income 111; missing 114; in parliament 115; participation rate 111; political participation 118; senior decision-making powers in government 118 Wood, A. 136–137 working age population 17

Index 169 work trips by public transport 109 World Bank 1, 6, 9, 42–43, 49, 64, 142; study ‘The East Asian Miracle’ 41, 141, 143 World Development Indicators 6, 8, 11, 17, 64, 71, 74, 80, 87, 102, 105–108, 124–125, 134–135 World Development Report 15, 84, 88

world growth rate of real GDP per capita 22 World Investment Directory 147 World Tables 12 Young, A. 32–33, 38–40, 42–43, 49, 53, 55–56, 141–143 youth literacy 115 Zurawicki, L. 146

Economic Development in Pacific Asia

Economic and Social Development in Pacific Asia

Political Development in Pacific Asia

Regionalism, Trade And Economic Development in the Asia-Pacific Region

Economic Growth in the Asia Pacific Region

Economic Dynamism in the Asia-Pacific (Pacific Studies (London, England).)

Asia-Pacific in Figures 2006 (Economic and Social Commission for Asia and the Pacific)

Asia Pacific Financial Deregulation (Pacific Trade and Development Conference Series)

Asia-Pacific Development Journal, December 2006

Towards Recovery in Pacific Asia (ESRC Pacific Asia Programme Series )

Governance in the Asia-Pacific

Reshaping the Asia Pacific Economic Order (Pacific Trade and Development Conference (Papers))

Entrepreneurship and Economic Development in Hong Kong (Routledge Advances in Asia-Pacific Business, 5)

Asia-pacific Trade and Investment Review, 2008 (Economic and Social Commission for Asia and the Pacific)

Innovation: Asia-Pacific Readings

Islamic Banking And Finance in South-east Asia: Its Development And Future (Asia-Pacific Business) (Asia-Pacific Business)

Games in Economic Development

Games in economic development

Games in Economic Development

Asia-Pacific Transitions

Energy Security in Asia (Routledge Security in Asia Pacific Series)

Asia-Pacific Constitutional Systems (Cambridge Asia-Pacific Studies)

Asia-Pacific Constitutional Systems (Cambridge Asia-Pacific Studies)

Europe and the Asia Pacific (Esrc Pacific Asia Programme (Series).)

Central Asia in Transition: Dilemmas of Political and Economic Development

Culture and Society in the Asia-Pacific (Pacific Studies)

Economic and Social Survey of Asia and the Pacific 2006

AIDS in East Asia and the Pacific

Sustainable Agriculture and Food Security in Asia and the Pacific (Economic and Social Commission for Asia and the Pacific)

The Environment in Asia Pacific Harbours

Comparative linguistics in Southeast Asia (Pacific linguistics)

Economic Development in Pacific Asia

Economic and Social Development in Pacific Asia

Political Development in Pacific Asia

Regionalism, Trade And Economic Development in the Asia-Pacific Region

Economic Growth in the Asia Pacific Region

Economic Dynamism in the Asia-Pacific (Pacific Studies (London, England).)

Asia-Pacific in Figures 2006 (Economic and Social Commission for Asia and the Pacific)

Asia Pacific Financial Deregulation (Pacific Trade and Development Conference Series)

Asia-Pacific Development Journal, December 2006

Towards Recovery in Pacific Asia (ESRC Pacific Asia Programme Series )

Governance in the Asia-Pacific

Reshaping the Asia Pacific Economic Order (Pacific Trade and Development Conference (Papers))

Entrepreneurship and Economic Development in Hong Kong (Routledge Advances in Asia-Pacific Business, 5)

Asia-pacific Trade and Investment Review, 2008 (Economic and Social Commission for Asia and the Pacific)

Innovation: Asia-Pacific Readings

Islamic Banking And Finance in South-east Asia: Its Development And Future (Asia-Pacific Business) (Asia-Pacific Business)

Games in Economic Development

Games in economic development

Games in Economic Development

Asia-Pacific Transitions

Energy Security in Asia (Routledge Security in Asia Pacific Series)

Asia-Pacific Constitutional Systems (Cambridge Asia-Pacific Studies)

Asia-Pacific Constitutional Systems (Cambridge Asia-Pacific Studies)

Europe and the Asia Pacific (Esrc Pacific Asia Programme (Series).)

Central Asia in Transition: Dilemmas of Political and Economic Development

Culture and Society in the Asia-Pacific (Pacific Studies)

Economic and Social Survey of Asia and the Pacific 2006

AIDS in East Asia and the Pacific

Sustainable Agriculture and Food Security in Asia and the Pacific (Economic and Social Commission for Asia and the Pacific)

The Environment in Asia Pacific Harbours

Comparative linguistics in Southeast Asia (Pacific linguistics)

Recommend Documents