4;d 1,~~~~~~~~~~~~~~~~~~~~~~~~~~~~T-- ~ *4

.4 40vi~

-. ~~~~~~~~~~~~~ .,*'%* w~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~';

Go~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~PW

VERNMENTPO~~~~ w,"ANtR0UTVI.ROI*

1~~~~~~ast Asia ~~~~~~~~~~~~~~~~~~~~~~~ ~~~-a

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

THE LESSONSOF EAST ASIA

Government Policyand Productivity GrowthIs East Asia an Exception?

Vinod ThomasYan Wang

The World BankWashington, D.C.

Copyright © 1993The International Bank for Reconstructionand Development/THE WORLD BANK

1818 H Street, N.W.Washington, D.C. 20433, U.S.A.

All rights reservedManufactured in the United States of AmericaFirst printing October 1993Second printing September 1994

The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s)and should not be attributed in any manner to the World Bank, to its affiliated organizations, or tomembers of its Board of Executive Directors or the countries they represent. The World Bank does notguarantee the accuracy of the data included in this publication and accepts no responsibility whatsoeverfor any consequence of their use. Any maps that accompany the text have been prepared solely for theconvenience of readers; the designations and presentation of material in them do not imply the expressionof any opinion whatsoever on the part of the World Bank, its affiliates, or its Board or member countriesconcerning the legal status of any country, territory, city, or area or of the authorities thereof orconcerning the delimitation of its boundaries or its national affiliation.

The material in this publication is copyrighted. Requests for permission to reproduce portions of it shouldbe sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bankencourages dissemination of its work and will normally give permission promptly and, when thereproduction is for noncommercial purposes, without asking a fee. Permission to copy portions forclassroom use is granted through the Copyright Clearance Center, 27 Congress Street, Salem,Massachusetts 01970, U.S.A.

The complete backlist of publications from the World Bank is shown in the annual Index of Publications,which contains an alphabetical title list (with full ordering information) and indexes of subjects, authors,and countries and regions. The latest edition is available free of charge from the Distribution Unit, Officeof the Publisher, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or fromPlublications, The World Bank, 66, avenue d'1ena, 75116 Paris, France.

ISSN: 1020-0924

Vinod Thomas is chief economist in the World Bank's Office of the Regional Vice President for East Asiaand the Pacific. Yan Wang is an economist in that office.

Library of Congress Cataloging-in-Publication Data

Thomas, Vinod, 1949-Government policy and productivity growth: is East Asia an

exception? / by Vinod Thomas, Yan Wangp. cm. - (The Lessons of East Asia)

Includes bibliographical references.ISBN 0-8213-2615-51. East Asia-Economic policy. 2. East Asia-Economic conditions.

I. Wang, Yan, 1953- . II. Title. III. SeriesHC460.5.T48 1993338.95--dc2O 93-31252

CIP

CONTENTS

FOREWORD ............................................... vACKNOWLEDGMENTS ............................................. viiEXECUTIVE SUMMARY . ........................................... ix

1. INTRODUCTION ............................................. I

II. EAST ASIA'S OUTSTANDING PERFORMANCE ........................ 1

m. THE ANALYTICAL APPROACH .................................. 2

IV. DISTORTIONS, NTERVENTIONS, AND GROWTH ...................... 6

V. COUNTRY DIFFERENCES . ...................................... 13

VI. EAST ASLN POLICIES AND EFFECTS ............................. 16

VU. CONCLUSION .............................................. is

APPENDIX I Description of Variables used in the Analysis ........ .. ............. 19

APPENDIX II A List of Economies in the Analysis ............................ 21

BIBLIOGRAPHY ............................................... 23

iii

FOREWORD

Policymakers everywhere are searching for lessons from East Asia's enormous success ineconomic development. A number of recent cross-country and thematic studies have sought to identifyand analyze the policies behind this success. Among them is The East Asian Miracle, a recent WorldBank publication, which draws in part on the Lessons of East Asia project. Study teams, including in-country nationals, examined in some depth the experiences of the highly successful East Asian economiesand the public policies underpinning them.

Several clear contributions emerge from this set of country studies. The research:

* Highlights considerable variation in approaches within the group of East Asianeconomies. For example, some economies chose a substantial degree of governmentintervention; others did not. The studies dispel the notion that there is a single oruniform East Asian model of success.

* Demonstrates that a core set of good economic policies -- such as macroeconomicdiscipline, outward orientation, and human resource development -- laid thefoundation for East Asia's success. Pragmatic policymaking -- understood as beingnonideological and reversible -- seems to be at the heart of these policies and meritsreplication.

Dispels some of the myths about the more idiosyncratic interventions, such as"picking winners" in industry, which sometimes produced the desired result andsometimes did not. Because presence or absence of institutional features seems tohave affected the outcomes of these interventions, applications to other regionalcontexts must be approached cautiously. A dominant finding of the studies is thatserious diversions from macroeconomic equilibrium were largely avoided, even bystrong interventionists. At the same time, the later generation of industrializers weremore successful when they avoided these industrial policies.

A question not easily answered is why East Asian governments adopted fundamentally soundpolicies and were apparently able to achieve better results from their active policies and to incur lowercosts from errors. In this connection, the studies touch on such dimensions of policymaking as the roleof the state, leadership, and the bureaucracy. It is one thing to describe the institutional featuresaccompanying a successful episode, however, and quite another to know why and how those featurescame about. For instance, why did East Asian leaders apparently hold themselves more accountable foreconomic performance than has been the experience elsewhere? How did the governments manage togain sufficient national consensus to put difficult policies into effect? These aspects of political economycannot be ignored. Our analytic tools, however, are severely limited in penetrating these issues, inassessing their impacts, and in assigning credit to them. These country studies are only one step,although a significant one, in deepening our understanding of the experience of East Asia. It is hopedthat they will prompt additional work on the institutional foundations of rapid growth.

Gautam KajiVice President

East Asia and Pacific Region

v

ACKNOWLEDGMENTS

This paper and all the papers produced as part of the country studies project have benefitedfrom the insights and observations offered by discussants and commentators at a conference held at theEast-West Center in Honolulu, November 19-21, 1992. The participants included the following regionaland country experts: Duck-Soo Han, Hal Hill, Chalmers Johnson, Wolfgang Kasper, Hyung-ki Kim,Paul Kreisberg, Chung H. Lee, Manual Montes, Seiji Naya, Takashi Nohara, John Page,Tambunlerthchai Somsak, Wanda Tseng, Wing Thye Woo, Ippei Yamazawa, and Zainal Aznam Yusof.The papers also benefited from a Bank-wide review of the project in August, 1993.

The country studies team included Amar Bhattacharya, Leung Chuen Chau, ScottChristensen, Carl Dahblman, David Dollar, Kim Kihwan, Saha Dhevan Meyanathan, Mari Pangestu, PeterPetri, Ismail Salleh, Ammar Siamwalla, Teck-Wong Soon, C. Suan Tan, and Vinod Thomas. Thecountry authors would like to acknowledge fruitful dialogue with our country counterparts and theanalytic work prepared in World Bank country departments as part of Bank economic and sector work.The country studies were edited by Rupert Pennant-Rea. The project assistant was Jason Brown. Thisproject was undertaken with the support of Lawrence H. Summers, Nancy Birdsall, John Page, andGautam Kaji, Callisto Madavo, Marianne Haug, and Vinod Thomas.

Danny M. LeipzigerCountry Studies Director

vii

ACRONYMS

DPM - Deputy Prime MinisterEPB - Economic Planning BoardFDI - Foreign direct investmentFYR - Five Year PlanGDI - Gross Domestic InvestmentGDP - Gross Domestic ProductGIE - Government-invested enterprisesGTC - General Trading CompanyHCI - Heavy and Chemical IndustriesHUD - Harvard Institute for International DevelopmentKDI - Korean Development InstituteKEPCO - Korea Electric Power CorporationKIET - Korea Institute of Economics & TechnologyKIST - Korea Institute of Science & TechnologyKNR - Korean National RailwayKOTRA - Korean Trade Promotion CorporationKTA - Korean Traders AssociationMOF - Ministry of FinanceNIF - National Investment FundODA - Official Development AssistancePOSCO - Pohang Steel CompanyREER - Real effective exchange rateVAT - Value-added taxWDR - World Development Report

viii

EXECUTIVE SUMMARY

East Asian developing nations have consistently outperformed other developing regions in the last25 years. Between 1965 and 1990, average incomes irr East Asia and the Pacific increased by 5.3 percentper year, more than double the rate for all low- and middle-income countries. Although there areconsiderable differences between individual East Asian economies, their superior performance justifiesa focus on the collective East Asian experience in analyzing the effects of government policies oneconomic growth.

Total factor productivity (TFP) is known to be an important contributor to economic growth.Thus, explanations of East Asia's above-average increase in TFP are crucial to an understanding of theregion's impressive economic performance. Previous work has suggested a strong link betweenproductivity growth and government policies in all developing countries, a connection confirmed bystudies of individual countries in East Asia. This paper extends that work by highlighting the linkbetween policies and productivity increases within a cross-country framework.

Two models are used to explore the relationship between policy design and productivityperformance. The first follows the tradition of neoclassical growth theory, using a quasi-Solow approach.It applies a Cobb-Douglas production function, emphasizing familiar factor inputs. TFP appears as theunexplained residual of an economy's growth rate and is assumed to be a function of government policies,a regional dummy for East Asia and some interaction terms. The second model uses the recent literatureon endogenous growth theory. It assumes that government policies affect growth through their influenceon the real rate of return to capital, which in turn affects the rate of investment and economic growth.

Because there is no objective way to measure "policy," it is not possible to draw firm conclusionsfrom an analysis of individual government actions. Instead, two composite policy indices are constructed;the first (INDEX 1) to capture incentives and macroeconomic stability and the second (INDEX 2) tomeasure government expenditures. INDEX I comprises seven variables, including trade policy, outwardorientation, inflation and real interest rates. INDEX 2 contains public sector investment, totalgovernment expenditure, fixed capital formation and a range of productive expenditures, all as shares inGDP.

Three equations are estimated for the two indices (GDP per capita growth, TFP growth I and TFPgrowth 2), using different specifications and based on country-average data for ten East Asian nations and58 other developing economies. INDEX 1 is positive and significant in all equations, indicating a linkbetween incentives (or lack of price distortion) and stability and growth. Further tests indicate that therelationship is stronger among the subset of East Asian nations, suggesting that they are more effectivein conducting trade policies and maintaining macroeconomic stability. Although linear specifications failto reveal a significant association between INDEX 2 and growth, there is evidence of a curvilinearrelationship. In particular, it seems that government expenditures are positively associated with GDP andproductivity growth up to a point, but then they become harmful. East Asian nations fall within the rangewhere intervention has a positive association.

The analysis is then disaggregated to assess the extent of policy variation within the two countrygroupings. Although East Asia as a whole enjoyed a superior policy framework, by many measures,country differences within the region were as large as those elsewhere. These include sectoral distortions,where Korea's protection of agriculture stands out, and the state's share of total investment, which wassignificantly higher in China, Taiwan (China), Malaysia and Singapore than the regional average. Theseresults confirm that there was no single, region-wide approach adopted by all of the successful East Asiannations.

Nevertheless, certain broad policy features were shared by most East Asian nations and played animportant role in explaining their rapid growth. These include macroeconomic stability, outwardorientation, low price distortions and heavy investments. In particular, the results confirm the positive

ix

contribution of macroeconomic stability, which was especially pronounced in East Asia and involved anabsence of trade imbalances and discipline over public borrowing among other factors.

It is now widely accepted that East Asian governments intervened in their economies, and moreefficiently than governments elsewhere. However, the region's greater success with market reforms Inthe 1980s is less well-known. Various country studies explain why these reforms were more effectivein East Asia. Policy changes were deeper than in other developing nations and pursued more vigorously.A simple average of exchange rate and trade restrictiveness fell by half in East Asia from 1979 to 1990,far more than the average for other developing nations. It also seems likely that complementary factors -Including Institutions, labor skills and flexibility and infrastructure -permitted a greater supply response.

In conclusion, the study suggets that growth is damaged by trade and price distortions and bymacroeconomic instability. However, there is no simple relationship between government expenditureand economic performance. In East Asia, price distortions and macroeconomic instability were lowerthan in other developing nations, so did less to hamper growth. East Asia has also been more successfulIn undertaking reforms designed to eliminate price distortions. These findings have a powerful relevancefor all developing countries. The essential ingredients of East Asia's success can be replicated, althoughthe region does seem to be an exception in making public expenditures work to promote growth ratherthan hinder it.

x

I. INTRODUCTION

Developing countries have diverged sharply over the past quarter of a century in rates of outputand productivity growth. Some have experienced stunning success, others disappointing failure. As agroup, the East Asian economies consistently outperformed other developing regions, and theirachievement has attracted the attention of policy makers everywhere. Although there are considerabledifferences among the East Asian economies, their generally superior performance justifies an East Asianfocus in the analysis of government policies and growth.

Governments the world over have tried to speed up economic growth by intervening in themarketplace. The failures have far outnumbered the successes. But some of the successes have beenspectacular. East Asian economies, such as Japan, Korea, and Taiwan (China) made remarkable progressunder moderately interventionist policies. And now China, with only gradual market reforms, isachieving the most rapid and sustained growth of any large country in the world. The region's successraises three crucial questions: was the nature of government interventions very different in East Asia thanin other regions? Were the effects of government interventions very different in East Asia? What mightaccount for these differences?

In considering these questions, this paper distinguishes government interventions from marketdistortions. Intervention does not necessarily lead to growth-reducing distortions. Whether they dodepends on their scope, quality and scale. The paper shows that the main reason for East Asia's superiorperformance was not that governments there intervened less -- in fact, on average, they intervened asmuch as anywhere else in some areas such as public expenditures. Rather, the East Asian governmentsintervened efficiently and in ways that contained and minimized overall price, trade, and macroeconomicdistortions. The key to the region's better results was not so much any unique cultural factors as thedevelopment of effective policies, institutions and practices. That key is available for other countries tocopy.

II. EAST ASIA'S OUTSTANDING PERFOMANCE

Between 1965 and 1990, average incomes in East Asia and the Pacific (as the developing regiondefined in World Development ReRort 1992) rose at 5.3 percent a year (Table 1), which was more thandouble the rate for all low- and middle-income countries (LMICs). Over the period as a whole,economies, such as Korea and Taiwan (China) saw average incomes rise more than fivefold. In theshorter period of the 1980s, China experienced a dramatic 8 percent annual growth in average incomein the 1980s. The slowest growing country in the East Asian group was the Philippines, whose per capitaincome rose at about I percent a year in 1965-1990.

East Asia's high rates of investment in physical and human capital are often (and rightly) cited,as major contributors to its spectacular performance. These economies have consistently invested a largershare of output than other developing countries -- nearly 50 percent higher in 1990. As for humancapital, in 1990 their primary enrollment rate was 25 percent higher and their infant mortality rate 50percent lower than the average for all developing countries.

2

Table 1. Growth, Investment, and Social Developmentin East Asia and

All Low- and Middle-Income Countries (LMICs)

East Asia ALL LMICs1965 1990 1965 1990

Growth Rates (%)GDP per capita, 1965-90 5.3 2.5Exports, 1965-90 9.0 4.1

Structure of Demand (% of GDP):Gross Domestic Investment 22 37 20 26Gross Domestic Saving 22 35 20 24

Education a!Primary enrollment ratio 88 129 78 105Secondary enrolLment ratio na 46 22 43

HeaLth b!infant mortality per 1,000 95 34 117 63Adult Life expectancy at birth - 68 - 63

a/ The most recent year is 1989, nor 1990.b/ Based on incompLete country coverage and presented purely for illustration.Source: World Development Report 1992, (World Bank 1992).

Total factor productivity is known to be a powerful contributor to economic growth. East Asia'sabove average increase in total factor productivity is a key to its high growth rates. Although it hadhigher investment rates than others, it had even higher productivity growth: several times higher, as thispaper will show later. Previous work has suggested a strong link between productivity growth andgovernment policies in all developing countries (World Bank 1991), a link confirmed by studies ofindividual countries in East Asia. For instance, the acceleration in growth in Korea after 1959 and againafter 1982; in China after 1979; and in Indonesia after 1985: all those changes can be attributed to policyimprovements.

This paper highlights the link between policies and productivity growth within a cross-countryframework. It first presents a formal framework for assessing this link, and then uses that to derivecross-country averages. It examines country differences within East Asia to give a sense of howrepresentative the cross-country averages are. It also reviews some of the tangible and intangibleingredients of East Asian policies.

III. THE ANALYTICAL APPROACH

To assess the link between policies and productivity growth, we use two alternative models. Onefollows the tradition of neoclassical growth theory, using a quasi-Solow model. The other comes out ofthe recent literature on endogenous growth. We apply these models to explore, on the one side, the linkbetween economic and productivity growth; and on the other, the extent of distortions and governmentintervention. To explain the effects that are specific to East Asia, we examine the region's 10 economiesclosely and compare them with other countries in the sample, differentiating them according to policy

3

"deepening" (extent of trade and price reforms, shares of expenditures) and policy "productivity" (abigger coefficient of the policy variable).

The first approach applies a Cobb-Douglas production function (CD model), which emphasizescertain familiar production inputs. The unexplained residual of an economy's growth rate is consideredto be the growth rate of total factor productivity, which is assumed to be a function of governmentpolicies, a regional dummy for East Asia, and some interaction terms between East Asia and thesepolicies. That is,

aYit =o + yP, + 8PV, D + OD + , Ox&nXi, (1)X X X=k,h,e

where yi is the output of i country in year t and Xk represents such inputs as physical capital (k),labor (l ), land (h), and educational capital (e), all measured by log differences. The term P., is policyindicator i for country i in year t ; the indicator could be an individual policy variable or a compositeindex. The term D is a regional dummy variable for East Asia, and the terms p., D are interaction termsbetween East Asia and policies and distortions.

Empirically, the model can be estimated in two ways: by estimating (a) equation 3 with all inputvariables and government policies, or (b) the production function first and then taking the unexplainedresidual as total factor productivity growth (TFPG) and estimating:

TFPGa, = so + + r 8Q D + OD, (2)

This paper highlights the results of the second approach, since its focus is on the effects of governmentpolicies and regional factors rather than on productive inputs. Because TFPG is measured after theeffects of traditional and nontraditional (education) inputs are accounted for, TFPG can be related topolicy and to some factors specific to the region.

In the second approach, which is closely related to recent models of endogenous growth, we beginwith a simple linear output-capital model. Output (y) is proportional to a broadly defined capital (K)that incorporates both physical and human capital, and labor enters the model through K -' Assumingthat the economy is closed to capital flows (saving = investment), the growth rate of output per capita(g) is a function of the rate of accumulation of new physical and human capital (i ), the capital-output

' The analysis here has benefited from papers by Rebelo (1991), King and Rebelo (1990), and Barro(1990) and from a research project at the Bank: "How do national policies affect long-run growth?". (SeeEasterly and others (1991) and draft papers by Barro and Lee (1992), De Long and Summers (1992),Easterly and Rebelo (1992), and Fischer (1992).

4

ratio (parameter a) and the depreciation rate (8). Government policies affect growth through theirinfluence on the real rate of return to capital, which in turn affects the rate of investment and economicgrowth. We can express this function as:

gt m a00 + Yit + 8Pt° D sED (3)

where ,, is the growth rate of GDP per capita for country i in year t, and p., D, and

Pv D are defined as above.

The data set is derived from pooled cross-country time-series statistics compiled for the WorldDevelopment Report 1991 and GDP data from the national accounts, as reported in the World Bank database; figures on government expenditure are from the IMF government finance statistics database. TheWDR 91 data set contains informnation on GDP growth rates (1960-90), capital, labor, land, education,and other inputs for 68 developing countries for 1960-87 (following Bhalla and Lau, forthcoming). Italso contains variables reflecting a wide range of public policies -- outward orientation, price distortions,foreign exchange premiums in the parallel market, fixed capital forrnation. And it quantifies public sectorinvestment, as well as government expenditures in education, health, economic services. (Data on thesepolicy variables are not available for all countries and all years, however. See Appendix I for sourcesand descriptions.)

Table 2 provides some descriptive statistics on growth and government policies for 10 East Asianeconomies and 58 other developing countries.' In growth of income per capita, the East Asian grouphas outpaced the others during the past three decades. Total factor productivity growth (TFPGI andTFPG2),3 calculated as the unexplained residuals in two estimated CD production functions, shows asimilar trend. East Asia averages more than 2 percent productivity growth a year, compared with lessthan 1 percent for the other developing countries.

Indicators of government interventions are categorized in two ways: measures of distortion, andmeasures of government action. Among the distortion indicators, we look at trade outward-orientationand macroeconomnic and price stability. Again, East Asia does better than the others: it was moreoutward-orientation, had a stronger record of positive real interest rates, and achieved lower inflationrates and exchange rate premiums in parallel markets. These results indicate that the East Asian grouphas had a superior policy framework. But the question remains whether the policies were also moreproductive.

' The 10 East Asian economies are China, Hong Kong, Taiwan, Indonesia, Japan, Korea, Malaysia,the Philippines, Singapore, and Thailand. We include Japan because it made the transition to a high-income country in the past 30 years. The total factor productivity growth rates (TFPGs) of these 10economies are listed individually in Table A-1. See appendix II for a list of the 68 countries.

' TFPGI is calculated based on a CD production function with traditional inputs only (equation 3)in Table A-1, while TFPG2 is based on a CD production function with both traditional and nontraditionalinputs--education (equation 4 in Table A-1).

5

We next consider the category of goverunent interventions measured by the shares of GDP goingto public sector investment, government consumption, total and decomposed expenditures, andgovernment fixed capital formation. The share of public investment is similar for East Asian and otherdeveloping economies, but the share of govermnent consumption in GDP is smaller in East Asia. Thedifferences in these categories are not particularly marked. But again the key question, investigated later,is whether public expenditures in East Asia were more productive.

Table 2Economic Growth and Government Policies:

East Asia andOther Low- and Middle-Income Countries (LMICs)

East Asia Other LMICGrowth or policyIndicator 10 economies 58 economies

Performance Indicator (1987 dollars)

Average annual growth inGDP per capita (X, 1960-90) 4.68 ( 4.9) 1.26 ( 5.4)

Total Factor Productivity Growth (X, 1960-87)TFPG1 2.58 ( 1.15) 0.92 ( 1.14)TFPG2 2.22 ( 1.11) 0.40 ( 1.27)

Incentives and Stabilitv

Trade policy index (TPI; 1977-88) 2.08 C 1.0) 1.69 ( 0.9)Parallel market premium (X, 1960-90) 14.16 (48.2) 52.70 (155.9)Inflation rate (X, 1961-89) 8.92 (15.0) 17.98 C 35.2)Real Interest Rate (X, 1970-88) 3.71 ( 7.9) -3.01 ( 19.6)

Govermnent Exoenditures ( X of GDP. 1960-89)

Public sector inveatment 8.53 ( 4.3) 8.97 ( 5.8)Consumption 7.01 1 3.4) 8.88 ( 5.1)Total expenditures 19.50 ( 5.6) 24.90 ( 11.5)Fixed capital formation 2.86 C 2.0) 3.40 ( 3.0)Social expenditures 8.62 ( 2.7) 10.18 ( 5.4)

Education 3.12 C 1.4) 3.13 ( 1.8)Health 0.79 ( 0.5) 1.51 ( 1.3)

Note: 1) Siwple group means are presented. Standard deviations are in parentheses. Goverrment expendituresare for consolidated goverrment. Government consumption here is measured by (gconx-ed"cation-health). Seeappendix I for a description of variables and time periods covered.

2) The 10 East Asian economies are: China, Hong Kong, Taiwan (China), Indonsia, Japan, Korea, Malaysia,the Philippines, Singapore, and Thailand.

3) Medians differ only slightly from the means for most variables.

6

IV. DISTORTIONS, INTERVENTIONS, AND GROWTH

Distortions and interventions are multidimensional concepts, just as the concepts of economicwell-being or the quality of life. It is now fully recognized that measuring the quality of life usingindividual variables such as GDP per capita is not sufficient. That is why many attempts have been madeto construct composite measures of the quality of life. Similarly, there is no completely objective wayto measure "policy," due to its complexity and multi-dimensionality. We cannot draw any firmconclusions from an examination of individual governnient actions. One way around the problem is tolook at as many individual policy measures as possible. But not all measures are available for allcountries for the same time periods -- a genuine weakness, which could result in selectivity bias. In thispaper, therefore, we construct composite policy indexes drawing on all data available on complementarypolicy indicators. This approach allows for international comparisons and, we believe, is more objectivethan relying on individual policy measures.

We constructed two composite indexes, one for incentives and macroeconomic stability (INDEXI)and one for government expenditures (INDEX2). INDEXI includes seven variables: four that havealready been mentioned -- the trade policy index, inflation, real interest rates, and the parallel marketpremium -- plus an index for outward-orientation (Dollar 1991), an index for trade liberalization(Papageorgiou, Michaely, and Choksi 1990), and an index for agricultural disprotection (Krueger, Schiff,and Valdes 1992). (These three were not shown in Table 1 because their coverage is incomplete.)INDEX2 includes public sector investment, total government expenditure, fixed capital formation, andproductive expenditures (including education, health, and economic services), all as shares in GDP.

To construct the composite index, we used the Borda ranking technique. It is based on publicchoice theory and has been used in evaluating welfare across countries.4 The Borda rule provides amethod of rank-order scoring that allows policy indicators to be aggregated even though they wereoriginally measured in different units and for different countries and periods. Each country receives apoint equal to its rank for each criterion of ranking for policy variables. The points are then added foreach country to obtain its aggregate score, and the countries are re-ranked on the basis of their totalscores.' The Borda rule gives an equal weight of one to every non-missing policy variable. This methodallows us to aggregate and construct a composite indicator which reflects different dimensions of publicpolicies. The missing value problem is partially alleviated. Even if a policy variable is missing for a

4 Dasgupta (1991) used the Borda rule in constructing a composite index of well-being. The strengthsand limitations of the Borda rule have been investigated by Goodman and Markowitz (1954), Smith(1973), and Fine and Fine (1974). The Borda rule has also been used in evaluating country performnancefor IDA allocations by the Bank.

I We first rank the countries by individual policies. Then we take each country's rank score, say i,j, k, and I for the selected policies, and calculate the mean to get the country's Borda score:(i+j+k+l)/4. Finally, we re-rank each country's average Borda score, which results in a single rank --- the composite policy index.

7

country throughout the period, we still have an observation: The rank-score of the country would dependon the average rank-score of the non-missing policy variables.

Figure I shows the positive association between incentives and stability (INDEX 1) on the one sideand GDP growth on the other. Figure 2 does the same for productivity growth. A few countries appearas outliers. For instance, in East Asia incentives (or less distortion) seem to be associated with lowergrowth in the Philippines; the same is true of Cote d'Ivoire, Gabon, and Togo in Africa. All the EastAsian high-performing economies for which data are available appear in the top right of the graphs,indicating better incentives and more stability with higher growth in GDP per capita and in productivity(TFPG). In both figures these countries appear in a cluster slightly above and separated from the others,again suggesting that the policies reflected by INDEX1 appear to be more effective in these economiesthan in others.

Figures 3 and 4 show a possible nonlinear relationship between government expenditures andgrowth. Most of the East Asian economies appear in the upper-middle range of the graph -- the topportion of an inverted U curve. Five of the six East Asian economies with data have moderateexpenditures and high growth of output. However, moderate intervention alone is not necessarilyassociated with high performance: only 7 of the 31 countries with moderate rates of governmentexpenditures have performed well. Performance probably depends on the quality of implementation. thequality of human capital, the instrument applied (price subsidies, say, or favorable credit/tax treatment),the sectors chosen, as well as some intangible factors discussed.

8

Figure 1.GDP Growth and Openness Index

Annual Growth Rate of GDP per caplta0.08 Korea

0.06. * Thi lhand

Indonosiad.04 Malaysia

0.02_.

O P h~~~~~~~~~~~~~~~~~l llppinos

-0.02 **.

-0.04

-0.06

0 10 20 30 40 50Openness and Stability Index

Figure 2.Productivity Growth and Openness Index

Total Factor Productivity O rowth0.06

Hong Kong0.04

Koret. Thall AdSingapore.

0.02 . . *

O .,, , ... . . .

* .tPhilippinee* * * * * Indonesia

-0.02 * *

-0.04

-0.06

-0.08~~~~~~~~~~~~~~~~~~~-0.08 II " "'i "'i ' II i t l l 1 ' 1111 1 IjI 11 11111 111lll llll

0 10 20 30 40 s0Openness and Stability Index

9

Figure 3.GDP Growth and Government Expenditure

Annual Growth Rates of GDP per capita0;08

KoroaSingapore

0.06 Thailand

Indonesla Malaysia0.04

0.02 .Philippines

0

-0.02 ,

-0.04

-0.06

-0.08 £ II IIIIIIIII111Il

0 10 20 30 40 s0Government Expenditure Index

Figure 4.Productivity Growth

and Government Expenditures

Total Factor Productivity Growth0.04

KoreaThailand Singapore

0.02

I indonesisPhilippines

-0 02 .

-0.04

-0.06

-0.08 II l i l I I I I I I I I I I l I I I 1 I

0 10 20 30 40 sOGovernment Expenditure Index

10

We then estimate three equations (GDP per capita growth, TFPG1, and TFPG2) for the twocomposite indexes, based on country-average data using different specifications and with and without theregional dummy for East Asia and the interaction terms. Tables 3 and 4 summnarize the results of theseregressions. INDEX1 is positive and significant in all equations using both specifications, indicating alink between incentives (or lack of price distortion) and stability and growth (Table 3). It is also able toexplain as much as 45 percent of the variation in growth in income per capita and 23 percent of that intotal factor productivity growth. Up to 52 percent of the variation in GDP growth rate can be explainedby this index, a regional dummy, and the interaction terms. A quadratic function form is attempted butnot shown since the coefficients are not significant in all of the three equations.

East Asia appears to be more efficient in trade policies and in maintaining macroeconomicstability. In all three equations, the interaction term between East Asia and INDEXI is positive andsignificant (at 90 percent confidence), indicating a bigger slope coefficient. And this relationship holdswhen the specifications are changed. An F-test is also conducted to compare a restricted model -- withoutthe dummy and the interaction term -- with a full model. The results suggest that the full model issignificantly "better" in its specification than the restricted model; in other words, East Asia has a slopecoefficient that is different from that of other LMICs.

Three specifications are estimated for INDEX2: a linear form, a quadratic form, and a "kinked"specification. None of the three equations using linear specification shows a statistically significantassociation between INDEX2 and growth, and its explanatory power is negligible. That specification isnot shown in Table 4. The nonlinear quadratic specification fits the data better: in all three equations,INDEX2 is consistently positive, with a negative coefficient for the quadratic term and t-ratios rangingfrom 3.0 to 4.9. These results suggest that government interventions, as measured by expenditures, havea positive and significant association with growth, but that the effect gradually weakens as countries getricher. This finding is consistent with those we found using panel data and those of previous studies byBarro (1990) and Easterly (1992).

The merit of the "kinked" specification is that it is easier to interpret the coefficients and tointeract with the regional dummy variables. The cut-off point is where INDEX2 equals 30. We alsotried other cut-off points, and the results were not sensitive for those between 20 and 30. In all 6equations, with or without interaction and dummy variables, the first segments of the fitted lines(INDEX2kl) have positive and significant coefficient and negative intercepts. The second segments(INDEX2k.2) have negative and significant slope coefficients and positive intercepts (as shown by thedummy for INDEX2k2). This suggests that government intervention as measured by expenditures ispositively associated with GDP and productivity growth up to a point, but it then becomes harmful.

Is government spending more efficient in East Asia? The answer is ambiguous. To investigatethis question, we included interaction terns between East Asia and INDEX2 in the kinked specification.The t-ratios show that interaction terms, individually, are insignificant in all three equations. All the EastAsian countries are clustered around the top portion of the "kinked" line. The results of F-tests conductedto compare the restricted model -- the one without the East Asia dummy and the interaction term -- andthe full model reject the restricted model only in the equation for GDP growth. This suggests that EastAsia dummy and interaction terms, as a group, should not be ignored.

When both INDEX1 and INDEX2 are included in the regressions (not shown), the coefficientsof the two indexes are robust. In both quadratic and kinked specification for INDEX2, the coefficients

11

of INDEX1 are positive and significant, with t-ratios ranging from 6.1 to 2.2. INDEX2 still has thekinked relationship with growth. The signs remain the same, and the t-ratios improve. Interaction termsfor East Asia and INDEXI remain positive and significant. Interactions for East Asia and INDEX2 aremostly insignificant, except for two cases. The magnitude of the coefficients does not change much. Thetwo indexes have a correlation coefficient of 0.16 and are insignificant (probability> I RI =0.26). Thisconfirmns the notion that government intervention does not necessarily lead to market distortion.

Table 3Growth and a composite index on Incentives:

Using average data for 1977-90

(1) (2) (3)Dependent Variable: zgdpcap TFPG1 TFPG2

Soecification 1without regionaL dummv

INDEXI .0010 .0005 .0006(6.71) (4.32) (4.38)

constant -.0225 -.0135 -.0185(-4.60) (-3.21) (-4.14)

N 59 59 59Adjusted R2 .4316 .2330 .2384F Value 45.04 18.62 19.15

Specification 2with a dumw and an interaction ter

INDEX1 .0007 .0005 .0005(4.39) (3.18) (3.19)

EA.INDEXI .0018 .0021 .0020(1.50) (1.76) (1.60)

East Asia -.0719 -.1049 -.1007(-1.06) (-1.67) (-1.50)

Constant -.0172 -.0119 -.0167(-3.65) (-2.70) (-3.59)

N 59 59 59Adjusted R2 .5230 .2528 .2529F Value 22.20 7.54 7.54

Note: Equation 1 concerns averages for 1977-90, while (2) and (3) are for 1977-87.

1) Dependent variable, TFPG1 is calculated based on equation (3), and TFPG2 based on equation (4) in Table A-1.2) INDEXI is constructed, using Borda ranking technique, by Trade policy index, real interest rates, PMP andinflation, and three other indexes: cmpindex, Dollar, and total. See Appendix I for descriptions.

12

Table 4Growth and a composite index on govenment expenditures:

Using average data for 1977-90

(1) (2) (3)Dependent Variab!e: zgdpcap TFPGI TFPG2

Nonlinear soecificationINDEX2 .0028 .0025 .0027

(3.02) (4.09) (4.35)

INDEX2sq -.0001 -.0001 -.0001(-3.12) (-4.38) (-4.62)

constant -.0171 -.017 -.0217(-1.6) (-2.50) (-3.17)

Adjusted R2 .1359 .2637 .2862F VaLue 4.85 9.78 10.822

Kinked specificationINDEX2k1 .0009 .0009 .0009

(1.70) (2.57) (2.73)

INDEX2k2 -.0023 -.0016 -. 0017(-2.4) (-2.5) (-2.6)

Dummy for INOEX2k2 .1035 .0711 .0759(2.62) (2.71) (2.85)

constant -.0051 -.0077 -.0118(-.56) (-1.3) (-1.9)

Adjusted R2 .1048 .1947 .2113F value 2.912 4.949 5.376

Kinked specification with recionat dummv and interaction terms

INDEX2k1 .0007 .0008 .0009(1.42) (2.33) (2.52)

INDEX2k2 -. 0018 -.0017 -.0018(-2.05) (-2.6) (-2.7)

Dummy for INDEX2k2 .0856 .0805 .0856(2.28) (2.86) (3.01)

EA.INDEX2k1 .0010 .0001 .0001(.76) (.06) (.03)

EA.INDEX2k2 -. 0001 -. 0007 -. 0008(-.18) (-1.2) (-1.3)

East Asia .0279 .0148 .0174(1.21) (.86) (.99)

constant -. 0080 -. 0095 -. 0139(-.94) (-1.5) (-2.2)

Adjusted R2 .3112 .2122 .2354F value 4.689 3.200 3.514

Note: Number of observationsw5O. Equation (1) concerns averages for 1977-90 white (2) and (3) are for 1977-87.INDEX2 is constructed, using Borda ranking technique by the shares of public sector investment (invpub),government fixed capital formation (gfixcap), productive expenditures (prodexp) and total expenditures(gexptot). See Appendix I for descriptions.

13

To test the robustness of our results, we also built two composite indexes based on annual data.INDEXla was constructed using only four variables, the trade policy index, real interest rates, parallelmarket premium, and inflation rates. The other three policy indexes included in the INDEX1 based onaverage data were excluded because of incomplete coverage. INDEX2a was built on the same set ofpolicy variables as the INDEX2 based on average data. There were 222 observations for INDEXla and488 for INDEX2a.

INDEXIa is positively and significantly associated with GDP and productivity growth, with t-ratios ranging from 6.1 to 2.0. Interaction terms for East Asia and INDEXI are always positive andsignificant, with t-ratios ranging from 2.1 to 2.5. INDEX2a, again, shows some features of non-linearity:in the quadratic specification, the index itself has a positive coefficient, with a negative coefficient forits quadratic term; both are significant. In the kinked specification with the cut-off point atINDEX2a=250, the first segment (INDEX2akl) of the line is positive and significant; the second segmentis negative and significant in the equation for GDP per capita. These results are consistent with thosefound using indexes based on country-average data, and hence are not shown.

Our hypotheses are supported by both the quasi-Solow model and the endogenous growth model.The regression results for the policy indexes based on average data and those based on annual data arehighly consistent. Overall, the results are not sensitive to different approaches and specifications.

V. COUNTRY DIFFERENCES

So far this paper has presented only a group portrait of East Asian performance and policies.When the group is disaggregated, it turns out that every one of the 10 has done better than the averagefor 58 other developing economies, whether in 1960-90 or in 1975-90 (Table 5). The variations inperformance are also smaller for the 10 East Asian countries than for the other 58. Nevertheless, withinEast Asia, country differences are substantial, as indicated by the standard deviation.

The story is similar for country policies. Table 6 shows that East Asia as a whole had a policyframework that was superior to that of the other counitries. However, by some measures countrydifferences are as large within the region are as they are elsewhere. An exception is macroeconomicstability. Not only was it more pronounced in East Asia, but variations within the region were also lessthan in the other 58 countries. For a measure of sectoral distortions through agricultural protection, EastAsia even has a higher variation than other regions. Korea's direct protection of agriculture stands incontrast to the practice of the other countries. These differences reflect the fact that there was no single,regionwide approach adopted by all the successful East Asian economies.

14

Table SGrowth Performance: East Asia and other LMICs, 1960-90 and 1975-90

Country GDP Per capita growth TFPG1 TFPG21960-90 1975-90 1960-90 1975-90 1960-90 1975-90

China 4.13 6.00 2.71 3.06 1.U8 2.84Hong Kong 5.97 5.72 4.66 4.28 3.96 4.03Taiwan (China) 5.21 4.67 3.88 4.08 3.63 3.85Indonesia 3.20 3.88 0.82 -0.71 0.33 -0.91Japan 5.28 3.29 2.46 1.56 2.33 1.47Korea, Rep. 6.61 7.03 2.38 3.26 2.13 3.05Mataysia 3.95 3.78 2.21 1.08 1.95 0.80Philippines 1.40 0.76 1.02 0.20 0.81 0.01Singapore 6.25 5.79 3.10 2.21 2.84 1.98Thailand 4.81 5.31 2.56 2.63 2.34 2.42

East Asia 4.68 4.63 2.58 2.16 2.22 1.95(4.9) (3.8) (4.2) (3.2) (4.2) (3.2)

Other LMICs(58 Economie3) 1.26 0.26 0.91 0.09 0.39 -0.32

(5.4) (5.3) (4.9) (4.8) (4.9) (4.8)

Note: Simple country means and group means are presented. Standard deviations are in parentheses.

Table 6Distortion: Trade Incentives and Stability

East Asia and Other LMICs

Incentive resime Macro stability Sectoral incentives

Country TPI Cmpindex Dollar PMP Inflation Real Agricultural disprotection'rate interest Direct Indirect

China 65.59 5.87Hong Kong . . 101.90 -0.53 6.10Taiwan (China) . . . . .Indonesia 1.25 10.61 101.43 6.13 28.51 1.18Japan . . . 2.34 5.49Korea, Rep 3.38 13.36 101.27 18.15 10.88 2.93 0.52 -0.26Malaysia . . 101.33 1.17 3.29 4.11 -0.10 -0.08Philippines 1.28 10.56 101.47 10.85 10.60 1.95 -0.05 -0.23Singapore . 14.08 101.15 0.85 3.19Thailand 2.67 . 101.95 0.25 5.26 8.37 -0.29 -0.15

East Asia 2.08 11.89 101.30 12.26 8.93 3.71 0.3 -0.18(1.02) (3.51) (0.47) (37.32) (15.0) (7.94) (0.39) (0.09)

Other LMICs,58 economies 1.69 9.18 99.83 42.30 17.98 -3.01 -0.13 -0.25

(0.93) (3.81) (1.29) (74.04) (35.15) (19.60) (0.24) (0.17)

This is based on Krueger, Schiff and Valdes (1992). See description for "Total" in Appendix I.Note: Simple country and group means. Standard deviations are in parentheses. See Appendix I for a

description of variabLes and time periods.

Table 7Interventions: Public Enterprises, Government Expenditures, Freedom and Liberty

East Asia and other LMICs

Piblic Sector GeneraL Consolidated governmentas a X of GDP goverrnent Exoenditures as a X of GDP Freedom and libertybInvest- Value- consLuxption Total Fixed capital Education Health Transportation Civil Political

Country ment added formation and Communication Liberty liberty

China 14.28 6.23 6.31Hong Kong 3.86 6.91Taiwan(China) 11.48 16.48 4.92 5.23Indonesia 8.76 19 9.88 19.51 6.18 1. 76 0.43 2.46 5.23 5.00Japan 5.41 8.82 1.00 1.54Korea, Rep. 6.73 3-10 10.60 16.18 1.31 2.84 0.24 0.68 5.46 4.77Malaysia 11.56 15.30 28.51 2.52 4.47 1.16 2.23 4.08 2.92Philippines 5.16 2-4 8.56 14.02 1.16 2.24 0.65 2.50 4.62 4.69Singapore 10.97 10.78 21.06 3.03 3.74 1.37 5.00 4.62ThaiLand 7.11 5-6 10.89 17.77 3.14 3.57 0.88 1.50 4.08 4.15

East Asia 8.53 9.5 10.92 19.46 2.87 3.12 0.79 1.84 4.51 4.36(4.27) (3.40) (5.6) (2.04) (1.43) (0.52) (0.41) (1.52) (1.46)

Other LMICs (58 Ecnmmies)

8.97 10.9 13.52 24.90 3.40 3.07 1.50 1.85 4.60 4.74(5.75) (5.14) (11.6) (2.96) (1.78) (1.30) (1.57) (1.55) (1.96)

* From Nair and Filippides (198). This estimate for 1984 is for 28 developing countries.b Based on Gastil (1989).Note: Simple country nd group means. Standard deviations are in parentheses. See Appendix I for a description of variables and time periods.

16

The same is true of government interventions (Table 7). In several East Asian economies (China,Taiwan (China), Malaysia, Singapore), the public sector's share in investment was significantly higherthan the average for the region and for the 58 countries elsewhere. Within East Asia there was also widevariation in the public sector's share in value added (although data were available for only four countries).For the region, the share of public investment in GDP was about the same as that elsewhere; forgovernment consumption, however, the East Asian share was significantly smaller.

Political and civil incentives (measured by Gastil's index of liberty, Gastil 1989) was no greaterin East Asia than in other developing countries. The pattern varied from country to country, with somein East Asia showing significantly less openness than the average for other developing countries.

Finally, East Asian countries have been investing heavily in infrastructure. Although the sharesof government expenditures in transportation and communication are similar, the averages for the roaddensities (unpaved and paved) is much higher for East Asia than other LMICs (290.4 and 251.9 km/area,respectively using available data), so is its rail density. Cross country variations are substantial: The roadand rail densities in Korea, for example, are at a similar levels as in Spain. Whereas, the rail density inChina (5.5 km/area) is much lower than that of India (18.8 kn/area).6

VI. EAST ASIAN POLICIES AND EFFECTS

These differences among countries in the region suggest that it would be a mistake to think thatthey all adopted a single, uniform model of development. There are, nonetheless, certain broad policyfeatures shared by the successful East Asian economies that have contributed to their rapid economicgrowth. Chief among these have been outward-orientation, low price distortions, macroeconomicstability, and heavy investment.

With trade and price policv, the better policy framework in East Asia (as measured by variousindexes) was associated with better performance. Governments did not intervene much less thanelsewhere, but their interventions were "market friendly," and involved few distortions. In trade, forexample, government intervention was subjected to the tests of international competition. Interventionsthat failed such tests were changed or abandoned.

This work highlights the importance of macroeconomic stabilitv, as reflected by inflation, interestrates, and the parallel market premium for the exchange rate. The region's macroeconomic stabilityaccounts for its particularly low overall distortion index. Its current account deficit for the past quarterof a century has remained below 1 percent of GDP, a remarkable achievement. China's influence on thisstatistic is strong; but even without China, the current account deficit was small.

Discipline in public borrowing was also a beneficial factor in East Asia. Until 1987, outstandinggovernment bonds in Taiwan (China) were legally restricted to 40 percent or less of the centralgovernment's annual budget. Thailand limits its budget deficits to 20 percent of total expenditure. In

6 Calculated based on data from the World Road Statistics and International Railway Statistics.Available on request.

17

Indonesia, a balanced budget law and a paucity of exchange controls serve as checks on irresponsiblefiscal behavior that could precipitate currency speculation and crisis. Malaysia has run large deficits attimes (as high as of 19 percent of GNP in 1982), but has cut them sharply (to 5 percent in 1990) whenthey threatened to damage economic growth.

These constraints on public borrowing did not always apply to public spending. In Taiwan(China), governmnent spending was 26 percent of GNP in 1980 and 35 percent in 1989, well above thelevel in many developing countries. But the public sector has been a net saver since 1964, nevercrowding out the private sector. State enterprises, contributed 49 percent of value added in industry in1955; their share is now down to 11 percent, but not because of privatization. These state enterprisesgrew in real value, but the private sector grew much faster. This pattern may help explain why publicinvestment is not significantly related to growth (in our earlier analysis) whereas private investmentplayed a major role. Meanwhile, government consumption has hampered growth less in East Asia thanelsewhere, probably because it did little to crowd out private investment.

Though our paper did not consider the effects of investment in education, this has long beenrecognized as vital to competitiveness and growth. Total factor productivity is larger in East Asia partlybecause of the larger contribution of education. It also seems likely, from the evidence presented hereand elsewhere, that East Asia's investment in education was a key element in its better policy framework.There was a strong positive interaction between incentives and low price distortions on one side andeducation on the other.

It is now widely accepted that East Asian governments have intervened in the economy(particularly in the 1960s and 1970s), but that they did so more efficiently than other governments. Lesswell-known is the region's even greater success with market reforms in the 1980s compared with otherdeveloping countries, including those that were also reforming. The gap in average income growthbetween East Asia and other developing regions (except South Asia) roughly tripled after 1982 comparedwith the period before. So the famniliar question -- why were East Asia's interventions more successful? -- needs to be matched by another: why were its reforms more successful as well?

Various country studies provide some answers. First, policy changes in East Asia were deeperthan those of other developing countries. A simple average of nine measures of exchange rate and traderestrictiveness (for example, advance import deposits or surrender requirements for export proceeds) fellby half in East Asia from 1979 to 1990, far more than the average for other developing countries. Norwere its reforms offset by new impediments: for instance, many developing countries imposed referenceprices on imports while tariffs were being reduced.

Second, the reforms were carried out vigorously. Indonesia, Korea, Malaysia, and Thailand tooknearly 30 percent of direct foreign investment in developing countries in 1990, primarily, it seems,because they took measures to change the business climate as well as policies. Indonesia and Koreaintroduced the "one-stop" offices for ministerial clearances, Malaysia removed numerous impedimentsto industrial licensing, and Thailand emphasized the warmth of its welcome. Such intangibles mayexplain why changes in policies alone cannot explain differences ir. results.

Third, the supply response to reforms depends critically on complementary conditions -- manpowerskills and flexibility, infrastructure (Table 8), and institutional factors. Where incentives, investments,and institutions are missing or inadequate, output responds weakly to reform. Much depends on the

18

building of consensus, public-private partnership, and the bureaucracy's capacity for policy planning andImplementation. While certain policies and reforms in East Asia brought larger benefits than elsewhere,the reforms themselves are insufficient to explain the payoffs. Again, unquantifiable factors played apart.

Yet these intangible influences were not there from the start; they had to be developed andnurtured. It is striking that Singapore's savings rate (including mandatory savings) rose from I percentof GNP in 1965 to about 40 percent today, perhaps the highest in the world. Cultural and institutionalchanges triggered Korea's success; an example is the way Confucianism elevated the status of thebureaucracy while at the same time inspiring it to nurture business values.

VII. CONCLUSION

The broad findings of our paper are quite clear. Growth is damaged by trade and price distortionsand by macroeconomic instability. However, it is not obviously affected, one way or the other, bygovernment expenditure. In East Asia, price distortions and macroeconomic instability were relativelylower than those in other developing countries, so did less to hamper growth. By contrast, governmentexpenditure in East Asia was broadly in line with the rest of the developing world, as a share of GDP.

When it comes to reforms aimed at reducing price distortions, these were more productive in EastAsia than in developing countries overall. Some of East Asia's success is attributable to a superior policyframework, and some to the greater returns achieved by a given policy framework or set of reforms. Apositive and significant interaction between East Asia and its policy framework suggests that the regionhas been more efficient in getting payoffs from a superior policy framework. Government expenditureshave a less clear and nonlinear relation with economic growth, with most East Asian economies belongingto the middle range of an inverted U curve.

These findings have a powerful relevance for developing countries everywhere. There was notsomething uniquely or inherently East Asian about the region's economic success. The essentialingredients -- avoiding serious price and trade distortions, and establishing macroeconomic stability -- arekey for generating growth anywhere. Numerous countries have recognized this fact and have beenreforming their policy regimes over the past decade: the East Asian experience is a basis for thisapproach, not an exception.

But the region does seem to be an exception in getting the most from its reforms and in makingpublic expenditures work to promote growth rather than hinder it. East Asian countries have beeneffective in blending the roles of market and state. The combination of modest distortions,macroeconomic stability, and effective government spending, together with various intangibles such asconsensus building and efficient bureaucracies, has made possible the rapid productivity and economicgrowth of the region. The intangible factors may not be linked to immutable cultural traits, and the bulkof the East Asian experience is one that can be copied elsewhere.

19

Appendix IDescription of Variables used in the Analysis

Variables Descriptions

Demendent variableszgdpkd Growth rates of GDP at constant 1980 prices, US dollars, 1960-1989. Source: IECNA, ESD.

zgdpcap Growth rates of per capita GDP at constant 1987 prices, US dollars, 1960-1990. Calculatedbased on GDP_MP in IECNA and population figure in SOCIND, 9ESD.

TFPG1 Total factor productivity growth, residual from a Cobb-Douglas production functionwith capital, labor, land as Inputs for 1960-87 (equation 3 in Table A-1).

TFPG2 Residual from a Cobb-Douglas production function with physical capital, labor, land, andeducation as inputs (equation 4 in Table A-1).

1. Indicators of Incentives and Stability (or distortions)

Trade policy index (TPI) Index of trade restrictiveness from most restrictive (1) to least restrictive(5), for 38 countries in 1977-88. Source: Thomas, HaLevi and Stanton(1991).

PMP Parallel market foreign exchange premia based on the differences between official exchangerates and black market rates: PMP=C(BHER-0ER)/OER)*100, where BMER is the parallel marketexchange rate and OER is the official end of period exchange rate. Source: SBER--InternationalCurrency Analysis, Inc. World Currency Yearbook. OER--IHFIFS, BESO.

Inflation Inflation rates calculated as log differences of consumer price Index when it exists, else togdifferences of wholesale price index. (1961-1989 for 84 countries). Source: IMFIFS, SES.

Real Real interest rates for 34 countries in 1970-88. Inflation was subtracted from the nominalinterest rate. Source: Gelb (1989)

Cmpindex Index of trade Liberalization from one (Least liberalized) to twenty (most liberalized) for 19countries in 1960-85. Source: Papageorgiou, Michaely, and Choksi (1990).

Dollar Purchasing power parity-based outward orientation index for 92 countries, calculated as theweighted average of mean price distortion in the period of 1973-85 and of its standarddeviation. Source: Dollar (1990).

Total Total disprotection of agriculture for 18 countries in 1960-86. It has two components: directand indirect disprotection. The first one measures the direct taxes/subsidies on agriculture.The second measures the impact of economy-wide policies on agricultural incentives. Source:Krueger, Schiff and Valdes (1992).

11. Goverrint simenditures

invpub Ratio of public sector investment to GDP. The public sector Investment is defined as capitalexpenditure of the consolidated general government pLus that of public corporate entities.Source: DEC analytical Data Base (IEC, World Sank).

gconx Share of general goverrnent consumption in GDP for 1960-89. Source: IECHA, SESD (CON-GOV,GDP_HP).

gexptot Share of consolidated government expenditure in GDP for available years in 1960-1990. Source:IHFGFS, 9ESD (TOTEXP TOTAL).

gfixcap Share of gross fixed capital formation for consolidated government in GDP. Source:IHFGFS, BESO(GFCF).

prodexp Share of productive expenditures including education, health and economic and infrastructureservices. Source: IHFGFS, SESD (FNTOTEXP_EDUC, FHTOTEXP HEALTH, FNTOTEXP_ECON).

20

continued.

Variables Descriptions

III. Cupposite indexes

indexl An index for incentives and stability based on average data for 1977-87 for sevenvariables: three measures of trade restriction (TPI, Cmpindex, Dollar), and nmeasures ofsectoral distortion (Total), real interest rate (real), the PMP and Inflation rate. Eachvariable enters in an ascending order from less to more stability (or large to smalldistortions). See text for the Borda ranking method.

index2 An index for government intervention based on average data for 1977-87 for four variables:public sector investment (invpub), goverrvnent fixed capital formation (gfixcap), productiveexpenditures (prodexp, including expenditures on education, health and economic services), andthe total government expenditures (gexptot); each as a share of GDP.

index2sq index2 squared.

index2kl =index2 if index2<=30, =0 otherwise.

index2k2 =index2 if index2>30, =0 otherwise.

dindex2k2 =1 if index2k2>0, =0 otherwise.

EA. Interaction terms between East Asia and indexes.

IV. irout variables

zcapital (zkpn) Instrumented growth of capital variable. Source: 8halla-Lau (1991). For a description, seeWorld Develooment Reoort 1991 Supplementary Data.

aland Growth rates of arable land in thousands of square kilometers. Source: SOCIND, BESO (AREA AR).

zlabor Growth rates of total labor force, interpolated. Source: SOIND, BESD (LABOR).

edt Estimated average years of education of the population of working age group (15-64). Based onUNESCO data on enrolLment rates for the period 1960-1988 and on mortality and birth statistics.Source: Lau, Jamison and Louat (1991) and Louat, Bhalla, and lau (1992). For a description,see WDR 1991 Supplementary Data.

edt60 edt in 1960.

deO3 =first difference of edt if edt is between 0 to 3 years, 0 otherwise.

de39 =first difference of edt if edt is between 3 to 9 years, 0 otherwise.

V. RegionMl Dumies

East Asia =1 if the country is under the East Asia and Pacific Vice-Presidency, =0 otherwise.

21

Appendix I1A List of Economies in the Analysis

Economies Value of Value of Economies Value of VaLue ofINDEXI INDEX2 INDEXI INOEX2

1. Argentina 26 13 36. Morocco 44 442. Burundi 18 na 37. Madagascar 22 no3. Benin na na 38. Mexico 32 214. Bangladesh 24 4 39. Mali 39 85. BoLivia 2 9 40. Malta na 406. Brazil 21 14 41. Mauritania 12 na7. Cen.Afri.Rep. 40 na 42. Mauritius 48 248. ChiLe 53 22 43. MaLawi 19 439. China na na 44. Malaysia 58 3710.Cote d'lvoire 42 na 45. Nigeria 8 na11.Cameroon 36 26 46. Nicaragua na na12.Congo na no 47. Taiwan (China) no no13.Colombia 51 17 48. Pakistan 49 2014.Costa Rica 27 23 49. Panama 46 2915.Algeria 11 na 50. Peru 15 1216.Egypt 33 39 51. Philippines 47 617.Spain 52 15 52. Portugal 43 3618.Ethiopia 20 28 53. Rwanda 13 1619.Gabon 38 49 54. Sudan 3 na20.Ghana 5 7 55. Senegal 37 na21.Greece 30 35 56. Singapore 57 3022.Guatemala 25 3 57. EL Salvador 6 523.Hong Kong 56 na 58. Syria 16 5024.Haiti 29 2 59. Togo 54 4725.Hungary na 48 60. Thailand 59 1926.Burkina Faso 35 18 61. Turkey 34 3427.Indonesia 45 31 62. Tanzania 7 4128.1ndia 41 no 63. Uganda 1 129.1srael 31 na 64. Venezuela 9 3230.Jamaica 23 27 65. Yugoslavia 14 1131.Japan na na 66. Zaire 4 2532.Kenya 28 33 67. Zambia 10 4533.Korea, Rep. 55 10 68. Zimbabwe 17 3834.Liberia na 4635.Sri Lanka 50 42

Note: INDEXI is constructed for 59 economies, INDEX2 for 50 economies. Those for which both indexes are notavailable due to missing policy variabLes are used only in the descriptive analysis and the estimationof Cobb-DougLas production functions.

22

Table A-iTotal Factor Productivity Growth by Region and by economy:

Estimated Cobb-Douglass Production Function using LSDV approach

Dependent Variabte: zgdpkd

Re4ressions by reion Regressions by individual economywithout with without witheducation education education education

Variables (1) (2) (3) (4)

zcapital .3862 (17.8) .3917 (18.0) .3909 (17.1) .3972 (17.3)

zlabor .3137 (2.68) .3643 (3.04) .0978 (0.59) .1677 (1.00)

zland .0509 (1.57) .0476 (1.47) .0561 (1.68) .0529 (1.59)

Education:

deO3 .0671 (2.48) .0906 (3.12)

de39 -.0020 (-.92) -.0031 (-1.4)

edt6O .0014 (1.60) .0009 (1.5)

Total Factor Productivity Growth

by reion TFPG1 TFPG2 by economy* TFPG1 TFPG2

Africa .0012 (0.36) -.0057 (-1.3) China .0271 .0188(2.66) (1.79)

Hong Kong .0466 .0396(4.30) (3.57)

East Asla .0206 (4.43) .0114 (1.66) Indonesia .0082 .0033(0.82) (.33)

Japan .0246 .0233(2.54) (2.40)

S. Asia .0075 (1.38) -.0001 (-.02) Korea .0238 .0213(2.26) (2.02)

MaLaysia .0221 .0195(2.09) (1.84)

EMENA .0149 (4.12) .0081 (1.56) Taiwan(China) .0388 .0363(3.56) (3.41)

Philippines .0102 .0081(1.00) (.80)

LAC .0023 (0.57) -. 0046 (-.86) Singapore .0310 .0284(2.90) (2.65)

ThaiLand .0256 .0234(2.48) (2.26)

N 1826 1826 1826 1826Adjusted RX .2032 .2063 .2391 .2440F Value 236.47 173.03 28.11 27.63

Note: Least Square Dummy Variable approach is used here on data from 68 countries for the period of 1960-1987.* In equations (3) and (4), 68 dummies for economies are incLuded, only 10 East Asian economies are presented.* t ratios are in parentheses.

23

Bibliography

Amsden, Alice H., 1991, "Diffusion of Development: The Late-Industrialization Model and Greater EastAsia", American Economic Review, May, 81:282-286.

Amsden, Alice H., 1989, Asia's Next Giant: South Korea and Late Industrialization, New York: OxfordUniversity Press.

Barro, Robert J., 1991, "Economic Growth in a Cross Section of Countries", Quarterly Journal ofEconomics, 106(2):407-433.

Barro, Robert J., 1990, "Government Spending in a Simple Model of Endogenous Growth", Joumal ofPolitical Economy, 98(5,pt 2):S103-125.

Barro, Robert J., 1989, "A Cross-Country Study of Growth, Saving, and Government", NaER WorkingPaper No. 2855, February 1989.

Barro, Robert J., and Jong-Wha Lee, 1992, "International Comparisons of Educational Attainment",Working Paper, The World Bank.

Bhagwati, Jagdish N., 1988, "Export-Promoting Trade Strategy: Issues and Evidence", The World BankResearch Observer, January.

Bhalla, Surjit, Lawrence Lau and Louat, (forthcoming), "Human and Physical Capital Stock inDeveloping Countries - Construction of Data and Trends", Working Paper, The World Bank.

Bhalla, Surjit, and Lawrence Lau, (forthcoming) "Determinants of Growth and Productivity inDeveloping Countries", Working Paper, The World Bank.

Collins, Susan M., 1990, "Lessons from Korean Economic Growth", American Economic Review, May1990, 80(2):104-107.

Dasgupta, Partha, 1990, "Well-Being and the Extent of its Realization in Developing Countries",Economic Journal 100,4: supplement.

De Long, Bradford, and Lawrence H. Sunmmers, 1992, "How Robust is the Growth-Machinery Nexus?",Working Paper, The World Bank.

Dollar, David, 1991, "Outward Orientation and Growth: An Empirical Study Using a Price-BasedMeasure of Openness", Working Paper, The World Bank.

Dowrick, Steve and Norman Gemmell, 1991, "Industrialization, Catching up and Economic Growth: AComparative Study Across the World's Capitalist Economies", The Economic Journal, March, 101:263-275.

24

Easterly, William, Michael Kremer, Lant Pritchett, and Lawrence H. Summers, (1992), "Good Policyor Good Luck? Country Growth Performance and Temporary Shocks", Working Paper, The World Bank.

Easterly, William, Robert King, Ross Levine and Sergio Rebelo, 1991, "Do National Policies AffectLong-Run Growth? A Research Agenda", Working Paper, The World Bank.

Easterly, William, and Sergio Rebelo, 1992, "Fiscal Policy and Economic Growth: An EmpiricalInvestigation", Working Paper, The World Bank.

Fine, B. and K. Fine, 1974, "Social Choice and Individual Rankings, I and II", Review of EconomicStudies, 44.

Fischer, Stanley, 1992, "Macroeconomic Factors in Growth", Working Paper, The World Bank.

Gastil, Raymond, 1989, Freedom in the World. New York: Freedom House.

Gelb, Alan H., 1989, "Financial Policies, Growth and Efficiency", Working Paper, The World Bank.

Goodman, L.A. and H. Markowitz, 1952, "Social Welfare Functions Based on Individual Rankings",American Journal of Sociology, 58.

Greenaway, David, and Chris R. Milner, 1987, "'True Protection' Concepts: Their Role in EvaluatingTrade Policies in LDCs", Journal of Development Studies, January, 23(2):200-219.

Grossman, Gene M. and Helpman, Elhanan, "Trade, Innovation, and Growth", American EconomicReview, May 1990, 80(2):86-91.

Harrison, Ann E., 1991, "Openness and Growth: A Cross-Country, Time Series Analysis for DevelopingCountries", Working Paper, World Development Report 1991, The World Bank.

Kaufmann, Daniel, 1991, "The Forgotten Rationale for Policy Reform: The Productivity of Investment",Working Paper, World Development Report 1991, The World Bank.

Kaufmann, Daniel, and Yan Wang, "The Impact of Macroeconomic Policies on Project Performance inSocial Sectors", Working Paper Series WPS #939, The World Bank.

King, Robert G., and Sergio Rebelo, 1990, "Public Policy and Economic Growth: DevelopingNeoclassical Implications", Journal of Political Economy, 98(5,pt 2):S126-144.

Krueger, Anne O., Maurice Schiff, and Alberto Valdes, 1992, "The Political Economy of AgriculturalPricing Policies", in vol.4, Synthesis: The Economics in DeveloRing Countries, by Schiff and Valdes,John Hopkins University Press, forthcoming.

Krueger, Anne O., 1990, "Asian Trade and Growth Lessons", American Economic Review, May 1990,80(2): 108-112.

25

Krueger, Anne O., 1978, "Liberalization Attempts and Consequences", in Foreign Trade Regimes &Economic Development, vol.X, National Bureau of Economic Research, New York.

Krueger, Anne O., 1980, "Trade Policy as an Input to Development", American Economic Review,Papers and Proceedings, May.

Krueger, Anne O., 1984, "Comparative Advantage and Development Policy 20 years Later", inEconomic Structure and Performance, Syrwuin, Taylor and Westphal Eds., Academic Press.

Lau, Lawrence, Dean T. Jamison and Frederic F. Louat, 1991, "Education and Productivity inDeveloping Countries: An Aggregate Production Function Approach", PRE Working Paper 612, TheWorld Bank.

Lucas, Robert E., Jr. 1988, "On the Mechanics of Economic Development", Journal of MonetarvEconomics, 22:342.

Mankiw, N. Gregory, David Romer and David N. Weil, 1992, "A Contribution to the Empirics ofEconomic Growth", The Ouarterlv Journal of Economics, May 1992, p407437.

Nair, G. and A. Filippides, 1988 "How Much Do State-Owned Enterprises Contribute to Public SectorDeficits in Developing Countries - and Why", PPR WPS 45, The World Bank, December.

Pack, Howard and Larry E. Westphal, 1986, "Industrial Strategy and Technological Change", Journalof Development Economics, 22:87-128.

Papageorgiou, D., M. Michaely, and A.M. Choksi, 1990, "Liberalizing Foreign Trade in DevelopingCountries: Lessons of Experience", The World Bank.

Park, Yung Chul, 1991, "Liberalization in Korea and Taiwan (China)", in Foreign EconomicLiberalization: Transformations in Socialist and Market Economies, Eds. by Koves and Marer, WestviewPress.

Rebelo, Sergio, 1991, "Long-Run Policy Analysis and Long-Run Growth", Journal of Political Economy,June 1991, 99(3):500-21.

Romer, Paul M., 1990, "Endogenous Technological Change", Journal of Political Economv, 1990,98(5,pt2):S71-102.

Romer, Paul M., 1986, "Increasing Returns and Long-Run Growth", Journal of Political Economy,94(5): 1002-37.

Smith, J.H., 1973, "Aggregation of Preferences with Variable Electorate", Econometrica, 41.

Srinivasan, T.N., 1991, "Development Thought, Strategy, and Policy: Then and Now", a backgroundpaper for the World Development Report 1991, The World Bank.

26

Thomas, Vinod, Nadav Halevi and Julie Stanton, 1991, "Does Policy Reform Improve Performance?",a background paper for the World Development Report 1991, PRE, The World Bank.

Thomas, Vinod, John Nash and Associates, Best Practices in Trade Policy Reform, a World BankPublication, Oxford University Press, 1991.

Wade, Robert, 1990, Governing the Market: Economic Theory and the Role of Government in East Asianlndustrialization, Princeton University Press, Princeton, New Jersey.

Westphal, Larry, 1990, "Industrial Policy in an Export-Propelled Economy: Lessons from South Korea'sexperience", Journal of Economic Perspectives, Summer, 4(3): 41-59.

World Bank, 1991, The World Development Report 1991: The Challenge of Development, New York,NY: Oxford University Press, published for the World Bank.

World Bank, 1987, The World Development Report 1987, New York, NY: Oxford University Press,published for the World Bank.

Distributors of World Bank PublicationsARGENTINA The Middl East Observr KENYA SOUTH ARIUCA, iOTSWANACarlo. Hinch, SRL 41. Shenl Strft AMfa Bock Sev ie A.) d Fl, sPr mgig hitlavCalria Cueme Cairn Quaran Hous. Manut Swe Oxford Urnv ty rPiFlorida 16,, 4th FlerOf.c 453/46 P.O. Box 45245 Southen Afnc1333 Buenos Air FINLAND Nairobi P.O. box 1141

Akatanminenruruakauppx Cape Town Soe

AUSTltALIA, PAPUA NEW GUINEA. P.O. Box 128 KOREA, REPUBLIC OFFiJI, SOLOMON ISLANDS. SF-40101 Halinki 10 Pan Kort Biook Corporsion For smrM*m aeise:VANUATU, AND WESTEN SAMOA P.O. Box 101, Kwangehemun Intemnbnal Subsiptn SevimDA. Wonndras n Sevic FiANCE Seoul P.O. Box 41060 Whitehorse Road World Bank Publication Cr aighiUMictham332 64 AvenuedUlna MALAYSIA JhainnsabwSg224Victoria 75116 Pans Univaty d Malaya Coopeadve

Bookahop, Lniited SPAINAUSTRIA CERMANY P.Q Box 1127, Jlan Panta Bea Itundi-Prnur Lbrem L.A.CGrold and Co. UNO-Veriag 59700 Kuala Lumpur Cuosilo37Graben 31 Poppelsd orer Ale SS 28001 MadridA-1011 Wien D-S300 Bonn I MEXICO

INITEC Librwr InateIonadl AED06BANCLADESH HONGC ONC, MACAO Apardo Potxl 224160 Conall de Cnt391Mitco Indusatrie Development Asia 2000 Ld. 14060 Tlelpan Mexice D.J. O009 Barlona

Aiatnca Society 0IlDAS) 46-48 Wyndham SwettHouoS. Road 16 Winnrng Centzv NETHERLANDS SRI LANKA AND THI MALDIVUDhanmondi R/Aras 2nd Floor DP LindehoomAnOr-Publikeige Lake House BookahopDhaka 1209 Central Hong Kong P.O Box 200 P.O. Box 244

74i0 AE Haskaerpn 10. Sir OiznpelI& A.Buxh afic INDIA Cardian trMwSthvPine Vie.w. I Fleoor A8id Publishe Privato Led. NEW ZEALAND Colomebo 2100 Agrbed Commecial Area 751 Mount Road EbSCO NZ Ltd.Chittagong14100 Madras- 600 002 Private Mail Bg 99914 SWEDEN

New Market Fer sixlie htil76, K DA Avenue heady ofltca: Auckland Fntzs FackbektoraingetKulna 9100 1IS)N Haradla Marg Repwingpgan 12. Be 16386

alard Estate NICERIA S-10327 SoxcholmBELCIUM Bombay - 400 038 Univerity Pres Limitedlean De Lannoy Three Crowns Building Jeiche For eskavtie. em,Av. du Roi 202 13/14 Ass! AL Road Private MA Bag 5095 Wennergan-W11"ma AS1060 Bruset New Delhi -110 002 ibadan P.O oxa L30S

5-171 255ehinCANADA 17 Chittaranjan Avenue NORWAYLe Ditfuseur Calcu tt- 700072 Narveon Iniomotion Guniar SWITZERLANDCP. *5 1501 rume Arpar Book Department For siasg btileBeucherville, Quder layadeva Hoel Building P.o. Box 61S Bi;enbd Libraln Payotj4B SE6 5th Main Road Candhiragar N-06t2 OCae 6CAs poetrl 3212

Bengalors - 560 009 CH i0M LaunmwCHILE PAKISTANtnveruc ICTS.A. 3-5-1129 Kachiguda Mira Book Agency Fo sabcria doa:

Av Santa Maria 6400 Co Road 65, Shahrah-e-Qw4id.-Azem Librairie PayrtEdificiNTEC. i0 201 Hyderabad - 500 m7 P.Q Box No. 72 Swrvice don Ab tnnebSantago L.ahor 54M CA poele 33U

Prarthana Flaw, 2nd Roor CH 1002 LauanmCHiNA Nerb Thakom aug,Naranogpura PERUChina Financial & Economic Ahmnedabad - 360 009 Editoral DPrreloSA THAILAND

Publishing House Aparado3 24 Centrl DPepartm SieP D Fe Si Dong Jie Patala House Lime 1 306 51iom Roed

Beijing 16-A Ashok Marg BrngjokLucknow - 226 001 PIDUPPiNES

COLOMBIA Intenuabonr Book C n_ TRINIDAD & TOBAGO, ANTICUAineenlae Lde. Central BzeazrRed Suite 17D3 tyland 10 bAIBUDA, BAIBADOSApanado Aerso 34270 60 Baiaj Napt Condominiui Towel DOMINICA. iLENADA, GUYAANABogta DE Naur440 010 Ayla Avenue. NV. dale JAMAICA. MONTSELRAT, ST.

Cosak EAatnmaen KITTS & NEVI. ST. LUCA,COTE DIVOIRE INDONESIA MakAd, Meto Maia ST. VI NCEIT & CRENADINESCente d'Edibon d de Piifueion Pt indira Limited Systenadc Studle Unit

Atrcaina(CEaDA) J)Ian Borobudur20 POLAND 89 Watts Stwed04 LP. 541 P.O. Box 181 Interuttional PublshingServiet Cusapa'Abidjan 04 Plateu Jk" 10320 UL Pkxna 31/37 Tetiidd. Wet rdlA

00-677WamawaCYPRUS IRELAND TURKEYCnioref Applied Rard GovemmentSuppiageAps-y Fera xodfh .vxe: neonCyprusCollage 4-5HarcourtRoad IPSJoumtaiz NarLabakhSoLNo.15tr Dlogen Sket Engeal Dublin 2 UL OCrena 3 CagaegluP.O. Box 206 02-916 Werawa IstnbulNloia ISRAEL

Yoamot Uterature Ltd. PORTUGAL UNITED KINGDOMDENMARK P.O. Oo 56055 UvLis Portugal Micoinfe LtLShLiudAklttura TelAviv IS6 Rua DoCArme 70-74 PO BoxB 3Roznoamns A 11 1200 LAbon Alton, Hampahira U34 2PGDK-1970 FredzriltbcgC ITALY gland

Llcoaa Com mnuionaie Semet SPA SAUDI AIRABIA QATARDOMINICAN REPUBl9C Vie Duce Di Calabria, 1/ Jat&ir Book Stt VNEZUULAEditorm Taller C par A. Casub Posawl 552 P.O Box 3196 Lbrria del EeR_wurscitn a Isael I Caktic 309 50125 Flrenux Riyadh 11471 Aptdo. 60D37Apertado de Carem 2190 Z-1 Caaca 1040-ASanD Demingo JAPAN SINCAPORE TAIWAK

Easten look Service MYANMARLIRUNNIEGYPT ARABREPUWBLCO0 Hotgo 3-0-tome, Banku 113 Inormation PubliadeAl Ahram Tokyo Private. Ltd.Al Celas Stred Colden Wheel BuildingCair 41. Kallang Pudding, 00A-03

Singpore 1334

f'dh a

--""U

-~~~~~~~~ I *S.

I, . ,'

9 .. ,

_~~~~~~~~~~ 3 9

|_~~~~~~~~ L

*! 6 3; .5

0~~~~~~ 5

A"lSXaS 1* . *1| *

; N 1911 eIIt i

__~~~~~~~~~~~~~~~~~~~~~~ S

_~~~~~ - *U m a

_~~~~~~~~~~~~I I t

_~~~~ .A - - -e ' *

_ -

! ~ ~~~~~ 0 0 6 I LM! i!.1'i.!!Si!, E m 5*

| ] ; ~~~~~~~~~~~~~~~~~~12615 DEV 100~0-8213-2615-5

GOVERNMENT POLICY & GROW

_Di~~~~~~~~~~~~~~~~~~~~~~~~~~~40 0 0