Global Economic Prospects 2000

GlobalEconomicProspectsand the Developing Countries

2000

ii

G L O B A L E C O N O M I C P R O S P E C T S

Copyright © 2000 by the International Bankfor Reconstruction and Development/The World Bank1818 H Street, NW, Washington, DC 20433, USA

All rights reservedManufactured in the United States of AmericaFirst printing December 1999

This publication has been compiled by the staff of the Development Prospects Group of theWorld Bank’s Development Economics Vice Presidency. The World Bank does not acceptresponsibility for the accuracy or completeness of this publication. Any judgments expressedare those of World Bank staff or consultants and do not necessarily reflect the views of theBoard of Executive Directors or the governments they represent.

The material in this publication is copyrighted. The World Bank encourages dissemination ofits work and will normally grant permission promptly.

Permission to photocopy items for internal or personal use, for the internal or personal useof specific clients, or for educational classroom use is granted by the World Bank, providedthat the appropriate fee is paid directly to the Copyright Clearance Center, Inc., 222 Rose-wood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470. Pleasecontact the Copyright Clearance Center before photocopying items.

For permission to reprint individual articles or chapters, please fax your request with com-plete information to the Republication Department, Copyright Clearance Center, fax 978-750-4470.

All other queries on rights and licenses should be addressed to the Office of the Publisher,World Bank at the address above or faxed to 202-522-2422.

ISBN 0-8213-4550-8ISSN 1014-8906Library of Congress catalog card number: 91-6-440001 (serial)

iii

Contents

Foreword vii

Summary ix

Abbreviations, Acronyms, and Data Notes xv

Chapter 1 Prospects for Growth and Poverty Reduction in Developing Countries 1The external environment for developing countries is improving 3The outlook for developing countries in 1999–2001 suggests

significant acceleration 17Projections for growth in developing countries in the long term are lower 22Recent trends and prospects for poverty in developing countries 28Risks to the forecast and a low-case scenario 36Notes 42References 44

Chapter 2 External Shocks, Financial Crises, and Poverty in Developing Countries 47External shocks and poverty in developing countries 48Income poverty and inequality during the East Asian crisis 51Beyond current income effects of the East Asian crisis 62Fostering sustained growth and reducing the social costs of volatility and crises 65Notes 67References 68

Chapter 3 Asian Restructuring: From Cyclical Recovery to Sustainable Growth 73The uneven recovery 74The focal point of restructuring: the financial sector 84Corporate restructuring: some progress, but a long way to go 91Notes 98References 99

Chapter 4 Managing the Recent Commodity Price Cycle 103Key issues confronting primary commodity exporters 104The savings response to commodity price cycles 108

iv


The savings response by oil exporting countries to the recent swing in oil prices 110The savings response by non-oil commodity exporting countries of

Sub-Saharan Africa 119Notes 127References 130

Appendix 1 Regional Economic Prospects 133

Appendix 2 Global Economic Indicators 151

Classification of Economies 169

Figures1.1 World industrial production 41.2 U.S private savings rate and current account balance 61.3 G-3 consumer prices, 1997–99 81.4 Short-term interest rates, 1998–99 91.5 World import volume growth 101.6 Spreads on Brady bonds 151.7 Real GDP growth for developing countries 171.8 Index of financial conditions since June 1997 181.9 Growth of U.S. employment and productivity in manufacturing 221.10 Long-term trends in commodity prices, 1990–2008 252.1 Terms of trade and GDP growth volatility, 1961–97 492.2 Unemployment in East Asia during the crisis 592.3 Poverty and spending on social safety nets in East Asia, 1996–98 602.4 Change in gross domestic savings, 1997–98 622.5 Relative price of foods during the crisis 633.1 Nontradable production before and after crises 753.2 Industrial production before and after crises 763.3 Production index in Korea by industry 773.4 Thai nonperforming loans and interest rates, June 1998–September 1999 793.5 Change in bank deposits and lending to the private sector,

December 1998–July 1999 813.6 Total investment in East Asia 813.7 Cross-border mergers and acquisitions, Q1 1997–Q2 1999 954.1 Primary commodity prices versus manufactures unit value index, 1960–98 1054.2 The decline in commodity prices due to the East Asian crisis 1084.3 Current and previous price declines of nonenergy commodities 1094.4 Mexico’s current account, 1970–82 1154.5 Shares of merchandise exports of Sub-Saharan Africa non-oil commodity

exporters, 1990–97 1194.6 Agricultural exporters’ trade prices, 1993–98 1204.7 Metals and minerals exporters’ trade prices, 1993–98 1214.8 Real income, savings, and aid during booms 1254.9 Change in savings rates during boom compared with base periods 126

v

Tables1.1 Global conditions affecting growth in developing countries, 1998–2001 21.2 World output growth, 1998–2001 51.3 Contributions to world import growth 101.4 U.S. import prices of manufactured goods, 1999 111.5 Annual percentage change in oil and non-oil commodity prices 141.6 Average monthly gross capital market flows to developing countries 141.7 World growth, 1981–2008 231.8a Population living below $1 per day in developing and transition

economies, 1987–98 291.8b Population living below $2 per day in developing and transition

economies, 1987–98 291.9 Projected growth rates in real per capita private consumption and changes in Gini

coefficients for 1998–2008 301.10a Population living below $1 per day in developing and transition economies for

1998–2008 under scenarios of slow growth and rising inequality(Scenario A) and inclusive growth (Scenario B) 33

1.10b Population living below $2 per day in developing and transition economies for1998–2008 under scenarios of slow growth and rising inequality(Scenario A) and inclusive growth (Scenario B) 33

1.11 Population estimates and projections, 1998 and 2008 351.12 World, industrial, and developing countries in the low-case scenario 372.1 Growth, poverty rates, and Gini coefficients in East Asia, 1996–98 532.2 Employment and real wages in East Asia during the crisis 572.3 Public spending on health and education 643.1 Financial distress, past and projected, 1995–2002 773.2 Ration of nonperforming loans to total loans, December 1998-September 1999 783.3 Public debt and recapitalization costs as share of GDP, 1998 843.4 Institutional arrangements for corporate and financial restructuring 853.5 Structural changes in the financial system 853.6 Estimated recapitalization costs for commercial banks, mid-October 1999 873.7 Restructuring: out-of-court and in-court progress, August 1999 913.8 Illustrative postcrisis policy reforms in crisis countries 963.9 FDI flows in East Asia, 1992–99 974.1 Savings, investment, and real income changes, selected country groups, 1996–97 1114.2 Capital flows to oil exporters and energy prices, 1970–97 1144.3 Ratios of public and private savings to GDP, 1996–98 1154.4 Economic performance of major oil exporters and other countries, 1980–97 1174.5 Share of oil and non-oil commodities in merchandise exports, 1980 and

most recent year 1184.6 Policy performance and GDP, savings, and real income during boom periods 1224.7 Exports and terms-of-trade changes, boom compared to base period 1234.8 Changes in savings and real income relative to base periods 1234.9 Economic performance 1264.10 Decomposition of real income changes 127

C O N T E N T S

vi


Boxes1.1 Sectoral and regional effects of the East Asian crisis 12–131.2 Prospects for a new round of multilateral trade negotiations 26–271.3 Technical discussion of assumptions 341.4 Can the international development target for reducing income

poverty be achieved? 371.5 Failing to forecast the severity of the East Asian crisis 38–391.6 The possible impact of the Y2K bug on developing countries 40–412.1 Volatility, growth, and poverty 512.2 External shocks and fluctuations in poverty in Mexico 52–533.1 Why distress can persist 803.2 Redeployment of assets: lessons from Japan 92–934.1 Counterfactual scenarios 1124.2 Public sector expenditures during the oil price boom 1164.3 Fiscal adjustment in Saudi Arabia and Venezuela 1174.4 Savings and real incomes during the commodity price cycle 1244.5 Real incomes in Benin during the commodity price cycle 127

vii

Foreword

Developing countries are now recovering from the worst ravages of the financial crisis of1997–98. But the recovery is uneven and fragile. Growth remains well below the precrisistrends in many countries, so much so that the average per capita income of developing

countries outside of Asia is expected to fall in 1999. Long-term projections for growth in devel-oping countries (excluding the transition economies) suggest that it is likely to be lower in2002–2008 than in the precrisis 1990s. The experience of the past year has underscored howfinancial volatility can increase poverty significantly in the short to medium term. As a result,there is a growing consensus that in order to maximize the positive effects of growth that cancome with openness, the international community must find ways to reduce the frequency andseverity of economic crises.

For the first time, this year’s Global Economic Prospects analyzes the trends in povertylevels in developing countries. Progress on poverty reduction in many developing countries islikely to remain slow and below poverty-reduction targets recently adopted by the internationalcommunity.

This year’s report also considers three areas where the crisis has had a major impact ongrowth and welfare in the developing world.

First, the crisis has increased poverty in the East Asian crisis countries, Brazil, and Russia.Not only has the increase in poverty been significant, whether measured by levels of income orconsumption, but the crisis has engendered large costly movements of populations and sharpdeclines in standards of living for the middle classes. Urban poverty increased in all countries,particularly the Republic of Korea. Although efforts were made to maintain spending on socialservices, real public expenditures on health and education fell in the crisis countries, with aparticularly severe impact on access to services in Indonesia.

Second, though the East Asian crisis countries are experiencing a cyclical recovery, severestructural problems remain. The level of nonperforming loans remains high, and a large shareof firms are insolvent. Weak firms have operated on thin margins and their inability to payinterest, following the onset of the crisis, has added to their debt burden. Such firms constitutea significant portion of each of the crisis economies and the appetite to invest in them isextremely limited. They will continue to act as a drag on investment and growth until suchtime as the financial claims on them are resolved and either their operations return to ad-equate profitability or their assets are redeployed. Without vigorous corporate and financialrestructuring, the return to a sustainable growth path will likely take longer, the costs of thecrisis could rise, and these economies will remain vulnerable to new external and internalshocks.

Finally, the exchange rate depreciations and declines in demand in East Asia have exacer-bated the fall in primary commodity prices that began in 1996. Countries that depend on pri-mary commodities have faced an enormous challenge in smoothing consumption in the face ofbooms and busts in commodity prices during the 1990s. Some commodity-dependent countries

viii


cope with the wide swings in commodity prices more successfully than others. In the oil export-ing countries, weak policy environments led to mixed savings performance and lower invest-ment over the oil price cycle. These countries have generally been unsuccessful in reducing theirdependence on oil revenues, and the fall in investment will further impede progress. By con-trast, the commodity price cycle of the 1990s does not appear to have adversely affected theprospects for growth in the non-oil exporting countries of Sub-Saharan Africa. Changes in realincomes were generally smaller than in the oil exporting countries, and improvements in poli-cies in several countries enabled them to increase savings and investment rates during bothcommodity-price booms and busts.

Overall, then, Global Economic Prospects 2000 shows a mixed picture, with a number ofextraordinary challenges confronting developing countries in their efforts to further economicprogress and reduce poverty. We hope that this report will serve both to sharpen the WorldBank’s work in supporting our clients, and to inform the international community about thecritical development issues of the day.

Joseph E. StiglitzSenior Vice President and

Chief EconomistThe World Bank

November 1999

This report was prepared by the Development Prospects Group, and drew from resourcesthroughout the Development Economics Vice Presidency, the Poverty Reduction Board, theEast Asia Regional Vice Presidency, and other World Bank regions. The principal author of thereport was Mustapha Nabli, assisted by William Shaw, with direction by Uri Dadush. Thechapter authors were Mick Riordan (chapter 1), Mustapha Nabli (chapter 2), Ashoka Mody(chapter 3), and William Shaw (chapter 4). The report was prepared under the general directionof Joseph Stiglitz.

The report drew on inputs by other staff of the Development Economics Vice Presidencyand from throughout the Bank. Caroline Farah, Himmat Kalsi, Robert Keyfitz, Annette I. DeKleine, Robert Lynn, Dominique van der Mensbrugghe, Dilip Ratha, and Bert Wolfe contrib-uted to the analysis of global economic trends and prospects in chapter 1. Giovanna Prennushi,Shaohua Chen, Martin Ravallion, and Michael Walton wrote the section on poverty. CarolGabyzon and Alan Winters contributed to chapter 2, and Peter Fallon and Jeni Klugman pro-vided background papers. Eung Ju Kim, Shoko Negishi, and Vivak Suri contributed to chapter3. And Ibrahim Al-Ghelaiqah, Gholam Azarbayejani, John Baffes, Betty Dow, Shamya S.Jayasuriya, Robert Keyfitz, Donald Mitchell, and Shane Streifel contributed to chapter 4.

Many others from inside and outside the Bank provided comments, guidance, inputs andsupport at various stages of the report’s publication. Paul Collier, Nora Lustig, Frank Lysy, andJohn Page served as discussants at the Bankwide review. We would particularly like to thankSara Calvo, Gerard Caprio, Constantijn Claessens, David Dollar, Alan Gelb, Arvind Gupta,James Hanson, Homi Kharas, Daniela Klingebiel, Ira Lieberman, Panayotis Varangis, and Joachimvon Amsberg for their helpful comments. The Development Data Group contributed to theappendix. Lawrence MacDonald served as the External Affairs task manager, Robert Kingmanaged dissemination from the Development Prospects Group, and Phil Hay managed mediaarrangements. Sydnella Kpundeh served as the principal assistant to the team. Book design,editing, and production were directed and managed by the Production Services Unit of theWorld Bank’s Office of the Publisher.

ix

DEVELOPING COUNTRIES ARE NOW RECOVER-ing from the worst ravages of the fi-nancial crisis of 1997–98. The East

Asian economies are rebounding from lastyear’s collapse in output. Improved prospectsand an easing of monetary conditions in manyparts of the developing world have boostedemerging equity markets and reduced interestrates from the sky-high levels of mid-1998.Developing countries also are benefiting fromthe acceleration of growth and interest ratereductions in industrial countries.

However, the recovery is both uneven andfragile, and many countries continue tostruggle in the aftermath of the crisis. Severalcountries in Africa, Latin America, and East-ern Europe face declines in output in 1999,and outside of Asia developing countries’ percapita income is expected to fall. Continuedimbalances in industrial countries markedlyincrease the risks presented by the internationaleconomic environment. Furthermore, the cy-clical recovery in East Asia has not addressedsevere difficulties that were either caused orexacerbated by the crisis. In addition to a re-view of international economic developmentsand prospects, Global Economic Prospects2000 considers three areas where the crisis hashad a major impact on growth and welfare inthe developing world.

First, the crisis has increased poverty inthe East Asian crisis countries, Brazil, and theRussian Federation, and elsewhere. Chapter2 reviews the evidence on the crisis’ socialimpact on East Asia and other developingcountries and addresses the broader issue ofthe impact of external shocks on poverty indeveloping countries.

Summary

Second, though the East Asian crisis coun-tries are experiencing a strong cyclical recov-ery, severe structural problems remain, notablythe banking systems’ high levels ofnonperforming loans and the large share ofinsolvent firms. Chapter 3 outlines the depthof the problems faced by the corporate andfinancial sectors of these economies, analyzesthe challenges facing the restructuring process,and discusses the appropriate role of govern-ment in supporting restructuring and reduc-ing systemic risk.

Third, exchange rate depreciations anddeclines in demand in East Asia exacerbatedthe fall in primary commodity prices that be-gan in 1996. Countries that depend on pri-mary commodities have faced an enormouschallenge in smoothing consumption in theface of booms and busts in commodity pricesand adjusting to the secular decline incommodity prices relative to manufactures.Chapter 4 examines how the most commod-ity-dependent economies in the world—themajor oil exporting countries and the non-oilexporters of Sub-Saharan Africa—have ad-justed to the commodity price cycle.

Prospects for growth and povertyreduction in developing countries

The effects of the crises of 1997–99, fromEast Asia to Russia and Brazil, persist in

many aspects. In most developing countriesgrowth remains weak and well below theprecrisis trends. Social dislocations are severeand progress in poverty reduction has stalled.At the same time, recent developments in theglobal economy have been largely encourag-

x


ing, with signs of strong initial recovery in theEast Asian crisis economies and continued ex-pansion in the industrial countries leading toa bottoming-out of world industrial produc-tion and trade.

Recent events have confirmed the impor-tance of the factors identified in Global De-velopment Finance (March 1999) as shapingthe global recovery, notably the easing ofmacroeconomic policies in industrial countries,early signs of recovery in the East Asian crisiscountries, and easier financial conditions indeveloping countries. But the magnitude ofthese effects has been much larger than an-ticipated, and recent evidence has yielded somesurprising developments: adjustment in someof the worst-hit countries, such as Russia andBrazil, has been much more favorable thanexpected in March, and a sharp increase inoil prices, following the decision of the Orga-nization of Petroleum Exporting Countries(OPEC) in April 1999 to curtail oil supplies,has benefited developing countries that dependheavily on oil exports.

The positive evidence has been strongenough to support an upward revision of theMarch projections for growth. Growth for theG-7 countries this year is likely to register 2.6percent, 0.9 percentage points higher than theforecast made six months ago. Continuedstrong growth in the United States is the prin-cipal factor in the revision, but Japan’s per-formance in the first half of 1999 (3.2 percentannualized GDP growth), which was muchbetter than anticipated, also contributes to thechange. Europe, which had been hampered byinventory overhang, is now showing signs ofa strong revival. Reflecting these develop-ments, world industrial production appears tobe on an accelerating path. For developingcountries, GDP growth for 1999 is expectedto be 2.7 percent—a revision of 1.2 percent-age points from the March forecasts—and theoutlook for 2000 has been upgraded by 0.5percentage points.

Positive as these revisions are, they maskthe considerable fragility of developing coun-tries, which have yet to recover fully from the

financial crises of 1997–99, nor do they re-flect the markedly different patterns of growthand recovery among regions. Except for EastAsia and South Asia (regions bolstered bygrowth in China and India), aggregate real percapita incomes in 1999 are expected to de-cline or stagnate in several developing regions.Further, the news since March has not beenall good. The tightening of oil supply hasmeant higher import bills for many develop-ing and industrial countries. And the favor-able financial conditions have not madeinternational investors less risk-averse, asshown by the high levels of interest ratespreads. International capital flows to devel-oping countries have fallen much more sharplythan anticipated.

Although improving, the external environ-ment for developing countries remains sub-ject to a high degree of uncertainty.

The underpinnings of growth, especiallyin the developing countries, remain fragile.Capital flows to emerging markets continueto be scarce and expensive. In such an envi-ronment, the prospective unwinding of largeimbalances in the industrial countries presentspotential risks for these projections. Chiefamong these risks are the consumption boom(driven by the stock market) and wideningexternal deficit in the United States, and theuncertain outlook for Japan.

One potential scenario assumes a tight-ening of monetary policy in the United States(in response to signs of increased inflation),which sharply reduces equity prices, resultingin slow growth in the United States and Eu-rope and a relapse into recession in Japan. Fordeveloping countries, effects are transmittedthrough a further slowing in export marketgrowth, declines in oil and non-oil commod-ity prices because of deteriorating demandconditions, and increased risk aversion in fi-nancial markets. Although policy responses tothese external circumstances would varywidely across developing countries dependingon current conditions, most countries wouldbe obliged to adjust through a compressionof domestic demand and imports. An assumed

xi

closure of the financing gap on the demandside (almost $100 billion) results in a loss of 2percentage points of growth for developingcountries as a group in both 2000 and 2001,implying a loss of nominal GDP of some $260billion.

Long-term projections for growth in de-veloping countries have been downgraded bysome 0.3 percentage points, suggesting thatgrowth for the group (excluding the transi-tion economies) in 2002–2008 is likely to belower than in the precrisis 1990s.

This estimate reflects several factors, in-cluding a somewhat less favorable external en-vironment and, importantly, prospects for aprotracted work-out of structural weaknessesin developing countries—particularly in finan-cial systems and fiscal positions—which havebecome more apparent in the wake of the cri-sis. One implication of lower long-term pro-jections is that progress on poverty reductionwill be slower. For some regions, includingSub-Saharan Africa and Latin America and theCaribbean, reductions in poverty are likely toremain below the targets recently adopted bythe international community. Effective policyactions to encourage rapid and equitablegrowth are essential to reduce poverty.

External shocks, financial crises,and poverty in developingcountries

The financial crisis has underlined how glo-balization, especially financial integration,

exposes developing countries to externalshocks.

External shocks can reduce the gains inpoverty reduction from openness and increasepoverty significantly in the short to mediumterm. This fact underscores the importance ofaddressing the issue of volatility in order tomaximize the positive effects of growth onpoverty reduction. The countries most affectedby the East Asian crisis illustrate the asym-metric impact of changes in per capita incomeon poverty and the negative effects of volatil-ity on growth.

Any development strategy for stable andsustainable growth must include both adequatesafety nets and appropriate policies and insti-tutions designed to prevent financial crises, andto respond when crises occur. Prospects forpoverty reduction depend not only on futuregrowth but also on countries’ capacity to man-age volatility and reduce growth fluctuations.

Though less dramatic than early predic-tions suggested and very heterogeneous, thenegative social impact of the East Asian crisisand consequent crises in Russia and Brazil hasbeen enormous.

The increase in income or consumptionpoverty has been significant. In addition, thecrisis has engendered costly, large reallocationsof people and sharp declines in middle-classstandards of living. Unlike the situation inLatin America, where income inequality in-creased significantly during crises, in East Asiathe effects on income distribution have beensmall and highly differentiated. The extent ofthese effects depends on the country’s incomelevel and the impact of the crisis on differenteconomic sectors.

Urban poverty increased in all countries,particularly Korea, where total employmentdeclined and open unemployment grew morethan in other countries in the region. Fallingreal wages in the urban formal sector affectedmostly high-income groups. In Thailand theimpact was felt mostly in rural areas becauseof the large inflows of workers from urbanareas and the relatively small increases in ag-ricultural prices.

The severity of the crisis in Indonesia isreflected in the strong responses of householdsto increase consumption as a share of income,adjust their asset holdings, and increase theshare of staple foods in their consumptionbaskets. In Korea and Malaysia the responseof households was to increase the savings rate.The composition of consumption expenditureschanged significantly. Households spent more,primarily on essential items such as food, fuel,housing, health, and education.

The crisis demonstrated the flexibility oflabor markets in developing countries. These

S U M M A R Y

xii


markets help absorb the effects of shocksthrough reduced wages and labor mobilitywithin and between urban and rural areas.

Wages fell sharply during the East Asiancrisis, with particularly spectacular declines inIndonesia. Wage declines moderated the im-pact of the recession on employment. Thus thedecline in total employment in Thailand andMalaysia was limited, and employment actu-ally rose in Indonesia. Labor was reallocatedfrom the formal (urban) sector to other ac-tivities, particularly the informal sector andagriculture, where exchange rate depreciationsimproved incentives.

Real public expenditures on education andhealth fell in the crisis countries, although ef-forts were made to increase spending on safetynets.

The extent to which households were ableto adjust their spending to offset this declinevaried across countries as well as incomegroups. In Thailand, families and governmentprograms acted to cushion the impact of thecrisis in order to avoid declines in school en-rollment rates or in access to health services.In Indonesia, however, the severity of the cri-sis led to significant declines in poor house-holds’ access to both education and healthservices, particularly in urban areas. Such set-backs can have irreversible effects on humandevelopment.

Even where public spending on safety netsincreased significantly, the impact on povertywas limited for several reasons. These includedthe absence of safety nets before the crisis, re-sponse lags, institutional problems, and lowlevels of spending relative to the scale of pov-erty. In some cases, evidence suggests that well-functioning programs were underfundedrelative to the potential impact of shocks onpoverty.

Asian restructuring: from recoveryto sustainable growth

The aftereffects of the externally triggeredliquidity crisis in Indonesia, Korea, Ma-

laysia, and Thailand indiscriminately sub-

merged both strong and weak producersand financiers. The rising tide is lifting thestrong, but the financially weak continue tostruggle because of both crisis-induced andlongstanding vulnerabilities.

Since the onset of the East Asian crisismore than two years ago, the corporate sec-tors and financial systems in the crisis econo-mies have remained in severe distress. Thebanking systems’ nonperforming loans haveskyrocketed to unprecedented levels:nonperforming loans range between approxi-mately 30 percent of GDP for Korea andMalaysia to 60 percent of GDP for Thailand.In contrast, nonperforming loans in othermajor emerging market crises (Chile in theearly 1980s and Mexico in 1995) were lessthan 20 percent of GDP. In the Scandinavianbanking crises during the early 1990s,nonperforming loans amounted to approxi-mately 5 percent of GDP.

East Asia’s heavy reliance on bank-basedfinancial systems and the high debt-equityratios of corporations have made the economicdistress especially acute. Weak firms in EastAsia operated on thin margins in the yearsleading up to the crisis, and their inability topay interest following the onset of the crisishas added to their debt burden. Such firmsconstitute a significant portion of the corpo-rate sector in each of the crisis economies, andthe appetite to invest in them is extremely lim-ited. They will continue to act as a drag oninvestment and growth until the financialclaims on them are resolved, and either theiroperations return to adequate profitability ortheir assets are redeployed.

Without vigorous corporate and financialrestructuring, the return to sustainable growthwill likely take longer, the fiscal costs of thecrisis could rise, and the economies will re-main vulnerable to new external and internalshocks.

Recognizing the urgency, East Asian gov-ernments were quick to create an institutionalstructure for corporate and financial restruc-turing; they also earmarked funds for bankrecapitalization. The political momentum for

xiii

reform has, however, slowed down, in partbecause the deeper structural problems nowneed to be addressed. Experience from othereconomies, including Japan’s, shows that aslackening of the reform effort can undoprogress.

Government restructuring policies need tobe guided by two principal considerations: lim-iting the likelihood of systemic disruptionwhile also containing fiscal costs; and clarify-ing financial claims and building an environ-ment conducive to asset reallocation. Basedon these two principles, fiscal costs come prin-cipally from the government’s social contractto protect bank depositors and to prevent sys-temic failure. Government funds are not re-quired for corporate restructuring.

Bank restructuring is important becauseit contributes to both policy objectives. Expe-ditiously restoring the health of the bankingsystem is required, but the process of restruc-turing itself can be disruptive.

Restructuring is necessary because apoorly capitalized banking sector creates con-tinued systemic risks and growing fiscal liabili-ties for governments. Healthy banks are alsobest positioned to enforce claims and to pur-sue corporate restructuring. But restructuringshould be undertaken in a manner that en-sures the integrity and the organizational capi-tal of the financial system so that prudentlending to businesses and households maycontinue. Achieving this objective requiresmaking difficult choices. Having providedimplicit or explicit guarantees, governmentscan either move ahead rapidly by taking fis-cal responsibility for the costs of the crisis, orthey can encourage private resolution of thedistress while applying regulatory forbear-ance. Waiting to resolve problems is likely tomake them worse. However, expeditious andtransparent action should be accompanied bymarket-based measures to recoup fiscal costsand to signal a credible commitment to se-verely restrict guarantees and bailouts in thefuture.

Corporate restructuring needs to deal firstwith the delineation and allocation of losses.

Improvements in accounting standardsand bankruptcy regimes can help support thisprocess. However, in the absence of effectivebankruptcy procedures, out-of-court proce-dures offer a mechanism for resolution. Oncefinancial claims are resolved, corporate restruc-turing can be expected to occur through natu-ral market forces, except where a majorimpediment prevents such forces from work-ing. Governments can facilitate asset mobil-ity by creating the framework for effectivedomestic and cross-border mergers and acqui-sitions. The Japanese experience cautions that,without an adequate infrastructure for resolv-ing claims and for fostering asset mobility, fun-damental corporate restructuring can beinterminably deferred at a high economic costeven in a sophisticated economy.

Managing the recent commodityprice cycle

Primary commodity prices have undergonea pronounced cycle since the mid-1990s,

driven by both temporary and secular factors.Primary commodity prices continue to be

more volatile than the prices of manufactures.Energy prices have been especially volatile.Crude oil prices rose 74 percent from early1994 through the end of 1996, then fell 56percent by the end of 1998, and in 1999 re-covered nearly the entire decline of the previ-ous two years. Average non-oil commodityprices rose by 46 percent from the monthlylow in mid-1993 to mid-1997, and thendropped 30 percent by late 1999. The cycle inprimary commodity prices was driven bychanges in global demand, weather-relatedsupply shocks, supply responses to the highprices of the early 1990s, technological inno-vations that have reduced production costs,and exchange rate depreciations among largecommodity exporters linked to the East Asiancrisis.

Such volatility poses real challenges todeveloping countries that depend on primarycommodities for a substantial share of theirexport revenues. Countries where consump-

S U M M A R Y

xiv


tion rises with real incomes during commod-ity price booms may face either painful reduc-tions in consumption or declines in investmentthat reduce long-term growth when prices fall.Countries’ savings and investment behaviordiffered markedly over the recent commodityprice cycle; these differences primarily reflectedthe quality of policies rather than shifts in theterms of trade.

In the oil exporting countries weak policyenvironments led to mixed savings perfor-mance and lower investment over the oil pricecycle. On average, countries allocated to in-creased consumption about half of the aver-age 5 percent of GDP improvement in realincomes during the upswing in oil prices(1996–97). During the 1998 drop in oil prices,however, consumption did not decline, im-plying that savings fell by the full amount ofthe decline in real incomes. Countries’ perfor-mances varied greatly, depending on their spe-cific political and economic circumstances.

Oil exporting countries’ investment fellrelative to output over the commodity pricecycle. The decline in investment was actuallygreater than the decline in domestic savings,so the current account deficit fell. The majoroil exporting countries have generally failedto reduce their dependence on oil revenues,and the fall in investment will further impedeprogress. At the same time, several of these

countries face high levels of unemployment,continued slow growth, and rapidly expand-ing populations. They need to strengthen theirpolicies to encourage greater private sector(and non-oil) activities and to improve theinstitutional environment.

The commodity price cycle of the 1990sdoes not appear to have adversely affected theprospects for growth in the non-oil exportingcountries of Sub-Saharan Africa. Changes inreal incomes were generally smaller than inthe oil exporting countries because the priceof their commodity exports changed by lessthan the price of oil, and the losses from de-clining export prices were partially offset bygains from lower import prices, particularlyenergy prices. More important, however, im-provements in policies in several countriesenabled them to increase savings and invest-ment rates during both commodity pricebooms and busts. Many countries cut theirfiscal deficits in an effort to rein in the growthof debt and to reduce inflation, while privatesavings rose in response to improved policiesthat increased the return to investment, par-ticularly in export sectors. Countries withbetter policies, as measured by the World Bank,achieved larger increases in savings and highergrowth of GDP than countries with worsepolicies, despite smaller increases in real in-comes in the former group.

xv

Abbreviations, Acronyms,and Data Notes

ASEAN Association of Southeast Asian Nations

ASEAN-4 Indonesia, Malaysia, Philippines, and Thailand

BIS Bank for International Settlements

CFA Communauté Financière Africaine

CIS Commonwealth of Independent States

CPI Consumer price index

East Asia-5 Indonesia, Malaysia, the Philippines, the Republic of Korea, and Thailand

ECB European Central Bank

ECLAC United Nations: Economic Commission for Latin America and theCaribbean

EMU European Monetary Union

ERM Exchange rate mechanism

EU European Union (formerly the EC)

EU-4 France, Germany, Italy, and the United Kingdom

EU-12 Belgium, Denmark, France, Germany, Greece, Ireland, Italy, Luxembourg,the Netherlands, Portugal, Spain, and the United Kingdom

EU-15 Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland,Italy, Luxembourg, the Netherlands, Portugal, Spain, Sweden, and theUnited Kingdom

FDI Foreign direct investment

G-3 Germany, Japan, and the United States

G-5 France, Germany, Japan, the United Kingdom, and the United States

G-7 Canada, France, Germany, Italy, Japan, the United Kingdom, and theUnited States

G-8 Canada, France, Germany, Italy, Japan, Russia, the United Kingdom, andthe United States

G-10 Belgium, Canada, France, Germany, Italy, Japan, the Netherlands, Sweden,the United Kingdom, and the United States (and sometimes Switzerland isinvolved)

G-22 Argentina, Australia, Brazil, Canada, China, France, Germany, Hong Kong(China), India, Indonesia, Italy, Japan, Korea, Malaysia, Mexico, Poland,Russia, Singapore, South Africa, Thailand, the United Kingdom, and theUnited States

xvi


GCC Gulf Cooperation Council

GDP Gross domestic product

GTAP Global Trade Analysis Project

HIPC Heavily indebted poor countries

ILO International Labour Organisation

IMF International Monetary Fund

LIBOR London interbank offered rate

M2 A measure of broad money supply in the United States

Mercosur Latin America Southern Cone trade bloc (Argentina, Brazil, Paraguay,and Uruguay)

MUV Manufactures unit value index

NIE Newly industrializing economy

ODA Official development assistance

OECD Organisation for Economic Co-operation and Development

OPEC Organization of Petroleum Exporting Countries

UNCTAD United Nations Conference on Trade and Development

Data notesThe following norms are used through-

out.• Billion is 1,000 million.• All dollar figures are U.S. dollars.• In general, data for periods through 1997

are actual, data for 1998 are estimated,and data for 1999 onward are projected.

The “classification of economies” tables at theend of this volume classify economies by in-come, region, export category, and indebted-ness. Unless otherwise indicated, the term“developing countries” as used in this volumecovers all low- and middle-income countries,including the transition economies.

1

1

The effects of the crises of 1997–99, fromEast Asia to Russia and Brazil, persistin many aspects. In most developing

countries growth remains weak and well be-low precrisis trends. Social dislocations aresevere, and have increased not only in Asiabut also in other affected countries. Progressin poverty reduction has stalled in the devel-oping world at the end of the 1990s, and thenumber of poor is rising in most regions. Atthe same time, recent developments in the glo-bal economy have been encouraging, withsigns of strong initial recovery in the East Asiancrisis economies and continued expansion inthe industrial countries leading to a bottom-ing-out of world industrial production andtrade.

For countries at the epicenter of the crisisin East Asia, corporate restructuring and a res-toration of sound financial institutions willneed to be addressed in an aggressive mannerfor growth to broaden and attain sustain-ability. As well, many commodity and oil-de-pendent countries, especially in Sub-SaharanAfrica, have been affected by extreme volatil-ity in commodity prices and more recent weak-ness in prices of tropical products. This reportreviews the evidence of the social impact ofthe recent crisis, and more broadly of exter-nal shocks, with a particular focus on the poorin chapter 2. The intertwined issues of corpo-rate and financial restructuring and their sig-nificance for growth recovery and itssustainability in East Asia are assessed in chap-

ter 3. Chapter 4 explores how two of thegroups of countries most affected by recentboom-bust cycles in commodity prices, themajor oil exporting and Sub-Saharan coun-tries, have adjusted and assesses the implica-tions for their growth prospects.

In forecasts prepared in March, thegrowth rate for developing countries was es-timated at 1.5 percent for 1999—which wouldhave marked the weakest performance forthese countries since the debt crisis of the early1980s.1 But a gradual build-up in the momen-tum of developing country growth through2001 was anticipated to be driven by severalfactors: the effects of policy measures under-taken in major industrial countries (includinginterest rate reductions in the United Statesand Europe, and large fiscal stimulus measuresin Japan); early signs of recovery in the crisiscountries of East Asia, especially apparent inthe Republic of Korea, Malaysia, and Thai-land; and more broadly, an easing of finan-cial conditions in developing countries,evidenced in widespread declines in domesticinterest rates and rising equity markets. Pri-vate capital flows to emerging markets wereanticipated to fall well below levels of 1998,but to show a gradual improvement over thecourse of the year.

The importance of the factors identifiedin the March forecasts in shaping the globalrecovery has generally been confirmed—butthe magnitude of these effects have been muchlarger than anticipated, especially with respect

Prospects for Growth andPoverty Reduction inDeveloping Countries

2


to the strength of Asian recovery. Addition-ally, evidence from the second and third quar-ters of 1999 has revealed two surprisingdevelopments. First, the adjustment in someof the worst-hit countries, including Russia andBrazil, has been much more favorable than ex-pected. Second, a sharp increase in oil pricesfollowed the decision of the Organization ofPetroleum Exporting Countries (OPEC) inApril 1999 to curtail oil supplies, benefitingdeveloping countries that depend heavily onoil exports.

The positive evidence has been strongenough to support a substantial upward revi-sion of the March projections for growth, not

only for developing countries but also for theworld economy. Developing country growthis expected to be significantly higher for 1999and moderately higher for 2000. The growthrate for 2001 is expected to remain at aboutthe March level (table 1.1). Global growth isnow expected to accelerate to 2.6 percent in1999 and 2.9 percent in 2000, marking up-ward revisions of 0.8 and 0.5 percentage pointsrespectively from earlier projections. For de-veloping countries, GDP growth for 1999 isexpected to be 2.7 percent—a revision of 1.2percentage points from the March forecasts—and the outlook for 2000 has been upgradedby 0.5 percentage points.

Table 1.1 Global conditions affecting growth in developing countries, 1998–2001(percentage change from previous year, except LIBOR)

Global DevelopmentCurrent Finance 1999

Estimate Forecasts Forecasts1998 1999 2000 2001 1999 2000 2001

World GDP growth 1.9 2.6 2.9 2.8 1.8 2.4 2.8G-7 countriesa 1.8 2.6 2.4 2.1 1.7 1.9 2.1Low- and middle-income countries 1.6 2.7 4.2 4.5 1.5 3.7 4.6Excluding Eastern Europe and CIS 2.1 3.0 4.5 4.8 2.1 4.0 4.8Memo itemLow- and middle-income countries per capita GDP 0.1 1.2 2.8 3.1 0.0 2.2 3.1

World trade volume 4.2 5.0 6.4 6.3 4.2 5.8 6.2

Inflation (consumer prices)G-7 countriesa,b 1.4 1.3 1.6 1.8 1.4 1.8 2.0United States 1.6 2.2 2.5 2.5 2.3 2.6 2.7

Commodity prices (nominal U.S. dollars)Commodity prices, except oil –15.7 –11.2 2.8 4.2 –6.3 1.7 5.1Oil price (weighted average) –31.9 37.8 2.8 –2.7 –8.2 25.0 6.7Manufactures export unit valuec –3.9 –0.6 2.5 2.5 1.3 2.6 2.7

Interest ratesSix-month LIBOR (U.S. dollars, percent per annum) 5.6 5.5 6.0 6.0 5.0 6.0 6.0

Exchange ratesd

U.S. dollar per German mark –1.7 — — — — — —U.S. dollar per Japanese yen –7.6 — — — — — —

— Not applicable.a. Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States.b. In local currency, aggregated using 1988–90 GDP weights.c. Unit value index of manufactures exports from G-5 to developing countries, expressed in U.S. dollars.d. Positive figure is appreciation of local currency versus dollar.Note: The Republic of Korea has been reclassified as an upper-middle income country.Source: World Bank, Global Development Finance 1999; World Bank Development Prospects Group, November 1999.

3

Positive as these revisions are, they maskthe considerable fragility of developing coun-tries, which have yet to recover fully from thefinancial crises of 1997–99. Nor do the aggre-gates reflect the very uneven patterns of growthand recovery among regions. Save for East Asiaand South Asia (regions bolstered by growthin China and India), aggregate real per capitaincomes in 1999 are expected to decline or stag-nate in several developing regions.2 Further, thenews since March has not been all good. Thetightening of oil supply has meant higher im-port bills for many developing countries andincreased import prices in some industrialeconomies, where growth has been outstrip-ping estimated potential. And the favorablefinancial conditions have not made interna-tional investors less risk averse, as evidencedin interest rate spreads at high levels. Interna-tional capital flows to developing countrieshave fallen much more sharply than anticipated.

Long-term projections for growth in de-veloping countries (presented at the end of thechapter) have been downgraded by 0.3 per-centage points from forecasts of one year ago,suggesting that growth in developing countriesas a group in 2002–2008 is likely to be lowerthan in the precrisis 1990s. This estimate re-flects several factors, including a somewhatless favorable external environment; and, im-portantly, prospects for a protracted work-outof structural weaknesses in developing coun-tries—particularly in financial systems andfiscal positions—that have become more ap-parent in the wake of the crisis. One implica-tion of lower long-term projections is thatprogress on poverty reduction will be slower.And for some regions, including Sub-SaharanAfrica and Latin America and the Caribbean,reductions in poverty are likely to remain be-low the targets recently adopted by the inter-national community.

The external environment fordeveloping countries is improving

The external environment for developingcountries is expected to improve further

in 2000 and 2001. Growth among the threeindustrial-country centers is expected to bemore balanced. Policies geared toward achiev-ing a soft landing in the United States, stem-ming deflationary tendencies and shifting theimpetus for growth from the public to the pri-vate sector in Japan, and fostering businessactivity in the European Union (EU) will sup-port this shift. One result of these transitions—should they occur smoothly—would be abroadening of the locus of ‘locomotive power’for the world economy from the United Statesto include Europe. A modest increase in inter-national interest rates is likely to accompanyefforts to slow the momentum of the U.S.economy. But for developing countries, thebenefits of a smooth transition to more sus-tainable growth in the United States clearlyoutweigh the potential cost of a hard landing,and would likely be reflected in reduced riskperceptions.

Exchange rate developments that occurredin 1998–99 will exert some influence on theglobal distribution of demand in 2000 and2001. In particular the initial weakening ofthe euro against the dollar and the yen shouldprovide a boost to European exports, supportmanufacturing in Germany and Italy, andstimulate growth more broadly in the EuroArea. On the other hand, the strengthening ofthe yen poses some risk to the nascent recov-ery in Japan.

Stronger growth in the EU and stabiliza-tion in Japan will help firm up export mar-kets for many developing countries. Inparticular, Central and Eastern Europe, theMaghreb countries, and Sub-Saharan Africawill benefit from an upturn in European im-port demand. Resumed growth in Japaneseimports will further boost expandingintraregional trade in East Asia. Althoughgrowth in U.S. imports is likely to slow fromits recent highs, it should remain strong enoughto provide Latin American and Asian export-ers with opportunities for continued marketexpansion.

Developments in global commodity mar-kets are likely to have strongly differentiated

P R O S P E C T S F O R G R O W T H A N D P O V E R T Y R E D U C T I O N

4


effects across developing countries. Non-oilcommodity prices, which bottomed out dur-ing 1999, should undergo a moderate firmingthat will help bolster export revenues for manycountries in Sub-Saharan Africa, LatinAmerica, and East Asia. In addition, the like-lihood of only modest near-term changes indollar-based G-7 manufactures prices will tendto improve the terms of trade for all low- andmiddle-income countries. But if the sharp risein petroleum prices is sustained, it will par-tially or completely offset these developmentsin many oil-importing countries. Oil export-ers, many of which are in crisis, are the clearestbeneficiaries of recent trends.

Growth in the industrial countries hasbeen faster than expectedGrowth for the G-7 countries this year is likelyto register 2.6 percent, 0.9 percentage pointshigher than the forecast made six months ago(table 1.2). Continued strong growth in theUnited States is the principal factor in the re-vision, but Japan’s performance in the first halfof 1999 (3.2 percent annualized GDP growth),which was much better than anticipated, alsocontributes to the change. Europe, which hadbeen hampered by inventory overhang, is now

showing signs of a strong revival. Reflectingthese developments, world industrial produc-tion appears to be on an accelerating path,reaching 2.5 percent year-on-year advances inAugust, from declines at the start of 1999 (fig-ure 1.1).

United States. The remarkable growthperformance of the United States played amajor role in preventing a world recessionfollowing the outbreak of the East Asian fi-nancial crisis. GDP increased an average of 4percent during 1997–98, driven by privateconsumption and investment spending.Coupled with a strong dollar this growth sup-ported a 25 percent cumulative increase in U.S.import volumes over the same period and pro-vided a foundation for world exports as theAsian markets sank. The U.S. expansion wasbolstered by three factors:

• the restraining influence on inflation offalling prices for commodity and manu-factures imports;

• lower long-term interest rates, which re-flect a return of private capital fromemerging markets; and

• the 75 basis-point reduction in policy in-terest rates in late 1998.

Figure 1.1 World industrial productionPercent year-on-year

12

8

4

0

–4

Jan. 1997 July 1997 Jan. 1998 July 1998 Jan. 1999 July 1999

Developing countries

World

G-7

Note: Figures are three-month moving averages. Figure based on sample of countries comprising 85 percent of world GDP.Source: Datastream; World Bank staff estimates.

5

During 1999 the U.S. expansion has con-tinued to show resilience, with a recent strongpickup in exports complementing robust con-sumer and investment spending. But theeconomy also shows the widening imbalancesthat have resulted from the sustained periodof rapid expansion, including a negative house-hold savings rate (–1.5 percent)3 , a currentaccount deficit approaching 4 percent of GDP,potentially overvalued equity markets, and

modest signs of incipient inflationary pressures(figure 1.2).

Fears that higher oil prices and improv-ing conditions in foreign markets would rein-force incipient inflationary pressures in thedomestic market may have contributed to arecent rise in long-term interest rates. Partlyoffsetting these fears was evidence of acceler-ating growth in productivity, which couldenable—if sustained—a higher path of output


Table 1.2 World output growth, 1998–2001(percent)

Global DevelopmentCurrent Finance 1999

Estimate Forecasts Forecasts

1998 1999 2000 2001 1999 2000 2001

World total 1.9 2.6 2.9 2.8 1.8 2.4 2.8

High-income countries 2.0 2.6 2.5 2.3 1.8 2.0 2.2OECD high-income 2.0 2.6 2.5 2.3 1.8 2.0 2.2

G-7 1.8 2.6 2.4 2.1 1.7 1.9 2.1G-4 Europe 2.2 1.6 2.7 2.8 1.6 2.6 2.4

Other industrial 3.7 2.9 3.1 3.1 2.7 2.7 2.9Euro Area 2.7 2.0 2.9 2.9 2.2 2.7 2.7

Non-OECD high-income 1.2 3.0 4.0 4.6 1.3 3.6 4.7Asian NIEs 1.8 3.6 4.9 5.2 2.0 4.3 5.2

Low- and middle-income countriesa 1.6 2.7 4.2 4.5 1.5 3.7 4.6Excluding Central and Eastern Europe

and CIS 2.1 3.0 4.5 4.8 2.1 4.0 4.8Sub-Saharan Africa 2.4 2.3 3.1 3.4 2.5 4.0 4.0

Excluding South Africa and Nigeria 3.6 3.2 3.6 3.9 3.6 4.5 4.4Asia and Pacific 1.6 5.4 6.0 5.9 3.8 5.1 5.8

East Asia and Pacific b 0.1 5.5 6.2 6.2 3.6 5.2 6.0Excluding China –7.6 4.3 5.3 5.1 0.3 3.4 4.5

South Asia 5.1 5.4 5.5 5.3 4.4 4.8 5.2Europe and Central Asia –0.2 0.3 2.5 3.3 –1.5 2.3 3.6

Central and Eastern Europe 2.3 1.0 3.2 4.3 2.3 3.8 4.6CIS –2.7 0.7 1.3 2.3 –5.5 0.6 2.4

Middle East and North Africa 3.2 2.0 3.2 3.5 0.6 2.5 3.3Maghreb 4.8 2.6 4.0 4.1 2.8 3.9 3.9Mashreq 3.1 3.2 4.6 4.2 3.2 3.6 4.1Developing GCCc 1.0 –0.3 1.7 2.1 –2.4 0.6 2.0

Latin America and the Caribbean 2.1 –0.6 2.7 3.5 –0.8 2.5 3.9

Memo itemsEast Asia Crisis-5d –7.9 4.4 5.3 5.1 0.3 3.5 4.5Low- and middle-income, excluding East Asia Crisis-5d 3.3 2.4 4.0 4.4 1.7 3.8 4.7

a. Including Central and Eastern European countries and states of the CIS.b. Including the Republic of Korea.c. Gulf Cooperation Council (Bahrain, Oman, and Saudi Arabia).d. Indonesia, Malaysia, the Philippines, the Republic of Korea, and Thailand.Source: World Bank, Global Development Finance 1999; World Bank Development Prospects Group, November 1999.

6


growth with restrained inflation in the con-text of tight labor markets.

Against the background of three modestinterest rate increases in 1999, GDP growthis expected to slow from 3.8 percent in 1999to 2.8 percent in 2000. As the pace of domes-tic demand growth eases in response to fur-ther interest rates increases, GDP could slowto 2.2 percent by 2001. Inflation is expectedto rise only modestly. The current accountdeficit will widen further, rising from $315billion to $350 billion as import demand slowsgradually, exports respond with some lag tooverseas recovery, and higher oil prices inflatethe import bill. The risk of a sharper slow-down remains substantial, however. The po-tential for more dramatic corrections of currentimbalances—through a sharp adjustment ofthe dollar or a fall in equity markets, for ex-ample—are elements of concern in the nearterm, which would also have significant ef-fects on developing countries most dependenton U.S. markets and international capital flows(see the section Risks to the forecast).

Japan. Starting in the last months of 1997,Japan’s economy registered five consecutivequarters of negative growth. Uncertainties inthe country’s financial system, rising unem-

ployment, and falling wages placed a damperon consumer spending. Business investmentoutlays continued to decline in the face of fall-ing domestic demand and massive over-capac-ity. Japan’s largely domestic-based recessionand the resulting decline in imports of some7.7 percent in 1998 significantly aggravatedthe regional crisis (see box 1.5).

During the first quarter of 1999 the ef-fects of large-scale public investment spend-ing bore fruit, however. GDP growth of 8.1percent (annualized) was driven by a 10 per-cent quarter-to-quarter surge in public invest-ment outlays. But private consumption,residential investment, and fixed business in-vestment also increased at rates that were muchstronger than had been anticipated. Surpris-ing second-quarter growth of 0.4 percent (an-nualized) was clearly led by households, asboth public and private investment spendingdeclined sharply. More recently Japan, like theUnited States, has been benefiting from thestrong recovery in growth in East Asia. A pro-nounced increase in export shipments has sup-ported gains in Japanese production.

Although recent developments could markthe end of Japan’s second-longest postwar re-cession, recovery in private demand is not yet

Figure 1.2 U.S. private savings rate and current account balance

2

1

0

–1

–2

–3

–4Q1 1998 Q2 1998 Q3 1998 Q4 1998 Q1 1999 Q2 1999

Private savings as a share of personal disposable income

Current account balance as a share of GDP

Percent

Source: International Monetary Fund; U.S. Survey of Current Business; Datastream.

7

self-sustaining. The risk of relapse remains, forseveral reasons:

• Retail spending continues to display vola-tility, and consumer sentiment may falterin the face of declining incomes and fur-ther weakness in labor markets.

• Public loan guarantees have propped upinvestment in housing and business mar-kets, helping increase liquidity in thesesectors.

• Investment remains concentrated amongsmall- and medium-sized enterprises whilelarger corporations are saddled with sub-stantial excess production capacity.

• The sharp appreciation of the yen mayrestrain prospective export performance.

• Deflation has emerged across all segmentsof the Japanese economy, adding to debtburdens.

These factors suggest not only that fur-ther recovery will likely be of modest propor-tions but also that it is contingent on continuedpublic sector support—which is becoming in-creasingly difficult to sustain. Against thisbackground, it appears likely that Japan willregister positive growth on the order of 1.3percent this year—a marked improvement overthe March assessment. But growth may fadeto 0.9 in 2000 if fiscal stimulus is largely with-drawn from the economy. The private sectorrecovery will need to become self-sustainingsoon since the consolidated public deficit willapproach or exceed 10 percent of GDP in 2000.

Europe. Although consumer sentimentand spending have been buoyant in Europe,weakness in the industrial sector (particularlyin Germany and Italy) checked earlier expec-tations for stronger growth during 1999. Inpart, this weakness is a further manifestationof the East Asian and Russian crises. Sharpdeterioration in the performance of Germanand Italian exports in emerging markets damp-ened business sentiment, increasing inventoryoverhang, depressing industrial production,and ultimately restraining trade within the EU.Uncertainty in the business sector may have

contributed to the 10 percent fall in the valueof the euro against the dollar through June1999, and this prompted the European Cen-tral Bank (ECB) to implement a 50 basis-pointeasing of policy interest rates. Despite a ratereversal in the autumn, the euro continued todecline against the dollar, to near parity byearly December. GDP in the Euro Area slowedfrom 2.7 percent in 1998 to 1.8 percent an-nualized during the first half of 1999.

Business surveys that were conducted inthe first half of 1999 reported widespreadimprovements in current and prospective con-ditions. Figures for industrial production inFrance (1.5 percent month-to-month in July)and Germany (1.2 percent in August) are nowconfirming these expectations. The potentialfor a rebound in production in the second halfof the year is now much stronger, as evidencedby rising export orders and shipments, con-tinued firm consumer demand, and diminish-ing inventory overhang. The growth slowdownin the United Kingdom in late 1998 and early1999 appears to have passed, with consumerspending rising at an impressive 4 percentannual rate in recent months, production upby 1.7 percent, and unemployment at recordlow levels. On the external front, German andU.K. exports to countries outside the EU arenow rising fairly rapidly. The UnitedKingdom’s 15 percent annualized export vol-ume growth rate during the summer monthsof 1999 reflects not only continued strengthin the United States but also an upswing inemerging market demand. These factors au-gur well for a rebound in EU growth that couldreach 2.9 percent in 2000–2001. An accelera-tion of import demand from recent lows of2.5 percent to 6.5 percent over the period willhelp pick up some slack from slowing U.S. de-mand and support the advance of world tradeat rates near the 20-year average.

Inflation and interest rates toremain moderateInflation slowed significantly in the industrialcountries in 1998 and 1999, falling from aG-7 average of 2.1 percent in 1996 to 1.4 per-


8


cent in 1998. More recently, inflation trendsacross the major countries have displayed dif-ferent dynamics, the result of developmentsin the global environment as well as in thedomestic arena (figure 1.3). Continued mod-erate inflation in Europe coupled with defla-tionary trends in Japan is likely to keep anyincrease in international prices in 2000–2001to modest proportions.

In the United States several elements con-verged to slow and eventually stabilize infla-tion over 1997–98. These included:

• the onset of the Asian crisis, which inter-acted with the strong dollar to cause sub-stantial declines in the prices of importedgoods;

• the resulting increase in competition fromimported products, which restrained theability of domestic producers to raise prices;and

• underlying improvements in productivitygrowth (as well as large additions tomanufacturing capacity), which kept unitlabor costs under control.

As some of the favorable internationalinfluences waned and signs of moderately in-

creasing wage pressures emerged, U.S. con-sumer price inflation accelerated. Inflationclimbed above the 2 percent mark in mid-1999and may increase toward 2.5 percent in thenext years.

In Germany, and more broadly acrossEurope, similar international influences havebeen at work. But continued modest inflation(less than 1 percent in Germany during 1999)owes more to the sluggishness of economicactivity and large output gaps within the EuroArea. Imminent recovery in activity, togetherwith the euro’s near-term weakness, shouldplace some modest upward pressure on Euro-pean inflation moving in 2000–2001. At thesame time, Japan has been beset with a boutof deflation that may require some time todissipate. Falling consumer demand, excesscapacity in industry, and the strength of theyen all contribute to this phenomenon.

Shifts in interest rates in the industrialcountries are likely to be moderate over thenext 12 months, reflecting expected develop-ments in economic activity and inflation. Thecumulative 75 basis-point tightening by theFederal Reserve over the second half of 1999may be followed by further modest incremen-tal increases in rates beyond the current 5.5

Figure 1.3 G-3 consumer prices, 1997–99Percent year-on-year

4

United States

Germany

Japan

3

2

1

0

–1Jan. 1997 July 1997 Jan. 1998 July 1998 Jan. 1999 July 1999

Note: Figures are three-month moving averages.Source: IMF, International Financial Statistics.

9

percent level for Fed Funds (figure 1.4). Fol-lowing its recent increase in policy rates theECB may wait for stronger confirmation ofthe rebound in Euro Area activity before imple-menting further rate increases, designed to shiftthe stance of monetary policy from loose toneutral. Japan will likely maintain effectivezero levels of the uncollateralized overnightcall rate until private demand stabilizes andbegins to grow sustainably.

World trade is likely to continueto accelerateWorld trade growth has accelerated fromabout 4 percent in 1998 to 5 percent in 1999and is projected to average nearly 6.5 percentin 2000–2001. But these apparently modestchanges in average growth rates mask boththe sharp intrayear downturns of the last twoyears, and significant shifts in the geographicdistribution of demand.

Throughout 1998 world trade slowedsharply as the combined effects of the EastAsian, Russian, and Brazilian crises contrib-uted to a steep downturn in import demandin Asia, Latin America, and the economies ofthe Commonwealth of Independent States

(CIS) (figure 1.5). During the second half of1998 world import growth averaged 2 per-cent at annual rates. Increases of 8–10 per-cent in North American and EU imports wereoffset by a 5 percent drop in demand in devel-oping countries.

Both the pace and pattern of growth haveshifted dramatically during 1999. Recentmomentum indicators suggest a movementtoward 6 percent growth, amid further changesin the geographic mix of demand in late 1999.Decline in Latin America is being offset byfirming trends in Western Europe, the newlyindustrializing economies (NIEs) in Asia, andthe oil exporting countries, whose foreign cur-rency receipts have surged in recent months.

The rapid recovery in East Asia has beenthe most important factor in the accelerationof world trade. The shift from a large declinein import volumes in Asia (including Japanand the NIEs) in 1998 to moderate growthprojected in 1999 will add about 4.6 percent-age points to world imports this year (table1.3). The Asian import recovery has helpedstimulate activity around the world, especiallyin those economies in Western Europe andwithin Asia itself that are highly dependent


Figure 1.4 Short-term interest rates, 1998–99

6

5

4

3

2

1

0

–1

Jan. 1998

U.S. Fed Funds rate

Euro interbank rate

Japan call money rate

April 1998 July 1998 Oct. 1998 Jan. 1999 April 1999 July 1999 Oct. 1999

Percent

Source: International Monetary Fund.

10


on exports. Growth in world import volumeis expected to rise toward 6.5 percent over thenext two years. This rate is near the 20-yeartrend but is still below what might be expectedin a cyclical recovery period following a sharpslowdown.

As world trade volumes increase in 1999,the steep decline in dollar-based trade pricesof 1998 is likely to slow, partly reflecting thedecline of the U.S. currency vis-à-vis the yen.The sharp fall in prices of both manufacturesexports and commodities in 1998 caused adecline in the dollar-based global trade de-flator of some 7 percent. But the large up-swing in oil prices and improvement in pricesof manufactures exports could boost theworld deflator to gains of 2 percent in 1999.If the U.S. dollar softens moderately in thelast months of 1999 (as suggested by futuresmarkets), prices for manufactured goods mayflatten out. Table 1.4 shows the performanceof these prices in the first three quarters of1999. A flattening out of import prices wouldallow global trade flows expressed in dollarsto rise for the first time since the onset of theAsian crisis. This possibility is good news formany of the heavily indebted developing coun-tries, whose debts are denominated primarilyin dollars. However, primary commodity ex-porters will no longer benefit from declinesin the prices of imported manufactures—declines that have helped support their terms

Table 1.3 Contributions to world import growth(percent year-on-year)

1998 1999 2000

World import growth 3.7 5.8 6.6

Contributions to world growtha

G-7 countries 3.9 3.4 2.9North America 2.1 2.1 1.5Japan –0.4 0.4 0.2G-4 Europe 2.1 0.8 1.2

Other industrial countries 1.6 1.2 0.9Asian NIEs –0.5 0.2 0.6

Developing countries –1.3 1.0 2.2East Asia –1.6 1.3 1.0Latin America 0.8 –0.2 0.5Eastern Europe 0.7 0.0 0.2

CIS –0.2 –0.5 0.2

a. Share of world trade times import growth.Source: World Bank staff estimates.

Figure 1.5 World import volume growthPercentage change, year-on-year

18

12

6

0

–6Q1 1996

G-7 countries

Projection

World

Developingcountries

Q1 1997 Q1 1998 Q1 1999 Q1 2000

Note: Estimate for developing countries is based on U.S. dollar data deflated by world import price index.Source: World Bank staff estimates.

11


of trade during the recent prolonged weak-ness in commodity prices.

As noted in the introduction, the EastAsian crisis had important adverse effects onglobal trade volumes and commodity prices,which are now unwinding. Box 1.1 offers ananalysis of the price, sectoral, and regional ef-fects of the crisis.

Commodity prices are unlikely torebound sharplyThe decline in commodity prices during thepast three years has been one of the sharpeston record, with nonenergy commodity pricesfalling 29 percent from their peak and petro-leum prices falling 56 percent in the two yearsthat ended in December 1998 (see chapter 4).Since the lows of late 1998 and early 1999,some commodity prices have begun to recover,but mostly because supplies have been reduced.To date, there has not been a substantial re-covery in demand. Faced with collapsing pe-troleum prices, OPEC agreed to sharp cuts inproduction in March and has shown its re-solve to adhere to these cuts. This develop-ment has caused petroleum prices to recoverfully from their decline. Metals prices haverecovered about one-fourth of their declinesince early 1999 as high-cost mines and smelt-ers either closed or began operating at reducedcapacities. Demand for primary commoditieshas begun to rise somewhat, spurred by the

unexpectedly positive economic outcomes forthe East Asian countries, sustained rapidgrowth in the United States, and signs of re-viving economic activity in Europe. Stockpilesof metals have now begun to decline, and de-mand for metals has shown signs of increas-ing in Asia.

Agricultural prices have yet to show clearsigns of recovery. Supplies continue to increase,adding to already large stocks. Unlike metalsor petroleum, supplies of agricultural com-modities depend on the weather, and in 1999the weather has helped generate large outputgains. Sugar production is expected to exceedconsumption for the fourth consecutive year,and soybean and total oilseed production areexpected to reach a record levels. Grain pro-duction has fallen in response to low prices,but not enough to tighten markets and raiseprices.

The near-term outlook for commodityprices is somewhat mixed, with energy pricesexpected to remain strong because of OPEC’sdetermination to limit supplies while othercommodity prices are expected to remain weak(table 1.5). Real nonenergy commodity pricesare expected to fall 10.5 percent in 1999 andthen begin to recover moderately in 2000.Agricultural prices are likely to recover moreslowly than prices for other commodities be-cause of the abundant stocks that have accu-mulated and a continued weak supply response

Table 1.4 U.S. import prices of manufactured goods, 1999(percentage change, year-on-year)

Percentageof total U.S.

Imports from: imports Dec. 1998 April 1999 July 1999 August 1999

Japan 16.6 –3.0 –1.1 0.6 1.0Canada 16.5 –1.8 –1.6 0.2 0.2European Union 16.4 0.5 0.9 –0.1 0.3Asian NIEs 10.9 –8.4 –6.5 –3.6 –3.4Latin America 10.4 –3.8 –1.7 –0.6 0.7World 100.0 –3.0 –2.0 –0.7 –0.3

Memo itemsU.S. import price –6.3 –1.7 1.3 2.7World import price –4.0 –3.3 –0.9 . .

. . Not available.Source: U.S. Bureau of Labor Statistics; World Bank staff estimates.

12


• The developing country export sectors mostadversely affected by the crisis include com-modities (other crops, mining, and crude oil),commodity-dependent downstream sectors(processed foods, refined oil, and iron andsteel), and manufacturing sectors character-ized by fierce price competitive pressures(textiles and apparel, machinery) and someservices.

Trade effects. The crisis simulation has itsstrongest impacts on investment and import de-mand in the EA-5 countries, with only a modestincrease in exports. The drop in investment affectsmostly demand for capital goods, including ma-chinery and transportation equipment, and con-struction, with multiplier effects in the nonmetal-lic minerals sector. Income-sensitive sectors, suchas motor vehicles, are also hit hard, with the leastaffected sectors being food and clothing and, to alesser extent, energy—that is, basic necessities.8

When translated into changes in import demandthe largest reductions fall on capital goods andmotor vehicles. Other developing countries sufferan unambiguous drop in the volume of exports tothe EA-5 countries and Japan, on the order ofmagnitude of $30 billion and $13 billion, respec-tively.

In the simulation, aggregate import volumeinto the industrial countries increases by over2 percent, more than the increase in industrialcountry aggregate demand (0.5 percent). Changesin relative prices play a large role in fostering de-mand for imports. In percentage terms, some of thelargest increases in import demand in the industrialcountries are in rice, sugar, oil, textiles and ap-parel, light manufacturing, motor vehicles, othertransportation equipment, and electronics. Exportvolumes to industrial countries from the non-crisisdeveloping countries rise some $40 billion, largelycounterbalancing the drop in export volume to-ward the EA-5 and Japan.

Price effects. The loss in net exports is small involume terms ($1.3 billion) for noncrisis develop-

Box 1.1 Sectoral and regional effectsof the East Asian crisis

A global general equilibrium model4 was em-ployed to trace the real impacts of the crisis in

East Asia, with a focus on regional and sectoraleffects.5 The exercise was designed to isolate theimpacts of the observed output declines in the fiveEast Asian crisis economies (EA-5) and Japan,including the current account reversals, which onaggregate were over $100 billion. These reversalswere allocated to the industrial economies, withthe United States accounting for the largest shareof the increased current account deficit—on theassumption that credit to developing countries wasconstrained at the precrisis level. Thus the financialshock for developing countries—in the form ofchanges in capital flows—was confined to the EA-5countries, with an explicit assumption that theother developing countries were able to maintainprecrisis current account balances. The results ofthe crisis simulation are compared to a baselinesimulation that calibrates growth rates in the crisiseconomies and Japan to precrisis growth forecastsand fixed current account balances at their pre-crisis levels.6

The analysis suggests the following conclu-sions:

• The main transmission channel of the demandshock in East Asia and Japan on developingcountries—isolated from other observedshocks—is through a large fall in exportprices,7 reflecting the devaluation of currenciesand the decline in domestic demand, with thestrongest negative effects on commodity prices.The analysis in some sense provides an upperbound on price effects owing to the assump-tion of freely flexible prices embedded in themodel. The demand shock could have been acatalyst for new crises in other countries, par-ticularly commodity-dependent economies andthose with fragile finances.

• Export volumes from developing countries toEast Asia decline unambiguously, though thiseffect is largely offset by growth in importdemand in the industrial countries.

13

Box 1.1 (continued)

ing countries. However, in value terms, the reduc-tion is much larger ($34 billion), making the effectof the Asian crisis equivalent to a capital accountreversal of $34 billion. While the decline in exportprices is much higher for the EA-5 countries thanfor the other developing countries, the latter none-theless suffer from an average decline in exportprices of 2.1 percent. The largest declines occur inthose sectors where commodities are most substi-tutable (for example, crude and refined oil), orsectors where the two regions are significant com-petitors (such as textile and apparel, iron and steel,and machinery).9

Net impact on exports. The figure in this boxdepicts the decomposition of net trade effects forthe two developing regions into four volume ef-fects—one for each region of destination—and oneprice effect, capturing the essential story. The vol-ume effects clearly show the negative impact ofimport demand in the EA-5 and Japan and thepositive effect in industrial countries. (The figurealso shows the increased market share of the

EA-5 in other developing countries and the lackof change in intraregional trade.) The balance ofthe volume effects are positive for the EA-5 coun-tries (+5.7 percent) and mildly negative for theother developing countries (–0.1 percent). However,the negative price effects lead to an increase inexport value of only 0.8 percent for the EA-5 coun-tries and a decline of –2.1 percent for the otherdeveloping countries. On balance the primarytransmission channel for noncrisis developing coun-tries is a drop in export prices, so that the negativeimport volume effect in the crisis countries islargely offset by positive import volume effects inthe industrial countries.

The simulation is useful in illustrating the priceeffects on sectoral and regional demand under theassumption of flexible prices and instantaneousadjustment of both prices and volumes. In reality,in 1998 export volumes of developing countries(excluding the EA-5), slowed sharply, as the rise inimports of the industrial countries was not nearlysufficient to offset the demand collapse in Asia.


Decomposition of change in export values

6.0

4.0

2.0

0.0

–2.0

–4.0

–6.0

Percent

Note: The first four columns represent the impact of export volume changes on the value of exports and the last column represents the aggregate price impact.Source: World Bank.

Other developing countries

East Asia–5

East Asia–5 Japan Other industrialcountries

Region of destination

Other developingcountries

Price

14


to falling prices. Real energy prices are antici-pated to increase 38 percent in 1999 and tostabilize in these terms in 2000. The forecastassumes that OPEC will raise production offi-cially or unofficially during the winter andprevent prices from being sustained much above$20/bbl. But until such time as OPEC does raiseproduction, speculative demand on futuresmarkets could take prices well above $20/bbl.

Private capital flows to developingcountries are likely to recover onlymodestly in the short termThe environment for private capital flows toemerging markets is worse than was expected

six months ago. Gross private flows for Janu-ary–August 1999 were almost a third lowerthan they were a year earlier (table 1.6). Flowsgenerated by bond issues were 25 percentlower. Issues from Europe and Central Asia(ECA) and Latin America plunged 56 percentand 17 percent respectively. Bank lendingcommitments dropped 25 percent during theperiod, with ECA and Latin America againregistering large drops (more than 50 percent).International equity placements did increaseover this period, largely reflecting pri-vatization receipts in Korea and Thailand.East Asian equity and other markets haverisen sharply from their lows in late-1998.

Table 1.6 Average monthly gross capital market flows to developing countries(billions of U.S. dollars)

1997 1998 1999

Total Total Jan.–July July–Dec. Jan.–June July August

All developing countries 24.2 16.2 19.9 12.5 13.9 10.3 5.7East Asia and Pacific 6.2 3.0 3.2 2.7 3.6 3.4 2.2Europe and Central Asia 4.3 3.9 4.4 3.3 2.0 1.4 0.9Latin America and the Caribbean 10.1 7.4 9.8 5.0 6.5 3.6 1.5Middle East and North Africa 1.7 1.0 1.0 1.1 0.7 0.3 0.3South Asia 1.0 0.4 0.7 0.1 0.4 0.1 0.0Sub-Saharan Africa 0.9 0.5 0.7 0.4 0.7 1.6 0.8

Note: Gross capital market flows include bond issues, syndicated loan commitments, and equity issuses, based on reports from theinternational capital markets.Source: Euromoney; World Bank staff estimates.

Table 1.5 Annual percentage change in oil and non-oil commodity prices(World Bank commodity price indexes, nominal U.S. dollars)

Trends Forecasts

Commodity group 1981–90 1991–95 1997 1998 1999 2000 2001

Non-oil commodities –2.3 4.1 2.2 –15.7 –11.2 2.8 4.2Food –3.3 3.2 –6.1 –9.6 –14.0 2.9 4.0

Grains –2.9 3.8 –13.9 –9.6 –13.9 5.7 7.8Beverages –5.8 8.6 35.2 –17.6 –25.7 –0.7 4.9Raw materials –0.5 6.2 –10.5 –23.2 1.2 3.9 4.7Fertilizers –2.5 0.7 –0.1 2.0 –6.4 –4.2 –2.2Metals and minerals 0.5 0.3 1.2 –16.1 –3.2 6.0 4.0

Petroleum –4.7 –5.6 –6.2 –31.9 37.8 2.8 –2.7

G-5 manufactures unitvalue index 3.3 3.6 –5.1 –3.9 –0.6 2.5 2.5

Source: World Bank Development Prospects Group.

15

But the primary reason for their recoveryappears to be the response of domestic inves-tors to rising liquidity and falling interestrates, and to some degree a stemming of capi-tal flight. So far there is little evidence of largeinflows from abroad.

While the recovery in East Asia helpedrevive confidence in emerging markets to amodest extent, offsetting events occurred.Following the Brazilian devaluation in Janu-ary 1999, secondary market bond spreads in-creased for most of the major developingcountry borrowers. However, contagion fromBrazil was restricted and relatively short-lived.The average spread on Brady bonds—an in-dex that is heavily weighted toward LatinAmerica—dropped to 890 basis points in Aprilfrom over 1,300 basis points at the end ofJanuary (figure 1.6). With some signs of sta-bilizing market conditions in the second quar-ter, gross capital flows increased significantly,bolstered by a 33 percent increase in bond is-suance and a near doubling of bank lendingcommitments compared with the first quar-ter. The East Asian crisis economies showedincreasing strength in the bond markets, andLatin American borrowers secured significant

amounts of commercial bank credit. Still,market access remained restricted to the mostcreditworthy borrowers.

During the third quarter, tightening U.S.monetary policy, the Daewoo restructuring,and Ecuador’s difficulties again deterred in-vestors. Gross capital flows fell sharply, andsecondary bond markets became illiquid as thevolume of transactions dropped sharply. Ec-uadorian Brady bond spreads soared to around4,000 basis points, but Latin American spreadswere not affected significantly. By the end ofAugust average secondary market spreads oninternational bonds stood at lower levels thantheir 1998 averages but remained well above1997 levels. Investor concern manifested it-self both in limited access and high interestrate spreads.

Despite falling volumes of private marketflows, foreign direct investment (FDI) has re-mained resilient and is likely to be the primarysource of finance for developing countries forthe foreseeable future. The global crisis hascontributed to reduced growth prospects evenfor FDI, however. Preliminary data indicatesome reduction in FDI flows in 1999. In EastAsia new FDI projects approved in 1999 have


Figure 1.6 Spreads on Brady bondsBasis points

August

1,800

1,400

1,000

600

200

1991 1992 1993 1994 1995 1996 1997 1998 1999

Source: Bloomberg; World Bank.

16


fallen sharply in Indonesia, China, and Thai-land.10 Although the downturn may be par-tially offset by an increase in Korea, overallflows in East Asia may continue to declinemoderately in the medium term. FDI to LatinAmerica may increase somewhat in 1999owing to Brazil’s privatization projects, butthese levels may be difficult to maintain in2000–2001. Flows to ECA should remainstable in the near term and are likely to growin the medium term as major recipients pro-ceed with privatization programs and inves-tors are attracted by the prospect of EUaccession. On balance these trends suggest thattotal FDI flows to developing countries mayremain below the peak levels of 1997 over thenext few years.

On the back of the economic recovery inAsia and an expected bottoming-out of therecession in Latin America, flows from inter-national capital markets (bonds, equity, andbank loans) are expected to stage a modestrebound during 2000–2001 but to remain wellbelow precrisis levels. Several factors suggestthat these flows to emerging markets will in-crease only slowly. Financing requirements inAsia have been cut drastically as countries havemoved to large current account surplus. In-vestors are likely to continue to shy away fromthe perceived risks of emerging markets in thenear term, and this tendency will improve onlygradually in the medium term. Since the EastAsian crisis, pegged currencies have given wayto flexible exchange rates in many cases, in-creasing the risk of borrowing in foreign cur-rency. The restructuring of corporations andbanks has not progressed as expected in manyAsian countries, and other problems (such asthe situation with Daewoo) have resurfacedto dampen investor enthusiasm. Other emerg-ing economies also have structural problemsthat have proven to be politically intractable,leading to difficulties with—or defaults on—international payments (as in Ecuador, Paki-stan, and Russia). Policymakers are encour-aging greater supervision of financial marketsand banks and actively debating the notion ofburden-sharing by private investors.

The outlook for official flows improvesNet official flows received by developing coun-tries rose from $39 billion in 1997 to $48billion in 1998, largely because of noncon-cessional disbursements under rescue packagesfor countries most affected by the financialcrisis. Preliminary data reported by donorcountries from the Development AssistanceCommittee of the Organisation for EconomicCo-operation and Development (OECD) showsome welcome relief from the declining trendin net official development assistance (ODA)flows. Net ODA flows to developing coun-tries and multilateral development agenciesregistered an increase of 8.9 percent in realterms in 1998, compared with a 21 percentcumulative decline from 1992 to 1997. Therise in net development assistance is the resultin part of increased concessional flows to coun-tries affected by the financial crises. But it alsoreflects increased awareness on the part ofseveral donor countries of the need to sustainaid levels in order to achieve long-term devel-opment objectives.

Significant enhancements were made tothe Debt Initiative for Heavily Indebted PoorCountries (HIPC) following a broad consul-tative review by the World Bank, InternationalMonetary Fund (IMF), and the Cologne Sum-mit. The enhanced initiative calls for deepen-ing debt relief by targeting it to bring the netpresent value of debt down to 1.5 times ex-port earnings (as of the decision point). Thistarget is much more accommodative than thetarget in the original initiative (2.5 times ex-ports, as of the completion point). The en-hanced HIPC also lowers the thresholds forthe fiscal burden of debt and offers faster debtrelief.11 Debt relief will now be initiated assoon as a country qualifies at the decision pointas well as when it reaches the completion point.The number of eligible countries will also risefrom 29 to 36 or more.

The new framework is expected to pro-vide $27 billion in relief (in net present valueterms)—more than double the $12.5 billionof the original framework. More than $5 bil-lion will be provided on debts owed to the

17

World Bank. Relief on debt owed to theIMF is estimated at around $3 billion, whichwill be funded by a revaluation of a portionof the IMF’s gold reserves. Donor countrieshave indicated that they will be steppingup their contributions to help fund debtrelief under the enhanced initiative. In addi-tion G-7 leaders agreed at the Cologne Sum-mit to cancel their official developmentassistance debt, and some countries have an-nounced that they could go beyond this toinclude some nonconcessional loans for quali-fying HIPCs.

The outlook for developingcountries in 1999–2001 suggestssignificant acceleration

Improvements in the external environment,favorable policy trends in emerging mar-

kets, and increasing confidence among bothdomestic and foreign investors should helpspur growth in developing countries over thenext two years. However, growth will varymarkedly across regions (table 1.2 above).Growth of 2.7 percent for developing coun-tries as a group is projected in 1999, acceler-

ating to 4.2 percent in 2000, and further to4.5 percent by 2001—still below the precrisislevel in the non-transition developing coun-tries (figure 1.7).

Capital flows to developing countries haveremained very low compared with the levelsthat prevailed in the mid-1990s. But manyemerging markets have seen a substantial loos-ening of liquidity conditions since the turn ofthe year, including lower domestic interestrates, rising equity prices, and in some in-stances, upward pressure on currencies. Theseimprovements are highlighted in figure 1.8,which displays changes in an index of finan-cial conditions—a simple average of the per-centage change in interest rates, equity prices,and exchange rates—for a sample of emerg-ing markets.12

Improvements in domestic investors’ con-fidence reflect in part the more favorable glo-bal outlook (in contrast to the near-panic inworld markets following the collapse of Rus-sia and the hedge fund Long Term CapitalManagement last fall). As well, in many coun-tries appropriate adjustment policies are be-ing adopted, including measures aimed atrecapitalizing banking systems.


Figure 1.7 Real GDP growth for developing countriesPercent

5

4

3

2

1

01998 1999 2000 2001

Source: World Bank, Global Development Finance 1999; World Bank baseline forecast.

March forecast

Current forecast

18


Important policy assumptions are builtinto the projections for each developing region:

• In East Asia declining interest rates andother measures to support domesticdemand, including fiscal stimulus andsteps to recapitalize banks, have helpedrestore investor confidence. These policymeasures are expected to combine withimproved export performance (based onthe 20–25 percent real depreciation incurrencies since the onset of the crisis andthe incipient revival of world trade) toconsolidate the recovery in the near term.However, a sustainable growth pathrequires further and more determinedbank and corporate restructuring (seechapter 3).

• Fiscal consolidation will be essential in anumber of Latin American countries inorder to reassure investors, encourage aresumption of capital flows to the privateand public sectors, facilitate exchange ratestability, and bring interest rates downfurther.

• Russia is expected to enhance its stabili-zation by making further progress on

both the fiscal and inflation front in thenear term. This scenario would improveprospects for several countries in theECA region, particularly neighboringstates of the CIS. Several Central andEast European countries are working tomeet EU accession criteria and are ex-pected to adopt policies that will raisestandards and institutions to EU norms,which will have positive spillover effectson the region.

• In the Middle East and North Africa, re-cent events—including the 1998 drop inoil prices (now reversed)—and the needto progress on implementation of Euro-Mediterranean Agreements—have createda sense of urgency with respect to growthand efficiency-enhancing strategies. As aresult many countries are expected to in-tensify their efforts to promote both tradeand financial integration. These effortsinclude more extensive privatization pro-grams, the liberalization of capital mar-kets, and various measures designed tostrengthen the productivity of domesticindustry in order to enhance export per-formance.

Figure 1.8 Index of financial conditions since June 1997

Mean of percentage changes in currency, stock market, interest rates

–90 –60

Better off Worse off

–30 0 30 60 90

PolandArgentinaHungary

The PhilippinesCzech Republic

ThailandIndia

South AfricaMalaysia

Republic of KoreaBrazil

MexicoIndonesia

TurkeyRussian Federation

Source: World Bank staff estimates.

December 1998–November 1999

June 1997–December 1998

19

• In Sub-Saharan Africa structural reformshave become more widespread since themid-1980s, including tax and public ex-penditure reforms, the restructuring andprivatization of parastatals, trade liberal-ization, and deregulation of internal mar-kets. Such reforms have worked topromote exports and investment in manycountries. The result has been faster out-put growth already, especially when policyreform has been supported by debt relief.

• In South Asia the resolution of politicaluncertainties in many countries may helpto sustain growth in the near term. Do-mestic business confidence in India isstrong, and both major political partiesare committed to a broad range of eco-nomic reforms. But the fiscal deficit re-mains a serious concern. The projectionsassume progress in macroeconomic sta-bilization and conflict resolution through-out the region

Near-term growth is expected to improveEast Asia. Although most East Asian econo-mies experienced virtually no growth in 1998,GDP is projected to increase by 5.5 percentin 1999. This increase includes a dramatic turn-around for the five crisis economies, whichshould see growth rise from –7.9 percent toaround 4.5 percent. Currency depreciationand the continued strength of the yen have con-tributed to strong export activity for the group.Stronger growth in Japan and intraregionaltrade are reinforcing this recovery. A suppor-tive policy environment is beginning to gener-ate gains in private consumption. Expectationsfor capital spending are less optimistic becauseof significant excess capacity, the widespreaduncertainties associated with bad loans in thebanking system, and the need for fasterprogress on corporate restructuring.

The economic problems in China havebecome more apparent. Consumer prices havebeen in a steady decline for some time. Con-sumer demand is weak (in part because ofuncertain employment conditions), and inven-tories of unwanted goods are rising. The bank-

ing sector’s weakness, which has grown moreevident, is related to the continued poor per-formance of state-owned enterprises (SOEs).Extensive public investment projects have sup-ported growth thus far, but the government’sresources are relatively limited, and the fiscalexpansion cannot be sustained indefinitely.After lagging through the first half of 1999,however, a combination of factors are nowbuttressing Chinese exports, including renewedgrowth in East Asia and continued importgrowth in the United States.

The outlook for 2000 and 2001 for EastAsia is positive—6.2 percent GDP growth—with more balanced advances across final de-mand components and countries. Withinflation in check, monetary policy will remainrelatively loose, supporting consumer demand.Looser monetary conditions, the narrowing ofthe output gap, an emerging recovery of cashflow and improving banking balance sheetsshould lead to a modest recovery of invest-ment activity, though the region is unlikely tosee a return to precrisis investment rates inthe foreseeable future. A pickup in world eco-nomic activity and strong intraregional link-ages should support continued export growth.Imports will rise further, leading to a narrow-ing of presently large current account sur-pluses. Failure to pursue financial andcorporate restructuring, however, coulddampen investors’ enthusiasm, possibly gen-erating another bout of financial and exchangerate difficulties (see chapter 3). But East Asianeconomies are in a much better position forthe near term than they were a year ago, withstrong current account balances, much lowershort-term external liabilities, and improvedreserve positions.

Latin America. Output for the LatinAmerica and Caribbean region has declined anestimated 0.6 percent in 1999, falling from a2.1 percent advance in 1998, despite favorablegrowth performance in Mexico. Economicdownturns deepened in many countries in thewake of the Russian crisis of August 1998. Inbroader terms, adverse terms of trade and areduction in capital flows to the region (the


20


result of concerns about the fundamentals inseveral large countries) contributed to a reces-sion in the last quarter of 1998. For some coun-tries, particularly Brazil and Ecuador, therecession was the reflection of home-grownimbalances, though the global crisis affectedtiming and severity. But for other countries(Argentina, Chile, Colombia) the causal factorswere tied more closely to the ongoing deterio-ration in the external environment. Althoughthe financial contagion was much less thanfeared, the Brazilian devaluation at the begin-ning of 1999 was a further shock to the re-gion. Prices of key commodity exports tumbled(especially for countries in Central America andthe Caribbean), intraregional trade slowedsharply, and private capital flows declined fur-ther from their 1997 and 1998 levels.

Brazil has emerged from its January 1999fiscal crisis and devaluation showing greaterresilience than was envisioned six months ago.But growth in a number of other countriesworsened as the deteriorating global environ-ment finally took its toll, yielding on net onlya marginal change in the region’s previousgrowth forecast of –0.8 percent for 1999. In-vestor uncertainty has increased recently withthe shift in monetary stance in the United Statesand changes in the political administrationsin several countries in the region, especiallyVenezuela (in 1998) and Argentina and Chilein late 1999. The beginning of a modest re-covery of growth in the region can be expectedaround the turn of the year and should reflectthe improvement in the global outlook, amoderation of political uncertainties, and amodest recovery in anticipated capital flowsfor 2000. However, the small island states inthe eastern Caribbean will face a difficult tran-sition period from their potential loss of pref-erential access to the EU for banana exports.

Europe and Central Asia. The region hasexperienced widespread repercussions fromRussia’s August 1998 devaluation and debtdefault. Russia’s sharp import compression hasaffected neighboring CIS countries and Tur-key, and to a lesser degree Poland and otherCentral European countries. Russian output

declines have been limited by gains in domes-tic production for import substitution, thoughnot yet by a recovery of non-oil exports. How-ever, export and government revenues havebeen greatly enhanced by the recent hike inoil prices. There has also been progress in debtrestructuring, and improvement on the policyfront has also paved the way for gaining anadditional round of funding from the IMF.GDP is expected to register positive growthof 0.7 percent in the CIS countries for 1999,an improvement over the projected 5.5 per-cent fall anticipated in March. In addition todevelopments in Russia, the increase is fur-ther supported by the beneficial effects ofhigher petroleum prices on hydrocarbon ex-porters in the Caucasus and Central Asia. Incontrast, growth in Central and Eastern Eu-rope has slowed more than expected (to 1percent), reflecting the earlier sluggishness ofthe EU economy and weakened import de-mand facing countries such as Poland, Hun-gary, the Czech Republic, and Turkey. Growthin Turkey has been particularly affected. Tradewith Russia has collapsed, and the country’sprospects have been dimmed temporarily bythe effects of the earthquake in August 1999.

The war in Kosovo had substantial effectson the wider Balkan region. Refugees streamedinto Albania and Macedonia. Both direct andtransit trade in the area were disrupted, im-pacting not only those countries engaged inthe conflict but also neighboring Romania andBulgaria. Tourism receipts also droppedsharply, especially in Croatia. Growth in 2000and 2001 is expected to recover markedly inCentral and Eastern Europe, rising to 3.2 per-cent in 2000 and 4.3 percent in 2001, spurredlargely by stronger growth in the EU. Amongthe states of the CIS the turnaround is expectedto be more gradual, as political uncertainty inRussia before the December 1999 parliamen-tary and June 2000 presidential elections con-strains the country’s ability to address furtherdebt restructuring and budgetary issues. Con-sequently, GDP for the region as a whole isanticipated to recover only gradually to 2.5percent in 2000 and 3.3 percent in 2001.

21

Middle East and North Africa. The regionhas been positively affected by recent devel-opments in oil markets. But it has also beennegatively affected by slow growth in the EU—the major export market for the region’s di-versified economies. Some deceleration ofgrowth from the 1998 rate of 3.2 percent isexpected, and 1999 growth is projected toreach only 2 percent. Earlier declines in oilprices had a significant negative impact on oil-exporting countries, forcing governments totighten fiscal policies, adjust or defend ex-change rates, and fund higher external imbal-ances with foreign reserves and new borrowing(see chapter 4). The tightening of OPEC quo-tas in April 1999 led to higher oil prices, butrestricted production (and export volumes),and this, combined with tighter fiscal con-straints, has contributed to lower 1999 growthrates for oil exporters.

In the short term the sharp rise in oil pricesin 1999 should continue to contribute to a re-covery and begin to ease the financial short-fall. However, oil exporters still face a fiscalconstraint and will need to exercise tight con-trol on spending, especially on subsidies. Im-proved growth in oil exporters favors thenon-oil exporting countries of the region thatreceive significant transfers in the form ofworker remittances. (These transfers have de-clined as GDP growth falls in oil-exportingcountries.) A notable positive trend, particu-larly among the Maghreb countries, is accel-eration of privatization and restructuringprograms that contributed to higher inflowsof foreign capital, particularly in the form ofdirect investment. These reforms and a resur-gence of growth in Europe in late 1999 andearly 2000 should lead to solid growth in ex-ports and stronger gains in private investment.Growth in the broader region is anticipatedto rise to 3.2 and 3.5 percent in 2000 and 2001respectively.

Sub-Saharan Africa. Despite continuedimprovements in political and economic fun-damentals, GDP growth for 1999 has beenrevised downward modestly to 2.3 percentfrom the 2.5 percent anticipated in March.

While stronger oil prices are benefiting a smallnumber of countries, they adversely affect theterms of trade for most. Meanwhile low pricesfor non-oil commodities, poor weather, civilstrife and turbulence related to the East Asiancrisis (particularly affecting South Africa) haveworsened near-term prospects for the major-ity of countries in the region. Recent perfor-mance falls well short of what is needed tomake inroads against poverty, and per capitaincomes will decline for a second successiveyear (–0.2 percent) in 1999.

At the same time, however, developmentsin 1999 augur well for near-term prospects,and growth is expected to rise to 3.1 percentin 2000. A pick-up in exports in 2000 and2001 should set the stage for a modest recov-ery. This expectation is based on several fac-tors, including higher prices for oil and othercommodities, the recovery in East Asia, andreviving demand in the EU. Following a de-cline of 0.3 percent in 1998, merchandise ex-port volumes are expected to rise by 3.8percent in 1999 and 6.3 percent in 2000. Ex-cluding South Africa and Nigeria, exportgrowth for the region is projected to average4.8 percent in 1999 and 2000.

More critical to medium- and long-termprospects, however, is the ongoing structuraladjustment throughout the region, which hasopened markets and had a major impact onproductivity, exports, and investment. ManySub-Saharan countries are likely to benefitfrom debt relief, allowing money to be spenton imports rather than on debt servicing. Ofthe 36 countries eligible for debt relief underthe enhanced HIPC Initiative, 30 are in Sub-Saharan Africa.

South Asia. Output gains in India havebeen better than anticipated in 1999, both inmanufacturing and in agriculture. These gains,plus the fact that the region’s largest countrysuffered few effects from the East Asian cri-sis, make it possible to upgrade projectionsby 1 percentage point in 1999 and 0.7 per-cent in 2000. Growth in the region is expectedto accelerate to 5.4 percent in 1999 (from 5.1percent) and to consolidate further in 2000,


22


largely because of developments in India. Thedirect effects of the East Asian crisis were lim-ited to the region’s smaller, more open econo-mies. Bangladesh, Pakistan, and Sri Lankashared in the global crisis during 1998–99,partly reflecting the weakness in world tradeand commodity prices. However, in thesecountries, too, the economic outcome owedmore to domestic than international factors.Fiscal and broader financial sector difficulties,sanctions, and the outbreak of hostilities inKashmir affected Pakistan, while severe flood-ing and political uncertainty took a toll onoutput growth in Bangladesh. In the near term,all countries of the region will be adverselyaffected by the surge in oil prices, but theyshould benefit from a jump in external de-mand, especially in Europe and, more gener-ally, in Asia.

Projections for growth indeveloping countries in the longterm are lower

The disruptions to global economic activ-ity, trade, commodity, and financial mar-

kets caused by the East Asian crisis are likelyto have diminished by 2001, with world out-put growth trending toward 2.8 percent from1.9 percent in 1998. But two issues that affectgrowth in developing countries remain. First,the external environment is projected to besomewhat less favorable than in the precrisisperiod and also more fragile—that is, it coulddeteriorate again. Second, the crisis has ac-centuated structural weaknesses in develop-ing countries, especially with respect to thefinancial sector and the government balancesheet (or at least has led to a more realisticappreciation of the problems). Principally forthese reasons, the long-term (2002–2008) fore-cast for growth in developing countries hasbeen reduced to 4.9 percent from 5.2 percent(see table 1.7 for the last long-term projec-tion, which was completed a year ago). 13 Evenwith this downward revision, the projectedfigures represent a significant increase from1991–98 levels for developing countries as agroup. However, this growth is fueled mainlyby increased growth in the transition econo-mies. Excluding the transition countries, de-veloping countries’ growth rates are projected

Figure 1.9 Growth of U.S. employment and productivity in manufacturing

Percent

8Productivity

(output per hour)

Employment

6

4

2

0

–2

–41996 1997 1998 Q3 1998 Q4 1998 Q1 1999

Annual growth Quarterly rates

Note: Quarterly growth rates are seasonally adjusted at annualized rates.

23

to decline from 5.3 percent in 1991–98 to 5percent in 2002–2008.

The external environmentLong-term projections assume slightly stron-ger growth for the industrial countries—2.5percent G-7 growth contrasted with 2.4 per-cent gains in earlier forecasts. Growth pros-pects for the United States and the Euro Areahave been upgraded, but a more substantialdowngrading of potential GDP growth in Ja-pan will offset these increases.

United States. Productivity growth haspicked up from 1 percent annually during the1980s to 1.5 percent over the last five years,rising to 2.2 percent in 1998. In the manufac-turing sector productivity gains have averaged

4.3 percent over the last three years—nearlydouble the 2.3 percent rate of the 1980s.Output per hour in the manufacturing sectorrose by an annual rate of 7.5 percent in thefourth quarter of 1998 (figure 1.9). While theevidence on the sustainability of these advancesremains open to question, it appears to justifya modest upgrading of long-term growth ratesto 2.7 percent from the 2.5 percent rate esti-mated in Global Economic Prospects 1998/99.

Europe. Several factors suggest thatgrowth may be considerably more robust thanthe 1.8 percent pace experienced during the1990s, and the 2.7 percent long-term growthrate forecast for the Euro Area in last year’sGlobal Economic Prospects has been revised


Table 1.7 World growth, 1981–2008(annual percentage change in real GDP)

GlobalEconomic

Global Economic Prospects 2000 Prospects1998/99

Estimate Forecasts

Region 1981–90 1991–98 1998 1999 2000 2001 2002–2008 2001–2007

World total 3.1 2.5 1.9 2.6 2.9 2.8 3.2 3.2

High-income countries 3.0 2.3 2.0 2.6 2.5 2.3 2.7 2.6OECD countries 3.0 2.2 2.0 2.6 2.5 2.3 2.6 2.5Non-OECD countries 5.2 5.7 1.2 3.0 4.0 4.6 5.2 5.1

Developing countries 3.3 3.2 1.6 2.7 4.2 4.5 4.9 5.2East Asia and Pacifica 8.1 8.5 0.1 5.5 6.2 6.2 6.3 6.3Europe and Central Asia 2.7 –4.0 –0.2 0.3 2.5 3.3 4.0 5.0Latin America and the Caribbean 1.6 3.6 2.1 –0.6 2.7 3.5 4.2 4.4Middle East and North Africa 0.7 2.9 3.2 2.0 3.2 3.5 3.7 3.7South Asia 5.7 5.8 5.1 5.4 5.5 5.3 5.1 5.5Sub-Saharan Africa 1.8 2.8 2.4 2.3 3.1 3.4 3.6 4.1

Memo itemsEast Asian crisis-5b 6.9 6.0 –7.9 4.4 5.3 5.1 5.3 5.2Transition countries of Europe

and Central Asia 2.5 –5.1 –0.8 0.8 2.1 3.1 3.7 4.8Developing countries, excluding

the transition countries 3.6 5.3 2.1 3.0 4.5 4.8 5.0 5.2

a. East Asia and Pacific includes the Republic of Korea.b. Indonesia, Malaysia, the Philippines, the Republic of Korea, and Thailand.Note: GDP is measured at market prices and expressed in 1987 prices and exchange rates. Growth rates over historic intervals arecomputed using least squares methodSource: World Bank staff estimates, November 1999.

24


upward 0.2 percentage points. The volatilityof European growth during the early 1990s,which was tied to the imbalances caused byunification of Europe’s largest economy, hasgradually faded. There has been considerableprogress in inflation reduction and fiscal con-solidation. Efficiency gains associated with theadvent of the euro (reduced transactions costs,deeper financial markets, and industry con-solidation to serve the larger market) are likelyto emerge with time. Labor market and otherregulatory reforms, in tandem with strongerinvestment in high technology—that will al-low Europe over time to catch up to the UnitedStates in this area—could support further pro-ductivity growth. And despite substantial tran-sition costs, the accession to the EU of five toten Central and Eastern European countrieswill bolster reforms in the new entrants andincrease the EU consumer market by some 100million people.

Japan. The prospects for a resumption ofrapid long-term growth in Japan appear dim.The ongoing efforts to overcome theeconomy’s deeper-seated structural problemsare unlikely to bear fruit for a number of years.Successive government stimulus packagesduring the 1990s have added to the debt bur-den but failed to spur growth: GDP has grownby an average of only 1.3 percent over the1990s to date, compared with 4 percent inthe 1980s. The fiscal balance has balloonedfrom a surplus of 2.9 percent of GDP in 1990to an estimated deficit of 9 percent in 1999,increasing gross government liabilities rela-tive to output by 40 percentage points to over100 percent. In the medium term fiscal con-solidation is unavoidable, especially in lightof the increasing demands being made on thepublic pension system by an aging popula-tion. Recent progress in stabilizing the pri-vate financial sector and the large-scalecorporate restructuring that will alleviate ex-cess capacity and restore financial viability willneed to continue. These adjustments are likelyto result in a period of sluggish output growthin both the public and private sectors, andprojections for trend growth have been revised

downward by 0.6 percentage points to 1.9percent.

Inflation in the G-7 countries is likely toremain low. Economy-wide efficiency gains inthe United States and Europe, expectations formoderation in fiscal balances (in the latter,owing to constraints imposed by Maastrichtcriteria and the Stability Pact), and moderateactivity in Japan should keep inflation below2.5 percent in the medium to long term. Intandem with low inflation and diminishingpressure on capital markets from the publicsector, real long-term interest rates are ex-pected to decline gradually from their currenthigh levels of 4–4.5 percent to near a secularaverage of 3–3.5 percent.

World trade growth. Despite strongeroutput gains in the United States and Europe,projections for world trade growth have notbeen revised markedly (an upgrade of 0.1 per-centage point to 6.4 percent). The projectionsreflect a low growth profile in East Asia rela-tive to precrisis trends and tighter external fi-nancing constraints on many emergingmarkets. These constraints may limit emerg-ing markets’ ability to import at rapid ratesover an extended period.

Long-term trends in trade will also beshaped by the eventual outcomes of the pro-spective Millennium Round of the WorldTrade Organization (WTO), for which frame-work discussions commenced in late 1999(box 1.2). Developing countries have substan-tially reformed their trade regimes in the lasttwo decades and now have a much greaterstake in advancing trade reforms than theyhave had in any previous round. Though theAsian crisis and its aftermath have had a pro-found effect on global trade and investment,it has not led to major reversals in trade policy.There is, however, little reason to be compla-cent. Significant trade imbalances still existthat could generate much deeper trade ten-sions if they are left unresolved. Moreover, ac-celerating the integration of theleast-developed countries into the globaleconomy will significantly enhance their pros-pects for long-term growth.

25

Commodity prices. The prospects for asharp recovery in commodity prices are notbright. Non-oil commodity prices will increaseonly an average of 3 percent in nominal terms(0.5 percent in real terms) and are not likelyto reach and maintain their 1995–96 peaksover the 10-year forecast period (figure 1.10).Increased demand in industrial countries willprovide some support for higher prices, butsupplies will continue to outpace demand.Technological improvements have reduced theproduction costs of many commodities, andcurrency devaluation in many important com-modity-exporting countries will contribute toincreased global supplies.

Even if OPEC continues to withhold out-put in the medium term to keep prices near$20/bbl, oil prices will be under almost con-stant downward pressure because of limitedgrowth in the demand for OPEC’s crude oiland increasing competition from both oil andnon-oil energy producers. Production costs areexpected to continue to decline because of fur-ther advances in technology (oil sands devel-opment, for instance), including those forcompeting fuels such as liquefied natural gas.This development effectively puts a ceiling on

oil prices of around $20 per barrel for the fore-seeable future.

Private capital flows. Through 2008 netprivate capital flows as a share of GDP areunlikely to exceed their precrisis highs ofaround 6 percent for middle-income develop-ing countries (“emerging markets”). But thereare a number of reasons private capital flowsshould recover from current levels. Techno-logical improvements in communications andinformation processing, financial innovation,and deeper financial intermediation shouldcontinue to be propitious for internationalcapital flows. A resumption of faster growthin developing countries will generate increasedopportunities for investment and hence de-mand for capital. Prospects for medium-termimprovement in the U.S. current account bal-ance, sustained current account surpluses indeveloped Europe, and a decline in real inter-est rates should increase the supply of fundsavailable for overseas investment, while Asiancurrent account surpluses are very likely todecline. In the long term the global supply ofsavings should increase with the aging of in-dustrial country populations.

Figure 1.10 Long-term trends in commodity prices, 1990–2008Current U.S. dollar indices; 1999=100

150

125

100

75

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

Source: World Bank.

Oil

Non-oil

Food

Metals


26


work for dealing with trade-related issues than itwas for the quantitative reductions it made in tradebarriers. And there has been some concern regard-ing backsliding—for example, “dirty tariffication”and back-loading of the Multi-Fibre Agreement(MFA) reforms. Moreover, developing countries,particularly the poorest, have been slow to imple-ment many elements of the agreement owing inpart to a lack of either human or financial re-sources. Replacing lost fiscal revenues, new report-ing requirements, and required modifications ofdomestic regulations and procedures have provento be a significant—and often overlooked—burden.However, progress has been made in some of thesectoral negotiations since the end of the UruguayRound, with agreements in telecommunications,financial services, and information technology. Anextension of the Information Technology Agree-ment has been postponed, along with the talks onliberalizing maritime services (in which developingcountries have a considerable stake).

A built-in agenda for future negotiations. Apotential agenda for the next round could easilyinclude two dozen topics but will necessarily bepared down to a shorter, more manageable list.The so-called built-in agenda was essentially deter-mined by the Uruguay Round Agreement. Specifi-

Box 1.2 Prospects for a new round ofmultilateral trade negotiations

The World Trade Organization (WTO) is on theverge of initiating a new round of multilateral

trade negotiations. The last round—known as theUruguay Round—was completed five years ago.While the resulting agreement marked some signifi-cant achievements, a wide range of trade-relatedissues remain to be discussed in a new round.

Past achievements and the role ofdeveloping countriesThe Uruguay Round Agreement led to four broadachievements:

1. further reforms of traditional trade barriers,notably in textile and apparel, services, andagriculture;

2. the establishment of frameworks for dealingwith less traditional trade barriers such as intel-lectual property rights and technical and sani-tary standards;

3. creation of the WTO, with its strong disputeresolution mechanism and strict discipline;and

4. increasing influence of developing countries.14

The Uruguay Round Agreement was perhapsmore important for improving the general frame-

The rise in flows from international capi-tal markets is likely to be gradual, however.The experience of the financial crisis and theshift from fixed to floating exchange rate re-gimes may encourage both borrowers andlenders to take a more cautious approach.Some developing countries are likely to moveslowly toward capital account liberalization,and those with full convertibility are likely toimplement prudential controls and regulations.Greater burden sharing with private investorsreduces moral hazard and increases investor’srisk perceptions with respect to investing inemerging markets. Stricter regulations forbanks in the BIS-area after the losses of recent

years, and diminished prospects of furtherderegulation of institutional investment indeveloping countries are also likely to dampenflows to emerging markets. Significant vola-tility in capital market flows can be expected,with the major risks arising from uncertain-ties in the world economic environment, par-ticularly the performance of the U.S. economyand developments in Japan.

Policies are likely to be more encourag-ing for foreign direct investment as countriescontinue to vie for foreign capital and skills.Efforts are being made to promote long-terminvestments rather than short-term flows.Some countries, including Indonesia and Ko-

27

Box 1.2 (continued)

cally, it calls for a renewal of negotiations in theagricultural and service sectors. There is wide-spread agreement—though with some dissentingvoices—that industrial tariffs will be included,both because they are still at significant levels inmany sectors and regions and because includingthem offers greater potential for trade-offs thatwill lead to broader, deeper reforms. A futureround will most likely concentrate on reducingtrade barriers rather than on reforming the insti-tutional framework set up in the previous round.

Reaching consensus on other agenda itemsmay prove more difficult. Developing countriesare particularly wary of entering into new agree-ments where the burden of implementation islikely to be heavy—for example, on customsvaluation. They are also disinclined to opennegotiations in areas that are perceived to intro-duce new barriers to subjects where obtainingagreement will be difficult to achieve for as-sorted reasons, for example, government pro-curement and state trading rules, trade-relatedinvestment measures (TRIMs), and trade-relatedintellectual property (TRIPS).

Developing country interests. The interestsof developing countries generally lie in a fewareas. Among these are:

• broad-based reduction in tariff and non-tariff barriers, though with a more specificconcern to limit the scope of escalating andpeak tariffs;

• opposition to any backtracking on the MFAphase-out;

• obtaining credit for past unilateral liberaliza-tion; and

• ensuring commitments to greater movement ofpersons (so called “mode 4” liberalization al-lowing, for example, easier movement of con-struction-related labor services).

Of greater relative importance for the least-developed countries will be technical assistance,capacity building, implementation of existingagreements, erosion of preferences, and specialand differential treatment. Differences are likelyto appear at the sectoral level, most particularlyin agriculture. Agricultural exporters, notablythose belonging to the Cairns Group, will want tosee further progress on agricultural trade reform.Food importers, however, will be concerned aboutthe impact of reduced agricultural export subsi-dies on the terms of trade.15


rea, have improved existing frameworks orestablished new ones to promote direct in-vestment and increase incentives for merg-ers, acquisitions, and corporate restructuring.Further progress can be expected inprivatization programs. Easing ownership re-strictions and opening up protected core sec-tors such as telecommunications and financeto foreign investors will further encourageforeign investment. In addition developingreal financial and regulatory infrastructureshould encourage long-term strategic invest-ments. For these reasons, foreign direct in-vestment is likely to increase in line with GDPgrowth.

Structural weaknessesThe financial crisis has unveiled profoundweaknesses not only in the financial systemsof developing countries but also in their abil-ity to intermediate domestic savings and largeforeign capital inflows. The crisis has contrib-uted directly to the rise in bad loans, weak-ened thousands of individual firms, andcreated many direct and contingent govern-ment liabilities. The size of these liabilities islikely to increase before it starts declining. Thefundamentals that support growth in somelarger developing countries thus offer morecause for concern today than they did beforethe crisis.

28


Developing countries face more than near-term adjustment measures in their efforts toestablish sound financial markets and a stron-ger underpinning for sustainable growth.Domestic saving and investment need to beencouraged. The fiscal costs of the crisis willneed to be recovered, while at the same timeresources will be required to boost the qualityof human capital and physical infrastructurenecessary to sustain growth, and institutionscapable of implementing reforms will need tobe strengthened. A review of these factors forseveral large developing countries has led tosome downgrading (0.3 percentage points) ofassessments for long-term growth from theprojections prepared a year ago. These revi-sions range from a full percentage point inECA, 0.5 points in Sub-Saharan Africa, and0.4 points in South Asia to 0.2 points in LatinAmerica. Compared with actual growth in theprecrisis period, the largest reduction in growthprospects is in East Asia, while most otherdeveloping regions are expected to see fastergrowth (table 1.7 above).

Although Russia achieved stabilizationmore quickly than anticipated in 1999, recentdynamics in the fundamentals required to sup-port long-term growth have deteriorated. Re-visions for Sub-Saharan Africa take intoaccount the likelihood of a more protractedpath to achieving long-term growth potentialin South Africa. The revised forecasts for LatinAmerica are based on increasing evidence of“reform fatigue” in a number of countries andthe need for longer-term fiscal adjustment inBrazil. More detailed information on the long-term forecasts for each region can be found inappendix 1.

Recent trends and prospects forpoverty in developing countries

Recent trends. The picture that emerges at theturn of the century is one of stalled progressfor the poor and of rising numbers of poorpeople in most developing regions. Accordingto recent World Bank estimates (World Bank

1999), the number of people in the world liv-ing on less than $1 per day in 1998—1.2 bil-lion—was virtually the same as in 1987.However, some changes did take place duringthis period. The number of poor rose to 1.3billion in the early 1990s, but then declinedto about 1.2 billion in 1996. The poor as ashare of population and the number of peopleliving on less than $1 per day had both de-clined substantially in the mid-1990s.16 Theglobal financial crisis halted this decline, andpreliminary estimates for 1998 indicate thatno further decrease has taken place since 1996(tables 1.8 and 1.9).17

The largest change in trend occurred inEast Asia and Pacific, the region at the centerof the crisis. The number of people in povertyin the region had fallen sharply before the fi-nancial crisis, from 432 million in 1993 to 265million in 1996. The crisis put an end to along period of rapid growth and led to sig-nificant increases in poverty in the most af-fected countries (see chapter 2). While the crisiscountries are now showing clear signs of re-covery, it is too soon to assess the implica-tions for poverty.

China—which accounted for 82 percentof East Asia’s poorest in 1996—continued togrow through 1998. The number of rural poorfell from an estimated 358 million in 1990 to208 million in 1997, an impressive achieve-ment. However, survey data for 1998 show aslight increase—from 211 to 213 million—inthe total number of poor in China between1996 and 1998.

In South Asia, the incidence of poverty(the share of the population living in poverty)declined moderately through the 1990s, butnot sufficiently to reduce the absolute num-ber of poor. The actual number of poor peoplein the region has been rising fairly steadily since1987. In India, home to almost half of theworld’s poor, the rate of poverty reductionappears to have slowed in the 1990s, particu-larly in rural areas. In addition, the gap be-tween some of India’s largest and poorest statesand the richer states is growing. India’s poor-est states exhibit slow progress in human de-

29

Table 1.8a Population living below $1 per day in developing and transition economies, 1987–98

Population Number of people living on less than $1 a day Headcount indexcovered by (millions) (percent)

at leastone survey 1987 1990 1993 1996 1998 1987 1990 1993 1996 1998

Region (percent) Estimate Estimate

East Asia and Pacific 90.8 417.5 452.4 431.9 265.1 278.3 26.6 27.6 25.2 14.9 15.3Excluding China 114.1 92.0 83.5 55.1 65.1 23.9 18.5 15.9 10.0 11.3

Eastern Europe and 81.7 1.1 7.1 18.3 23.8 24.0 0.2 1.6 4.0 5.1 5.1Central Asia

Latin America and 88.0 63.7 73.8 70.8 76.0 78.2 15.3 16.8 15.3 15.6 15.6the Caribbean

Middle East and 52.5 9.3 5.7 4.9 5.0 55 4.3 2.4 1.9 1.8 1.9North Africa

South Asia 97.9 474.4 495.1 505.1 531.7 522.0 44.9 44.0 42.4 42.3 40.0Sub-Saharan Africa 72.9 217.2 242.3 273.3 289.0 290.9 46.6 47.7 49.7 48.5 46.3

Total 88.1 1,173.2 1,276.4 1,304.3 1,190.6 1,198.9 28.3 29.0 28.1 24.5 24.0Excluding China 879.8 915.9 955.9 980.5 985.7 28.5 28.1 27.7 27.0 26.2

Table 1.8b Population living below $2 per day in developing and transition economies, 1987–98

Population Number of people living on less than $2 a day Headcount indexcovered by (millions) (percent)

at leastone survey 1987 1990 1993 1996 1998 1987 1990 1993 1996 1998

Region (percent) Estimate Estimate

East Asia and Pacific 90.8 1,052.3 1,084.4 1,035.8 863.9 892.2 67.0 66.1 60.5 48.6 49.1Excluding China 299.9 294.9 271.6 236.3 260.1 62.9 57.3 51.6 42.8 45.0

Eastern Europe and 81.7 16.3 43.8 79.4 92.7 92.9 3.6 9.6 17.2 19.9 19.9Central Asia

Latin America and 88.0 147.6 167.2 162.2 179.8 182.9 35.5 38.1 35.1 37.0 36.4the Caribbean

Middle East and 52.5 65.1 58.7 61.7 60.6 62.4 30.0 24.8 24.1 22.2 21.9North Africa

South Asia 97.9 911.0 976.0 1,017.8 1,069.5 1,095.9 86.3 86.8 85.4 85.0 84.0Sub-Saharan Africa 72.9 356.6 388.2 427.8 457.7 474.8 76.5 76.4 77.8 76.9 75.6

Total 88.1 2,549.0 2,718.4 2,784.8 2,724.1 2,801.0 61.0 61.7 60.1 56.4 56.0Excluding China 1,796.6 1,918.8 2,020.5 2,096.5 2,168.9 58.2 58.8 58.6 57.7 57.6

Note: The numbers are estimated from those countries in each region for which at least one household survey was available duringthe period 1985–98 (for many countries more than one survey was available). The proportion of the population covered by suchsurveys is given in the first column. Survey dates often do not coincide with the dates in the above table. To line up with the abovedates, the survey estimates were adjusted using the closest available surveys for each country and applying the consumptiongrowth rate from the national accounts. Using the assumption that the sample of countries covered by surveys is representative ofthe region as a whole, the numbers of poor are then estimated by region. This assumption is obviously less reliable in the regionswith the lower survey coverage. The headcount index is the percentage of the population below the poverty line. Further details ondata and methodology can be found in World Bank (forthcoming-a) and Chen and Ravallion (forthcoming).

velopment indicators, including health andeducation indicators; low growth rates, par-ticularly in the agriculture sector; inadequateinfrastructure; and weak and fragmented in-stitutions. Based on experience, these stateswill not see much impact on poverty even with

higher growth rates. If present trends continue,the bulk of the poor in these states will beunable to participate in future growth(Ravallion and Datt 1999).

The new World Bank estimates indicatethat Africa is now the region with the largest


30


share of people living on less than $1 per day,and prospects for improvement remain dim.While average growth rates rose during the1990s in many African countries, they remainbelow levels sufficient to reduce the numberof poor people. In other African countries eco-nomic growth remained low because of theproliferation of conflict, political instability,and, in some cases, adverse weather.

In Latin America and the Caribbean thepoverty rate has remained roughly constantin the 1990s, despite the acceleration in eco-nomic growth in many countries in the mid-1990s, and the number of poor increased.However, while the incidence of income pov-erty (material deprivation as measured by percapita consumption) has not declined, socialindicators have improved: adult literacy, lifeexpectancy, access to safe water, and infantmortality are now at levels consistent withwhat would be expected given the region’s levelof economic development.

In Brazil poverty fell by about 30 percentin the two years following economic stabili-zation in 1994. Poverty indicators then rosein the aftermath of the Asian crisis. This nega-tive trend continued after the Russian crisis,and will almost certainly have worsened dur-ing 1999. The most recent poverty measure-ments available from the monthly employmentsurvey show that the headcount rate in met-ropolitan areas in February 1999 was 10 per-

cent higher than before the onset of the Asiancrisis. One-third of the gains in poverty re-duction achieved after the Real Plan have beenundone. Brazil has a significant number ofsocial protection programs that provide somecompensation for many of those affected bythe crisis, but many of the most vulnerable,especially those in the informal sector, are notprotected. The unemployment rate showed aworrisome increase, reaching, and now level-ing off at, historically high levels of 7 to 8percent since 1998, compared with 3 to 4percent in 1993–96. Real wages, however, didnot decline significantly until the first quarterof 1999.

In Central and Eastern Europe and Cen-tral Asia the indications are that the upwardtrend in the incidence of poverty has leveledoff in line with the leveling off of the down-ward trend in GDP, although the estimates for1998 are tentative. In the countries of theformer Soviet bloc poverty rose markedly from1990 to 1996. Chronic poverty is emergingas a vital concern in the region, because evenin countries with a robust recent growth recordthe group of chronically poor appears to begrowing.

Real GDP per capita in Russia collapsedin the 1990s, declining by 41 percent from1990 to 1999. Furthermore, inequality—asmeasured by the Gini index—increased sharplyfrom 0.24 to 0.39 from the late 1980s to the

Table 1.9 Projected growth rates in real per capita private consumption and changes in Ginicoefficients for 1999–2008

Scenario A Scenario BSlow growth and rising inequality Inclusive growth

Growth rate Change in Growth rate Change in(percent per inequality (percent per inequality

Region annum) (percent) annum) (percent)

East Asia and Pacific 4.0 +10 4.9 0Eastern Europe and Central Asia 2.7 +20 3.7 0Latin America and the Caribbean 0.6 +10 1.7 0Middle East and North Africa 0.4 +10 1.5 0South Asia 2.4 +20 4.0 0Sub-Saharan Africa –0.1 +10 1.0 0

Source: World Bank.

31

mid-1990s (Rutkowski 1999). The number ofpeople living in poverty rose dramatically asa result of these two forces. The profile ofpoverty also changed during the transitionperiod, with large numbers of working andunemployed adults and their children joiningthe ranks of the “old poor” from before thetransition (Klugman and Braithwaite 1998;Milanovic 1999).

Prospects for poverty reduction. What arethe prospects for reducing poverty in the me-dium term? This section explores what mightoccur under different assumptions with re-spect to income poverty. The share of peoplewho will be living on less than $1 or $2 perday in the future depends on how much percapita consumption levels will change andwhether changes will affect people with dif-ferent levels of consumption equally or willaffect some groups more than others. Forexample, if average per capita consumptionlevels increase equally for all—the poor as wellas the rich—then the share of those consum-ing less than the threshold will decline. How-ever, if consumption levels increase for therich only, then the share of the poor will re-main unchanged. The processes that affecthow changes in aggregate consumption lev-els are distributed across the population arenot well understood, so forming a judgmenton how many people will be living in povertyin the future is difficult.

The World Development Report 1990 onpoverty (WDR 1990) (World Bank 1990)made projections for 2000 of the proportionsof the population that would be living on lessthan $1 per day in 1985 Purchasing PowerParity terms under the assumption that “thestrategy recommended in the report gainedwider acceptance” (World Bank 1990, 138).Under this assumption, the report forecast thatthe global poverty rate would fall from 32.7percent in 1985 to 18.0 percent in 2000,representing a compound rate of decline of3.9 percent per year. The Bank’s latest esti-mates indicate a fall in the poverty rate from28.3 percent in 1987 to 24.0 percent in 1998,implying a compound rate of reduction in

poverty of only 1.5 percent per year. So theWDR 1990 projections overestimated thesubsequent rate of poverty reduction, althoughthe report did state that “it would be pos-sible to do somewhat better—or much worse”(p. 138).

Where were the WDR 1990 projectionswrong? In terms of the aggregate numbers,China and India have the greatest weight, andboth experienced a slower pace of poverty re-duction than anticipated, even though growthwas actually higher than predicted. The mainreasons were rising inequality in China and adiscrepancy between growth rates in consump-tion as measured by the national accounts andby household surveys in India, with surveyconsumption growing much less than nationalaccounts consumption, and consequently farless poverty reduction than expected.18 In therest of the world Central and Eastern Europeand the former Soviet Union experienced nega-tive growth and rising inequality, while LatinAmerica and the Caribbean, the Middle Eastand North Africa, and Sub-Saharan Africaexperienced unexpectedly low growth.

The case of India is worth highlighting.During the 1990s the growth rates of con-sumption expenditure per person from theIndian National Sample Survey have beenappreciably lower than implied by the con-sumption component of the national accounts.Possibly this captures actual developments(reflecting the underlying differences in theconsumption concepts used by surveys versusnational accounts). More likely it stems fromdata problems in one or both sources. Forexample, if survey data fail to capture growthin expenditures at the high end of the distri-bution, they underestimate growth in bothaverage expenditure and inequality.

The scenarios for poverty in 2008 con-tained in this report follow a methodologysimilar to that employed in the WDR 1990,but the underlying assumptions incorporatethe lessons learned from that experience. Thesescenarios are also based on survey informa-tion for many more countries: 96 comparedwith 22 at the time of the 1990 projections.


32


However, the projections should not be treatedas forecasts, but rather as representing a plau-sible range of possible outcomes for povertybased on alternative assumptions aboutgrowth and inequality. Our understanding ofthe quantitative dynamics of changes in pov-erty and inequality remains incomplete (seethe end of this chapter for a discussion of theuncertainties inherent in projecting growth).There are also large uncertainties about therelationship between growth and inequalityand about changes in inequality because of thecomplexity of the forces at work. While mostcountries have experienced little change inaggregate inequality over time, this is gener-ally the result of powerful countervailingforces. To take one area of interaction, mostgrowing countries experience both a rise inthe relative demand for skills and a rise in therelative supply of skills as education expandsthat can lead to small or negligible changes ininequality as these effects balance out.19

We develop two scenarios for the nextdecade. Scenario A—slow growth and risinginequality—entails little progress in reducingthe total number of poor in keeping with theexperience of the last decade. Scenario B—inclusive growth—tries to capture what isachievable if the right combination of poli-cies and interventions leads to sustainedgrowth without increases in inequality.

In scenario A all regions experience rela-tively low growth rates because of cyclicalboom and bust episodes, and inequality in-creases. The recent experience of large vola-tility in growth rates in developing countries(see chapter 2), together with substantial near-term risks to the outlook, centered in the in-dustrial countries, suggest that longer-termgrowth rates could be lower than projected inthe base case discussed earlier, especiallyamong the developing countries. This growthscenario combines elements of the near-termrisk scenario—a “hard landing” for the U.S.economy and attendant spillover effects—witha pattern of medium-term recovery and sub-sequent relapse of global growth (table 1.10).The developing regions most adversely affected

are those with a high dependence on commod-ity exports (the Middle East and North Af-rica and Sub-Saharan Africa) or a reliance onforeign capital flows (Latin America and theCaribbean). Moreover, although less exposedto external developments, fiscal and financialdifficulties in China and India are heightened,spurring remedial policy responses thatdampen growth further during the downturns.

Scenario A also assumes that inequalityincreases from current levels (current levels arebased on the distribution of consumption fromthe latest available surveys for each country).While inequality has shown a marked recenttendency to increase in the transition econo-mies, this has not been true of all developingcountries: inequality has increased in some,but has fallen in others. However, scenario Acaptures the widespread concern about up-ward pressure on inequality across the devel-oping world by building in rising inequalityin varying amounts by region. The Gini coef-ficient is assumed to increase by 10 percent inall regions except Central and Eastern Europeand the former Soviet Union and South Asia,where it increases by 20 percent. In Centraland Eastern Europe and the former SovietUnion this is consistent with recent experience(although the increase in inequality appearsto be leveling off).20 In South Asia the 20 per-cent increase is in keeping with the rising in-equality observed in some countries in theregion, such as Bangladesh. For India we as-sume that the weaker effects of growth onpoverty observed in the 1990s also stem fromrising inequality that the survey data do notfully capture. Under this scenario, inequalityin South Asia reaches the levels found in re-gions of medium inequality, such as East Asiaand the Middle East and North Africa. LatinAmerica and the Caribbean and Sub-SaharanAfrica remain the two regions with the high-est average inequality. Note that this scenarioimplies increasing inequality also in regionswith very low or negative growth rates (LatinAmerica, the Middle East and North Africa,and Sub-Saharan Africa). As the recent ex-ample of Russia indicates, the social implica-

33

tions of rising inequality in the context ofworsening standards of living are much moreworrisome than if they are accompanied by ageneral increase in living standards.

Scenario B uses the growth forecasts inthe base case discussed earlier. The base caseposits a fairly smooth growth path for bothindustrial and developing countries, beyond

near-term recovery from the recent episode offinancial crises, toward potential growth ratesby the end of the 10-year forecast horizon.Contrary to current concerns, scenario B as-sumes that inequality remains unchanged, ashas been the case in many countries over longperiods, and even in the countries of the formerSoviet Union recent evidence suggests that

Table 1.10a Population living below $1 per day in developing and transition economies for 1998–2008under scenarios of slow growth and rising inequality (Scenario A) and inclusive growth (Scenario B)

Number of poor Headcount index(millions) (percent)

1998 2008 1998 2008

Region Estimate Scenario A Scenario B Estimate Scenario A Scenario B

East Asia and Pacific 278.3 182.8 72.1 15.3 9.2 3.6 Excluding China 65.1 58.3 18.2 11.3 9.2 2.9Eastern Europe and Central Asia 24.0 45.7 7.4 5.1 9.6 1.6Latin America and the Caribbean 78.2 130.8 74.7 15.6 22.9 13.1Middle East and North Africa 5.5 11.4 4.7 1.9 3.3 1.4South Asia 522.0 465.0 205.9 40.0 31.0 13.7Sub-Saharan Africa 290.9 406.2 329.8 46.3 51.5 41.8

Total 1,198.9 1,241.8 694.7 24.0 21.9 12.3 Excluding China 985.7 1,117.3 640.8 26.2 25.9 14.9

Table 1.10b Population living below $2 per day in developing and transition economies for 1998–2008under scenarios of slow growth and rising inequality (Scenario A) and inclusive growth (Scenario B)

Number of poor Headcount index(millions) (percent)

1998 2008 1998 2008

Region Estimate Scenario A Scenario B Estimate Scenario A Scenario B

East Asia and Pacific 892.2 632.0 482.7 49.1 31.8 24.3 Excluding China 260.1 218.3 169.8 45.0 34.5 26.8Eastern Europe and Central Asia 92.9 100.8 46.3 19.9 21.2 9.7Latin America and the Caribbean 182.9 227.3 183.9 36.4 39.8 32.2Middle East and North Africa 62.4 74.7 47.8 21.9 21.7 13.9South Asia 1,095.9 1,083.0 945.4 84.0 72.2 63.0Sub-Saharan Africa 474.8 604.2 568.0 75.6 76.6 72.0

Total 2,801.0 2,721.9 2,274.1 56.0 48.0 40.1 Excluding China 2,168.9 2,308.2 1,961.2 57.6 53.5 45.5

Note: Scenario A—slow growth and rising inequality—entails little progress in reducing the total number of poor, in keeping withthe experience of the last decade. Scenario B—inclusive growth—tries to capture what is achievable if the right combination ofpolicies and interventions leads to sustained growth without increases in inequality.Source: World Bank.


34


inequality is stabilizing. Thus the scenariodescribes what could be achieved if countriesadopted policies and interventions that fos-tered inclusion, so that all benefited equallyfrom growth. (See box 1.3 for more detail onthe assumptions underlying the two scenarios.)

The assumptions underlying the two sce-narios are reported in table 1.10. Tables 1.8and 1.10 show the resulting poverty projec-tions for $1 and $2 per day. For ease of refer-ence, table 1.11 reports population figures.

The results of the two scenarios are verydifferent. Under scenario A the number ofpeople living in poverty would remain virtu-ally unchanged, as in the experience of the pastdecade. In 2008, 1.2 billion people would stillbe living on less than $1 per day. The regionalcomposition, however, would change consid-erably. The number of poor in Sub-SaharanAfrica would increase dramatically from 291

to 406 million people—almost 52 percent ofthe region’s population—and numbers wouldalso increase elsewhere except in East andSouth Asia. Similarly, more than 2.7 billionpeople would still be living on less than $2per day, more than a billion of whom wouldbe living in South Asia alone, with about 600million more in Sub-Saharan Africa and 400million in China. In Latin America, Centraland Eastern Europe, and Central Asia, boththe incidence of poverty and the numbers ofpoor would increase, while in the Middle Eastand North Africa a minor reduction in inci-dence would be inadequate to reduce the num-bers of poor.

Scenario B yields a brighter picture. Thedifference is large: the number of people liv-ing on less than $1 per day declines to about700 million by 2008, and the number of thoseliving on less than $2 per day to about 2.3

can be sensitive to the way inequality changes.21

This can be a serious problem when the povertyrate is low, because estimates at the tails of thedistribution can naturally be quite sensitive. Forthe bulk of the developing countries, however, thepoverty rates are a safe distance from the tails.

Third, the impact of changes in growth andinequality on the incidence of poverty is projectedusing a model of the distribution of consumptionfitted to the data for each country at each basedate, rather than using fixed growth and inequalityelasticities of poverty.22 Researchers have foundthat these elasticities can change substantially overtime, and they vary from country to country; thus,the model used lets the elasticities vary over timeand between countries consistently with the data.For example, the elasticity of poverty to growthtends to be lower (in absolute value) in high-in-equality countries (Ravallion 1997). The rate ofpoverty reduction at any given rate of growth willthen be lower in countries with initially more un-equal distributions, and it will tend to fall wheninequality increases.

Box 1.3 Technical discussion of assumptions

This box provides comments on the assumptionsunderlying the poverty projections. First, the

projections do not allow for any correlation be-tween growth rates and changes in inequality, forexample, higher inequality accompanying highergrowth. Experience does not suggest that a statisti-cally significant correlation exists across develop-ing countries as a whole (Ravallion and Chen1997). In reality, most countries have grown with-out experiencing any long-term increase in inequal-ity. The World Development Report 2000/01(World Bank forthcoming-b) will explore in moredepth the linkages between inequality and growth.

Second, the projections are based on an as-sumption about the precise way inequality changes,namely, that the distribution of per capita expendi-tures (the Lorenz curve) shifts by an equal propor-tion at all points. This is only one way in whichinequality may change. For example, an increase ininequality may affect only the nonpoor, in whichcase the poor would maintain their share of in-come, and poverty would decline just as it wouldwith no change in inequality. Poverty projections

35

billion (a smaller decline than in the numberof those living on less than $1 per day). How-ever, even under the more optimistic assump-tions underlying this scenario, progress in LatinAmerica and the Caribbean, and especially inSub-Saharan Africa, is inadequate to make sig-nificant inroads into the numbers of the poor,with a continued increase in numbers in Sub-Saharan Africa in particular.

A comparison of the two scenarios illus-trates the high degree of uncertainty aboutwhether the future pattern of development willbe accompanied by serious progress in reduc-ing poverty and the significant risk that theinternational development target for reducingincome poverty will not be achieved (see box1.4). In addition, the range of what might rea-sonably occur is even larger than portrayed.One could argue that if structural and socialproblems are not effectively tackled in SouthAsia and East Asia the relatively robust growthnow forecast under the more pessimistic sce-nario would be reduced further. However,based on experience from across the develop-ing world, more rapid growth and reductionsin inequality are, in principle, achievable inLatin America and the Caribbean and Sub-Saharan Africa, as well as elsewhere.

The fundamental message of the scenariosis the centrality of effective public action atthe international and country levels to develop

the institutions and policies that will bringabout inclusive growth. The downside risksare devastating for the prospects of millionsof people in the developing world, those nowliving in desperately poor conditions; thosewho would be born into a life of poverty; andthose at risk of falling into poverty because ofthe national, local, and personal risks that willcertainly persist.

What policies and interventions could leadto a pattern of rapid and equitable growth?This is one of the central topics of the forth-coming World Development Report 2000/01(World Bank forthcoming-b) on poverty anddevelopment and the issues can only betouched on here. While significant uncertaintysurrounds the quantitative dimensions of fu-ture changes in poverty, we know a great dealabout the kinds of public action that are ef-fective in achieving inclusive development. Atthe international level measures to ensuresteady growth in demand for products pro-duced by the developing world are crucial. Thiscan be fostered through sustained growth inthe industrial countries; opening of trade, es-pecially in agriculture, but also in other la-bor-intensive activities; actions that reducevolatility; and, of great importance for low-income, aid-dependent countries, the effectiveimplementation of an enhanced HIPC Initia-tive within the framework of an overall pov-erty-oriented program of internationalassistance.

At the national level, rapid inclusivegrowth requires institutions and policies thatboth encourage high levels of private and pub-lic investment to create jobs, services, and theinfrastructure necessary to expand opportu-nities for the poor and leads to gains in theirhuman and physical assets. Examples are poli-cies that reduce disparities in growth ratesbetween urban and rural areas by fosteringthe development of the rural nonfarm sector,policies that ensure access to good qualityeducation for all and an equitable distribu-tion of productive assets such as land, mea-sures that tackle the economic and physicalinsecurity that the poor face, and policies that

Table 1.11 Population estimates andprojections, 1998 and 2008

Population(millions)

1998 2008Region estimated projected

East Asia and Pacific 1,817.1 1,987.4Excluding China 578.5 632.7

Eastern Europe and Central Asia 466.1 475.5Latin America and the Caribbean 501.9 571.1Middle East and North Africa 285.1 344.1South Asia 1,305.3 1,500Sub-Saharan Africa 628.3 788.7

Total 5,003.8 5,666.9Excluding China 3,765.2 4,312.2


36


foster mechanisms that give a voice to the poorat the local level and ensure that formal insti-tutions respond effectively to their demands.

Many observers would judge that India,for example, which is central to the globalpicture of poverty, is vulnerable to pressuresfor rising inequality even, or perhaps especially,if it undergoes rapid overall growth. For struc-tural and institutional reasons the poor maybe particularly ill-equipped to participate insuch growth, notably because of low levels ofeducation and health, and because many areliving in states with weak institutions, a heri-tage of distorted policies, and complex anddeep social divisions. To include the poor inthe growth process, the government wouldhave to confront the large differences betweenrapidly growing and laggard states and thedismal state of the education system and otherpublic services, especially in the poorer states,and ensure that local elites did not capturedecentralization processes. In China the gov-ernment would have to confront poverty inthe more backward regions and among mi-nority groups. If equitable growth wereachieved, the number of poor people in SouthAsia would be cut by more than half and inChina to a fourth of current levels.

As noted earlier, in Sub-Saharan Africaand Latin America even the combination of asmooth transition to potential growth and noincrease in inequality in scenario B would notlead to a reduction in the number of the poor.Faster growth rates and a reduction in inequal-ity would be needed. In Latin America a re-duction in inequality of the order of 10 percentcould reduce the number of people living onless than $2 per day from 184 to 142 million.In Sub-Saharan Africa that would not be suf-ficient: a decline in inequality of 10 percentor more and an increase in growth rates of 20percent or more above the assumptions of sce-nario B would be needed just to keep the num-ber of poor people constant.

If East Asia and South Asia achieve therates of growth and changes in inequality ex-plored above, the regional composition of pov-erty would change markedly, with a large rise

in Sub-Saharan Africa’s share of the world’spoor.

Note that we have only examined theimplications of the expected growth rates forthe total number and share of the income poor.This hides many issues. Even when the aggre-gate poverty rate is falling, there will typicallybe both losers and gainers among the poor,reflecting heterogeneity in the circumstancesof poor people. Even when the incomes of poorfamilies are rising rapidly, they may not beable to get adequate health care or schoolingfor their children, because not enough of theeconomy’s growth is being used to improvekey public services, with further implicationsfor the sustainability of the income povertyreduction. Conversely, even with slow growthin income poverty, effective public action canresult in gains in other dimensions of well–being. The results presented here give us abroad picture, but more detailed micro-economic analysis is needed to understandsuch diverse impacts and complete the overallpicture of how the living conditions of the poorare evolving.

Risks to the forecast and alow-case scenario

In light of the volatility of the internationalenvironment, the macroeconomic forecasts

discussed above are subject to various risks.An analysis of the forecasting errors that oc-curred in relation to the onset and eventualdepth of the crisis in East Asia illustrates thesignificance of these risks in an environmentof deeper financial integration (box 1.3). Alow-case scenario is developed to highlight theprincipal risks attached to the baseline fore-cast. The potential for economic disruptionsdue to the “millennium bug” is discussed inbox 1.4.

Recent signs of recovery from the globaleconomic crises of 1997–98 are encouragingbut obvious vulnerabilities remain. Asia is be-ginning to recover, but its structural weak-ness persists. Latin America is still in recession,and several crisis countries remain exposed

37

to a sudden change in sentiment. The mostlikely global scenario involves an upturn ingrowth for both industrial and developingcountries over the next few years. But the un-derpinnings of growth, especially in the de-veloping countries, remain fragile. Capitalflows to emerging markets continue to bescarce and expensive. In such an environment,

the prospective unwinding of large imbalancesin the industrial countries present the clearestpotential risks for these projections. Chiefamong these risks are the consumption boom(which is being driven by the stock market)and widening external deficit in the UnitedStates, and the continuing uncertain outlookfor Japan.

achieve the target. Only East Asia and Pacific (un-der the assumptions of relative fast growth in thatregion) would extend its great gains for the 1990sand clearly reach the goal. By contrast, if the sce-nario of more rapid and inclusive growth were tooccur (scenario B), the target could be also beachieved in South Asia and in the world as a whole(driven by the potential gains in Asia), but neitherLatin America nor Sub-Saharan Africa would be ontrack to reach the target. As noted, these findingsare not predictions, but are intended to underlinethe centrality of achieving inclusive development inall countries and the magnitude of the challenge inregions with weaker prospects, especially Sub-Saharan Africa.

Box 1.4 Can the international development targetfor reducing income poverty be achieved?

The international development target for incomepoverty, one of the targets of the International

Development Goals, is to reduce the proportion ofpeople in absolute poverty by half between 1990and 2015. 23 At a global level, this is interpreted asreducing the share of people living below $1 perday (at the national level using national povertylines would be appropriate). The projection exer-cise undertaken for this report does not explicitlyassess the achievability of this target (see Demeryand Walton 1998). However, the implications ofthe scenario exercise are illustrative. Under a sce-nario with a plausible, but pessimistic, range ofassumptions on growth and inequality changes(scenario A), the world would not be on track to


Table 1.12 World, industrial, and developing countries in the low-case scenario(annual percentage change)

Estimate Low-case scenario Base-case scenario

Indicator 1999 2000 2001 2002 2000 2001 2002

Real output (following adjustment)World 2.6 1.2 1.2 2.9 2.9 2.8 3.0

G-7 countries 2.6 0.6 0.4 2.2 2.4 2.1 2.4Developing countries 2.7 2.3 2.9 4.8 4.2 4.5 4.8

Sub-Saharan Africa 2.3 2.0 2.4 3.1 3.1 3.4 3.4East Asia 5.5 4.5 4.3 6.4 6.2 6.2 6.2

ASEAN-4a 1.6 2.4 2.5 5.6 4.4 4.8 5.2South Asia 5.4 3.9 3.9 4.6 5.5 5.3 5.3Europe and Central Asia 0.3 1.0 2.3 3.9 2.5 3.3 3.6Latin America and the Caribbean –0.6 –0.3 0.3 4.1 2.7 3.5 4.4Middle East and North Africa 2.0 1.4 2.1 3.5 3.2 3.5 3.6

a. Indonesia, Malaysia, the Philippines, and Thailand.Source: World Bank staff estimates, November 1999.

38


magnitude of change across a number of importantmacroeconomic dimensions (see table below). Chiefamong these were:

• the extent of the decline in domestic demand,especially the massive drop in investment;

• the unexpectedly severe 18 percent decline inimport volumes;

• weaker growth in export volumes 26 and thesharp decline in dollar export prices, which ledto overestimation of export values expressed indollars; and

• the large shift to current account surplus (theexpectation that they would come to nearbalance in 1998 turned out to be widely offthe mark).

Several of the crisis countries had enjoyed avirtually uninterrupted period of growth (in the 6–8 percent range) for nearly 30 years. Thus, a forecast that GDP would decline, albeit modestly, rep-

Box 1.5 Failing to forecast the severityof the East Asian crisis 24

Much has been written about the failure topredict the outbreak of the East Asian crisis,

but the virtually universal failure to anticipate theseverity of the crisis once it had erupted may havebeen an even greater shortcoming. This sectionattempts to identify the main sources of error inWorld Bank forecasts completed at the end of1997, when the crisis had already spread through-out Southeast Asia and Korea. These forecastsmesh together views developed by regional eco-nomic specialists with a more systematic large-scaleglobal macroeconomic model. A significant portionof the failure is assigned to the inadequate repre-sentation of financial markets in macroeconomicmodels, whether explicitly formulated or represent-ing the way economists interpreted events.

A January 1998 25 analysis of the likely macro-economic consequences of the financial crisis cap-tured the direction of change in most countries andwas able to anticipate some of the main features ofthe subsequent world crisis. But it badly missed the

Analysis of forecast errors for 1998, East Asia-5

Estimate Actual Difference

GDP growth (percent) –0.2 –7.7 –7.5Domestic demand –6.0 –15.0 –9.0

Total investment –13.5 –31.3 –17.8Exports GNFS volumea 18.0 10.6 –7.4Imports GNFS volumea –0.8 –18.0 –17.2

Current account balance (billions of U.S. dollars) 7.3 66.2 58.9Percentage of GDP 0.8 10.8 10.0

Merchandise exports (billions of U.S. dollars) 392.5 329.6 –62.9Merchandise imports (billions of U.S. dollars) 362.8 268.8 –94.0

Balance of trade (billions of U.S. dollars) 29.7 60.8 31.1Export price in U.S. dollars (percent) –5.9 –15.8 –9.9Import price in U.S. dollars (percent) 1.2 –8.2 –9.4Terms of trade (percent) –7.0 –8.3 –1.3

a. GNFS — goods and nonfactor services.Source: World Bank staff estimates.

39


Box 1.5 (continued)resented a serious deviation from the trend growthrate. It is clear in hindsight that the assessmentsfailed to incorporate some important features ofthe Asian economies and of the channels for trans-mitting the crisis to countries both within and out-side the region. The forecast errors may be tracedto three interrelated factors:

The interactions of foreign credit and the do-mestic financial system. The extent of the reversalof capital flows and the financial panic induced byfears of a currency devaluation were not adequatelyanticipated. Even if they had been, the implicationsof these large changes in the corporate and house-hold balance sheets on investment and consump-tion would not have been easy to quantify. Thusthe profound effects of the credit squeeze were notpredicted. Curtailment of working capital andexport credit—which were not reflected in themodel—appear to have contributed to the mutingof the export supply response.

Spillover effects. The extent and rapidity ofspillover effects from Thailand to neighboringcountries and Korea contributed to underestimatesof the depth of the recession in general and theweaker than expected performance of exports in

particular. The mechanisms through which some ofthese spillover effects affect international financialmarkets are not clearly understood, and they arenot modeled beyond the standard linkages throughinternational trade, prices, and interest rates. Evenwhere bilateral trade linkages were small, the Thaidevaluation was seen as a warning that countries insimilar position could be hit.

The regional downturn. First, in a region inwhich 50 percent of trade is conducted amongregional partners (including Japan), the simulta-neous downturn in important export markets andcurrency devaluation offset some of the anticipatedboost to competitiveness and export growth. Sec-ond, the extent of the decline in dollar-based exportprices from the region placed additional constraintson dollar revenues, pushing more adjustment of thecurrent account to the import side. Third, despiterenewed uncertainties, Japan was anticipated togrow at more or less 1 percent in 1998, when infact GDP contracted by almost 3 percent. Therecession in Japan is estimated to have resulted in aswing to the negative of about 5 percentage pointsin export growth for the five East Asian crisiscountries.

A heuristic decomposition of GDP growth forecast errors, 1998(percent)

Thailand Malaysia Rep. of Korea Indonesia Philippines

GDP growth forecast error –4.5 –10.4 –6.0 –12.3 –3.6Effects of Japan recession –0.8 –1.5 –0.5 –0.8 –1.0Effects of regional downturn –1.0 –2.8 –0.5 –0.8 –2.2Terms of trade (percentage of GDP) –0.7 –0.5 –1.0 –0.5 –0.1Residual effects –2.0 –5.6 –4.0 –10.2 –0.3

Memo itemsProportion explained by Japan,

regional downturn, and terms of trade 55 45 33 17 90Proportion attributable to balance sheet

and further contagion effects 45 55 66 83 10

Source: Dadush, Riordan, and Wolfe forthcoming.

40


correct affected software programs and imbeddedchips, and several recent papers report a low prob-ability of major disruptions in economic activity inthe new year. 31 The United States may have re-solved nearly all Y2K problems in financial institu-tions, electric utilities, aviation, telecommunica-tions, and the federal government. However,problems remain in local government, health care,education and small businesses. 32 Similarly, mostof the industrial countries report that by December1999 Y2K related problems will be fixed in mostsectors, particularly the key sectors of finance,energy and air transport. 33 However, the Interna-tional Y2K Cooperation Center reports that errorsin business operations (such as in billing) maycause backlogs and delays in operations and in-creased business costs. Forecasts of the impact ofthe Y2K bug are subject to considerable uncer-tainty, given the immense numbers of systems af-fected, the difficulties in finding all potential prob-lems and in testing systems that have been fixed,and the inevitable reliance on self-reporting ratherthan independent verification.

Box 1.6 The possible impact of the Y2Kbug on developing countries

One risk to the short-term forecast is the poten-tial impact of the Y2K bug, a computer mal-

function that may affect date-sensitive computerprograms. For two decades computer programmersused only two digits to refer to calendar years. Thispractice saved precious memory and increasedprocessing speed but made many programs suscep-tible to failure in the year 2000, since computersmight interpret the digits “00” as 1900 rather than2000. This problem could have a drastic impact onoperations supported by computers, causing any-thing from the transmission of incorrect data to thecomplete collapse of essential systems. The actualprovision of critical services such as power, water,and telecommunications is not very dependent ondates, and these services were the first targets ofremediation efforts. Severe interruptions in theseareas are not likely. Problems are more likely tooccur in support operations (such as billing andrepair systems) that could require time-consumingrepairs and degrade services over time. 30

Governments and private sector firms in theindustrial countries have made extensive efforts to

Many alternative scenarios for globalgrowth and financial flows are possible. Anextreme yet plausible “downside” case illus-trates potential changes in the financing re-quirements of developing countries within aglobal environment of crisis—triggered inthis case by developments in the industrialcountries.

One potential scenario envisages contin-ued rapid growth in U.S. domestic demandover the remainder of 1999 and early 2000,fuelled in large part by equity market gains.In the context of tight labor markets, globalrecovery, and firming commodity prices, thisgrowth clearly signals that inflation will ac-celerate. In the early months of 2000, the Fed-eral Reserve responds assertively to the

potential upturn in inflation by increasing theFed Funds rate by 100 basis points. Marketparticipants overreact in their reassessment ofequity valuation levels in light of changes inthe prospective growth environment, and eq-uity prices fall by some 30 percent. In a sec-ond and ensuing response, the Federal Reservelowers rates by some 200 basis points to re-store market confidence, and the dollar fallsby 15 percent against major partner curren-cies. In consequence:

• These developments are transmitted rap-idly to equity markets, and the effect oneconomic activity in Europe and Japan isimmediate. Wealth effects in all three blocsdampen consumption growth, especially

41

in the United States, and investment slowssharply.27 The incipient European recov-ery is muted, while Japan’s lack of fiscalheadroom leads to a relapse into re-cessionary conditions. Growth in the G-7 falls 1.8 and 1.7 percentage points belowthe baseline in 2000 and 2001 respectively,while G-7 import demand drops 4 and 3.5percentage points below base in the sameyears.

• For developing countries, effects are trans-mitted through a further slowing in ex-port market growth, declines in oil andnon-oil commodity prices because of de-teriorating demand conditions, and in-creased risk aversion in financial markets.Risk aversion reduces emerging markets’

access to financing from sources otherthan direct investment. For low- andmiddle-income countries, these develop-ments open an ex ante financing gap—that is, an increased need for additionalfinancing to maintain domestic demandin the face of adverse external shocks. Thatneed peaks at some $100 billion in 2001.28

Additionally, a simulation of the supplyside of international capital markets, us-ing the assumptions of the low-case sce-nario, suggests a reduction in flows of $75billion versus baseline levels.29 In effect,the imbalance between supply and de-mand in external financing for develop-ing countries widens to $175 billionagainst the baseline.


Box 1.6 (continued)Already there are indications that the millen-

nium bug is having some impact on financial mar-kets, and developing countries’ access to externalcapital has been adversely affected as the year endapproaches. The markets are pricing in a liquidityshortage in the new year, since forward curvesanticipate a rise of 50 basis points in the U.S. dol-lar LIBOR in January 2000. This situation mayreflect the potential for Y2K disruptions or concernthat public anxiety over the millennium bug couldincrease the demand for cash.

Remedial measures in developing countriesappear to be less complete. Awareness of the risksposed by the millennium bug has come later thanin industrial countries. In the last two years gov-ernments in a few of the larger countries and theWorld Bank (through the infoDev program) havehad some success in increasing awareness of theY2K problem in the developing world. Devel-oping countries are less reliant on computer sys-tems than industrial countries, but they havedevoted fewer resources to fixing their systems,in part because workers with the required exper-tise are scarce, and many have been attracted bybooming demand for their skills in industrialcountries.

In August 1999 most of the 72 developingcountries responding to questions on Y2Kreadiness reported only limited vulnerability.34 Allof these countries reported that Y2K-relatedproblems would be resolved by December 1999.The U.S. State Department has carried out anevaluation of potential Y2K problems in 106developing countries (which represent 87 percentof total GDP in the developing world), focusing onthe prospects for key sectors. Countriesaccounting for 32 percent of the sample GDP hada low risk of economic disruptions, countries with57 percent of sample GDP moderate risk, andcountries with 10 percent of sample GDP highrisk. 35 The regions with the most countries withhigh risk were ECA and Sub-Saharan Africa. Bycontrast several of the Latin American countrieshad low risk, and only a few moderate risk. It isimpossible to translate these subjectiveimpressions of the risk of economic disruptionsinto plausible forecasts of the impact of Y2K ondeveloping countries’ growth. The risks of Y2Kcausing an economic downturn in developingcountries as a group appear to be remote, butnevertheless it is prudent to anticipate significantproblems with the new year.

42


Although policy responses to these exter-nal circumstances would vary widely acrossdeveloping countries depending on currentconditions, most countries would be obligedto adjust through a compression of domesticdemand and imports. An assumed closure ofthe financing gap on the demand side (almost$100 billion) results in a loss of 2 percentagepoints of growth for developing countries asa group in both 2000 and 2001, implying aloss of nominal GDP of some $260 billion.Latin America and the Caribbean is hardesthit, and continues in recession with growthrates 3 percentage points lower, while the na-scent recovery in the East Asian crisis coun-tries slows considerably (table 1.12).

World GDP growth at 1.2 percent in 2000(in technical terms, nearly a global recession)under such a scenario would mark the weak-est performance since the crisis years of 1982–83, when high interest rates and debt stalledthe world economy. And recovery of worldtrade and activity would likely be protractedthrough 2002, as policy responses in the in-dustrial countries take effect with some lag.During this interval, and despite macroeco-nomic adjustment efforts among developingcountries to stem widening current accountdeficits, a $75 billion shortfall in external fi-nancing persists (the reduction of private flowsfrom the supply side). This gap would needbe filled by some combination of reserve draw-downs among developing countries, substan-tial increases in counter-cyclical funding fromofficial sources, both multilateral and bilat-eral, as well as additional ex post macroeco-nomic adjustment efforts by developingcountries.

Notes1. Global Development Finance. April 1999.

Development Prospects Group, World Bank, Washing-ton, D.C.

2. Aggregate real per capita incomes in 1999 areexpected to decline or stagnate particularly in LatinAmerica, Sub-Saharan Africa, and the former SovietUnion, as well as in selected oil exporting countries.

3. Though recent data and definitional revisionshave brought the household saving rate to positiveterritory.

4. Calibrated to the 1995 Global Trade AnalysisProject (GTAP) database.

5. See McKibbin and Martin (1999) for a simi-lar analysis that also includes the effects of financialtransmission.

6. The demand shock in EA-5 and Japan wassimulated by shocking factor productivity.

7. Relative price changes are with respect to theprice of industrial countries’ manufacturing exports.

8. Evidence suggests that there was more adjust-ment in so-called necessities than this analysis indi-cates. In particular, it appears that households adjustedconsumption patterns to preserve educational expen-ditures, which unfortunately are not identified sepa-rately in the household consumption basket.

9. Chapter 4 contains a broader discussion of theimpact of the crisis on commodity markets.

10. FDI approvals declined 80 percent from their1998 levels in Indonesia in the first quarter of 1999,21 percent in China during January–July 1999, and17 percent in Thailand during January–May 1999.

11. For more information, see the website for theinitiative at (www.worldbank.org/hipc).

12. In this index of financial conditions in emerg-ing markets, a currency appre ciation is taken as in-dicative of looser financial conditions, reflectingreduced capital outflows or increased inflows.

13. World Bank 1998, Chapter 1.14. Membership, as of November 15, 1999,

stands at 135 countries—over three-quarters of themdeveloping economies. Forty-five have joined since thestart of the Uruguay Round negotiations in the mid-1980s. With the recent landmark trade agreement be-tween China and the United States, the former isexpected to join the WTO by early 2000.

15. The World Bank is actively assisting develop-ing countries in their preparations for the next tradenegotiation round. For further information on theseactivities and the World Bank’s broader researchagenda on trade, see (www.worldbank.org/trade).

16. When shifting from 1985 to 1993 PurchasingPower Parity terms, the poverty lines had to be recalcu-lated. As in the past, the lower line was set at the levelof the lowest poverty lines in low-income countries. Thelines used for the new estimates are equal to $1.08 perday and $2.15 per day in 1993 PPP terms (referred tosynthetically as $1 and $2 per day in the text), corre-sponding to the median of the 10 lowest poverty linesin low-income countries, and double that level. See Chenand Ravallion (forthcoming) and the September 1999global poverty figures (World Bank 1999).

43

17. The figures for 1998 are preliminary estimatesbased on surveys for that year for a handful of coun-tries and on older survey data, updated using estimatedgrowth rates in real private consumption per capita,for the majority of countries.

18. Note that urban survey data for China donot include migrants from rural areas to the cities. Thisprobably leads to an underestimation of the rate ofpoverty reduction. Measurement methods in the offi-cial tabulations from the survey data for rural Chinaare also believed to overestimate the rate of increasein inequality (Ravallion and Chen 1999).

19. Moreover, the relationship between inequali-ties in the return to skills and inequalities in overallincome or consumption is also highly complex, anddepends on patterns of labor force participation, house-hold composition, and transfers, among other factors.

20. Not all countries in the region have experi-enced significant increases in inequality. Countries suchas Hungary and Poland have not, but others such asRussia have experienced significant increases.

21. To explain this assumption in more detail, weneed to consider the Lorenz curve, giving (on the verti-cal axis) the share of total income, L(p), held by thepoorest p fraction of people (on the horizontal axis).The projections assume that the Lorenz curve shifts inor out by the same proportion at all points, relative tothe line of equality (in which everyone has the sameincome, that is, L(p) = p). So the new Lorenz curve isgiven by L(p) – g.(p–L(p)) where L(p) is the old Lorenzcurve and g is the proportionate increase in the Giniindex (Kakwani 1993). This assumption, which iscomputationally convenient, is commonly made in dis-tributional analysis, but represents only one of the pos-sible ways in which inequality may change. Povertyprojections can be sensitive to seemingly subtle differ-ences in how the Lorenz curve shifts over time (Ravallion1999). As mentioned in the text, this can be a seriousproblem when the poverty rate is low because estimatesat the tails of the distribution can naturally be quitesensitive, but for the bulk of the developing countriesthe poverty rates are a safe distance from the tails. Howmuch of a difference this would make for aggregatepoverty numbers is unclear. Errors in one direction inone country could be offset to some extent by errors inthe other direction elsewhere.

22. The model uses flexible functional forms forthe Lorenz curve, and the models pass through a se-ries of tests to assure that they fit well and satisfy theproperties required of valid Lorenz curves.

23. For information on the International Devel-opment Goals and the other targets for improving well-being, see (http://www.oecd.org/dac/Indicators).

24. This box is based on “Some Lessons fromForecasting Errors in the Recent Crisis.” Dadush,Riordan, and Wolfe forthcoming.

25. Global Economic Prospects and the Devel-oping Countries Short-Term Update 1998, internaldocument.

26. Export performance is biased upward byKorea’s strong gain during the year. Export volumegrowth for the affected ASEAN countries averagedabout 5 percent.

27. There is evidence that the wealth effects ofstock market changes can be small. Some studies(Brayton and Tinsely 1996) have shown that for theUnited States the marginal propensity to consume thatis generated by a change in corporate equity wealth isonly 0.3, with a mean response lag of 2 years. Yet therise in the share of households owning equity (40 per-cent in 1995—and higher in 1999—contrasted with32 percent in 1989) clearly amplifies these effects.Moreover, the effects on confidence of a stock marketcorrection of the magnitude assumed in the scenario(within the global context) could be especially adverse.Concerns about the state of Japan’s economy and thepossibility of significant difficulties among financialinstitutions attach more systemic risk to the unfoldingof the scenario.

28. The ex ante financing gap is the measuredchange in aggregate current account balance for thedeveloping countries (from the baseline projections)before any offsetting policy response at the countrylevel.

29. The simulation of the behavior of private capi-tal flows in this low-case scenario is based on a modelof the past behavior of financial net flows (other thanforeign direct investment) to developing countries be-tween 1979 and 1998. They are also based on assump-tions from the global economic model about expectedgrowth in world trade, world GDP, developing coun-try GDP, world interest rates, and an index of riskaversion, relative to the baseline scenario.

30. McConnell 1999.31. This is by no means a unanimous view. Every

shade of opinion on Y2K can be found on the internet;compendiums of links on Y2K include (http://home.att.net/~year2k), and (www.y2klinks.net/Y2Kgen.htm).

32. Third Quarterly report on US Readiness forYear 2000 issued by the President’s Council on Y2KConversion is located on (www.y2k.gov).

33. The International Y2K Cooperation Center(at www.iy2kcc.org) and Gartner Group survey.

34. The survey was carried out by the Interna-tional Y2K Cooperation Center.


44


35. The grouping of countries was carried out bythe World Bank, based on the wording of the countryevaluations from the U.S. State Department. The sec-tors covered were transportation, energy, telecommu-nications, health, finance, local government, and waterand wastewater. The individual country informationcan be found at the U.S. State Department’s Bureau ofConsular Affairs internet site (http://travel.state.gov/y2kca.html).

ReferencesBergsten, C. Fred. 1999. “The Global Trading System

and the Developing Countries in 2000.” Work-ing Paper 99-6. Institute for International Eco-nomics, Washington, D.C.

Brayton, Flint, and P. A. Tinsley. 1996. “A Guide toFRB/US: A Macroeconomic Model of the UnitedStates.” Finance and Economics Discussion Pa-per 96–46. October.

Chen, Shaohua, and Martin Ravallion. Forthcoming.Global Poverty Measures 198798 and Projectionsfor the Future. Washington, D.C.: World Bank.

Croome, John. 1998. “The Present Outlook for TradeNegotiations in the World Trade Organization.”Policy Research Working Paper 1992. WorldBank, Washington, D.C.

Dadush, Uri, Elliot Riordan, and Bertram Wolfe. Forth-coming. “Some Lessons from Forecasting Errorsin the Recent Crisis.” Economic Notes.

Demery, Lionel, and Michael Walton. 1998. ArePoverty Reduction and Other 21st CenturySocial Goals Attainable? Washington, D.C.:World Bank. Also available on (http://w w w. w o r l d b a n k . o r g / p o v e r t y / l i b r a r y /demwalt.htm).

Finger, J. Michael, and Philip Schuler. 1999. “Imple-mentation of Uruguay Round Commitments: theDevelopment Challenge.” Policy Research Work-ing Paper 2215. World Bank, Washington, D.C.

Hertel, Thomas W., ed. 1997. Global Trade Analysis:Modeling and Applications. Cambridge: Cam-bridge University Press.

Kakwani, Nanak. 1993. “Poverty and EconomicGrowth with Application to Côte d’Ivoire.” Re-view of Income and Wealth 39: 121–39.

Klugman, Jeni, and Jeanine Braithwaite. 1998. “Pov-erty in Russia during the Transition: An Over-view.” World Bank Research Observer, 13(1):3758.

Krueger, Anne O. 1999. “The Developing Countriesand the Next Round of Multilateral Trade Ne-gotiations.” Policy Research Working Paper2118. World Bank: Washington, D.C.

Londoño, Juan Luis, and Miguel Székely. 1997. “Per-sistent Poverty and Excess Inequality: LatinAmerica 197095.” Working Paper 357. Wash-ington, D.C.: Inter-American DevelopmentBank.

Martin, Will, and L. Alan Winters, eds. 1996. TheUruguay Round and the Developing Countries.Cambridge: Cambridge University Press.

McConnell, Bruce W. 1999. “Y2K: The Texture of theImpact.” Press release, International Y2K Coop-eration Center. September 22.

McKibbin, Warwick, and Will Martin. 1999. “The EastAsian Crisis: Investigating Causes and Policy Re-sponses.” Policy Research Working Paper 2172.World Bank, Washington, D.C.

Michalopoulos, Constantine. 1999. “DevelopingCountry Goals and Strategies for the MillenniumRound.” Policy Research Working Paper 2147.World Bank, Washington, D.C.

Milanovic, Branko. 1999. “True World Income Dis-tribution, 1988 and 1993: First Calculation Basedon Household Surveys Alone.” World Bank,Washington, D.C.

Ravallion, Martin. 1997. “Can High Inequality De-veloping Countries Escape Absolute Poverty?”Economics Letters 56: 51-57.

_____. 1999. “On Decomposing Poverty into Growthand Redistribution Components.” World Bank,Development Research Group, Washington,D.C.

Ravallion, Martin, and Shaohua Chen. 1997. “WhatCan New Survey Data Tell Us about RecentChanges in Poverty and Distribution?” WorldBank Economic Review 11: 357-82.

_____. 1999. “When Economic Reform Is Faster thanStatistical Reform: Measuring and ExplainingInequality in Rural China.” Oxford Bulletin ofEconomics and Statistics 61: 33–56.

Ravallion, Martin, and Gaurav Datt. 1999. “When IsGrowth Pro-Poor? Evidence from the DiverseExperience of India’s States.” Policy ResearchWorking Paper, World Bank, Washington, D.C.

Rutkowski, Michael, ed. 1999. Russia’s Social Pro-tection Malaise: Key Reform Priorities as a Re-sponse to the Present Crisis. SP Discussion Paper9909. Washington, D.C.: World Bank.

Standard and Poors DRI. 1999. “The U.S. Forecast Sum-mary: Raising the Speed Limit.” David Wyss. May.

World Bank. 1990. World Development Report: Pov-erty. New York: Oxford University Press.

_____. 1998. Global Economic Prospects 1998/99.Washington, D.C.

_____. 1999. Poverty Trends and Voices of the Poor.Washington, D.C.: World Bank. Also available

45


on (http://www.worldbank.org/poverty/data/trends/index.htm).

_____. Forthcoming-a. Poverty Reduction and theWorld Bank: Progress in Fiscal Year 1999. Wash-ington, D.C.

_____. Forthcoming-b. World Development Report2000/01. New York: Oxford University Press.

47

DEVELOPING COUNTRIES HAVE BECOME

increasingly integrated into globalgoods and financial markets over the

last decade. Their export volume increased by9 percent per year during the 1990s, up from2 percent during the 1980s. Net long-termcapital flows, even after declining in 1998,remained almost three times the 1990 level.As discussed in previous issues of Global Eco-nomic Prospects (World Bank 1993, 1997,1999a), globalization provides developingcountries with significant benefits and spurseconomic progress. GDP growth in develop-ing countries (excluding the transition econo-mies) averaged 5 percent during the 1990s,compared with 3 percent during the 1980s.Poverty—the number of people living on lessthan $1 a day—fell from 29 percent in 1990to an estimated 24 percent in 1996. But thefinancial crisis of 1997–99 has also shownhow globalization, and in particular greateropenness to external capital flows, can exposedeveloping countries to increased volatilityfrom international financial and goods mar-kets. The poor are especially vulnerable to thisvolatility.

This chapter reviews the evidence aboutthe impact on poverty of the external shocksand volatility to which developing countriesare exposed. It then presents and assesses evi-dence of the impact of the 1997–98 financialcrisis on poverty in the most affected EastAsian countries. Finally, it discusses lessonsand policy conclusions.

The chapter reaches the following con-clusions:

• The financial crisis has underlined howglobalization, especially financial integra-tion, exposes developing countries to ex-ternal shocks. These shocks often reducethe gains in poverty reduction from open-ness and increase poverty significantly inthe short to medium term. This fact un-derscores the importance of addressing theissue of volatility in order to maximizethe positive effects of growth on povertyreduction.

• The countries most affected by the EastAsian crisis illustrate the asymmetric im-pact of changes in per capita income onpoverty and the negative effects of vola-tility on growth. Though less dramaticthan early predictions suggested and veryheterogeneous, the negative social impactof the East Asian crisis and consequentcrises in Russia and Brazil has been enor-mous. The increase in consumption pov-erty has been significant. In addition, thecrisis has resulted in large and costly re-allocations of people and sharp declinesin middle-class standards of living. Un-like the situation in Latin America whereincome inequality increased significantlyduring crises, in East Asia the effects onincome distribution have been small andhighly differentiated. The extent of theseeffects depends on the country’s income

2External Shocks,Financial Crises, and Povertyin Developing Countries

48


level and the impact of the crisis on dif-ferent economic sectors.

• Urban poverty increased in all countries,particularly the Republic of Korea, wheretotal employment declined and open un-employment grew more than in othercountries in the region. Falling real wagesin the urban formal sector affected mostlyhigh-income groups. In Thailand the im-pact was felt mostly in rural areas becauseof the large inflows of workers from ur-ban areas and the relatively small increasesin agricultural prices.

• The crisis demonstrated the flexibility oflabor markets in developing countries.These markets help absorb the effects ofshocks through reduced wages and labormobility within and between urban andrural areas. Thus the decline in total em-ployment in Thailand and Malaysia waslimited, and employment actually rose inIndonesia. Labor was reallocated from theformal (urban) sector to other activities,particularly the informal sector and agri-culture, where exchange rate depreciationsimproved incentives.

• Even where public spending on safetynets increased significantly, the impact onpoverty was limited for several reasons.These included the absence of safety netsbefore the crisis, response lags, institu-tional problems, and low levels of spend-ing relative to the scale of poverty. Insome cases evidence suggests that well-functioning programs were underfundedrelative to the potential impact of shockson poverty.

• The severity of the crisis in Indonesia isreflected in the strong responses of house-holds to increase consumption as a shareof income, adjust their asset holdings, andincrease the share of staple foods in theirconsumption baskets to cope with theshock. In the Republic of Korea and Ma-laysia the response of households was toincrease the savings rate. The composi-tion of consumption expenditures changedsignificantly. Households spent more, pri-

marily on essential items such as food,fuel, housing, health, and education.

• Real public expenditures on education andhealth fell in most countries. The extentto which households were able to adjusttheir spending to offset this decline var-ied across countries as well as incomegroups. In Thailand families and govern-ment programs acted to cushion the im-pact of the crisis in order to avoid declinesin school enrollment rates or in access tohealth services. In Indonesia, however, theseverity of the crisis led to significant de-clines in poor households’ access to botheducation and health services, particularlyin urban areas. Such setbacks can haveirreversible effects on human development.

• Any development strategy for stable andsustainable growth must include bothadequate safety nets and appropriatepolicies and institutions designed toprevent financial crises and respond whencrises do occur. Prospects for povertyreduction depend not only on futuregrowth but also on countries’ capacityto manage volatility and reduce growthfluctuations.

External shocks andpoverty in developing countries

Discussions of the link between growth andpoverty reduction in developing countries

implicitly take the view that long-term growth(and therefore poverty reduction) is a stableprocess. But as the financial crisis of 1997–99shows, the process of growth is neither smoothnor linear and is often subject to sharp changes(especially major slowdowns and recessions)from a variety of external or internal shocks(World Bank 1999a). The asymmetric effectsof income growth on poverty during expan-sions and downturns, however, imply thatthese changes often have profound, long-last-ing effects on the poor. A decline in per capitaincome tends to have a negative effect on pov-erty that is much greater than the improve-ment generated by an equivalent increase.

49

While economic crises hurt both poor andrich, the poor have less leeway to respond tothe crises. If domestic capital markets wereperfect and the economic downturn tempo-rary, all economic agents could borrow tosmooth consumption and maintain welfare.But capital markets are imperfect and seg-mented. Credit or insurance is typically notavailable to the poor. With few savings andlow or subsistence incomes, the poor becomeeven more vulnerable to shocks. Crises andrecessions can result in irreversible negativeeffects on the poor through their impacts onhealth, schooling, and nutrition. Volatility ingrowth also tends to create more uncertaintyand risk for investors. That fact alone tendsto reduce the rate of economic growth, fur-ther dimming prospects for poverty reduction.Thus the volatility of the growth process indeveloping countries matters a great deal forboth immediate and long-term poverty reduc-tion and income distribution.

In general, the growth process is muchmore volatile in developing countries thanin industrial countries. Sudden reversals andother changes in international financial flows

are only one source (albeit an important one)of external shocks that can lead to crises andrecessions in developing countries. Fluctua-tions in the terms of trade are another im-portant and long-standing source, reflectingdeveloping countries’ reliance on primarycommodity exports and price variability ininternational markets. Volatility in the termsof trade was almost three times greater indeveloping countries than in industrial coun-tries during 1961–97 (Pritchett 1998; East-erly, Islam, and Stiglitz 1999) (figure 2.1).Volatility is particularly significant for theMiddle East and North Africa, LatinAmerica, and Sub-Saharan Africa. Usingsimulation models that replicate the range ofobserved economic fluctuations, Mendoza(1995) finds that disturbances in the termsof trade account for about one-half of theobserved variability in GDP and real ex-change rates, and that the share is greaterfor developing countries than for industrialcountries. Policies to mitigate and cope withvolatility in growth and the consequent ef-fects on the poor are therefore essential inall developing countries.

Figure 2.1 Terms of trade and GDP growth volatility, 1961–97

14

Standard deviations (percent)

3.9

11.3

7.4 8.1

9.7

11.413.1

2.5

5.34.2

2.5

8.6

4.75.6

12

10

8

6

4

2

0Terms of trade Real GDP

High-income

Low- and middle-income

East Asia and Pacific

South Asia

Middle East and North Africa

Latin America and the Caribbean

Sub-Saharan Africa

Note: Unweighted average of countries' standard deviations of relative distance from Hodrick-Prescott filter trend.Source: World Bank staff calculations.

E X T E R N A L S H O C K S , F I N A N C I A L C R I S E S , A N D P O V E R T Y

50


External shocks, long-term growth, andpovertyExternal shocks, such as variations in theterms of trade, volume of trade, and externalfinance, are highly correlated with variationsin GDP growth. They account for a signifi-cant share of the volatility in developing coun-tries (Easterly, Islam, and Stiglitz 1999).According to Hausmann and Gavin (1995)external shocks explain 30 percent of cross-country variation in GDP volatility in LatinAmerica. When terms of trade, export vol-umes, external finance, and interest rateshocks are taken into account, developingcountries experience more and larger exter-nal shocks than industrial economies. Theincidence of small and medium-size shocks isabout the same for both (World Bank 1993).During the 1970s and 1980s it was not un-usual for developing countries to suffer unfa-vorable shocks equivalent to 4 percent of GDPor more.

Volatility of growth and other macro-economic variables is also much larger indeveloping countries than in industrial coun-tries (Pritchett 1998; Easterly, Islam, andStiglitz 1999). Figure 2.1 shows that volatil-ity in GDP growth is more than twice as highin developing countries as it is in high-in-come countries of the Organisation forEconomic Co-operation and Development(OECD). The volatility of GDP growth ishigher for all developing regions, except forSouth Asia, and it is more than three timeshigher for the Middle East and North Af-rica. GDP growth in developing countries ishighly unstable, with large shifts over timeand low correlation of per capita growth ratesacross decades (Easterly and others 1993;Pritchett 1998).

Volatility has a negative impact on pov-erty in part because it reduces long-termgrowth (box 2.1). For instance, a large degreeof volatility makes “stop and go” policies morelikely, slowing growth and leading to low-quality policies such as those in Sub-SaharanAfrica, especially during the 1970s and 1980s.(Guillaumont, Jeanneney, and Brun 1999). Ex-

ternal negative shocks can also interact withsocial conflicts and weak domestic institutionsfor conflict management to produce growthcollapses (Rodrik 1998). After controlling forother factors, Hausmann and Gavin (1995)find that a higher standard deviation of realGDP is associated with higher rates of pov-erty. They estimate that if Latin Americancountries had the same GDP volatility as in-dustrial countries, poverty would decrease by7 percentage points.

External volatility and fluctuations inpovertyVolatility does more than simply increase pov-erty. Short-term fluctuations in income growthalso cause sharp variances in the incidence ofpoverty, even in the short to medium term.For example, in Venezuela poverty decreasedby 10 percentage points between 1989 and1991, rose by 20 percentage points between1991 and 1994, then fell again in 1995 androse in 1996 (Lustig and Deutsch 1998).Mexico is another striking example of this ef-fect, as box 2.2 describes.

Fluctuations in commodity prices may in-duce short- to medium-term changes in bothgrowth and poverty. During the boom yearsgrowth is faster and poverty declines, but dur-ing busts, which are usually more sudden, pov-erty increases. Fluctuations in commodityprices have a significant, direct impact on per-sonal incomes and an indirect impact on gov-ernment social expenditures and GDP. Earlierstudies argued that commodity price boomsdo not significantly affect real GDP(Cuddington 1988; Gelb and Associates1988). But more recent empirical work hasshown that changes in terms of trade havesignificant effects on real output growth.1 De-clining trends in real commodity prices havea negative effect on real income growth in thelong term in developing countries (see chap-ter 4).2 In addition, a slowdown in growth inthe bust years may become more severe as in-vestments made during the boom years areoften less productive (Collier and Gunning1996).

51

Empirical findings also support the no-tion that economic cycles have an asymmetriceffect on poverty. They show that a contrac-tion will have a greater impact on the povertyrate than an expansion of the same size(Morley 1994; Londoño and Székely 1997a;De Janvry and Sadoulet 1998).10 It has beenestimated that a 1 percent decline in per capitaincome during recessionary episodes in LatinAmerica in the 1980s reduced earlier gains by3.4 percent of per capita income growth inurban areas and 2.2 percent in rural areas (DeJanvry and Sadoulet 1998). One explanationfor this phenomenon is that during recessionsthe unskilled are the first to lose their jobs,because firms tend to hoard their skilled em-

ployees. As a result, income distribution be-comes more inequitable, amplifying the effectof declining incomes on poverty (Agénor1998).

Income poverty and inequalityduring the East Asian crisis

The recent crisis in East Asia has under-lined the risks for developing countries

of reversals in private capital flows and thedramatic social impact of the resulting finan-cial crises. The East Asian crisis had a sub-stantial impact on output and poverty in1998, although these effects began to lessenin 1999.


irreversibilities or asymmetric adjustment costs ininvestments increase uncertainty and lower invest-ment (Pindyck 1991; Aizenman and Marion1999).7 Second, costs increase because productivefactors move among sectors in response to morefrequent shifts in price signals. And third, the riskof inappropriate monetary, fiscal, trade, and finan-cial policies increases.

Terms-of-trade volatility has been found tohave a negative effect on long-term growth in de-veloping countries. Commodity price uncertainty,as measured by the standard deviation of forecasterrors from some statistical models, reduces growthrates (Dehn and Gilbert 1999).8 Most empiricalstudies have used direct volatility of terms of tradeas a proxy for uncertainty and have found negativeeffects on long-term growth (Mendoza 1994;Hausmann and Gavin 1995; Guillaumont,Jeanneney, and Brun 1999; Easterly and Kraay1999).9

The overall evidence also indicates that overthe long run the dependence of many developingcountries on commodities with volatile prices has anegative impact on long-term growth and thereforeon poverty. Dehn and Gilbert (1999) find a signifi-cantly negative effect of commodity price uncer-tainty on poverty, as measured by infant mortality.

Box 2.1 Volatility, growth, and poverty

W hen growth proceeds smoothly over timeand income inequality improves (or at least

does not worsen dramatically), poverty declines asper capita income and real wages rise. The elastic-ity of poverty, as measured by the headcount in-dex—for example, with respect to the growth ofper capita income3 —is estimated to be between–1.5 and –3.5.4 The size of the effect is greater incountries where income is more evenly distributed(Ravallion 1997).5

To the extent that volatility creates uncer-tainty, it has negative effects on growth and there-fore on poverty. Recent empirical evidence supportsthis view and contradicts the early literature. Usingbalanced panel data for a sample of 92 countriesfor 1960–85, Ramey and Ramey (1995) find that aunit increase in the standard deviation of innova-tion in GDP (innovation to GDP growth is used asa measure of uncertainty) implies a lower GDP percapita growth of 0.2.6 Similarly, from a growthregression of 130 countries for 1960–95, Easterlyand Kraay (1999) find that the standard deviationof growth has a strong negative effect (-0.18) onaverage per capita growth (after controlling forother variables).

There are three likely explanations for thenegative link between volatility and growth. First,

52


The impact of the crisis on povertyIncome poverty almost invariably increasesduring a crisis. Household surveys conductedin Latin America during recessionary periodsin the 1980s and 1990s provide evidence ofthis effect. They show that the incidence ofpoverty increased during the first year of therecession in 9 out of 11 cases, and remainedhigher for one or more years after the reces-sion in 19 out of 21 episodes (Lustig 1999).

Poverty also increased during the first yearof the crisis in the most affected East Asian

countries (table 2.1). Evidence from Korea il-lustrates the asymmetric impact of crises onpoverty. During stable growth in 1990–97, theestimated elasticity of the percentage of poorwith respect to per capita GDP was –3.5(Kakwani and Prescott 1999). But during thecrisis in 1998 the incidence of poverty in-creased by 123 percent. Real per capita GDPdeclined by 6.7 percent, and consumption percapita declined by 10.4 percent. In Indonesiaas well, the rate of increase in poverty wasabout 10 times the rate of the decline in con-

ing a Gini index of 0.54 in 1989 that remainedunchanged until 1996 (Székely 1998). Total pov-erty rose from its lowest point of 28 percent in1984 to 36 percent in 1989, or from 20.7 to 29.6million poor. Infant and preschool mortality causedby nutritional deficiencies increased from 1982onward, and educational indicators for the poordeteriorated (Lustig 1998).

From 1989 to 1994 growth resumed, largelydue to economic and financial liberalization, realper capita GDP growth averaged 2 percent.Although the total poverty headcount index haddeclined slightly to 34 percent by 1994, thenumber of poor had increased to 30.7 million(Lustig and Székely 1998; Székely 1999a).According to Székely (1999b) 86 percent of therise in poverty trends in Mexico from 1984 to1994 resulted from the increase in inequality,while the rest was the result of the drop in GDPper capita.

From 1989–94 poverty rose among ruralworkers in the primary sector and in the southernand southeastern regions. This increase was theresult of the appreciation of the peso and thedecline in institutional support for agriculture,including the loss of subsidies, the collapse ofguaranteed prices for major crops, and high inter-est rates (Lustig and Székely 1998). The financialcrisis that hit Mexico at the end of 1994 hadconsiderable repercussions for growth and total

Box 2.2 External shocks and fluctuationsin poverty in Mexico

Mexico’s experience since the 1970s shows howpoverty declines during periods of economic

growth and increases during periods of crisis andadjustment. External shocks contribute to thesevariations.

During the 1970s Mexico experienced rela-tively high and sustained growth. The increase inreal GDP per capita averaged 3.8 percent per year,despite the short-lived financial crisis of 1976 (seebox figure). Income inequality declined: the Giniindex dropped from 0.58 in 1970 to 0.51 in 1977(Londoño and Székely 1997b). Total poverty fellsignificantly, dropping from 49 percent in 1968 to34 percent in 1977, or from 23.3 to 21.3 million.Further gains were realized during the second halfof the decade (reflected in the numbers for 1977–84), spurred partly by favorable terms-of-tradeshocks and rising oil production.

In the early 1980s the international environ-ment became unfavorable for Mexico. Thecountry’s terms of trade declined, and real interna-tional interest rates increased. The resulting debtcrisis, the adjustment of the 1980s, and the col-lapse of oil prices in 1986 resulted in a sharp de-cline in incomes. Between 1982 and 1988 real GDPper capita growth averaged a negative 1.9 percentper year, and real wages fell by 36–46 percent from1983–88. This decline in incomes contributed to adramatic increase in poverty (Lustig 1998). In-equality increased sharply in the late 1980s, reach-

53


Table 2.1 Growth, poverty rates, and Gini coefficients in East Asia, 1996–98

Indonesia Malaysia Rep. of Korea Thailand

Real per capita GDP growth (percent)

1997 2.9 5.4 4.5 –1.41998 –15.1 –9.2 –6.7 –10.3

Real per capita consumption growth (percent)1998 –5.5 –12.4 –10.4 –11.6

Headcount poverty indexa National National Urban National1996 11.3 — 9.6 11.41997b 11.0 8.2 8.6 9.81998 16.7 — 19.2 12.9

Gini index1996 0.380 — — 0.4771997 — 0.496 0.290 —1998 0.370 — 0.294 0.481

— Not available.a. Figures for Indonesia are based on consumption expenditures, with a national poverty line equivalent to about $1 a day in 1985international purchasing parity (IPP) dollars; data are from February 1996 and December 1998. Figures for the Republic of Koreaare based on consumption expenditures, with a national poverty line equivalent to about $4 a day in 1985 IPP dollars. Figures forThailand reflect national income poverty, measured at around $2 a day. Figures for Malaysia reflect income poverty.b. The 1997 figures for Indonesia and Thailand are estimates based on precrisis trends in declines in poverty.Source: Kakwani 1999; Kakwani and Prescott 1999; World Bank staff calculations.

Box 2.2 (continued)poverty. Real GDP growth declined to a negative6.2 percent in 1995 and averaged 2.6 percentfrom 1995 to 1998. Total poverty increased dra-matically, rising to 45 percent in 1996, or 41.7

million (Székely 1999a). Mexico has adjusted tothe crises primarily through downward flexibilityin real wages rather than through increases inunemployment.

Changes in poverty, inequality, and per capita GDP in Mexico

15

1968–77

1977–84

1984–89

1989–92

1992–94

1994–96

10

5

0

–5

–10

–15

Percent

Change in total poverty

Average per capita GDP growth

Change in Gini coefficient

54


sumption per capita—much higher than theusual elasticity during expansions. While theselosses have been reversed somewhat since1999, the extent and sustainability of the re-covery remains to be seen (see chapter 3). Evenreturning to the precrisis level of poverty, how-ever, is likely to require more time and incomegrowth.

The severity of the impact of the EastAsian crisis varied across countries. Differencesin national poverty levels and the distributionof the income of the poor around these levelsmay explain some of the variances. For in-stance, in Korea the poverty line is around $4per day, while in Indonesia it is around $1 perday. If the individuals whose incomes droppedsignificantly are clustered above the povertyline in Korea and below the poverty line inIndonesia, the impact of the crisis on povertymay well appear lower in Indonesia. But otherfactors are also responsible for these differ-ences.

Korea’s experience was strikingly differ-ent from the others. Korea had the largest in-crease in open unemployment, a decline in theeconomically active population, and a largedrop in real wages that was second only toIndonesia’s. Labor mobility from the informalsector was also more limited than in othercountries. Korea is also the most urbanizedEast Asian country, and the negative impactof recessions has been found to be most dev-astating for poor urban dwellers (Morley1994; Lustig and Deutsch 1998; De Janvry andSadoulet 1998). The increase in urban pov-erty in 1998 was huge in Korea: the headcountindex, based on consumption expenditures,reached 19.2 percent, an increase of more than10 percentage points.11 The increase was evengreater (15 percentage points) between the firstquarter of 1997 and the third quarter of1998—the lowest (7.5 percent) and highestpoints (23 percent), respectively (Kakwani andPrescott 1999). The incidence of poverty de-clined to 15.8 percent in the last quarter of1998.

In other countries the increases weresmaller than had been anticipated, given the

magnitude of the crisis (table 2.1). In Indone-sia the impact of the crisis on poverty was stillsignificant. Estimates for Malaysia are notavailable, but welfare declines were wide-spread, presumably leading to increases inpoverty in both urban areas and traditionallypoor rural states.

Urban and rural poverty. Urban povertyincreases during crises owing to a combina-tion of lower real wages, higher unemploy-ment, and increases in the relative price offoods. The impact of a crisis on poverty willbe smaller if workers can move easily fromthe formal sector to other activities, particu-larly agriculture, and if exchange rate depre-ciations lead to improved incentives foragriculture. Even under those conditions, how-ever, it is still likely that urban poverty willincrease.

The relative impact on urban and ruralareas was different in Indonesia and Thailand.In Indonesia the crisis had a strong urban bias,even though the percentage changes in pov-erty rates were similar in urban and rural ar-eas. Poverty in urban areas rose from 9.7percent in 1996 to 15.4 percent in 1998, andin rural areas climbed from 12.3 percent to17.6 percent. Average per capita spending inurban areas fell 34 percent in real terms,whereas rural expenditures fell only 13 per-cent. A survey of expert respondent views sug-gests that urban areas were, on average, muchharder hit than rural areas (Poppele, Sumarto,and Pritchett 1999). Of the 20 hardest-hit ar-eas, 14 were urban, while of the 20 that suf-fered the least impact, 13 were rural. In nearlyevery province, region, and island, the nega-tive impact of the crisis was consistently higherfor urban than for rural areas. In Thailand,however, the impact on poverty was more se-vere in rural areas than in urban. Poverty ratesrose from 11.8 to 17.2 percent in rural areas,but only from 1.2 to 1.5 percent in urban set-tings.12 One possible explanation for the dif-ference in the impact of the crisis on the twocountries is the higher price incentives for ag-ricultural production in Indonesia, whichstimulated production.

55

Regional effects. The impact of the crisisvaried considerably across subnational regions.In the northern region of Thailand, for ex-ample, the poverty ratio actually dropped from10.2 percent in 1997 to 9.2 percent in 1998.In the northeastern and southern regions it rosedramatically, climbing from around 15 per-cent to 23.2 percent and from 8.6 to 14.8percent, respectively. In Indonesia per capitareal expenditures declined by 42 percent inWest Java and by 30 percent in Jakarta, re-gions that were better off before the crisis. Butreal expenditures declined between 10 and 20percent in other regions. These differences aremost likely linked to the behavior of producerprices. In Indonesia, areas that produced ex-port crops benefited from the sharp exchangerate depreciation. This fact combined withseveral reforms (such as clove marketing) toput more benefits in the hands of farmers.13

Similarly, in Thailand the poor performanceof the southern region may be linked to thefall in rubber prices during the crisis period.An important aspect of the crisis in Indonesiais that it does not appear to have affected poorareas disproportionately. Rather, the impactvaried in both well-off and poor areas.

The impact on income distributionGiven the significant drop in GDP normallyassociated with economic crises, poverty rateswill increase unless there is a massive reduc-tion in inequality. But income distributiontends to worsen during crises. Inequality inhousehold incomes or consumption increasedin most of the countries in Latin America dur-ing crises and recessions in the 1980s (WorldBank 1999a). For 10 recession episodes forwhich data are available in Latin America,inequality rose in 6 cases during the recessionyear (Lustig 1999). In Argentina the Gini co-efficient for the greater Buenos Aires area in-creased from 0.44 to 0.53 during the recessionof 1989. Inequality was higher after the re-cession than it had been before in 15 out of22 episodes. In Chile the Gini coefficient ontotal household incomes is estimated to haveincreased from 0.52 in 1979 to 0.55 in 1984

because of the impact of high open unemploy-ment and (by some measures) deep real wagecuts. However, the Gini coefficient in Chilehad declined to 0.53 by 1988 (Riveros 1994).

In Brazil income inequality increased, de-spite a successful defense of real wages in theformal sector and little increase in open un-employment, in part because of inflation anddeclining incomes in the informal and agri-cultural sectors (Fox, Amadeo, and Camargo1994). In Argentina average real wages oscil-lated wildly with episodes of inflation duringthe 1980s, though unemployment was nothigh. However, the gap in earnings betweenthe top and bottom deciles of income earnersin Buenos Aires widened steadily from 1980through 1988 as younger adults increasinglyentered the informal sector (Riveros andSanchez 1994).

Analyses of the effects of crises on incomedistribution suggested that the impact differedin middle- and low-income countries(Bourguignon, de Melo, and Suwa 1991). Dur-ing most economic crises and subsequent struc-tural adjustments in middle-income countries,income distribution worsens because wage cutsand layoffs in the formal sector tend to be bi-ased toward unskilled workers. The impactof crises on inequality in low-income coun-tries is more difficult to predict. Wage and em-ployment losses in the urban formal sectoraffect workers with relatively high incomes,and the rise in food prices hurts the urban poor.But the rural areas where most of the poorlive tend to gain because of currency depre-ciation and higher prices for agriculturalgoods. Bourguignon, de Melo, and Suwa(1991, 359) find, from simulations, that “inthe standard adjustment package, inequalityincreased significantly for the Latin Americanarchetype but decreased significantly for theAfrican archetype.” A major reason for thisdifference is that there are few formal sectorwage earners in the bottom half of the incomedistribution ladder in very poor countries—for instance, those in much of Sub-SaharanAfrica. Because crises hit the formal sectorhardest, the poor are less affected. In Latin


56


America, however, where formal sector work-ers come from all income brackets, poor peopleare hit more directly in a crisis.

Compared with Latin America in the1980s, the distributional impact of the EastAsian crisis was limited for high-income coun-tries (Korea), upper-middle-income countries(Malaysia), and lower-middle-income coun-tries (Indonesia and Thailand). Changes inoverall inequality, as measured by the Gini co-efficient, were minor between 1996 and 1998(table 2.1).14 In Thailand there may have beenweak redistribution from middle- to high-in-come groups (Kakwani 1999). But early stud-ies for Korea and Thailand suggest that thoseat the bottom of the income distribution lad-der—the “ultrapoor”—were hit harder thanothers with incomes below the poverty line(Kakwani 1999; Kakwani and Prescott 1999).The evidence is more mixed for Indonesia(Poppele, Sumarto, and Pritchett 1999).

Labor incomesTo a large extent the impact of the crisis onconsumption poverty reflects changes in thereal incomes of households. The channelsthrough which the impact of a crisis reacheshouseholds can be traced to the sources ofhousehold income—that is, wages, returns onassets, profits from self-employment, andtransfers (Ferreira, Prennushi, and Ravallion1999). These sources tend to vary with house-hold income level—for example, poor house-holds tend to depend on self-employmentincomes and transfers, whereas the rich receivemuch of their income from assets. For thisreason, changes in the overall composition ofnational income can move households up ordown the distribution ladder.

Labor markets have the most profoundeffects on poverty, however. Labor demandshocks hurt households by lowering realwages, increasing unemployment, and reduc-ing self-employment earnings. While reducedlabor demand almost always raises the inci-dence of poverty, different kinds of labor de-mand–shocks have different effects on incomeinequality. In a recession, real wages fall.

Households at the low end of the distributionladder in developing countries are affected theleast, because they receive little or no wageincome. But labor demand shocks have astrong impact on those formal sector workerswith the lowest skills, who are more likely tolose their jobs than their more skilled coun-terparts.15 They then either become unem-ployed or move to the informal sector, wheretheir earnings are likely to be lower. As a re-sult households at the middle to lower-middlerange of the income distribution ladder arepushed further down, swelling the numbersof households with low incomes.

The crises in East Asia followed a patternsimilar to those seen earlier in other countriesfaced with sharp reversals of external capitalflows. A comparative analysis of the impactof similar crises on labor markets offers thefollowing conclusions (Fallon and Lucas1999):

• Wages fall sharply during the crisis or inensuing years, usually by more than theGDP. In 22 recessionary episodes in LatinAmerica during the 1980s and 1990s, realwages fell in 16 cases during the year ofrecession, and in 18 cases remained lowerthan precrisis levels after two years (Lustig1999). This wage drop was also a strik-ing feature of the East Asian crisis.

• Total employment growth drops in thecrisis year, but usually by less than thedecline in GDP growth.

• Employment in manufacturing is alwaysadversely affected, though less spectacu-larly than are wages.

• The effects on agricultural employmentare more muted. In some cases (for ex-ample, Indonesia in 1998 and Turkey in1994), employment increased despite anabsolute decline in GDP.

• Rising unemployment is an important fea-ture of many crises. In Latin America un-employment increased during the year ofrecession in 24 of 31 episodes and re-mained higher in 24 cases two years intothe recession (Lustig 1999). The most sig-

57

nificant increases were in Argentina in1995 (6 percentage points) and Chile in1982 (11 percentage points).

Experience thus far in East Asia broadlysupports these conclusions. Real wage growthdropped sharply in 1998 and became nega-tive in all affected countries (table 2.2). Indo-nesia saw particularly spectacular wagedeclines that were broadly similar across sec-tors. Although wage cuts can moderate theimpact of a recession on employment, duringthe East Asian crisis nonagricultural employ-ment fell in all countries. Only in Indonesia,where agricultural employment increased con-siderably, did overall employment rise. Theconstruction sector was the most affected, witha dramatic drop in employment of 15 to 35percent, but manufacturing employment alsofell significantly. With the exception of Ko-rea, however, falls in employment in 1998 werenot large despite substantial decreases in GDP.(Korea saw a large decline in employment aswell as in real wages.) The inactive popula-tion increased by 9 percent between the sec-ond quarter of 1997 and the fourth quarter

of 1998, with women representing three-fourths of the increase. In Thailand 18.5 per-cent of the overall decline in per capita incomewas the result of wage cuts, whereas only 2.7percent was attributable to higher unemploy-ment (Kakwani 1998).

Labor force mobility. To some degree, theimpact of the crisis on employment was less-ened by the mobility available to individualworkers within and between the urban andrural sectors. In the first year of a crisis sig-nificant real exchange rate depreciation usu-ally results that can raise the price of tradablegoods relative to those of nontradables, withimportant implications for real household in-comes and poverty. Crises hit the urban for-mal sector first and, as noted above, can leadto a reallocation of labor from the urban tothe rural sector. But in the absence of any in-centive to increase agricultural production, thisreallocation of human resources may do littlemore than raise rural poverty instead of ur-ban poverty. In principle exchange rate depre-ciation can supply the needed incentive. In theabsence of intervention in domestic markets,it raises the price of export crops relative to


Table 2.2 Employment and real wages in East Asia during the crisis(percentage change)

Indonesia Malaysia Rep. of Korea Thailand1997 1998 1997 1998 1997 1998 1997 1998

Employment, totala 1.8 2.6 4.6 –2.7 1.4 –5.8 1.8 –3.0Agricultural –4.7 13.3 –0.6 –5.3 –3.4 0.0 1.3 –1.8Nonagricultural 6.8 –4.7 5.8 –2.2 2.4 –6.5 2.2 –3.9Manufacturing 4.1 –9.8 7.6 –2.9 –4.3 –13.1 –0.1 –1.9Construction 10.6 –15.9 8.9 –13.4 1.7 –26.4 –5.6 –33.6

Real consumption wage, totalb 8.6 –41.0 — — — — 5.7 –1.5Agricultural 4.1 –35.0 — — — — 10.0 –8.9Nonagricultural 9.9c –42.0c — — 2.6 –10.0 5.0 –0.5Manufacturing 11.1 –44.0 6.0 –2.4 0.7 –10.6 7.1 –4.5Construction 8.5 –42.0 — — 3.3 –14.7 3.8 –2.2

— Not available.a. Figures for the Republic of Korea are from Q4 to Q4. Figures for Thailand are calculations by World Bank staff based on theLabor Force Survey, National Statistical Office, and surveys of national employment for February and August.b. Figures for Indonesia are for August 1998 and average 1997. Figures for the Republic of Korea are seasonally adjusted. Figuresfor Thailand are calculations by World Bank staff based on the Labor Force Survey, National Statistical Office, and average wages(excluding fringe benefits) of February and August surveys.c. Urban areas.Source: Employment – Islam and others 1999; Mansor and others 1999; Shin 1999; World Bank staff calculations. Real wages –Datastream; Islam and others 1999.

58


the prices of other commodities. In rural ar-eas higher food prices spur agricultural pro-duction, and the impact on poverty will dependon the strength of the link between agricul-tural production and the poor. Insofar as smallfarmers benefit, the link to poverty may bestrong. But insofar as export crop productionis concentrated among large farmers, the im-pact depends on whether the demand for ag-ricultural workers increases sufficiently tooffset the growing supply of rural labor.

In Indonesia around 2.5 million workers,or 3 percent of the total work force, were dis-placed by the crisis in the first year. Job lossesoccurred in all sectors of the economy exceptagriculture and the small transportation andcommunication sectors. The manufacturingsector accounted for nearly half of all joblosses, followed by construction. Losses weresomewhat smaller in the mining, trade, andservice sectors. About three-quarters of thejobs lost in these sectors were in rural areas.In urban areas many workers displaced fromthe manufacturing and construction sectors en-tered the trade and other service sectors. Ur-ban employment actually grew from 29.4 to30.3 million persons (Islam and others 1999).In contrast, displaced workers in rural areashad fewer opportunities, and many wereforced to take up agricultural employment.16

Agricultural employment rose in Indonesia in1998 despite severe drought conditions insome areas. While labor reallocation wasgreater in rural than in urban areas, the in-crease in poverty was greater in urban areasdue to the incidence (although limited) ofunemployment, the greater decline in wagesin manufacturing and construction, and thefall in informal sector incomes as more crowd-ing occurred.

In Thailand the crisis greatly affected theflow of labor between urban and rural areas.The principal reason for the increase in pov-erty in the rural areas, particularly in the north-east, was the integration of the rural andBangkok labor markets through migration.The crisis dramatically curtailed the regularflow of workers to Bangkok, particularly from

the northeast, increasing the rural labor sup-ply beyond what it would otherwise have been.The number of recent migrants to Bangkok(those arriving in the past year) had dropped50 percent by February 1998 from the levelsof a year earlier, and the share coming fromthe northeast fell from 68 percent to 38 per-cent. The types of workers who moved toBangkok also changed dramatically. In Feb-ruary 1997, 25 percent of all migrants toBangkok had less than an elementary educa-tion and only 28 percent had a secondary orhigher education. One year later these num-bers were 13.5 percent and over 41 percentrespectively. These results indicate that un-skilled workers stopped going to Bangkok,whereas those with higher skills continued tomigrate. In Indonesia there is evidence of sig-nificant return urban-rural migration. Around1 million urban workers entered the agricul-tural sector, although their families stayed inurban areas.

Other labor market developments. Laborforce mobility and fewer work opportunitieswere accompanied by other labor market de-velopments. First, hours worked per week fellas workers crowded into the urban informaland rural sectors. In Indonesia, the share ofemployees working fewer than 35 hours perweek increased from 30.6 percent in 1997 to34.3 percent in 1998, with the trend towardshorter hours greater in urban areas. In Ko-rea, average hours dropped from 46.6 hoursper week to 46.0, with the shorter workingweek most prevalent in manufacturing. Sec-ond, the composition of employment changed,shifting away from wage employment. In In-donesia, the proportion of workers outsidewage employment rose from 55.1 percent in1997 to 58.9 percent in 1998. Third, in Ko-rea at least, the number of highly skilled work-ers rose as a share of employment. Forinstance, the share of managers and profes-sionals had increased from 17.1 percent at theend of 1997 to 21.1 percent by the end ofMarch 1999.

With the exception of Korea, where theagricultural sector is much smaller than in the

59

other East Asian economies and total employ-ment decreased substantially, the increase inunemployment in 1998 was not particularlylarge (figure 2.2). In Korea, the unemploymentrate peaked at 8.7 percent in February 1999,an increase of 6.4 percentage points over thelow of 2.3 percent in June 1997.17 But the ratehad fallen to 6.2 percent by June of the sameyear. Underutilization of labor was greater thanthe data on unemployment indicated, as for-mal sector working hours fell in all countriesduring the recession. In Thailand total unem-ployment increased by 2.5 percentage pointsduring the crisis and remained high during thefirst half of 1999. In Malaysia open unem-ployment increased by much less than ex-pected, rising from 2.7 to 3.2 percent. It peakedat 4.5 percent in March 1999, both becauseproductivity-based wages allowed real wagecuts and because migrant workers employedin construction, the hardest-hit sector, left thecountry.

It is still too early to assess how the crisisaffected the incidence of unemployment in dif-ferent groups. In Korea unemployment seemsto have risen more among men than amongwomen, possibly because the female partici-

pation rate dropped as the crisis intensified.The number of regular female employees fellby around 20 percent between October 1997and October 1998. Layoffs in the formal sec-tor initially raised unemployment among olderage groups, but unemployment among theyoung is undoubtedly rising in the absence ofjob creation. The less educated and less skilledwere the hardest hit. For those with no highschool diploma, unemployment increased from1.2 percent in June 1997 to 5.8 percent in June1998, and for those with high school diplo-mas it climbed from 2.8 percent to 8.4 per-cent (Na and Moon 1999).

Government safety nets andpoverty alleviationRaising transfers can offset increases in incomepoverty caused by declines in labor demand.For this reason some governments have triedto strengthen safety nets, or income transferprograms. The success of these efforts dependson several factors: the existence of well-func-tioning programs, the institutional and deliv-ery capacity of central and local agencies, thesize of budget allocations and the severity offiscal constraints, and the political economy


Figure 2.2 Unemployment in East Asia during the crisis

8

Percent

7

6

5

4

3

2

1

0

1.6

4.1

Thailand

2.6

7.4

Republic of Korea

4.75.5

Indonesia

2.73.2

Malaysia

Note: Figures for the Republic of Korea are for urban unemployment, and those for 1998 are from the second half of the year. Figures for Indonesia are from August of each year. Figures for Thailand reflect unemployment as a percentage of the current labor force and are an average of February and August figures for each year.Source: Shin 1999; Islam and others 1999; Malaysian Labor Force Survey; World Bank 1999d.

1997

1998

60


affecting redistributive and poverty alleviationefforts.

It is still too early to assess fully the im-pact of efforts to provide safety nets duringthe crisis. However, the available evidencesuggests that even though levels of publicspending on safety nets increased significantly,the impact on poverty was limited. An analy-sis of country experiences suggests that thislimited effect can be traced to a range of fac-tors, including response lags, institutionalproblems, and low levels of spending relativeto the scale of poverty.

The governments of East Asia generallydid increase the budgetary share of incometransfers significantly in response to the cri-sis. However, spending as a share of nationalincome remains low by international standardsand has increased in only two countries(Klugman 1999) (figure 2.3). Korea had thelargest proportionate increase, with spendingon safety nets rising from zero to 5 percent ofthe budget, followed by Indonesia, where thebudgetary share rose from zero to 3.6 percent.In Malaysia the safety net as a share of gov-ernment expenditure held steady at a low 0.16

Figure 2.3 Poverty and spending on social safety nets in East Asia, 1996–98

2.5

2.0

1.5

1.0

0.5

0.0

24

19

14

9

4

-11996

Percent Percent

Percent Percent Percent Percent

Percent Percent

Indonesia

1998 1996 1998

1996 1998 1996 1998

6 20

15

10

5

5

4

3

2

1

0

Republicof Korea

1.0 9

8

7

6

5

4

3

2

1

0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

Malaysia

3.0

2.5

2.0

1.5

1.0

0.5

0.0

14

12

10

8

6

13

11

9

7

5

Thailand

Expenditures (left axis) Headcount index (right axis)


61

percent during the period. Spending on safetynets as a share of national income rose quitesteeply in Korea and Indonesia, though expen-ditures were low to begin with. These trendscontrast with those observed in Europe andCentral Asia, where expenditures on safety netsdeclined significantly across all countries(Milanovic 1998).

Korea. The central response of the Ko-rean government was to introduce a publicworks scheme that grew enormously duringthe crisis period, rising to 200,000 participantsin January 1999 and 410,000 in mid-1999.The scheme paid wages lower than the pre-vailing wage for unskilled workers in order toattract only those truly in need of employment.The scheme also was supposed to guarantee ajob for all who wanted one, although therewas significant excess demand for the num-ber of places available by late 1998. (Therewere 700,000 applicants in January 1999.)

The budgetary share of safety nets in Ko-rea did increase during the crisis, but recentanalyses have shown that the incremental bud-get and program coverage in 1998 were inad-equate to meet the country’s needs. Safety netprograms covered only 7 percent of the “new”poor in 1998. Overall coverage of the poor(old and new) dropped from almost one-thirdprior to the crisis to about 17 percent in 1998,and it is expected to fall further (to 16 per-cent) during 1999 (Subbarao 1999).18 Thereis evidence that women have been excludedfrom public works, and there are accusationsof mismanagement and a lack of useful out-put. Still, what has been done reflects twoimportant lessons: keeping the wage sharehigh—around 70 percent—creates more jobsper won spent; and diversifying the jobs menuto include more than engineering works (forinstance, work in libraries) increases job op-portunities. Some other East Asian govern-ments are just beginning to learn these lessons.

Thailand. In Thailand government safetynets did not fill much of the gap left by infor-mal transfers at least until 1998. Overall safetynet expenditures increased during the crisis,especially during 1999, though some income

transfers appear to have been procyclical—thatis, contracting with the economy rather thanexpanding. This situation was the result of con-flicting pressures and the government’s reluc-tance to undermine the informal safety net.For example, social pensions and family al-lowances in rural Thailand appear to be welltargeted, but they are underfunded (Prescott1999). The benefit value is less than one-thirdof subsistence requirements and reaches onlyone-third of the target group. Further, the realvalue of the benefit transfers in Thailand hasbeen falling over time because of the govern-ment’s failure to adjust them for inflation. Oneprogram that expanded in response to the cri-sis in Thailand was the Ministry of Health’sprogram providing low-income groups withaccess to public health services. Under thestimulus package, job creation for the poorunemployed was a priority.

Indonesia. The major new safety net pro-grams introduced in Indonesia as a result ofthe crisis were a rice distribution scheme(known as OPK) and a public works scheme(Padat Karya). Early evidence suggests thattheir coverage of the poor and their impacton poverty have been limited. The OPK makes10 kilograms of medium-grade rice availableto selected households every month at subsi-dized prices. On average, this amount repre-sents less than 30 percent of the income of asingle individual living at the poverty line andless than 6 percent of the income of a house-hold of five. OPK uses an indicator-based tar-geting system with minimum standards forfood intake, housing, clothing, and medicalexpenditures.

Evaluations of these programs suggest thatjust over one-third of all poor households haveparticipated in the Indonesian public worksprogram. The leakage of benefits to the non-poor has been significant, however. In Jakarta,9 percent of the poor households worked onceon the scheme, compared with 30 percent ofthe middle income households. In Medan only5 out of over 400 poor households partici-pated in Padat Karya, because the contractorused his own workers.


62


Other experiences. Declines in publicspending and a failure to reach the poor haveundermined safety net programs in Russia andCentral Asia. Russia provides a striking ex-ample of the failure of safety net programs toalleviate poverty. Spending on social assistancedeclined throughout the transition, amount-ing to only 4 percent of the poverty gap in1997, while the incidence of poverty more thantripled. This shortfall can be attributed in partto the concurrent collapse of the tax revenuesystem (UNICEF 1998), but it is also the re-sult of an apparent failure to identify andimplement programs that reach the poorand that have sufficient political and electoralsupport.

In addition, the safety net in Russia is notwell targeted. Most Russian households re-ceive some type of government transfer, but asignificant proportion of the very poor (al-most 3 out of 10) and of the poor (1 out of 5)receive no benefits. At the same time almostfour out of five households that are not poordo receive public transfers (Foley andKlugman 1997). Even so, the decline in bud-get allocations for public transfers, coupledwith widespread delays in payments, hasmeant a clear weakening in the impact oftransfers on poverty over time. The reductionin the poverty headcount attributable to pub-lic transfers fell from 29 to 24 percentagepoints between 1994 and 1996 (Klugman andKolev 1999).

Beyond current income effectsof the East Asian crisis

Standard poverty measures based on in-comes and household expenditures capture

only some aspects of the social impact anddistress crises cause. Households use variousmechanisms to cope with shocks from crises.These responses may help mitigate the imme-diate impact, but they may also have impor-tant implications for future poverty andvulnerability to shocks. Crises also create pres-sures on governments for fiscal austerity whichmay exacerbate the negative social impacts,

but appropriate fiscal policy may also help al-leviate some of these effects.

Behavioral responses and assetaccumulation during crisesFinancial crises affect not only current incomesbut also the value of household assets. Infla-tion, for instance, has been found to be one ofthe most significant determining factors ofpoverty (Datt and Ravallion 1997; Agénor1998; Easterly and Fischer 1999). It erodesthe value of fixed-denomination assets suchas money, which is the primary asset of thepoor and near-poor. These groups have littlescope for hedging.

As they do to labor market shocks, house-holds respond in variety of ways to the in-come, wealth, and relative price effects of acrisis. These responses include consumptionsmoothing, changing the composition of theconsumption basket, selling existing physicalassets, and acquiring fewer new ones.

Savings behavior during the East Asiancrisis. Changes in household savings patternsduring the crisis varied significantly acrosscountries. The savings rate changed little inThailand. In Indonesia the savings rate de-clined sharply, falling about 8 percentagepoints of GDP (figure 2.4). The savings re-sponse helped reduce the impact of the crisis

Figure 2.4 Change in gross domesticsavings, 1997–98

6

Percentage of GDP

4

2

0

–2

–4

–6

–8

ThailandIndonesiaMalaysia Republic

of Korea

Source: World Bank staff calculations.

63

on consumption among poor households, butthe reductions in capital accumulation reflectthe severity of the crisis.

In Korea and Malaysia the decline in percapita consumption was much greater than thedecline in per capita GDP between 1997 and1998 (table 2.1). In Korea the savings rate rose,but the increase reflects primarily the behav-ior of high-income groups.19 Total gross do-mestic savings increased by more than 4percentage points of GDP. Private savings in-creased at an even greater rate, especially inlight of the increase in the government deficit.Savings declined from 16.3 percent to 11.6percent among the 20 percent of householdswith the lowest incomes, however (Kakwaniand Prescott 1999). Because of the decline inthe consumption ratio and the varied effectson income distribution in Korea, the incidenceof income poverty increased much less thanpoverty based on consumption (discussedabove, table 2.1), rising from 2.6 percent in1997 to 7.3 percent in 1998.

Changes in asset holdings and the com-position of the consumption basket. Informa-tion about changes in asset holdings in thecrisis countries is limited. Some evidence is

available for Indonesia showing that peoplein the most affected regions, such as Java, soldsome of their assets (Poppele, Sumarto, andPritchett 1999). But more complete evidenceis available on changes in the consumptionbasket. Rising food prices are especially im-portant in determining changes in the compo-sition of consumption, because higher pricesreduce the real incomes of households withrelatively high food expenditures—mainly theurban poor. Food prices rose relative to othercommodities in all countries after the crisis(figure 2.5). However, the effect was smallexcept in Indonesia, where relative food pricesrose by 40 percent between mid-1997 and mid-1998. The effect is reflected in the dramaticchanges in the composition of expenditures:the share of staple foods increased from 23.1to 31.7 percent, while that of meats and non-food items (including health and education)declined (Poppele, Sumarto, and Pritchett1999).

Between 1996 and 1998 households inThailand, particularly those with low incomes,increased essential real expenditures such asfood, fuel, medical supplies, shelter, and edu-cation but reduced other expenditures (World


Figure 2.5 Relative price of foods during the crisisIndex of food prices relative to total consumer price index (1997=100)

140

130

120

110

100

90

80Q3

1999Q2

1999Q1

1999Q4

1998Q3

1998Q2

1998Q1

1998Q4

1997Q3

1997Q2

1997Q1

1997Q4

1996Q3

1996Q2

1996Q1

1996Q4

1995Q3

1995Q2

1995Q1

1995

Source: Datastream.

Indonesia

Malaysia

Thailand

Republic of Korea

64


Bank 1999d). Korean households had a dif-ferent response. The shares of food, clothing,and furniture in total expenditures actually de-clined, but spending for education and healthincreased (Kakwani and Prescott 1999).

Fiscal austerity and household demandfor health and educationThe East Asian economies eventually widenedtheir fiscal deficit targets to counter the re-cessionary effects of the crisis. Yet real gov-ernment consumption expenditures fell in allcountries except Thailand in 1998. In Indone-sia the decline outpaced the fall in GDP. As aproportion of GDP, health expenditures re-mained relatively unchanged during the last halfof the 1990s, including the first year of the re-cession (table 2.3). Education expenditures fellrelative to GDP when the crisis struck in Ma-laysia and Korea, but rose in Thailand.

Changes in public spending on educationand health affects both the availability of theseservices and, because the services may becomemore expensive, households’ decisions to usethem. The full impact of the crisis on publicservices remains unclear because of the differ-ences in fiscal policy responses and the lim-ited information available on changes in thecomposition of public expenditures. Some dataare available for Korea, Thailand, and Indo-nesia, however.

Korea. In 1998 households spent less onitems such as clothing and recreation butmaintained spending levels on education andhealth. The rates of decline for education (9percent) and health (14 percent) expenditureswere lower than the decline in total expendi-tures (18 percent) (Kakwani and Prescott1999).

Thailand. Families and government pro-grams acted to cushion the impact of the cri-sis on education and health (World Bank1999c). So far the crisis has not had a nega-tive effect on education. In the year followingthe onset of the crisis, total gross enrollmentsin primary and general upper secondaryschools increased from 74.8 percent to 75.5percent. The dropout ratio—children of schoolage not attending school—continued to declinebetween 1996 and 1998. Families employeda variety of strategies to keep their children inschool. First, households spent less on nones-sential consumables such as alcohol, tobacco,clothing, footwear, household goods, transpor-tation, and communications. Second, familiesused savings or borrowed from informalsources to finance education, so that realspending on education rose for both the poorand nonpoor. Third, for nontertiary educationfamilies shifted their children from private topublic schools as the relative cost of attend-ing private schools increased. Finally, families

Table 2.3 Public spending on health and education(percentage of GDP)

1994–95 1995–96 1996–97 1997–98 1998–99

HealthIndonesia 0.7 0.6 0.6 0.6 0.6Korea, Rep. of 0.5 0.5 0.5 0.6 0.6Malaysia 1.3 1.3 1.4 1.4 1.3Thailand 1.1 1.1 1.2 1.4 1.3

EducationIndonesia 1.4 1.2 1.4 0.7 0.7Korea, Rep. of 5.0 5.0 5.1 4.3 4.0Malaysia 5.3 4.9 4.0 4.7 4.3Thailand 3.4 3.3 3.1 3.4 4.2

Note: Public expenditures include national and local government.Source: Baptist 1999.

65

made greater use of government scholarshipand loan programs.

Government policies and programs alsosupported continued investment in education.Real expenditures on education remained con-stant between 1997 and 1998, and the shareof education increased in total expenditures.A number of programs and measures were in-troduced to protect educational opportunitiesfor the vulnerable. These included allowingparents to pay tuition fees in installments, per-mitting schools to waive tuition fees on a case-by-case basis, introducing scholarships,expanding the education loan program, en-couraging private schools to extend paymentdeadlines, and providing vouchers to privateschool children (in the Bangkok metropolitanarea). These achievements are remarkable butmay be difficult to sustain, as the recoveryremains weak and the effects of the crisis con-tinue to be severe. Households, particularlythe poor, may be less able to shift more re-sources to education and sustain higher debts.

There is also no evidence of a negativeeffect on national health outcomes. For in-stance, the number of reported cases of mal-nutrition continued on a downward trend in1998. Households’ real expenditures on bothprivate and public health services declined sig-nificantly. Out-of-pocket real expenditures onmedical and institutional care were 36 percentlower in 1998 than in 1996, whereas spend-ing on self-medication increased by 12 per-cent. The decline in expenditures was lowerfor the poor, who undoubtedly tried to sus-tain essential health expenditures and who alsobenefited from public health services. Thegovernment maintained its level of investmentin health, with real expenditures on healthdown by 5 percent in 1998 from 1997 levelsbut still 11 percent higher than they were in1996. The decline mainly affected investmentexpenditures. The government enlarged itshealth safety net by increasing the coverageof public health insurance. Use of public healthservices increased between 1996 and 1998,with the number of outpatient visits rising by22 percent.

Indonesia. The severity of the shock andfalling living standards led to a decline inschool enrollment rates (Frankenberg, Tho-mas, and Beegle 1999). This decline was muchlarger at the secondary level—some 4 to 5 per-centage points of enrollment rates20 —than atthe primary level. Consistent with the urbanbias in the effects of the crisis on incomes, thedecline in school enrollment was largest in ur-ban areas, particularly Jakarta. The popula-tion with the lowest per capita expenditureshad the highest rates of decline in school en-rollment: more than 5 percentage points forthe 13–19 age group for the lowest twoquartiles, and more than 6 percentage pointsfor the 7–12 age group for the lowest quartile.Because they had to increase food expendi-tures during the crisis, families had a difficulttime maintaining expenditures on education,and its share in total expenditures declinedfrom 3.5 percent in 1997 to 2.9 percent in1998.

The effects of the crisis on health werecomplex and heterogeneous but clearly nega-tive. The share of household expenditures onhealth declined from 1.4 percent in 1997 to 1percent in 1998. The use of public health ser-vices following the crisis declined by 1.8 per-centage points (from 7.2 to 5.4 percent) foradults and by more than 7.1 percentage pointsfor children. The proportion of visits made totraditional practitioners nearly doubled. In1997 nearly one-half (46.7 percent) of all chil-dren under the age of five had visited a com-munity health post in the month before thesurvey, but this rate declined to about one-quarter (27.7 percent) in 1998.

Fostering sustained growth andreducing the social costs ofvolatility and crises

Financial crises have large social costs andtend to retard or even reverse gains in pov-

erty reduction for significant periods of time,even in the most successful countries. Policiesand institutions that reduce these risks, helpprevent financial crises, and minimize their ef-


66


fects when they do occur can help smooth thegrowth process and maximize the positive ef-fects of growth on poverty alleviation (WorldBank 1999a; Ferreira, Prennushi, andRavallion 1999; Lustig 1999). Realizing thelong-term benefits of openness, reducing pov-erty over the long run and avoiding tempo-rary setbacks in poverty reduction requiresappropriate national and international poli-cies. These policies must minimize the risksof external volatility and improve the capac-ity to manage it at both levels. Safety nets arepart of any broad-based strategy for limitingthe impact of crises and negative shocks onpoverty.

Preventing crises. Avoiding crises is clearlythe most effective way to achieve stable andsustainable growth. Macroeconomic and fi-nancial policies that avoid profligate fiscal andmonetary policies, seriously overvalued ex-change rates, and unsustainable current ac-count deficits are necessary to prevent crises(World Bank 1999a; Lustig 1999). More flex-ible exchange rates, greater reliance on fiscalpolicy, and better and tighter domestic finan-cial regulation are often needed to reduce ex-cessive capital inflows and domestic lendingbooms (World Bank 1999a). Financial sectorliberalization must proceed carefully and instep with the capacity of countries to enforcetighter regulation and supervision. Efforts toimprove prudential safeguards and bankingoperations need to be accelerated in most de-veloping countries. The opening of the capi-tal account, particularly to the more volatilecapital flows, needs to be carefully orches-trated to match the capacity of countries tomanage risk (World Bank 1999a; Stiglitz andBhattacharya 1999). Other important policiesmust focus on improving corporate gover-nance, increasing transparency, and buildingsupportive institutions.

Socially sensitive crisis management. Mac-roeconomic policy responses to crises need tobe designed to minimize the social costs andavoid large declines in aggregate demand andemployment (World Bank 1999a; Lustig1999). Monetary policy should always avoid

high inflation as well as excessive increases ininterest rates, both of which worsen any con-traction in aggregate demand. Policy responsesaimed at reducing the social impact of crisesshould make fiscal policy countercyclical inorder to reduce the extent of contraction.However, developing countries typically haveprocyclical fiscal policies that tend to aggra-vate the impact of downturns (Easterly, Islam,and Stiglitz 1999). This phenomenon is theresult of the high sensitivity of tax receipts tochanges in incomes, underdeveloped domes-tic financial markets, limited access to foreigncapital markets, and the risk of losing inves-tor confidence. These factors make pursuingcountercyclical fiscal policies difficult. Estab-lishing effective countercyclical fiscal policiesrequires that public finances be managed wellduring good times, so that there is room forexpansionary policies during negative shocks.The adequate well-institutionalized use of sta-bilization funds may also be helpful (Lustig1999). But even East Asian countries that hadresponsible fiscal policies before the crisis havefound it difficult to achieve the looser fiscalobjectives.

Fiscal adjustments should also protect theexpenditures that are most important for thepoor, such as employment and human devel-opment programs and targeted subsidies.Where crises result in high unemployment, thefiscal stimulus needs to be directed to labor-intensive activities.

Managing volatility. In the long run de-veloping countries stand to make gains ingrowth and to reduce poverty through opentrade policies and integration into the worldeconomy. But external shocks, such as capitalflow reversals and collapses in commodityprices, may cause temporary increases in pov-erty that are difficult to reverse. Developingcountries must develop the capacity to man-age increased external and internal volatilitythrough better economic management, morerobust institutions for managing risks (suchas banks), and improved safety nets.

Safety nets. Before a crisis, safety nets canspur productivity and growth by providing the

67

insurance necessary for households to makerisky choices with higher potential returns.Safety nets also help ensure that crises do nothalt development. They help maintain essen-tial household investments in education andhealth and eliminate the need for the poor todivest themselves of physical capital. Settingup safety nets during times of economic growthmay be the only effective way to protect thepoor during crises (Ferreira, Prennushi, andRavallion 1999). The need for redistributioninevitably increases during crises, even if mostpeople’s incomes fall. Although the newly poorgenerate most of the increase in need, policiesneed not distinguish between the old and newpoor.

As a response to a temporary shock, in-creases in spending on social safety nets areideally financed over time, both past and fu-ture. Despite this logic (or indeed, because ofit) countries usually have to adopt policies withthe lowest budgetary costs. Targeting benefitsis one desirable means of keeping costs low.Self-selection mechanisms such as public worksare an important means of reducing the bud-getary costs of redistributive policies and ofestablishing institutions to deliver transfers.Establishing well-functioning institutions maybe one of the largest setup costs countries in-cur in response to deep crises. Ideally, ofcourse, such investments precede a crisis. Re-cent international experience confirms thatopen and transparent institutions operating ina noncorrupt way are as important to the es-tablishment of safety nets as they are to otherareas of public action.

The possibility of making a guarantee oflow-wage work on community-initiatedprojects the central element of a safety net maybe limited. If there is an institutional basis forsignificantly expanding workfare programs,with central and local agencies operating in atransparent and noncorrupt fashion, then theseschemes could play a significant role in allevi-ating poverty. The contribution of publicworks to poverty reduction tends to be largerin countries such as Korea, where the socialcosts of crises have primarily taken the form


of high unemployment. Public works are moreeffective at reducing poverty in these coun-tries because the opportunity costs of partici-pating in public works projects are lower forthe jobless than for the working poor. Publicworks are an important and useful option forreducing poverty during crises, especially indeveloping countries in need of infrastructureinvestments. However, they are best imple-mented alongside other programs for both theable-bodied population and those unable towork.

Notes1. Easterly and others (1993); Lutz (1994);

Mendoza (1994); Hausmann and Gavin (1995);Spatafora and Warner (1995); Deaton and Miller(1995); Collier and Gunning (1996); Guillaumont,Jeanneney, and Brun (1999); Lundberg and Squire(1999). The evidence concerns GDP and output growthand does not account for the direct real income growththat results from increased purchasing power on in-ternational markets.

2. Deaton and Miller (1995) suggest that in coun-tries that are marginal producers of commodities,where labor accounts for a significant share of costs,poverty itself may explain why real commodity pricesdo not increase in the long run. These prices cannotrise as long as there are unlimited supplies of labor atthe subsistence wage. Long-term marginal costs areset by poverty in tropical countries, and commoditytrade cannot contribute to reducing it. Incentives fortechnical progress are weak, and even when progressoccurs it tends to make prices fall while real wagesremain at the subsistence level.

3. The headcount index is the proportion of in-dividuals in the population whose income or consump-tion expenditures fall below the poverty line.

4. The elasticity for other measures of poverty,such as the poverty gap, are usually higher. See Liptonand Ravallion (1995).

5. For a lower-end Gini index of 0.25, the elas-ticity would be –3.3. It is –1.82 for a higher index of0.59 (Ravallion 1997).

6. Aizenman and Marion (1999) find similar re-sults.

7. This result holds under risk neutrality, but italso requires some degree of imperfect competition(Caballero 1991).

8. The authors also find that foreign aid and goodpolicies can offset this vulnerability.

68


9. This finding is contrary to earlier findings byMcBean (1966) that export instability has no effecton growth.

10. The asymmetry notion is also supported byfindings about the relationship between the headcountpoverty index and mean country income (Ravallion1997). More generally the poverty headcount will,for a given absolute decrease in income, increase morefor a poor than for a wealthy country (Milanovic1998).

11. Poverty figures for the rural areas are notavailable for Korea.

12. The figures used for 1997 are estimates basedon the declining poverty trends before the crisis.

13. Rural areas gained if they were not primarilyrice producing and were not affected by drought (pricesremained low in these areas until August 1998). Someareas also had natural disasters during the crisis. Thedrought of 1997–98 was not as bad on Java as hadbeen feared, but it did hit hard in the Eastern Islands,on the west coast of Sumatra, and in parts of Sulawesi.East Kalimantan suffered an ecological disaster whenthe drought interacted with wildfires.

14. The measured Gini coefficients are eitherbased on consumption expenditures or incomes. Inthe latter case they also may be biased, as they donot reflect adequately all incomes and changes of as-set values.

15. This result holds true even in industrial coun-tries. See Farber (1993), and Layard, Nickell, andJackman (1994).

16. There was a shift from rural nonagriculturalto agricultural jobs. Rural employment increased from56.05 to 57.37 million, while agricultural employmentincreased from 34.8 to 39. 4 million.

17. Among the crisis-hit countries, Korea is theonly one that had an unemployment insurance scheme,and had extended its potential coverage.

18. Budgetary allocations to support the unem-ployed in Korea were to double in April 1999 to in-crease the program’s coverage of the poor.

19. The savings response in Korea and Malay-sia, which have higher incomes than the other crisis-affected countries, may reflect greater wealth effectsfrom the crisis.

20. The average enrollment rate for the 13–19age group was around 60 percent in 1997.

ReferencesAgénor, Pierre-Richard. 1998. “Stabilization Policies,

Poverty and the Labor Market: Analytical Issuesand Empirical Evidence.” EDI, World Bank,Washington, D.C. December.

Aizenman, Joshua, and Nancy Marion. 1999. “Vola-tility and Investment: Interpreting the Evidencefrom Developing Countries.” Economica 66(262): 157–79. London. May.

Baptist, Jacqueline. 1999. “Public Expenditures onHealth and Education during the Crisis and Pre-Crisis Period in East Asia.” World Bank. August.

Bourguigon, Francois, Jaime de Melo, and AkikoSuwa. 1991. “Distributional Effects of Adjust-ment Policies: Simulations for Archetype Econo-mies in Africa and Latin America.” World BankEconomic Review 5: 339–66. May.

Caballero, Ricardo J. 1991. “On the Sign of the In-vestment-Uncertainty Relationship.” AmericanEconomic Review 81 (1): 279–88. March.

Collier, Paul, and Jan Willem Gunning. 1996. “PolicyTowards Commodity Shocks in DevelopingCountries.” IMF Working Paper 96/84. Interna-tional Monetary Fund, Research Department,Washington, D.C. August.

Cuddington, John. 1988. “Fiscal Policy in Commod-ity-Exporting LDCs.” Policy, Planning, and Re-search Working Paper (WPS) 33. World Bank,Washington, D.C. July.

Datt, Gaurav, and Martin Ravallion. 1997. “Macro-economic Crises and Poverty Monitoring: A CaseStudy of India.” Review of Development Eco-nomics 1 (2): 135–52.

Deaton, Angus S., and Ronald I. Miller. 1995. “Inter-national Commodity Prices, Macroeconomic Per-formance, and Politics in Sub-Saharan Africa.”Princeton Studies in International Finance 79.Department of Economics, International FinanceSection. Princeton, N.J.: Princeton University Press.

Dehn, Jan, and Christopher L. Gilbert. 1999. “Com-modity Price Uncertainty and Economic Growthand Poverty.” Processed. August.

De Janvry, Alain, and Elisabeth Sadoulet. 1998.Growth, Poverty and Inequality in Latin America:a Causal Analysis, 1970–94. Univeristy of Cali-fornia at Berkeley. September. (Presented at theInter-American Development Bank Conference onSocial Protection and Poverty. February 1999.)

Easterly, William, Roumeen Islam, and Joseph E.Stiglitz. 1999. “Shaken and Stirred: Volatility andMacroeconomic Paradigms for Rich and PoorCountries.” Michael Bruno Lecture, XIIth WorldCongress of the International Economic Associa-tion, Buenos Aires, August 27.

Easterly, William, and Art Kraay. 1999. “Small States,Small Problems.” Policy Working Paper 2139.World Bank, Washington D.C.

Easterly, William, and Stanley Fischer. 1999. “Infla-tion and the Poor.” Presented at the Annual Bank

69


Conference on Development Economics. WorldBank, Washington, D.C. April 28–30.

Easterly, William, Michael Kremer, Lant Pritchett, andLawrence H. Summers. 1993. “Good Policy orGood Luck? Country Growth Performance andTemporary Shocks.” Journal of Monetary Eco-nomics 32: 459–83.

Fallon, Peter, and Robert E. B. Lucas. 1999. “Losersand Winners During Economic Crises.” WorldBank, Washington, D.C. May.

Farber, Henry S. 1993. “ The Incidence and Costs ofJob Loss: 1982–91.” Brookings Papers on Eco-nomic Activity 1: 73–132.

Ferreira, Francisco H. G., Giovanna Prennushi, andMartin Ravallion. 1999. “Protecting the Poorfrom Macroeconomic Shocks: An Agenda forAction in a Crisis and Beyond.” Policy ResearchWorking Paper 2160. World Bank, Washington,D.C.

Foley, Mark, and Jeni Klugman. 1997. “The Impactof Social Support: Errors of Leakage and Exclu-sion. ” In Jeni Klugman, ed. Poverty in Russia.Public Policy and Private Responses. EDI Devel-opment Studies, Economic Development Institute.World Bank, Washington D.C.

Fox, M. Louise, Edward Amadeo, José MarcioCamargo. 1994. “Brazil.” In Susan Horton, RaviKanbur, and Dipak Mazumdar, eds. Labor Mar-kets in an Era of Adjustment Vol. 2. EDI Devel-opment Studies, World Bank, Washington, D.C.

Frankel, Jeffrey A., and David Romer. 1999. “DoesTrade Cause Growth?” American Economic Re-view 89(3): 379–99.

Frankenberg, Elizabeth, Duncan Thomas, andKathleen Beegle. 1999. “The Real Costs ofIndonesia’s Economic Crisis: Preliminary Findingsfrom the Indonesia Family Life Surveys.” Laborand Population Program Working Paper Series99–04, Rand, Santa Monica, Calif. March.

Gelb, Alan H. and Associates. 1988. Oil Windfalls:Blessing or Curse? New York: Oxford Univer-sity Press. (A World Bank Research Publication.)

Guillaumont, Patrick, Sylviane Guillaumont Jeanneney,and Jean-François Brun. 1999. “How InstabilityLowers African Growth.” Journal of AfricanEconomies 8 (1): 87–107.

Hausmann, Ricardo, and Michael Gavin. 1995. “Over-coming Volatility in Latin America.” InternationalMonetary Fund Seminar Series (International);No [1995-34]:1–86. August.

Islam, Rizwanul, and others. 1999. “Indonesia Coun-try Paper.” Paper presented at Seminar on Eco-nomic Crisis, Tokyo, 13–15 October 1999. ILO.Geneva. August.

Kakwani, Nanak. 1998. “Impact of Economic Crisison Employment, Unemployment and Real In-come.” Development Evaluation Division. Na-tional Economic and Social Development Board,Bangkok.

———. 1999. “Poverty and Inequality During the Eco-nomic Crisis in Thailand.” Office of the NationalEconomic and Social Development Board,Bangkok. NESB Newsletter. Volume 3, No. 1.

Kakwani, Nanak, and Nicholas Prescott. 1999. “Impactof Economic Crisis on Poverty and Inequality inKorea.” Processed. World Bank, Washington, D.C.

Klugman, Jeni. 1999. Social Safety Nets and Crises.PREMPOV. World Bank, Washington, D.C.

Klugman, Jeni, and Alexandre Kolev. 1999. “TheWelfare Repercussions of Single Parenthood inRussia in Transition. ” In Jeni Klugman andAlbert Motivans, eds. Family Structure in the NewRussia. London: Macmillan, forthcoming.

Layard, P. Richard G., Stephen Nickell, and RichardJackman. 1994. The Unemployment Crisis. Ox-ford; New York: Oxford University Press.

Lipton, Michael, and Martin Ravallion. 1995. “Pov-erty and Policy.” Chapter 41 in J. Behrman andT. N. Srinivasan, eds. Handbook of DevelopmentEconomics Vol. III, Elsevier Science.

Londoño, Juan Luis, and Miguel Székely. 1997a. “Dis-tributional Surprises After a Decade of Reforms:Latin America in the Nineties.” Working Paper352. Office of the Chief Economist. Inter-Ameri-can Development Bank, Washington, D.C. August.

———. 1997b. “Persistent Poverty and Excess Inequal-ity: Latin America 1970–1995.” Working Paper357. Office of the Chief Economist. Inter-Ameri-can Development Bank, Washington, D.C. Sep-tember.

Lundberg, Mattias, and Lyn Squire. 1999. “Growthand Inequality: Extracting the Lessons for Policy-makers.” World Bank, Washington, D.C. May.

Lustig, Nora. 1998. Mexico: the Remaking of anEconomy, Second edition. Washington, D.C.:Brookings Institution Press.

———. 1999. “Crises and the Poor: Socially Respon-sible Macroeconomics.” Sustainable Develop-ment Department, Poverty and InequalityAdvisory Unit. Inter-American DevelopmentBank, Washington, D.C. September.

Lustig, Nora, and Ruthanne Deutsch. 1998. “The In-ter-American Development Bank and PovertyReduction: an Overview.” Sustainable Develop-ment Department, Poverty and Inequality Advi-sory Unit. Technical Study No. Pov-101-R.Inter-American Development Bank, Washington,D.C. May.

70


Lustig, Nora, and Miguel Székely. 1998. “EconomicTrends, Poverty and Inequality in Mexico.” Sus-tainable Development Department, Poverty andInequality Advisory Unit. Technical Study No.Pov-103. Inter-American Development Bank,Washington, D.C. December.

Lutz, Matthias. 1994. “The Effects of Volatility in theTerms of Trade on Output Growth: New Evi-dence.” World Development 22 (12): 1959–75.

Mansor, Norma, Tan Eu Chye, Ali Boerhanordin,Fatima Said, and Saad Said. 1999. “MalaysiaCountry Paper.” Paper presented at Seminar onEconomic Crisis, Tokyo, October 13–15.

Matusz, Steven J., and David Tarr. 1999. “Adjusting toTrade Policy Reform.” Policy Research WorkingPaper 2142. World Bank, Washington, D.C. June.

McBean, Alasdair. 1966. Export Instability and Eco-nomic Development. London: Allen and Unwin.

Mendoza, Enrique G. 1994. “Terms of Trade Uncer-tainty and Economic Growth: Are Risk Indica-tors Significant in Growth Regressions?” Boardof Governors of the Federal Reserve System. In-ternational Finance Discussion Paper 491. De-cember.

———. 1995. “The Terms of Trade, the Real ExchangeRate, and Economic Fluctuations.” InternationalEconomic Review 36 (1): 101–37. February.

Milanovic, Branko. 1998. Income, Inequality, andPoverty during the Transition from Planned toMarket Economy. Regional and Sectoral StudiesSeries. Washington, D.C.: World Bank

Morley, Samuel. 1994. Poverty and Inequality in LatinAmerica: Past Evidence, Future Prospects. Over-seas Development Committee, Washington, D.C.

Na, Seong-Lin, and Hyung-Pyo Moon. 1999. “SocialImpact of Current Economic Crisis in Korea.”World Bank: World Development Report Con-sultative Meeting in Kuala Lumpur, May 10–12.

Pindyck, Robert S. 1991. “Irreversibility, Uncertaintyand Investment.” Journal of Economic Literature29 (3): 1110–48. September.

Poppele, Jessica, Sudarno Sumarto, and Lant Pritchett.1999. “Social Impact of the Indonesian Crisis:New data and Policy Implications.” World Bank,Washington, D.C. February.

Prescott, Nicholas. 1999. “Social Risk Management—Lessons from the Crisis.” Presented at the Insti-tute of Policy Studies on Asian EconomicRecovery: Policy Options for Growth and Stabil-ity, June 21–22.

Pritchett, Lant. 1998. “Patterns of Economic Growth:Hills, Plateaus, Mountains and Plains.” WorldBank Policy Research Paper 1947. Development

Research Group, World Bank, Washington, D.C.July.

Ramey, Garey, and Valerie A. Ramey. 1995. “Cross-Country Evidence on the Link Between Volatil-ity and Growth.” American Economic Review 85(5): 1138–51.

Ravallion, Martin. 1997. “Can High-Inequality De-veloping Countries Escape Absolute Poverty?”Economics Letters 56: 51–7.

Riveros, Luis A. 1994. “Chile.” In Susan Horton, RaviKanbur, and Dipak Mazumdar, eds. Labor Mar-kets in an Era of Adjustment Vol. 2, EDI Devel-opment Studies, World Bank, Washington, D.C.

Riveros, Luis A., and Carlos E. Sanchez. 1994. “Ar-gentina.” In Susan Horton, Ravi Kanbur, andDipak Mazumdar, eds. Labor Markets in an Eraof Adjustment Vol. 2, EDI Development Studies,World Bank, Washington DC.

Rodrik, Dani. 1998. “Where Did All the Growth Go?External Shocks, Social Conflict and GrowthCollapses.” NBER Working Paper 6350. NationalBureau of Economic Research, Cambridge, Mass.

Shin, Donggyun. 1999. “Statistical Overview of Cur-rent Labor Market Conditions and Trends.” Ko-rea Labor Institute. July.

Spatafora, Nikola, and Andrew Warner. 1995. “Mac-roeconomic Effects of Terms-of-Trade Shocks: theCase of Oil-exporting Countries.” Policy Re-search Working Paper 1410. World Bank, Wash-ington, D.C. January.

Stiglitz, Joseph E., and Amar Bhattacharya. 1999.“Underpinnings for a Stable and Equitable Glo-bal Financial System: From Old Debates to a Newparadigm.” Paper Presented at the Eleventh An-nual Bank Conference on Development Econom-ics. World Bank, Washington, D.C. April 28–30.

Subbarao, Kalinidhi. 1999. Financial Crisis and Pov-erty: Adequacy and Design of Safety Nets for theOld and New Poor in Korea. World Bank, Wash-ington, D.C. Draft.

Székely, Miguel. 1998. The Economics of Poverty,Inequality and Wealth Accumulation in Mexico.New York: St. Martin’s Press, in association withSt. Antony’s College, Oxford.

———. 1999a. Author’s unpublished calculations. Pov-erty calculated by adjusting income to match pri-vate consumption per capita from the NationalAccounts. Results for 1996 are preliminary. Inter-American Development Bank, Washington, D.C.

———. 1999b. “Poverty and Inequality in Mexico.”Presentation, Office of the Chief Economist. In-ter-American Development Bank, Washington,D.C. July.

71


UNICEF. 1998. Education for All? Regional Monitor-ing Report Number 5, UNICEF InternationalChild Development Centre, Florence, Italy.

World Bank. 1993. Global Economic Prospects andthe Developing Countries 1993. Washington,D.C.

———. 1997. Global Economic Prospects and the De-veloping Countries 1997. Washington, D.C.

———. 1999a. Global Economic Prospects and theDeveloping Countries 1998/99: Beyond Finan-cial Crisis. Washington, D.C.

———. 1999b. Panama Poverty Assessment: Priori-ties and Strategies for Poverty Reduction. Hu-

man Development Department, Latin Americaand Caribbean Region. Report No.18801 PAN.June.

———. 1999c. “Coping with the Crisis in Educationand Health.” Thailand Social Monitor. Washing-ton, D.C. June.

———. 1999d. Thailand Economic Monitor. Wash-ington, D.C. March.

———. 1999e. Russia: Targeting and the Longer-TermPoor. 2 Volumes. Poverty Reduction and EconomicManagement (ECSPE) and Human DevelopmentNetworks. Europe and Central Asia Region. Re-port No. 19377-RU. Washington, D.C. May.

73

SINCE THE ONSET OF THE EAST ASIAN CRISIS

more than two years ago, the corporatesectors and financial systems in the cri-

sis economies have remained in severe distress.Nonperforming loans (loans made by the fi-nancial system that are not being fully repaid)have skyrocketed to unprecedented levels: 19percent of all loans and 27 percent of grossdomestic product (GDP) in the Republic of Ko-rea, 20 percent of all loans and 30 percent ofGDP in Malaysia, 45 percent of all loans and60 percent of GDP in Thailand, and over 50percent of all loans and 25 percent of GDP inIndonesia. In contrast, nonperforming loansin other major emerging market crises (Chilein the early 1980s and Mexico in 1995) wereless than 20 percent of GDP. In the Scandina-vian banking crises during the early 1990s,nonperforming loans amounted to approxi-mately 5 percent of GDP. East Asia’s heavyreliance on bank-based financial systems andthe high debt-equity ratios of corporations havemade the economic distress especially acute.

Nonetheless, recovery has begun. Thisrecovery, along with the major policy measuresused to resolve the distress, raises the possi-bility that the process may now work in re-verse: a rising tide may lift all boats. Thatwelcome possibility, however, cannot be pre-sumed. While a strong cyclical recovery maycontinue, the aftereffects of the financial shockwill persist, and continued restructuring isessential both to reinforce that recovery andto reduce future vulnerabilities.

This chapter reviews the evidence on theextent of corporate and financial distress inEast Asia’s crisis countries and the significantprogress made to resolve that distress. It dis-cusses the relative roles of positive macroeco-nomic trends and financial restructuring forcontinued recovery and sustainable growth.It also draws policy lessons for managing cor-porate and financial distress.

The chapter reaches the following con-clusions:

• The ongoing recovery is still fragile anduneven. The externally triggered liquid-ity crisis during the second half of 1997indiscriminately submerged both strongand weak producers and financiers. Therising tide is lifting the strong, especiallythose benefiting from trade growth in elec-tronics products, but the financially weakcontinue to struggle on account of bothcrisis-induced and long standing vulner-abilities.

• Without vigorous corporate and financialrestructuring, the return to sustainablegrowth will likely take longer, the fiscalcosts of the crisis could rise, and the econo-mies will remain vulnerable to new exter-nal and internal shocks. Weak firms inEast Asia operated on thin margins in theyears leading up to the crisis, and theirinability to pay interest following the onsetof the crisis has added to their debt bur-den. Such firms constitute a significant

3Asian Restructuring:From Cyclical Recoveryto Sustainable Growth

74


portion of the corporate sector in each ofthe crisis economies, and the appetite toinvest in them is extremely limited. Theywill continue to act as a drag on invest-ment and growth until the financial claimson these firms are resolved, and either theiroperations return to adequate profitabil-ity or their assets are redeployed.

• Recognizing the urgency, East Asian gov-ernments were quick to create an institu-tional structure for corporate and financialrestructuring; they also earmarked fundsfor bank recapitalization. The politicalmomentum for reform has, however,slowed down, in part because the deeperstructural problems now need to be ad-dressed. Experience from other econo-mies, including Japan, shows that aslackening of the reform effort can undoprogress.

• Government restructuring initiatives—though required on many fronts—need tobe guided by two policy considerations:limiting the likelihood of systemic disrup-tion; and clarifying financial claims whilealso facilitating asset reallocation. To con-tain fiscal outlays, these initiatives shouldbe directed principally to honor the so-cial contract to protect bank depositorsand, where necessary, to preserve the pay-ments system and the orderly flow ofcredit. Government funds should notnormally be required for corporate re-structuring.

• Bank restructuring is important becauseit contributes to both policy objectives.Expeditiously restoring the health of thebanking system is required because apoorly capitalized banking sector createscontinued systemic risks and growing fis-cal liabilities for governments. Healthybanks are also best positioned to enforceclaims and to pursue corporate restruc-turing.

• The process of restructuring can itself bedisruptive if it is not carefully managed.Restructuring should be undertaken in amanner that ensures the integrity and the

organizational capital of the financial sys-tem so that prudent lending to businessesand households may continue. Achievingthis objective requires difficult choices.Having provided implicit or explicit guar-antees, governments can either moveahead rapidly by taking fiscal responsi-bility for the costs of the crisis, or theycan encourage private resolution of thedistress while applying regulatory forbear-ance. Waiting to resolve problems is likelyto make them worse. However, expedi-tious and transparent action should beaccompanied by market-based measuresto recoup fiscal costs and to signal cred-ibly a commitment to severely restrictguarantees and bailouts in the future.

• Corporate restructuring needs to deal firstwith the delineation and allocation oflosses. Improvements in accounting stan-dards and bankruptcy regimes can helpsupport this process. However, in the ab-sence of effective bankruptcy procedures,out-of-court procedures offer a mecha-nism for resolution. Once financial claimsare resolved, corporate restructuring canbe expected to occur through naturalmarket forces, except where major impedi-ments prevent such forces from working.Governments can facilitate asset mobil-ity by creating a framework for effectivedomestic and cross-border mergers and ac-quisitions. The Japanese experience cau-tions that, without an adequateframework for resolving claims and forfostering asset mobility, fundamental cor-porate restructuring can be indefinitelydeferred at a high economic cost even ina sophisticated economy.

The uneven recovery

Astrong cyclical recovery is taking placein the crisis economies of East Asia, rais-

ing the possibility that growth may alleviateor even eliminate the corporate and financialdistress. Although the sharp recovery from thedepression-like conditions is expected to con-

75

tinue over the short to medium term (see chap-ter 1), its transformation into high and sus-tainable growth will require more than thepresent temporary stimuli: the buildup in in-ventories, low interest rates, and gains fromcurrency depreciation. The recovery has beenuneven thus far, with rapid growth in the hightechnology sectors but more modest growth,and even continued decline, in important seg-ments of the Asian economies. Banking sys-tems, therefore, remain severely distressed. Thecorporate and financial distress can persist,absent vigorous restructuring, because the in-centives to accept and allocate losses are weak.That delay, in turn, can hamper growth byrestraining investment and raising the fiscalcosts of resolution.

The sources of unevennessThe observed unevenness in recovery is notsurprising. “Creative destruction” permits theatrophy of the weak and the shift of resourcesto higher productivity sectors (Harberger1998). For instance, with currencies still be-low precrisis levels, a period of slow growthin the nontradable sectors can be expected.The crisis has also emphasized weaknesses

in the competitive ability of traditional manu-facturing and has had disproportionate ef-fects on small- and medium-size firms.1

Though the unevenness is not surprising, itis important, since weak production perfor-mance contributes to the already massive fi-nancial sector distress and, in turn, hampersgrowth.

Nontradable sectors. The aftermath of thecrisis has seen a sharp decline in the nontradedsectors, where production remains belowprecrisis levels (figure 3.1). This is to be ex-pected because currency depreciations, whichfavor traded goods, reduce the incentive toinvest in the nontraded goods sectors. The poorperformance of nontraded sectors was also afeature of Mexico’s revival from its crisis(Krueger and Tornell 1999). Mexicannontraded production took almost three yearsto reach precrisis levels.

As discussed below, the share of firmsunable to pay their debts is significantly higherin the nontradable sectors than in the trad-able sectors. In Malaysia about three-quartersof the nonperforming loans are to enterprisesin the nontradable sectors. The high distressreflects endemic characteristics. Even prior to

A S I A N R E S T R U C T U R I N G : F R O M C Y C L I C A L R E C O V E R Y T O S U S T A I N A B L E G R O W T H

Figure 3.1 Nontradable production before and after crises

105

Index=100 at the start of the crisis

100

95

90

856543210–1–2–3–4–5–6

Note: The index is a three-quarter moving average.Source: Datastream.

Mexico

Republic of Korea

Malaysia

Precrisis Postcrisis

Quarters

76


the crisis, the nontradable sectors had beencharacterized by overcapacity and low pro-ductivity (Crafts 1999), reflecting local mo-nopolies in sectors such as retail trade anddistribution. The Japanese experience showsthat deregulation of domestic trade is impor-tant to spur competition and to increase pro-ductivity (Alexander 1999). Low productivityalso reflects excess capacity in the real estatesector.

Weaknesses in traditional manufacturing.Of all the crisis countries, Korea’s industrialproduction has recovered the fastest, risingabove precrisis levels (figure 3.2). The morerapid recovery in Korea reflects in part itsgreater strengths in sectors such as electron-ics, computers, and telecommunications (fig-ure 3.3). Korean firms have also done well inthe transport equipment sector, whereas Ma-laysian and Thai firms in this sector have suf-fered. Traditional manufacturing sectorswould have been expected to lead the way torecovery in the lower wage crisis countries.In Thailand the textiles sector grew rapidlyfollowing the depreciation, but output hasfallen back to precrisis levels as the currencyhas appreciated. Thai products are having a

hard time competing in export markets (EIU1999b). Traditional manufacturing in Korearebounded only slightly after the crisis, rein-forcing a secular decline that significantly pre-dates the crisis (figure 3.3).

Effects on small and medium-size firms.Small and medium-size firms are suffering dis-proportionately. While aggregate Korean in-dustrial production bottomed out in late 1998,production by small firms continued to fall inabsolute terms until July 1999, resulting in adecline of about one-third from precrisis pro-duction levels. In other countries, where smalland medium-size firms have a greater indus-trial presence, their financial inability to with-stand crisis has proved more of aneconomy-wide setback (see Domac and Ferri1998 for Korea; Domac 1999 and EIU 1999afor Malaysia; Mako 1999 for Thailand). Forexample, more than 50,000 small firms and400,000 households throughout Thailand ac-count for about 50 percent of the country’snonperforming loans (Mako 1999).The inabil-ity to restructure these debts effectively con-tributes to financial sector problems, whichfeed back into continued financial difficultiesfor small firms.2

Figure 3.2 Industrial production before and after crises

110

Index=100 at the start of the crisis

100

90

80

70-24 -18 -12 -6 0 6 12 18 24

PostcrisisPrecrisis

Indonesia

Thailand

Malaysia

Republic of Korea

Note: The index is a three-month moving average.Source: Datastream.

Months

77

Even though large firms have been thedrivers of recovery, they—especially the largeconglomerates—also pose systemic risks. InKorea the onset of the crisis was, in part, as-sociated with the collapse of two conglomer-ates, Hanbo Steel and Kia. In Thailand thefinancial troubles of Thai Petrochemicals sym-bolized overinvestment in capacity and exces-sive reliance on external debt. Throughout theregion diversified conglomerates were initiallyregarded as too big to fail, as demonstrated in

the Korean and Malaysian governments’ earlyefforts to support the survival of their largestbusiness groups. However, that perception maybe changing, especially in Korea, as troubleshave mounted at the chaebol Daewoo, wherea creditor-led restructuring is ongoing.

Continued high levels of corporate andfinancial distressThe uneven recovery is reflected in continuedcorporate and financial distress. Two measures


Table 3.1 Corporate distress, past and projected, 1995–2002(percentage of firms unable to meet current debt repayments)

1999 (Q2) 2000– 2000–

1995 1996 1997 1998 2002a 2002b

Country Total Total Total Total Total Manufacturing Services Real estate Total Total

Indonesia 12.6 17.9 40.3 58.2 63.8 41.8 66.8 86.9 52.9 60.8Korea, Rep. of 8.5 11.2 24.3 33.8 26.7 19.6 28.1 43.9 17.2 22.6Malaysiac 3.4 5.6 17.1 34.3 26.3 39.3 33.3 52.8 13.8 17.4Thailand 6.7 10.4 32.6 30.4 28.3 21.8 29.4 46.9 22.3 27.1

a. Estimate, based on the assumption that interest rates stay at their current level throughout the period.b. Estimate, based on the assumption that interest rates regain their 1990–95 averages.c. Malaysian firms in agriculture and utilities bring down the average for all firms in 1999.Note: Growth rates assumed through 2002 are based on IMF projections (IMF 1998).Source: Claessens, Djankov, and Klingebiel 1999; sectoral estimates provided by Claessens, Djankov, and Klingebiel for thispublication.

Figure 3.3 Production index in Korea by industry(1995=100, seasonally adjusted)

120

30

Q2 1996

Source: Datastream.

60

90

150

180

210

240

270

300

Q3 1996

Q4 1996

Q2 1997

Q1 1997

Q3 1997

Q4 1997

Q1 1998

Q2 1998

Q3 1998

Q1 1999

Q4 1998

Q2 1999

Q3 1999

Communication equipment

Transport equipment

Base metals

Paper and pulp products

Computers

ChemicalsFood

78


of distress are firms’ inability to meet their debtobligations and the mirror image of that in-ability in nonperforming loans on the balancesheet of banks. These measures rank the levelof country distress similarly. Indonesia has thehighest level of financial distress, whereasKorea and Malaysia have the lowest.3 Thedistress in all four countries is, however, his-torically severe when compared with othercountries that have experienced financialcrises, because of the high levels of bank credit-to-GDP ratios and high corporate debt-to-equity ratios (World Bank 1999). While thesharply lower interest rates should providerelief, several factors, which are likely to per-sist, have kept the distress at high levels.

Interest rates and fiscal distress. Lowerinterest rates are unlikely to suffice in elimi-nating distress. Based on financial statementsof firms listed on stock exchanges, the abilityof firms to meet their current interest paymentobligations can be estimated (table 3.1). Theseestimates need to be interpreted carefully be-cause they are typically based on a smallsample of listed firms for which the most com-plete information is available.

In all countries, the level of distress hadbeen building since 1995. In 1996, even whengrowth was still booming, more than 10 per-cent of firms (except in Malaysia) were alreadyunable to service their debt. The estimatesshow that in the second quarter of 1999 morethan a quarter of listed firms in Korea and

Malaysia were unable to service their currentdebt repayments. In Indonesia almost two-thirds of all firms were under severe liquiditystress (table 3.1). In all countries, distress wasespecially high in the nontraded sectors (ser-vices and real estate), as could be expectedfrom the trends in nontraded production de-scribed in the previous section.

The crisis of 1997 moved many marginalfirms into illiquidity. Moreover, such firmshave accumulated debt since the crisis becausethey have been unable to make interest pay-ments. This suggests that many firms that haverecently emerged from the worst effects of thecrisis are still in a precarious situation and arevulnerable to further shocks. Projections for2000–2002 show that, on current assumptionsof growth rates in the respective countries, asignificant portion of the firms will remain indistress. If interest rates rise from their presentlow levels to their 1990–95 averages, the dis-tress will be even greater.

Interest rates and nonperforming loans.Nonperforming loans increased in the first halfof 1999 despite declining interest rates, andare stubbornly high—at historically unprec-edented levels (table 3.2).4 In Thailand theproblems now center around the commercialbanks because, following their closure afterthe crisis, the assets in nonbank finance com-panies have shrunk to a small fraction of fi-nancial system assets. However, in Korea andMalaysia, the noncommercial bank sector (in-

Table 3.2 Ratio of nonperforming loans to total loans, December 1998–September 1999(percent)

Malaysia Rep. of Korea Thailand

Dec. 1998 June 1999 Dec. 1998 June 1999 Dec. 1998 Sept. 1999

Commercial banks 13.0 12.8 7.4 8.7 42.9 44.6Merchant banks 30.6 31.6 20.0 11.9 — —Other financial institutions 26.8 23.9 13.1 14.5 70.2 62.3Asset management companies 100.0 100.0 100.0 100.0 — —Total financial system 19.7 21.2 16.8 19.2 45.0 45.3

— Not applicable.Note: Nonperforming loans are measured on a gross, three-month basis and include assets carved out for sale by the assetmanagement companies, which by definition have 100 percent of their loans nonperforming. The steps toward sales ofnonperforming loans are discussed later in this chapter.Source: Financial Supervisory Services (Republic of Korea), Bank Negara (Malaysia), and Bank of Thailand.

79

surance companies and investment and trustcompanies in Korea, and finance companiesand merchant banks in Malaysia) continuesto account for about a quarter of thenonperforming loans. Projections of contin-ued growth in nonperforming loans stemlargely from the likely increase of such loansin the nonbank financial institutions (see, forexample, Xie 1999).

The decline in interest rates has not beensufficient to provide immediate relief. On thecontrary, especially in Thailand, but also inIndonesia and Korea, nonperforming loansrose even as interest rates fell (figure 3.4).5

Fragile firms, operating on thin margins, ex-perienced a severe decline in their net worthwhen interest rates rose sharply. The sharp fallin output further aggravated the problem.Recovery for the distressed firms will likelybe slow. Experience shows that economicdownturns associated with financial criseshave more enduring consequences than down-turns caused, for example, by inventory-drivenbusiness cycles (Furman and Stiglitz 1998).

Risks of a low-level equilibriumGiven enough time, financial institutions andcorporations can overcome their distress as

stakeholders resolve their claims on assets,even in the absence of formal bankruptcy pro-cedures, and as restructuring is induced bymarket pressures. Without additional shocks,the economies would then return to their newlong-term sustainable growth path, whichcould be lower than the precrisis level (WorldBank 1999). The important issue is: how muchtime? That is, can better management of therestructuring process reduce the costs of thecrisis and the length of the period in a “lowlevel equilibrium”?

A slowdown or mismanagement of therestructuring process raises two concerns. First,continued distress lowers investment, whichlowers growth, and in turn further contrib-utes to nonperforming loans and reduced in-vestment and growth prospects. While rapidrecovery may counteract this negative dy-namic, the evidence cautions against such apresumption. Nonperforming loans are likelyto remain high, and investment rates havefallen sharply, lowering growth prospects inthe short run. Second, strong incentives existfor all parties to wait rather than to resolvetheir problems (box 3.1). Failure to assess andallocate the losses could lead to their social-ization and rising fiscal costs.


Figure 3.4 Thai nonperforming loans and interest rates, June 1998–September 1999

50

45

40

35

30

25

20

18

16

14

12

10

8

6

4

2

0

PercentPercent

June1998

3-month moneymarket rate(right axis)

July1998

Aug.1998

Sept.1998

Oct.1998

Nov.1998

Dec.1998

Jan.1999

Feb.1999

Mar.1999

Apr.1999

May1999

June1999

July1999

Source: Bank of Thailand.

Nonperforming loans ratio(left axis)

Sept.1999

80


Nonperforming loans in East Asia. InThailand nonperforming loans have declinedmodestly from their peak levels. In other coun-tries, however, data up to June 1999 showcontinued growth in nonperforming loans. Anumber of factors contribute to the high levelof nonperforming loans and the possibility thatthey may actually grow in the short and me-dium term. First, the accounting methods inplace have not revealed the full extent ofnonperforming loans, especially in Korea andMalaysia, where nonbank financial companiesare especially important. Poor accounting andfaulty credit analysis masked borrowers whowere often connected to the lending institu-tions and who could not pay but were never-theless able to borrow repeatedly to meet theirdebt obligations. While many of thesenonperforming borrowers have been revealed,some observers believe, for example, thatKorean nonperforming loans could grow byover a third from their present levels beforethey start falling (Warburg Dillon Read 1999).

Second, many firms with thin operatingmargins and high debt levels are endemicallyweak, and thus have been unable to servicetheir debt despite the recovery. Because of thecapitalization of interest, their debt levels havegrown and will remain above precrisis levelsfor a number of years, even under optimisticgrowth scenarios.

Third, announcements of restructuringagreements led to the expectation thatnonperforming loans would fall as a share oftotal bank loans. However, most agreementsare just that—agreements in principle—andwill take time to become effective. More im-portant, and as discussed below, restructur-ing agreements have mainly taken the form ofdeferred debt payments, and there is little evi-dence to suggest that assets have been funda-mentally repositioned. As such, someagreements have proved unsustainable.

Fourth, except in Korea, there has beenvirtually no new lending, implying that banksare not likely to grow out of their bad debt

easily forthcoming. The problem is aggravatedbecause those in distress have an incentive toundertake further risky investments. Their risksare no greater than when they were merely wait-ing for the resolution of past problems. For bothcreditors and debtors, there is the hope that theproblems will just go away when growth resumesor if the government bails them out.

As firms and banks wait for their fortunesto improve, debt continues to accumulate. Thelonger that interest payments are deferred andcapitalized, the higher future growth must be fora company to emerge from negative to positivecash flows. The problem becomes systemic ascorporate distress reduces demand by loweringthe purchases of inputs and by reducing consumerconfidence, further increasing the strain on corpo-rate cash flow and balance sheets. Continuedrestructuring is required to counteract such anegative feedback loop.

Box 3.1 Why distress can persist

T he combination of high corporate financialdistress, unabated banking sector problems,

low levels of investment, and reduced growth ratescan cause distress to persist. In a good equilibrium,several factors go together, including high growth,high demand, rising property values, and an appro-priate level of credit from the financial system. In abad equilibrium, poor corporate performance con-tributes to banks’ nonperforming loans, reducingboth their capacity and willingness to lend. Whenproperty represents a large proportion of the cor-porate balance sheet or important collateral forloans, as in many parts of East Asia, a decline inproperty prices further weighs down recovery.

The private incentives to wait hamper theresolution of the losses. For debtors who haveexperienced large losses in their equity holdings,there is little downside to waiting. They cannotlose much more. For creditors, acknowledging thelosses requires raising new capital, which is not

81

problem soon. In Indonesia and Malaysia,deposits have grown significantly in 1999 (fig-ure 3.5), reflecting a confidence in the bank-ing systems, or at least in the governments’ability to meet their obligations to insureddepositors. Increased deposits, in turn, haveimproved banks’ liquidity, but that has not ledto increased lending to the private sector.

Korea, the only crisis country to experience acredit expansion, has also had the strongestrecovery. Such credit expansion will continueto be necessary as Korean firms stopdestocking and start rebuilding their invento-ries. However, with most of the banking sys-tem now owned and controlled by thegovernment, the quality of the lending remains


Figure 3.5 Change in bank deposits and lending to the private sector, December 1998–July 1999

20

10

0

-10

-20

Percentage change

-30

-40

-50

-60Indonesia Republic of KoreaMalaysia Thailand

Note: June 1999 is used for the Republic of Korea.Source: IMF, International Financial Statistics.

Total deposit

Total lending

Figure 3.6 Total investment in East Asia

45

40

35

30

25

20

15

10

5

0Indonesia Republic of KoreaMalaysia Thailand

Source: Datastream; IMF, International Financial Statistics.

Percentage of GDP

Q2 1999

Q1 1999

1998

1992–97 (average)

82


a concern. Korean banks hold significant de-posits in the financially fragile investment andtrust companies and have been under somepressure to retain these deposits.

Estimates also suggest that income gener-ated from the difference between lending anddeposit interest rates will not to be enough todeal with the nonperforming loans and therecapitalization of banks in Indonesia andThailand (Claessens, Djankov, and Klingebiel1999). In Korea and Malaysia, the better banksare positioned to grow out of their problems.However, even in those countries, significantdistressed segments will require continued re-structuring and infusion of capital. Koreanbanks, for example, are expected to lose 4 tril-lion won in 1999 to provision against strin-gent asset classification standards and becauseof increased exposure to nonperforming loans.

Fifth, some fraction of nonperformingloans are strategic—that is, borrowers canrepay, but choose not to because they cannoteasily be pursued by their creditors. Accord-ing to informal estimates, the share of strate-gic defaulters in Thailand is between a fifthand a third of all defaulters. This adds to theburden on banks.

Distress, investment, and growth. A highlevel of distress lowers investment and growthprospects. Investment rates have fallen sharplysince the onset of the crisis (figure 3.6). Rela-tive to the average of 1992–97, the investmentrate in the second quarter of 1999 was downby about 57 percent in Indonesia, 40 percentin Thailand, and 30 percent in Korea. Theextent of the fall in investment is greater wherethe ratio of nonperforming loans is higher.Thus, distressed firms, unable to meet theirdebt service obligations, have been unable toobtain credit and to undertake new investment.Healthy firms have continued to invest, includ-ing in restructuring to reposition and insulatethemselves from future crises. The extent ofthe decline in investment in Malaysia is greaterthan may have been anticipated by the rela-tively low level of nonperforming loans andby the low exposure of Malaysian banks andcompanies to foreign currency debt.

Where investment was initially excessiveand misdirected, the fall in investment is de-sirable. However, a prolonged drought in in-vestment could continue to depress growthin the short run and erode competitive abil-ity in the long run. Moreover, the evidenceshows that a decline in investment spendingis associated with reduced consumer confi-dence and reduced consumer spending, thecumulative effect of which is to reduce growthsignificantly.

The fall in interest rates from the highpostcrisis levels should help stimulate invest-ment, though short-term prospects are damp-ened by a recent rise in the costs of capital.During the second quarter of 1999 the invest-ment rate actually increased in Korea andMalaysia (figure 3.6). However, low levels ofinvestment can be expected to continue in theshort term, because significant capacity liesunused, corporate distress is still widespread,and continued uncertainties remain. In addi-tion, the large fluctuations in stock market in-dices since the peaks reached earlier in the year(see chapter 1) and the continuing upwardpressure on government bond yields imply arising cost of capital and a higher discountingof future growth prospects.

A negative feedback loop potentially op-erates: distressed firms weigh down growthprospects, but it is growth that helps the firmemerge from its financial troubles. For ex-ample, a firm that is just able to service itsdebt before a shock and then, following theshock, misses a year’s debt service finds itselfin deeper trouble at the end of the year as theinterest is capitalized. To meet its higher debtservice obligation, the firm must grow sig-nificantly faster than the rate of interest. Fora firm with high dependence on debt, agrowth rate double the interest rate, continu-ously for two or three years, may be requiredto achieve positive cash flows net of interestpayments.

International experience with restructur-ing. International experience suggests thatdelaying restructuring is costly. Studies of fi-nancial crises show that the average recovery

83

time back to trend growth rates is 2.8 yearsfor banking crises and 1.5 years for currencycrises (IMF 1998). The time required to re-solve banking crises in Latin America has beenlonger, approximately four to five years (Rojas-Suarez and Weisrod 1996).

The Mexican experience in 1995 and theChilean experience in 1982 offer useful les-sons for East Asia. The Mexican recovery ben-efited from the country’s participation in theNorth American Free Trade Agreement andfrom a favorable international economy.Mexico also took steps early on to resolve itsbanking crisis. The government carved out asubstantial fraction of bad loans in the sys-tem and placed them in a special agency, theFund for the Protection of Bank Savings(FOBAPROA), to be managed and sold.

However, despite its early and impressiverecovery from the crisis, Mexico’s growthperformance since that time has been mod-est, especially in 1999. Growth has beenweighed down especially by the sluggishnessin the domestic economy, including thenontradable sector, and by the unwillingnessand inability of the banks to lend.FOBAPROA was unable to sell virtually anyof the assets it had acquired. More important,the banking sector’s problems were not fullyresolved and nonperforming loans continuedto increase. Krueger and Tornell (1999, 33–34) find that “nonperforming loans are un-likely to disappear on their own, even undera high GDP growth scenario.” The contin-ued presence of nonperforming loans has hurtthe ability of the bank sector to perform itsfunctions adequately, with credit especiallyconstrained to producers selling in the domes-tic market. The authors draw three lessonsfrom their findings. First, the Mexican authori-ties could have been more ambitious in ex-tracting problem loans from the bankingsystem. Second, the government did not takesufficient steps to subsequently discipline thebanking sector, leaving open the prospect offurther bailouts. Finally, Mexico’s bankruptcyprocedures are still ineffective, rendering re-structuring problematic.

In comparison with Mexico’s recovery,Chile’s recovery was much slower. Yet, despitethe slow recovery, Chile has enjoyed robustand sustained growth. What explains the dif-ference? First, although initially slow to rec-ognize the full extent of the problem, Chileanauthorities persisted in their efforts to resolvethe problems of the banking sector, includingundertaking measures to discipline it. Second,Chile undertook far-reaching reforms to fos-ter capital market development and to encour-age greater competition in the economy,especially in nontradable sectors such as in-frastructure.

Mounting fiscal costsThe fiscal costs of the crisis are large (table3.3). These estimated costs are illustrative anddepend upon a number of assumptions, includ-ing the extent of nonperforming loans at theirpeak, the degree to which the nonperformingloans will have some future value, and theinterest rate that the governments will needto pay for the recapitalization funds (as de-scribed in more detail in the next section, table3.7). Keeping in view these limitations, in allcountries plausible scenarios indicate that thebank recapitalization costs are significantlylarge in relation to existing public debt. Onceagain, the extent of Korea’s problem is largebut still modest in relation to that of others,while Indonesia’s problem is the most severe.Korea has had low public debt and, while re-capitalization costs are significant and mayeven grow, the interest burden is modest. Thereported Malaysian recapitalization costs arerelatively small because it is assumed that sig-nificant repayments can be collected fromexisting nonperforming loans. Without theability to collect on nonperforming loans,however, Malaysian debt levels will actuallybe higher.

These higher debt levels can reducegrowth prospects through different channels.By increasing the government demand forfunds, they raise interest rates and, hence,crowd out private investment. Governmentflexibility to act in a countercyclical manner


84


is reduced. Higher income and corporate taxrates reduce incentives to invest while highertrade taxes can lead to misallocation of re-sources. Note also that recent analysis hascalled into question the reported budget defi-cit estimates in the crisis countries (Kharas andMishra 1999). The analysis indicates that thetrue budget deficits have been larger than re-ported because activities with significant fi-nancial implications have been undertaken offbudget. In such a context, the rise in the debtlevels could have a more serious impact oninterest rates, government flexibility, and taxrates.

The focal point of restructuring:the financial sector

In contrast to corporate restructuring, whichshould be led by the private sector, finan-

cial sector restructuring is more thegovernment’s responsibility. Major system-wide concerns (safeguarding the paymentssystem and restoring credit availability) andthe potentially large fiscal costs are centeredaround banks and other distressed segmentsof the financial systems. Moreover, throughrestructuring the financial system, govern-ments can facilitate corporate restructuring:healthy and soundly managed financial inter-

mediaries are better positioned to negotiatewith borrowers and to encourage corporaterestructuring than are governments. Govern-ments can also assist corporate restructuringby implementing policies that clarify and en-force financial claims, as discussed in the nextsection.

Governments in East Asia took early stepsto contain the crisis by extending insuranceto depositors and creditors. As the crisis spreadand deepened, this was followed by the estab-lishment of an institutional structure for man-aging the restructuring process and by closing,merging, and nationalizing several banks andnonbank financial companies. Significant re-capitalization funds were committed, some ofwhich have since been disbursed.

The lesson of this crisis, as well as of pastcrises, is that the momentum of governmentaction needs to be maintained, while ensur-ing, whenever possible, that the informationaland organizational capital of the financialsystem is preserved. Forbearance has been usedto permit the graduated attainment of pruden-tial standards, and this can have some impor-tant benefits. Judging by experience, however,continued generalized forbearance runs therisk of increasing the scale of future problems.Because governments have provided extensiveguarantees to bank depositors and creditors,

Table 3.3 Public debt and recapitalization costs as share of GDP, 1998(percent)


Public debt, 1996 23.9 35.3 8.0 3.7Public debt, 1998a 72.5 33.3 10.5 14.6Estimated recapitalization costs a, b 58.3 10.0 16.0 31.9

Funds disbursed 10.6 4.2 12.5 23.9Expected additional costs 47.7 5.8 3.6 8.0

Estimated debt after recapitalization 106.6 43.3 26.5 46.6Total interest payment 16.7 3.1 1.9 5.0

Portion for recapitalization 9.2 0.7 1.2 3.4Fiscal surplus/deficit, 1998a 1.4 –1.6 –2.9 –2.8Government bond yield (percent)c 15.7 7.3 7.2 10.8

a. 1997 data is used for Indonesia.b. For details on recapitalization costs, see table 3.6. For Thailand, the fiscal costs in this table include the net costs incurred forthe finance companies (B600 billion) less private resources raised (B250 billion).c. The bank lending rate is used for Malaysia.Source: IMF, International Financial Statistics.

85

on balance, it is desirable for governments toassume early responsibility for recapitalization.At the same time, measures to contain the fis-cal burden should be used to signal a commit-ment that the government is serious aboutlimiting further exposure through explicit orimplicit guarantees. These measures includesharing in the upside of nonperforming assetsthat are sold, making it worthwhile for banksto recoup from defaulting debtors whennonperforming loans are left with banks, andprivatizing acquired banks.

Early and stronggovernment involvementThe East Asian economies were quick to be-gin dealing with the banking sector crisis bycreating new institutions, reorganizing the fi-nancial sector, and creating mechanisms forasset resolution and the recapitalization ofbanks.

Institutions for restructuring. An impres-sive array of institutions has been put in placeto deal with corporate and financial restruc-turing (table 3.4). The agencies for voluntary


Table 3.4 Institutional arrangements for corporate and financial restructuring

Voluntary corporate workout Asset resolution company Agency for bank recapitalization

Indonesia Jakarta Initiative Indonesian Bank Indonesian BankTask Force Restructuring Authority Restructuring Authority

Korea, Rep. of Corporate Restructuring Korea Asset Management Korea Deposit InsuranceCoordination Committee Corporation Corporation

Malaysia Corporate Debt Danaharta DanamodalRestructuring Committee

Thailand Corporate Debt Restructuring Financial Sector Restructuring Financial RestructuringAdvisory Committee Authority and Asset Advisory Committee

Management Corporation (funded by the Financial(for nonbank finance Institutions Developmentcompanies) Fund)

Table 3.5 Structural changes in the financial system

Closures State takeovers Mergers

Indonesia 64 banks 12 commercial banks 4 of 7 state banks to be merged(18 percent) (20 percent) into a single bank (54 percent)

Korea, Rep. of 5 commercial banks, 4 commercial banks 9 banks and 2 merchant banks17 merchant banks, and more (25 percent) to create 4 new commercialthan 100 nonbank financial banks (15 percent)institutions (15 percent)

Malaysia None 1 commercial bank, 6 mergers of finance companies1 merchant bank, and 3 financial and commercial bankscompanies under central bank (2 percent)control (12 percent)

Thailand 57 finance companies 7 commercial banks 5 commercial banks and 13(11 percent) and 1 (13–15 percent) and 12 finance finance companies into 3commercial bank (2 percent) companies (2.2 percent) banks (20 percent)

Note: Figures in parentheses refer to percentage of assets in the financial sector.Source: IMF 1999; World Bank country reports.

Source: IMF 1999; World Bank staff.

86


corporate workouts and asset resolution haveall been established since the crisis. In Malay-sia the agency for bank recapitalization,Danamodal, is also new. Recapitalization agen-cies in the other countries have been adaptedto deal with the crisis. The contrast with Ja-pan is especially striking. In Japan, recogniz-ing the problem took much longer, but in thecrisis countries of East Asia, the problems weretoo severe to wait.

Reorganization of the financial sector. Eventhough the institutional structures for dealingwith the crisis are similar, the different coun-tries have chosen quite different restructuringoptions. These options range from closing non-viable financial institutions early on and dis-posing of their assets, to retaining the institutionsbut fostering strength through mergers.

The early closure option has been em-ployed in several countries. In Thailand vir-tually the entire segment of finance companieswas closed; in Korea select finance companiesand commercial banks, and many small bankswere closed; and in Indonesia several weakbanks were closed (table 3.5). However, in In-donesia more than 170 banks remain evenafter the closures. Malaysia has not closed anyof the financial institutions and is relying in-stead on extensive mergers of financial insti-tutions; the government expects mergers withgood banks to help resolve the problems ofthe poorly performing banks. The Malaysianplan to mandate the reconstitution of the en-tire financial sector into six groups has givenway to a more flexible, but as yet evolvingapproach.

While Thailand has largely dismantled itsnonbank financial institutions and Malaysiahas decided to merge them into more viable,often parent, banks, Korea has yet to developa strategy for this segment of the financialsector. The close ties between the nonbankfinancial institutions and the Korean chaebolshas complicated and aggravated the restruc-turing task. For example, the absence of earlyrestructuring at the second largest chaebol,Daewoo, led to a significant deterioration inthe financial status of nonbank financial com-

panies. These institutions, moreover, were notsubjected to the necessary discipline and con-tinued to lend to Daewoo, even while its vi-ability was in question. The revelation of thatdebt’s unsustainability casts the main shadowon Korea’s recovery.

Since governments have become substan-tial owners of the banking systems throughtheir direct takeovers and recapitalization ini-tiatives, the reprivatization of these institutionsposes a major challenge that will influence thelong-term structure and performance of thefinancial sectors. So far, efforts at privatizationhave encountered problems, partly as a resultof the continued growth of nonperformingloans, which new acquirers have difficultyvaluing. The recent experience suggests thatdifferences in perceptions of value can be large.The protracted negotiations for the sale ofKorea First Bank centered around the valua-tion of nonperforming loans that had not beencarved out or revealed and on the extent ofcontinued government obligations to assumenonperforming loans following theprivatization. With Daewoo as a principal cli-ent of Korea First Bank, the uncertainties invaluation were not altogether surprising. Thesale of Seoul Bank has, for the present, beendeferred. In Indonesia and also in Thailand,sales of banks have stalled for similar reasons.While several possibilities exist in Thailand,so far in 1999 two small nationalized bankshave been sold, with substantial governmentinjection of funds or commitment to assumeresponsibility for further growth innonperforming loans.

Asset resolution mechanisms. The twoextreme choices for asset management strate-gies include setting up a government agencywith the full responsibility of acquiring, re-structuring, and selling the assets or lettingbanks manage their own nonperforming as-sets. A specialized agency may be requiredwhen the task of dealing with nonperformingloans is fundamentally different from that ofmaking new loans and banks possess limitedmanagement capacity with a comparative ad-vantage in new lending. However, justification

87

for that agency to be government-owned and-operated arises either when extrajudicial pow-ers are required to deal with the nonperformingloans problem, or when there exist significanteconomies of scale in asset management thatcannot be realized by private contracting. In-termediate approaches include those employedin Thailand with the government principallyacting as an intermediary for the market-basedsales of nonperforming assets or facilitatingprivate asset management through tax incen-tives.

Governments removed a significant shareof nonperforming loans from the financialsystem and transferred them to governmentagencies: 26 percent ($37 billion) in Korea,66 percent ($28 billion) in Indonesia, and 50percent ($11 billion) in Malaysia (Claessens,Djankov, and Klingebiel 1999).6 In Thailand,the entire assets of 57 nonbank finance com-panies (over $20 billion) were transferred tothe Financial Sector Restructuring Authority,followed by the only significant subsequentresale of nonperforming assets. However, theThai government has not acquired thenonperforming assets of commercial banks.

Government ownership or managementof a specialized asset management companymay be justified if it is given administrativepowers that overcome the higher transactioncosts of the regular bankruptcy and judicialsystem. In Indonesia and Malaysia, the Indo-nesian Bank Restructuring Authority and

Danaharta, respectively, have extrajudicialpowers to receive compensation from debt-ors. Malaysia’s Danaharta also plans to takea more active role in restructuring assets be-fore selling them. Through restructuring of theassets prior to their sale, the expectation isthat value will be enhanced. The evidence onsuccessful restructuring by a government-runasset management company for loans otherthan for real estate is, however, weak(Klingebiel 1999). The realization rate byMexico’s FOBAPROA is expected to be in thelow single digits. The resources and skills re-quired for success restructuring of assets aredemanding. In fact, government agencies canreduce the effectiveness of market-based so-lutions. The terms offered by Danaharta tobanks—for example, special provisioning re-quirements and generous share of recoveredamounts—has implied that private buyers ofdistressed debt have essentially been priced out,thus reducing the space for market-orientedrestructuring.

In Korea, and especially in Thailand, al-ternative market-based approaches are beingattempted. The speed at which the FinancialSector Restructuring Authority in Thailandoperated was noteworthy. The realization rateof 25 percent was low, but there is no evidencethat waiting would have increased it. In Ko-rea, the Korea Asset Management Company,which acquires the nonperforming loans, ex-pects to delegate the task of managing and


Table 3.6 Estimated recapitalization costs for commercial banks, mid-October 1999

Amount disbursed Remainingfiscal costs

Percentage as percentageEstimated costs Local currency U.S. dollars of GDP of GDP

Indonesia 550 trillion rupiah 100 trillion 14 billion 11 48Korea, Rep. of 72 trillion won 56 trillion 47 billion 13 4Malaysiaa 31 billion ringgit 13 billion 3.4 billion 4 6Thailandb 1,121 billion baht 751 billion 11 billion 16 8

a. Estimated costs include those to be incurred by Danaharta for purchasing nonperforming loans (15 billion ringgit) andrecapitalization funds injected by Danamodal (16 billion ringgit).b. Amount disbursed includes significant private sector funding of recapitalization, as discussed in the text.Note: These are illustrative numbers based on varying assumptions of recovery of nonperforming loans, as discussed in the text.Source: Central bank data.

88


selling these loans to private contractors. Forcommercial banks in Thailand the governmentis providing special tax incentives to privatelyrun asset management companies.

Recapitalization of commercial banks. Allthe crisis countries have taken significant ini-tiatives toward injecting new funds into theirbanking systems (table 3.6). Except in Thai-land, governments have been the dominantsource of funds. In all countries, however,much remains to be done.

Reported estimates of recapitalizationcosts vary significantly, in part because thestrategies are complex, but also because im-portant judgments are required to arrive at afinal number. First, a judgment is required onwhether nonperforming loans will continue toincrease. Some estimates suggest that they willin both Korea and Malaysia, but have reachedtheir peak in Indonesia and Thailand (WarburgDillon Read 1999). Second, a judgment is re-quired on what fraction of the nonperformingloans will eventually be recovered. This judg-ment is perhaps more difficult to make thanthe first. The estimates of recapitalization costsin table 3.6, which are based on official coun-try sources, assume that the recovery rate onnonperforming loans will be around 50 per-cent in Korea and Thailand, about 70 percentin Malaysia, and less than 25 percent in Indo-nesia. The range reflects assumptions abouteconomic growth and interest rates and alsoabout the intrinsic worth of those assets. Auc-tions of distressed assets conducted in Koreaand Thailand caution that the realization ratesmay be lower than currently anticipated. Inboth countries real estate loans have sold forabout 50 percent of the original value of theloan. Loans based on automobile hire-pur-chase contracts have been similarly discounted.Commercial or business loans have fared sub-stantially worse, with realization levels typi-cally in the 20 percent range. When all loanssold are added up, the realization rate in Thai-land has been approximately 25 percent.

The pressure on Thai fiscal resources hasbeen mitigated through private efforts to raisecapital, which has amounted to about B250

billion, or about one-third the amount so fardisbursed. The Thai government has set asideB300 billion in a scheme to provide matchingfunds for privately raised capital. However,only B32.5 billion have been used by privatebanks because the government funding wastied to management changes, which has ledto private solutions. Privately funded recapi-talization through equity issues and innova-tive debt instruments has, therefore, been moreextensive in Thailand than elsewhere. Whilesuch private solutions are desirable, thus farthey have resulted in high costs of funding,which is unlikely to heal the cash flows of thealready distressed banks.

Government funds for augmenting thebalance sheets of banks have been the small-est in Malaysia, where about RM13 billion(4 percent of GDP) were disbursed by July1999 through Danaharta’s purchase ofnonperforming loans and through capital in-jections via Danamodal. Malaysia’s smallercosts in relation to GDP reflect both thesmaller shock Malaysia’s financial system facesrelative to other countries and the precrisislevels of capitalization, which were, on aver-age, higher than in the other crisis countries.However, costs will be greater if thenonperforming loans continue to rise, as someobservers expect, and also if the realizationrates on these loans are lower than those cur-rently assumed. For funding the recapitaliza-tion, Malaysia has relied on zero-couponbonds—that is, on bonds that pay no intereston an ongoing basis. While this alleviatesshort-term fiscal costs, the repayments will bebunched and, therefore, presume either sig-nificant economic growth or the continuedability to roll over the debt.

Complex tradeoffs for policymakersWhile the achievements thus far have beensignificant, major challenges remain. Largesegments of the banking systems remain un-dercapitalized and, except in Korea, bankshave been reluctant or unable to increase theirstock of loans. The institutional structure cre-ated to deal with the crisis has few “sticks” to

89

force the pace and has, consequently, not coun-teracted creditors’ and debtors’ natural ten-dency to wait (box 3.1). With fiscal costs risingand some of the more difficult problems stillahead, governments face the complex task ofmanaging their fiscal costs while also ensur-ing the continued integrity of, and in somecases the strengthening of, their countries’ fi-nancial systems. Yet experience has shown thata loss of political momentum in dealing withthe problems is only likely to aggravate them.

Forbearance versus recapitalization. Inthe wake of a crisis, the objective of bank re-structuring should be to maintain the flow ofcredit while ensuring that the new lending isprudent (Stiglitz 1999). Achieving this objec-tive presents different options. Lending couldbe encouraged by exercising regulatory for-bearance. However, the evidence does notsuggest that simply encouraging voluntaryworkouts while engaging in regulatory for-bearance leads to a resumption of lending. Infact, the experience from past systemic criseswarns that forbearance without tight over-sight could make matters worse. At the sametime, the alternative strategy of government-financed recapitalization could stimulate newlending, but may entail large fiscal costs andmay reduce incentives for prudent lending.Both forbearance and government bailoutsmay undermine the regulatory challenge ofbuilding a sound and competitive financialsystem.

While all countries have updated their fi-nancial sector regulatory systems to be morein line with international reporting standardsand prudential norms, varying degrees of regu-latory forbearance are in place to permit agraduated achievement of these norms. Theseapply, for example, to the following:

• Less stringent recognition of nonperform-ing loans (as in Malaysia, where loans areconsidered nonperforming if they have notbeen serviced for six months rather thanthree months elsewhere)7

• Relaxed provisioning against the non-performing loans (as in Korea, where

loans considered restructured have verylow provisioning when, in fact, they re-main extremely risky)

• Breathing room to achieve capital ad-equacy standards (especially in Indonesiaand Thailand).

There are important reasons for exercis-ing forbearance. First, bank restructuring andcorporate restructuring through a workoutprogram are inextricably tied to each otherwith respect to incentives. The banks need tohave incentives to take the debtors throughthe workout process, which is often difficult,protracted, and costly. Also, workouts oftentake place while the bank itself is undergoingrestructuring and is under severe pressure tomeet capital adequacy ratio requirements.Banks need to be encouraged to reach a re-structuring agreement with the debtor that isrealistic and that matches the debtor’s abilityto repay. Reassurances from regulatory au-thorities of capital adequacy forbearance canhelp. Otherwise, banks will be tempted topaper over their agreements with their debt-ors and not to recognize the true extent ofpotential portfolio losses. As such, initial for-bearance may have been the most realistic re-sponse to systemic crisis and simultaneousdistress among hundreds of large corporationsand thousands of smaller ones. Second, in theabsence of forbearance, the alternative maybe to close down an institution, which cre-ates a bankruptcy cost—that is, the constitu-ent elements of a closed institution may sellfor less than the institution’s value as an on-going entity.

However, the evidence suggests that, whileselective forbearance of relatively sound banksand securitized transactions may be appropri-ate, continued generalized forbearance couldultimately prove costly. The timely adoptionof more stringent accounting standards forrestructured debt is needed. Forbearance doesnot create stronger balance sheets, which arerequired to meet the new working capital needsof firms in distress. Rather, it dilutes the banks’incentives to negotiate more forcefully with


90


the controlling shareholders of distressed cor-porations and leads to an unrealistic assess-ment of recapitalization needs. The over-whelming international experience is thatforbearance, on balance, works to delay, butnot to heal (Kane 1989; Brinkmann, Horvitz,and Huang 1996; and Sheng 1996).

Government financed recapitalization alsopresents difficult tradeoffs. Early recapitaliza-tion can release capacity for new lending thatpermits a broad-based recovery. “An essen-tial element of banking reform is recapitaliza-tion of the banks with enough income earningassets to leave a prudential capital base in placeafter provisioning for bad loans” (vanWijnbergen 1998, 11). However, recapitaliza-tion of the financial system generates not onlyimmediate fiscal costs, but also creates a moralhazard for the future. Early resolution of theproblems is favored, because the dilemma cantypically be expected to worsen. Firms withheavy debt burdens, unable to obtain new fi-nancing and hence unable to grow, find theirdebt burden increasing over time, thereby in-creasing the extent and severity of nonper-forming loans at banks (Stiglitz 1999). Delaysmay also contribute to a culture of debt de-fault, further aggravating both the size of theproblem and the uncertainties in the timingof government outlays.

Under the circumstances, an early recog-nition of the governments’ fiscal obligationsis needed not just to honor their commitmentsto depositors, but also to create the basis formarket-led restructuring. To protect depositors,banks can, for example, be “paid” with gov-ernment bonds, the interest on which can becombined with a portion of the net earningsto service the interest claims of depositors.Bonds should also be tradable to permit re-payment where depositors decide to take theirsavings elsewhere. At the same time, as safeassets, the bonds on the balance sheet of thebanks would greatly improve the capital of thebanks. This should improve the incentives ofbanks to recognize losses without fear of de-pleting capital to an unsustainable level. Suchrecognition should, in turn, improve the effi-

ciency of the corporate restructuring process.In addition, if governments can credibly com-mit to refrain from further bailouts, incentivesfor further risky lending would decline.

Recovering recapitalization cost. The costof recapitalization can be lowered throughincentives to encourage increased reflows fromthe nonperforming assets. Government sup-port could be conditional on contractual pro-visions that share in the success if asset valuesrecover. For example, when Chrysler Corpo-ration was bailed out in the United States, thegovernment obtained warrants as a quid proquo that could be exercised at favorable valuein the event of a recovery. Similarly, when abank retains a nonperforming loan to benefitfrom government recapitalization, the pro-ceeds from any recovery could be shared withthe bank managing the loan. Such a provisionwas used in Chile. Finally, the privatizationof government-acquired banking institutionsremains a priority, both to recover fiscal out-lays and to lay the foundation for private riskbearing. Especially in Malaysia, which envis-ages an extensive merger program, but also inother countries the objective of containing fis-cal costs and restructuring the banking sectormay be combined. In place of administrativelymandated mergers, governments could use thesale of their ownership stakes to promotemarket-based mergers.

The role of deposit insurance role in cri-sis situations.8 An explicit system of depositinsurance put in place when the banking sys-tem is in sound health should, in the event ofa crisis, deal with its obligations early to con-tain the crisis. However, extending the systemof deposit insurance after the onset of a crisiscreates bad incentives (Garcia 1999). In Indo-nesia unconditional and comprehensive guar-antees were extended to all parts of the system,which was soon revealed to have deep-seatedproblems. Guarantees were even extended tosome depositors of the 16 banks that had beenclosed down before the guarantee scheme wasannounced. In Thailand a July 1997 cabinetdecision partially guaranteed depositors andcreditors of the 57 finance companies that were

91

subsequently closed down; this was followedby a blanket guarantee covering all deposi-tors and creditors of the remaining financecompanies and commercial banks (IMF 1999).Evidence from other postcrisis situations sug-gests that imposing losses on depositors andcreditors need not lead to panics and bankruns.9

Corporate restructuring: someprogress, but a long way to go

Government funds are not required for cor-porate restructuring, and their supply

may even hinder private resolution as stake-holders are induced to seek these subsidies.The proper role for governments is to facili-tate resolution of financial claims and fosterthe reallocation and mobility of assets. In theabsence of efficiently functioning systems toresolve financial claims, governments in all thecrisis countries have instituted out-of-courtmechanisms to encourage financial settle-ments. Beyond these immediate measures, butalso aiding in the short term, are ongoing ef-forts to achieve effective bankruptcy regimesand improved accounting standards. Once fi-

nancial property rights have been clarified, themarket system and the private sector shouldbe in a position to undertake the required re-allocations of productive assets, but govern-ments can play an important role in permittinggreater asset mobility. The Japanese experi-ence shows that without fundamental reformsto foster asset mobility through bankruptcyprocesses and mergers and acquisitions poli-cies, corporations may be slow to undertakesignificant restructuring (box 3.2). That ex-perience, though, also shows that success re-quires continuing procedural innovation andadaptation to meet the evolving needs of thecorporate sector.

Slow resolution of financial claimsImmediately following the crisis, governmentsin the crisis countries helped establish out-of-court mechanisms (see table 3.4) that couldspeed up the settlement of financial claims inthe absence of bankruptcy regimes able tohandle the large-scale distress. These mecha-nisms have been slow to produce results, inpart because they depend on moral suasion.However, progress has been achieved in Ko-rea and Malaysia. At the same time, account-


Table 3.7 Restructuring: out-of-court and in-court progress, August 1999


Out-of-court proceduresAll or the majority of financial institutions No Yes Yes Yes signed on to accordFormal process of arbitration exists, with deadlines No Yes No YesProvision of penalties for noncompliance No No Yes Yesa

Out-of-court restructuringsNumber of registered cases 234 53 92 825Number of cases started 157 27 83 430Number of restructured cases 22 10 46 167Percentage of restructured debt in total debt 13 32 40 22

In-court restructuringsNumber of registered cases 88 52 48 30Number of cases started 78 34 27 22Number of restructured cases 8 12 19 8Percentage of restructured debt in total debt 4 .. 8 7

.. Not available.a. In Thailand, penalties for noncompliance were introduced in August 1999 for creditors who had signed intercreditoragreements.Source: Claessens, Djankov, and Klingebiel 1999.

92


Bankruptcy. Until the early 1990s, approxi-mately 15 percent of broadly defined business fail-ures involved formal bankruptcy proceedings. Theproportion has risen in recent years to 30 percent, aslarger firms are now failing. These firms are morelikely to prefer the protections and safeguards ofcourts over the cheaper and faster proceedings ofprivate negotiations. For small firms, private proce-dures typically implied a freezing of credit and thecessation of operations, which is the worst possiblesocial outcome.

Gatekeeping procedures in Japanese law andregulations act as a barrier to court actions. Advancedeposits for court costs are especially onerous forsmall firms. Japanese courts can take a few monthsto determine if a bankruptcy petition meets certainconditions, and they do not issue an automatic stay,rendering firms vulnerable to raiding by creditors.

Several changes are being proposed to makebankruptcy procedures more efficient. In April 1998the Justice Ministry proposed revisions to integratethe five laws governing corporate bankruptcy into asingle law. Many of the changes are aimed at makingformal bankruptcy more accessible to small firms.Officials have sought to enhance the prospects ofreorganization rather than liquidation. The currentlaw permits applications for reorganization onlyafter a firm is virtually insolvent, favors change in

Box 3.2 Redeployment of assets:lessons from Japan

Restructuring in Japan is finally beginning underthe immense pressures of a long recession, but

it is being helped by improved regulations govern-ing bankruptcy and mergers and acquisitions. Japa-nese businesses are restructuring at a faster pacethan during past economic downturns. Mergers andacquisitions are occurring in numbers unprec-edented for Japan, while the debt associated withbankruptcies has also hit new highs. The unem-ployment rate is at a postwar peak, with more than1 million jobs lost in the last 18 months alone. Thiscontrasts with the 1990–91 recession, when Japa-nese firms were much more reluctant to reducetheir work force. Companies are also sheddingcross-held shares and major corporations are shut-ting down subsidiaries, closing operations, andselling off unprofitable businesses.

The Japanese experience offers several les-sons. Delays in restructuring weak banks andcorporations can contribute to long periods of lowgrowth. Absent active governmental effort, aninadequate institutional infrastructure for resolv-ing financial claims can persist even in an indus-trial country. Moreover, effective redeployment ofassets through bankruptcy processes and mergersand acquisitions requires an ongoing adaptationof regulations, and changes in business practices,as constraints are revealed.

ing standards and bankruptcy systems, whereneeded, have been reformed, which may alsohelp resolution of financial claims in the shortrun and may provide a sounder basis for im-proved corporate governance in the long run.

Voluntary workout mechanisms. Volun-tary mechanisms rely on the so-called LondonApproach and provide a framework withinwhich claims can be settled.10 While the de-tails vary across countries, their main featuresinclude: binding agreements on the part ofbanks to participate in and honor the agree-ments, with some possibility of penalties ifagreements are not adhered to; timetables to

achieve resolution; and standardized agree-ments between debtors and creditors and,equally important, between creditors them-selves. The main characteristics of such mecha-nisms across countries and their achievementsare described in table 3.7.

Korea and Malaysia appear to have ben-efited the most from these out-of-court mecha-nisms, in part because they have more bindingagreements among banks. In Korea, penaltiesexist for noncompliance, while Malaysia haswell-defined implementation schedules. Thaiprocedures, though similar to those of Ma-laysia, have achieved less, reflecting the deeper

93

problems of Thai restructuring and the rela-tive weakness of the bankruptcy regime in thatcountry. Stronger procedures recently includedin Thailand—penalties for noncompliance bycreditors who have signed intercreditor agree-ments—should help force the pace.

While the progress achieved is significant,it is premature to judge the quality of theserestructurings and the impact that they willhave on the debt resolution problems. A wait-ing strategy is reflected in the use of instru-ments that postpone the repayment of debtand has been especially evident in, though notconfined to, Malaysia. The Malaysian Corpo-

rate Debt Restructuring Committee, whichoversees the voluntary debt workout program,had, by June 1999, helped reschedule approxi-mately RM11 billion of debt (EIU 1999a). Ofthat amount, RM8.5 involved the issuance ofa seven-year zero-coupon bond to purchasethe existing debt of Renong Corporation andits subsidiary, United Engineers Malaysia. TheRatings Agency of Malaysia has assessed thatRenong and United Engineers Malaysia willbe unable to pay the debt and would, conse-quently, need to refinance 60 percent of thedebt outstanding at the time the bond matures.According to Claessens, Djankov, and


management, and requires a reorganization plan atabout the same time as the bankruptcy petition.Under the proposed approach, companies couldapply to the courts for protection with more oftheir assets intact, keep their management in place,and receive more time to draw up a turnaroundplan. Also included in the proposal is the shorteningof the period of asset assessment from the currentthree to seven months to one month. Other changescall for greater information disclosure and removalof barriers to selling parts of a company.

Mergers and acquisitions. Mergers and acquisi-tions are on the rise in Japan, though their impor-tance to the economy is still a small fraction of thatin the United Kingdom or the United States. Thevalue of foreign takeovers in Japan rose from $1.1billion in 1997 to $6.9 billion in 1998, and thenshot up to $7.1 billion in just the first quarter of1999. Similarly, domestic mergers and acquisitionshave also risen briskly.

For years, mergers and acquisitions were amark of failure associated with companies on theverge of bankruptcy, but there were also real eco-nomic barriers. The lack of transparency in thebooks of potential targets was a serious problem.For example, a department store with a strongfranchise and substantial real estate assets foundno buyers because the scale of off-balance sheetguarantees provided by the store was large anduncertain. Another problem was the widespread

system of cross-shareholding. The practice of mu-tual shareholding, initiated in the 1970s, was de-signed explicitly to ward off undesiredacquisitions.

However, economic forces are eroding barriersto mergers and acquisitions, aided by the decline incross-shareholdings and facilitated by regulatorychanges. Many Japanese companies, especially fam-ily-owned businesses established in the early post-war period, are seeking injections of capital, asoperating losses and write-offs of bad assets havebeen a drain. Cross-shareholdings are being elimi-nated, as the returns on these equity holdings havebeen persistently low or negative. At the same time,many regulatory constraints on business activitiesare being removed, and specific measures to facili-tate mergers and acquisitions are being instituted.For instance, a 1997 amendment of the CommercialCode by the Japanese Diet reduces the number ofshareholder meetings to approve mergers. TheHolding Company Law of 1997 removes constraintson carving out subsidiaries for sale and allows buy-ers more freedom in structuring their acquisitions.The securities transaction tax formerly requiredwhen an acquisition involved share purchases wasdiscarded in April 1999. In addition, the moves tointernational accounting principles and, in particu-lar, consolidated reporting, are bringing more trans-parency to the operation of subsidiaries.

Source: Alexander 1999.

Box 3.2 (continued)

94


Klingebiel (1999), two-thirds of the restruc-turing agreements in Korea involve a combi-nation of interest rate reduction, capitalizationof the interest rate, and longer grace periods.11

In Thailand the quality of the restructuringagreements has been such that 13 percent ofthe restructured debts have already revertedto nonperforming status. Note also that re-ported nonperforming loans do not yet regis-ter the impact of agreements. This may partlyreflect a statistical reporting lag, becausenonperforming loan data may not yet havebeen updated to reflect the agreements, orbecause weaknesses exist in the agreementsthemselves.

Bankruptcy reform. By creating a legalbasis and orderly mechanisms for the resolu-tion of debt defaults, bankruptcy procedurescan provide the stick that complements vol-untary restructuring initiatives in the crisiscountries. Effective bankruptcy systems shouldresolve the conflicting claims of stakeholderson the assets of insolvent corporations. Theyshould help preserve and quickly restructureviable firms as ongoing entities and shouldresult in the expeditious liquidation of nonvi-able firms.

Indonesia and Thailand have implementedsignificant legislative changes since the onsetof the crisis. Korea and Malaysia, in contrast,have relatively sophisticated bankruptcy codes.In Korea, the law, though well established, isalso thought to be complex in its implemen-tation and to favor debtors excessively (seeThe Economist 1999). Thus, except in Ma-laysia, the bankruptcy process is unlikely toplay a significant role in resolving the presentdebt overhang.

To a greater extent than in other coun-tries, procedural capacity in Indonesia remainsa bottleneck to the effective enforcement ofinsolvency laws. In August 1998 Indonesiaamended its bankruptcy legislation, creating,in particular, a specialized commercial courtwith jurisdiction over all bankruptcy-relatedmatters and subject to review only by the Su-preme Court (Mojdehi and Ito 1998). Theamendment also created expedited timetables

and introduced a stay provision similar to thatunder the U.S. Bankruptcy Code. At the sametime, the voluntary workout mechanism un-der the Jakarta Initiative Task Force (see table3.4) is helping to develop precedents for deal-ing with complex debt renegotiations. In prac-tice, however, the amended bankruptcy lawhas not succeeded in alleviating the widespreadcorporate and financial distress. Realistically,the bankruptcy court can help only modestlyin the present crisis by pronouncing in a con-sistent manner on a select number of cases.The vast majority of the 15,000 nonper-forming loans will be settled out of court.

The early workings of a new bankruptcyregime is expected to be frustrating. In adopt-ing a comprehensive bankruptcy law, Hungaryexperienced a crush in the tumult of the earlypostcommunist years. Due to an automatictrigger that required all firms with arrears ofmore than 90 days to file for either reorgani-zation or liquidation, Hungarian courts wereoverwhelmed by some 22,000 bankruptcycases soon after the law’s enactment (Gray,Schlorke, and Szanyi 1996, 425). While theexperience “indisputably spurred institutionbuilding in the courts, the trustee profession,and the banks,” during a crisis, the formaljudicial process will clearly be overwhelmedin most developing and transition countries.

The Thai experience, however, shows that,despite the many constraints, pushing aheadwith improvements in bankruptcy code pro-cedures can bring benefits. In Thailand a con-troversial piece of legislation has sought toenforce the rights of creditors more forcefully,including enforcing rights to personal guar-antees that served as collateral. A study ex-amined the relationship between the progressof Thai bankruptcy legislation and the equityvaluation of financial and nonfinancial com-panies (Foley 1999). Announcements indicat-ing a potential strengthening of bankruptcylaws enhanced the equity values for both debt-ors and creditors. In other words, market par-ticipants do not view a stronger bankruptcylaw as a zero-sum outcome where creditorsgain and equity holders lose. Rather, both can

95


gain, though creditors are likely to benefitmore. The positive gain for all parties impliesthat inefficient and protracted bankruptcy pro-ceedings have a real economic cost.

A short-term agenda for bankruptcy re-form requires further development of infor-mal, extrajudiciary processes to resolveproblems, in parallel with and as a comple-ment to, the formal insolvency process. Inaddition to the government-sponsored work-out mechanisms, prepackaged bankruptcyprocedures can speed things up. Under pre-packaging, the parties involved agree to theterms of the workout, and the court then usesan expedited procedure to bind a dissentingminority and to formalize the agreement. Per-haps more so than in industrial countries, out-of-court settlements may also include methodsfor auctioning the rights to the firm (seeBebchuk 1996; Hauch and Ramachandran1999). Auction procedures can help both tomaintain seniority among creditors and toreveal the value of the firm.

The bankruptcy process is especially proneto fraud, and the absence of adequate account-ing standards makes the difficult problem ofasset valuation even more complex. This un-

derscores the urgent need for accounting andcorporate governance reforms. A more imme-diate task may be to create greater transpar-ency in the appointment of judges and torequire the publication of decisions along withdetailed rationales for those decisions.

Enhanced asset mobilityOnce the financial claims on a company areresolved, market-driven mechanisms will prob-ably reallocate the resources to their best uses.Additional government interventions may,however, be justified by the existence of insti-tutional and market failures, such as monopo-lies and weak competition, cross-holdings andconnected lending, or labor market rigidities.These imperfections, which were often thesource of resource misallocations prior to thecrisis, also now hinder the required process ofreallocation.12

Of special importance are policies facili-tating mergers and acquisitions and encour-aging foreign direct investment. East Asiangovernments have taken several steps to en-courage mergers, both international and do-mestic. Also, foreign investment has beenliberalized in all the countries, though to vary-

Figure 3.7 Cross-border mergers and acquisitions, Q1 1997–Q2 1999

0Q1 1997 Q2 1997 Q3 1997 Q4 1997 Q1 1998 Q2 1998 Q3 1998 Q4 1998 Q1 1999 Q2 1999

2

4

6

8

10

12

14

16Number

Note: Cross-border transactions involve majority foreign ownership.Source: Securities Data Company.

Indonesia

Republic of Korea

Malaysia

Thailand

96


ing degrees. As with bankruptcy, success ofmergers and acquisitions depends heavily onprocedural simplicity and clarity (see box 3.2on the Japanese evolution in this respect).Reforms following the crisis also includedshort-term tax measures to facilitate assettransactions and, more importantly from along-term perspective, better accounting stan-dards, which should contribute to improved

corporate governance through better evalua-tion of financial assets and liabilities (table3.8).

Mergers and acquisitions. Since their cri-ses in 1997, Korea and Thailand have intro-duced various measures to encourage businessconsolidation, leading to a rapid rise in cross-border mergers and acquisitions in these twocountries (figure 3.7). The Korean government

Table 3.8 Illustrative postcrisis policy reforms in crisis countries

Corporate governance Loss allocation and transfer Factor mobility

Indonesia Presence of a corporate Tax exemptions for loan-loss Relaxation of foreign ownershipsecretary to improve disclosure reserves held by banks restrictions (September 1997)Bankruptcy Law updated (March 1998) Tax exemptions of up to(August 1998) 8 years for new investmentsCode of best practice for in 22 industries (January 1999)corporate governance(in progress)

Korea, Rep. of Restrictions on cross-debt Revaluation and adjustment Introduction of Foreignguarantees (April 1998) of capital and foreign Investment Promotion ActEnhancing institutional exchange losses (November 1998)voter rights (June 1998) (August 1999)Introduction of internationalaccounting standards(August 1999)Lowering the minimum equityholding requirement to exerciseshareholder’s rights (1999)

Malaysia Creation of High-Level Reduction of corporate tax Reduction of real property gainsFinance Committee on rate from 30 percent to tax rate from 30 percent toCorporate Governance 28 percent (October 1997) 5 percent for nonresidents onCode on takeovers and mergers Tax exemption on interest the sale of a property held forwith stricter disclosure from NPLs (effective a minimum of five yearsstandards (January 1999) for 1999 and 2000) (October 1997)

Exemption of real property gainstax on mergers of financialinstitutions (October 1998)

Thailand Financial statements of public Elimination/deferral of income Alien Business Law (Augustcompanies and financial tax and taxes on asset transfer 1998, revised in October 1999)institutions to be in accord with and unpaid interest Tax-free mergers and acquisitionsinternational best practices (January 1999) in cases of 100 percent mergers(1999) Introduction of new asset (January 1999)Requirement of board audit depreciation method Introduction of Equity Fund,committees (1999) (March 1999) Thailand Recovery Fund forBankruptcy and foreclosure large- and medium-scalelaws amended (March 1999) companies, and Venture Capital

Fund for small and medium-sizeenterprises (March 1999)Reduction of real estate transferfee from 2 to 0.01 percent of theappraised value (March 1999)

97

released a new legislative framework in July1999 to reduce transaction-related taxes in-curred in corporate mergers, acquisitions,and restructurings. For domestic transactions,the government has provided tax exemptionsand deferrals on capital gains from so-called “big deals”—that is, exchange of busi-nesses through the transfer of shares. Thai-land approved a set of new measures inJanuary 1999, including temporary measures(expiring on December 31, 1999) to lowertaxes on gains to debtors from the write-offof debts and on asset transfers from debtorsto creditors. Permanent measures in Thailandare designed to facilitate mergers and businessreorganization.

Unlike in Korea and Thailand where cross-border mergers have shot up, in Malaysiacross-border mergers and acquisitions havebeen low compared to just prior to the crisis.Malaysia has, however, had high levels ofdomestic mergers and acquisitions.13

Malaysia’s Promotion of Investment Act of1986 and other measures provide various taxincentives, including investment tax allow-ances in the services sector. The high level ofdomestic merger and acquisitions activity inMalaysia suggests that the policy regime isbasically a friendly one. Cross-border activityhas been relatively low, possibly on accountof restrictions on the repatriation of earnings(though the long-term effects of these restric-tions await further empirical examination).14

In contrast, the Indonesian system appearsnot to favor mergers and acquisitions. Gainsfrom transfers of assets in corporate reorga-nizations are taxable, and companies cannot

transfer tax losses in a liquidation process,merger, or acquisition (Asia Law 1998). Cer-tain exceptions apply only to banks, financialinstitutions, and companies going public.Merger and acquisitions activity has remainedat extremely low levels.

Foreign direct investment. Foreign directinvestment (FDI) inflows to Korea and Thai-land increased in 1998 by 82 percent and 26percent, respectively, and flows to these coun-tries continued at high levels in 1999 (table3.9). Both countries have taken effective mea-sures to deregulate and liberalize their foreigninvestment policies since late 1997. Note,however, FDI includes mergers and acquisi-tions involving existing enterprises as well asnew, or greenfield, investments. The FDItrends, therefore, reflect in part the trends incross-border mergers and acquisitions de-scribed in the previous section (figure 3.6).

Korea has opened several sectors to for-eign investors since April 1998, including vari-ous property businesses, securities dealings,and other financing businesses. The ceiling onforeign stock investment was abolished as ofMay 1998, granting foreign investors the rightto purchase all the shares of a domestic firm.Meanwhile, the Foreign Investment PromotionAct of November 1998 affords protection forforeign direct investment through nationaltreatment, the reduction and exemption ofcertain corporate taxes, the provision of fi-nancial support for local governments to at-tract foreign direct investment, and theestablishment of foreign investment zones.

In Thailand the Board of Investment haseased its regulations to promote foreign par-

Table 3.9 FDI flows in East Asia, 1992–99(billions of U.S. dollars)

1992 1993 1994 1995 1996 1997 1998 Q1 1999 Q2 1999

Indonesia 1.8 2.0 2.1 4.4 6.2 4.7 –0.4 –0.03 —Korea, Rep. of 0.7 0.6 0.8 1.8 2.3 2.8 5.1 1.0 1.8Malaysia 5.2 5.0 4.3 4.1 5.1 5.1 5.0 — —Thailand 2.1 1.8 1.4 2.1 2.3 3.8 6.8 1.0 2.2

— Not available.Source: World Bank Debtor Reporting System; IMF, International Financial Statistics.


98


ticipation in the economy. The 20-year-oldAlien Business Law was replaced in August1998 (and has since been revised again inOctober 1999) to incorporate sectoral liber-alization measures. Under the August 1998provisions, foreign firms are allowed to holdup to 100 percent equity in banks and in fi-nance companies for up to 10 years, and 39sectors have been opened for increased for-eign participation, including transportationand pharmaceuticals production. Policy lib-eralization includes a temporary measure in-troduced in November 1998 (expiring inDecember 1999) allowing foreign firms to owna majority stake in joint ventures that receivedfavorable policy treatment, and authorizingthem to distribute their products domestically.In the meantime, the proposed cutback ofimport tariffs is expected to help reduce pro-duction costs for both domestic and foreignfirms dependent on imported raw materialsand intermediate products.

Mirroring the trends on cross-bordermergers and acquisitions, foreign direct invest-ment inflows into Malaysia, though tradition-ally high, have not responded as they have inKorea or Thailand, and flows have fallensharply in Indonesia. In Malaysia new effortsto attract foreign investments have been un-dertaken; for example, restrictions on foreignholdings in new export-oriented manufactur-ing projects have been suspended until 2000and foreign ownership limits have been re-laxed. However, the value of approved projectsduring January–May 1999 at RM6.4 billionremained at the same annualized rate as in1998; the value of new FDI applications fellover the six month period to RM3 billion,compared with RM12.6 billion for calendar1998. In Indonesia the value of approved andrealized foreign direct investment declined by80 percent in the first quarter of 1999. Indo-nesia has recently started to implement newincentives to attract foreign investors. Foreignownership of up to 99 percent of banks hasformally been effective since May 1999. InJune 1999 a new decree was announced toallow shareholdings up to 100 percent in ex-

isting establishments and to provide a clearerlegal framework for the conversion of bondsissued locally into equity.

Notes1. Although East Asian firms, including those in

the crisis countries, have been adept at adopting newmanufacturing techniques, they have faced continu-ing challenges both from low-wage producers and fromJapan (see Mody, Suri, and Sanders 1992).

2. Such a financial accelerator has been docu-mented, for example, by Bernanke, Gertler, andGilchrist (1996, 2) who conclude: “A fall in theborrower’s net worth, by raising the premium on ex-ternal finance and increasing the amount of externalfinance required, reduces the borrower’s spending andproduction. This last result is the heart of the finan-cial accelerator: To the extent that negative shocks tothe economy reduce the net worth of borrowers (orpositive shocks increase the net worth), the spendingand production effects of the initial shock will be am-plified.” Small firms are especially prone to the down-ward spiral of the financial accelerator, but large,credit-constrained firms operating on thin margins maybe equally vulnerable.

3. While Thailand’s relative position on the twoscales is the same, the share of Thai firms unable topay debt, 28.3 percent, is much lower than the 45 per-cent of nonperforming loans in Thailand, unlike coun-tries where the ratios are quite similar. This may reflect,in part, the phenomenon of Thai strategic defaulters—that is, those able to, but not paying, their debts. Inaddition, small, unlisted firms contribute heavily toThai nonperforming loans. However, Mako (1999)reports that even among listed firms, about half wereunable to pay their debt, a ratio more consistent withnonperforming loans.

4. In the Scandinavian banking crises of thelate 1980s and early 1990s, the share of nonper-forming loans ranged from 5 to 7 percent, and theseloans mainly represented failed real estate lending.Even in Chile, at the onset of the early 1980scrisis, nonperforming loans were about 5 percent ofall loans.

5. Official numbers on nonperforming loans areless readily available for Indonesia. However, the per-centage is generally regarded as in the range of 50 per-cent and is expected to rise to more than 60 percentbefore falling.

6. In Malaysia, unlike in the other countries,Danaharta has not only purchased some of thenonperforming assets, but also is a management agentfor the restructuring of nonperforming assets.

99


7. The more stringent definition of nonper-forming loans is, however, still used for supervision.

8. The principles for creating a competitive butsound financial system center around the appropriaterole of deposit insurance, bank capital, and diversityand competition in the financial sector (see, for ex-ample, Greenspan 1998; Hellmann, Murdock, andStiglitz 1998).

9. See Dziobek and Pazarbasioglu (1998) for ex-perience in Côte d’Ivoire, Latvia, and Spain; Drees andPazarbasioglu (1998) for the Norwegian experience;Baer and Klingebiel (1995) for a variety of historicalepisodes.

10. The London Approach operates under theauspices of the Bank of England and has been used forcorporate workouts in recessionary periods when nor-mal bankruptcy procedures have proved insufficient(see Kent 1997).

11. As of early November 1999 discussions ofDaewoo’s debt restructuring also included a signifi-cant component of deferred debt payments (the WallStreet Journal, Novemer 2, 1999).

12. For a review of the postcrisis industrial policyagenda, see Mody (1999).

13. The total number of domestic mergers andacquisitions has been about 50 to 70 per quarter inMalaysia in 1997–99, while it remained low (in therange of 4 to 10) in the other countries (see SecuritiesData Company 1999).

14. Survey results of Japanese investors in theearly 1990s show them to be sensitive in their invest-ment decisions to restrictions on profit repatriation(see Mody, Dasgupta, and Sinha 1999).

ReferencesAlexander, Arthur. 1999. “Japan Confronts Corporate

Restructuring.” Background paper on Japan.Asia Law. 1998. Cross-border M&A: A Guide to Glo-

bal Strategic Direct Investment for Asian Com-panies. Asia Law and Practice, EuromoneyPublications.

Baer, Herbert, and Daniela Klingebiel. 1995. “SystemicRisk When Depositors Bear Losses: Five CaseStudies.” In Banking, Financial Markets, andSystemic Risk, edited by George Kaufman. Green-wich, Conn.: JAI Press.

Bebchuk, L. 1996. “Chapter 11.” Discussion Paper227. Harvard Law School, Center for Law, Eco-nomics, and Business, Cambridge, Mass.

Bernanke, Ben, Mark Gertler, and Simon Gilchrist.1996. “The Financial Accelerator and the Flightto Quality.” The Review of Economics and Sta-tistics 78(1): 1–15.

Brinkmann, Emile J., Paul M. Horvitz, and Ying-LinHuang. 1996. “Forbearance: An Empirical Analy-sis.” Journal of Financial Services Research 10:27–41.

Claessens, Stijn, Simeon Djankov, and DanielaKlingebiel. 1999. “Bank and Corporate Restruc-turing in East Asia: Opportunities for FurtherReform.” Financial Sector Discussion Paper 3.World Bank, Washington D.C.

Crafts, Nicholas. 1999. “East Asian Growth Beforeand After the Crisis.” IMF Staff Papers 46(2):139–66.

Domac, Ilker. 1999. “The Distributional Consequencesof Monetary Policy: Evidence from Malaysia.”World Bank, Washington D.C. Draft.

Domac, Ilker, and Giovanni Ferri. 1998. “The RealImpact of Financial Shocks: Evidence from theRepublic of Korea.” Policy Research Paper 2010.World Bank, Washington, D.C.

Drees, Burkhard, and Ceyla Pazarbasioglu. 1998. “TheNordic Banking Crises: Pitfalls in Financial Lib-eralization.” Occasional Paper 161. InternationalMonetary Fund, Washington, D.C.

Dziobek, Claudia, and Ceyla Pazarbasioglu. 1998.“Lessons from Systemic Bank Restructuring.”Economic Issues 14.

The Economist. 1999. “Death, Where is Thy Sting?”July 17.

EIU (Economist Intelligence Unit). 1999a. CountryReport: Malaysia Third Quarter. London.

————. 1999b. Country Report: Thailand ThirdQuarter. London.

Foley, C. Fritz. 1999. “Going Bust in Bangkok: Les-sons from Bankruptcy Law Reform in Thailand.”Draft. Harvard Business School, Cambridge,Mass.

Furman, Jason, and Joseph E. Stiglitz. 1998. “Eco-nomic Crises: Evidence and Insights from EastAsia.” Brookings Papers on Economic Activity2(1998): 1–135.

Garcia, Gillian G. H. 1999. “Deposit Insurance: A Sur-vey of Actual and Best Practices.” Staff WorkingPaper 99/54. International Monetary Fund, Wash-ington, D.C.

Gray, Cheryl W., Sabine Schlorke, and Miklos Szanyi.1996. “Hungary’s Bankruptcy Experience, 1992–1993.” The World Bank Economic Review 10(3):425–50.

Greenspan, Alan. 1998. “The Role of Capital in Opti-mal Banking Supervision and Regulation.” Eco-nomic Policy Review 4(October): 163–68.

Harberger, Arnold. 1998. “A Vision of the GrowthProcess.” American Economic Review 88(1):1–32.

100


Hauch, D., and S. Ramachandran. 1999. “CreditorAuctions to Reorganize Debts.” World Bank,Washington D.C. Draft.

Hellmann, Thomas, Kevin Murdock, and JosephStiglitz. 1998. “Liberalization, Moral Hazard inBanking, and Prudential Regulation: Are CapitalRequirements Enough?” Stanford GraduateSchool of Business, Stanford, Calif. Draft.

IMF (International Monetary Fund). 1998. WorldEconomic Outlook 1998. Washington, D.C.

————. 1999. “Financial Sector Crisis and Restruc-turing—Lessons From Asia.” August 12.

Kane, Edward. 1989. “The High Cost of IncompletelyFunding the FSLIC Shortage of ExplicitCapital.” Journal of Economic Perspectives 3(4):31–47.

Kent, Pen. 1997. “Corporate Workouts: A U.K. Per-spective.” In Crisi D’Impresa E Risanamento,edited by D. Masciandaro and F. Riolo. Milan:Edibank.

Kharas, Homi, and Deepak Mishra. 1999. “HiddenDeficits and Currency Crises.” Draft. WorldBank, Washington, D.C.

Klingebiel, Daniela. 1999. “The Use of Asset Man-agement Companies in the Resolution of Bank-ing Crises: Cross-Country Experiences.” WorldBank, Washington D.C.

Krueger, Anne O., and Aaron Tornell. 1999. “The Roleof Bank Restructuring in Recovering from Cri-ses: Mexico 1995–1998.” Working Paper 7042.National Bureau of Economic Research, Cam-bridge, Mass.

Mako, William P. 1999. “Thailand Corporate Restruc-turing.” World Bank, Washington, D.C. Pro-cessed. Draft.

Mody, Ashoka. 1999. “Industrial Policy after the EastAsian Crisis: From Outward Orientation to NewInternal Capabilities.” Policy Research WorkingPaper 2112. World Bank, Washington, D.C.

Mody, Ashoka, Rajan Suri, and Jerry Sanders. 1992.“Keeping Pace with Change: Organization andTechnological Imperatives.” World Development20(12): 1797–1816.

Mody, Ashoka, Susmita Dasgupta, and Sarbajit Sinha.1999. “Japanese Multinationals in Asia: Driversand Attractors.” Oxford Development Studies27(2): 149–64.

Mojdehi, Ali M. M., and Peter W. Ito. 1998. “Indone-sian Insolvency: New Reform Law Designed toRestore Confidence.” Los Angeles Daily Journal,December 22, p. 7.

Rojas-Suarez, Liliana, and Steven R. Weisrod. 1996.“Banking Crises in Latin America: Experiencesand Issues.” In Banking Crises in Latin America,edited by Ricardo Haussmann and Liliana Rojas-Suarez. Washington, D.C.: Inter-American Devel-opment Bank.

Securities Data Company. 1999. SDC Platinum 2.1—Mergers and Acquisitions Database.

Sheng, Andrew, ed. 1996. Bank Restructuring: Lessonsfrom the 1980s. Washington D.C.: World Bank.

Stiglitz, Joseph E. 1999. “Principles of Financial Regu-lation: A Dynamic, Portfolio Approach.” WorldBank, Washington, D.C. Draft.

van Wijnbergen, Sweder. 1998. “Bank Restructuringand Enterprise Reform.” Working Paper 29. Eu-ropean Bank for Reconstruction and Develop-ment, London.

Xie, Andy. 1999. “Korea: Another Financial Crisis?”(http://www.msdw.com)

Warburg Dillon Read. 1999. “Asian Economic andStrategy Perspectives.” September.

World Bank. 1999. Global Economic Prospects andthe Developing Countries 1998/99: Beyond Fi-nancial Crisis. Washington, D.C.

103

4

THE PAST FEW YEARS HAVE SEEN SHARP VARIA-tions in primary commodity prices.Energy prices were especially volatile.

Crude oil prices rose 74 percent from early1994 through the end of 1996, then fell 56percent by the end of 1998, and in 1999 re-covered nearly the entire decline of the previ-ous two years. Average non-oil commodityprices rose by 46 percent from the monthlylow in mid-1993 to mid-1997, and thendropped 30 percent by late 1999. The varia-tion of individual commodity prices wassharper still. Such volatility poses real chal-lenges for developing countries that dependon primary commodities for a substantial shareof their export revenues. Countries where con-sumption rises with real incomes during com-modity price booms will face a difficultadjustment when prices fall. The ability tosustain consumption by borrowing or runningdown assets may be limited. Cutting back oninvestment will depress long-term growth, andsharp reductions in consumption can be ex-tremely painful for the individuals most af-fected, often the poorest members of society,and may require costly reallocations of laborand capital.1 Thus, efforts to smooth consump-tion over the commodity price cycle can becritical to welfare, efficiency, and growth overthe long term.

Two groups of developing countries weremost affected by the commodity price cycle:the major oil exporting countries (countrieswhere oil accounts for more than 50 percent

of merchandise exports) and the non-oil ex-porting countries of Sub-Saharan Africa, wherenon-oil primary commodities, on average,make up 80 percent of exports. These coun-tries are among the most dependent on com-modity prices. This chapter will discuss howthese commodity-dependent economies haveadjusted to the swings in real incomes gener-ated by recent commodity price volatility, fo-cusing on their success in smoothingconsumption over the price cycle and the im-plications for growth prospects.

The main message is that countries’ sav-ings and investment behavior differed mark-edly over the commodity price cycle and thatthese differences primarily reflected the qual-ity of policies rather than shifts in the termsof trade. The policy environment was improv-ing in the non-oil exporting countries of Sub-Saharan Africa, and they achieved increasesin savings and investment over the commod-ity price cycle. In the oil exporting countries,weak policy environments led to mixed sav-ings performance and to lower investment overthe oil price cycle.

This chapter reaches the following con-clusions:

• The pronounced cycle in primary com-modity prices since the mid-1990s wasdriven by changes in global demand,weather-related supply shocks, supplyresponses to the high prices of the early1990s, technological innovations that

Managing the RecentCommodity Price Cycle

104


have reduced production costs, and ex-change rate depreciations among largecommodity exporters linked to the Asiancrisis. Primary commodity prices havebeen more volatile than the prices ofmanufactures in the last two decades, andboth oil prices and non-oil commodityprices have fallen relative to the prices ofmanufactures.

• The major oil exporting countries hadmixed success in smoothing consumptionover the recent oil price cycle. On aver-age, countries allocated to increased con-sumption about half of the average 5percent of gross domestic product (GDP)improvement in real incomes during theupswing in oil prices (1996–97). Duringthe 1998 drop in oil prices, however, con-sumption did not decline, implying thatsavings fell by the full amount of the de-cline in real incomes. Countries’ perfor-mances varied greatly, depending on theirspecific political and economic circum-stances.

• Oil exporting countries’ investment fellrelative to output over the commodityprice cycle. The decline in investment wasactually greater than the decline in do-mestic savings, so the current accountdeficit fell. The major oil exporting coun-tries have generally failed to reduce theirdependence on oil revenues, and the fallin investment will further impedeprogress. At the same time, several ofthese countries face high levels of unem-ployment, continued slow growth, andrapidly expanding populations. Theyneed to strengthen their policies to en-courage greater private sector (and non-oil) activities and to improve theinstitutional environment.

• The commodity price cycle of the 1990sdoes not appear to have adversely affectedthe prospects for growth in the non-oilexporting countries of Sub-Saharan Af-rica. Changes in real incomes were gen-erally smaller than in the oil exportingcountries because the price of their com-

modity exports changed by less than theprice of oil, and the losses from declin-ing export prices were partially offset bygains from lower import prices, particu-larly energy prices. More important, how-ever, improved policies in severalcountries enabled them to increase theirsavings and investment rates both dur-ing commodity price booms and busts.Many countries cut their fiscal deficits inan effort to rein in the growth of debtand to reduce inflation, while privatesavings rose in response to improvedpolicies that increased the return to in-vestment, particularly in export sectors.Countries with better policies, as mea-sured by the World Bank, achieved largerincreases in savings and higher growthof GDP than countries with worse poli-cies, despite smaller increases in real in-comes in the former group.

Key issues confronting primarycommodity exporters

The heavy dependence of many develop-ing countries on primary commodities for

the bulk of their export receipts can createsubstantial problems for economic manage-ment. Primary commodity prices tend to bevolatile, and are subject to long-term cyclesas well as to short-term booms and busts.Commodity prices have also declined relativeto manufactures’ prices during the past twodecades.

Dependence on primary commoditiesPrimary commodities accounted for 42 per-cent of developing countries’ total merchan-dise exports in 1997, compared with 19 per-cent for high-income countries. In Sub-SaharanAfrica, primary commodities accounted forabout 75 percent of total exports, and the shareof commodities in the exports of individualcountries often exceeded 90 percent. The highvolatility of commodity prices thus often leadsto high volatility of export earnings. However,many of these countries also import signifi-

105

cant amounts of primary commodities, andcommodity prices tend to be highly correlated.Deflated by the manufactures unit value in-dex, the correlation coefficients of the priceindexes of agriculture, energy, and metals from1960 to 1998 are all 0.79 or higher. Thus pri-mary commodity exporters’ terms of tradetend to be less volatile than the price of theircommodity exports.

The volatility of primarycommodity pricesThe volatility of primary commodity pricesincreased sharply following the collapse of theBretton Woods system in the early 1970s (fig-ure 4.1) and has remained high (Maizels 1994;Reinhart and Wickham 1994). The standarddeviation of the absolute value of the year-on-year changes in the nonenergy commodityprice index during 1970–98 was 11.5, com-pared with 5.7 for the manufactures unit valueindex during the same period.2 Energy priceshave been particularly volatile. Measured inU.S. dollars, the coefficient of variation ofenergy prices was at least twice that of manu-factures in each decade.

These measures of volatility may exagger-ate the extent of uncertainty caused by com-modity price changes because many seriesexhibit trends and some of the variation ispredictable. Results by Dehn and Gilbert(1999) show that oil exporting countries facethe greatest uncertainty.3 Exporters of theother three commodity groups (agriculturalfoodstuffs, agricultural nonfoods, and metalsand minerals) face only slightly smaller levelsof uncertainty.

Primary commodity prices are volatile fora number of reasons. Supplies of agriculturalcommodities are dependent on the weather,as is the demand for oil. The demand for mostcommodities is less price-elastic than for manu-factures, and this causes supply variability tolead to greater price variability. Supply is alsooften very inelastic in the short term (espe-cially when the commodity is perishable, thecosts of inventory are high, or the productiontakes considerable time, for example, treecrops) so that shifts in demand have a largeimpact on prices. In addition, producers’ or-ganizations, such as the Organization of Pe-troleum Exporting Countries (OPEC), have at

M A N A G I N G T H E R E C E N T C O M M O D I T Y P R I C E C Y C L E

Figure 4.1 Primary commodity prices versus manufactures unit value index (MUV),1960–98

180

Nominal dollar indexes (1990=100)

160

140

120

100

80

60

40

20

0

1960 1965 1970 1975 1980 1985 1990 1995

Energy

Nonenergy

MUV


106


times contributed to volatility through largeshifts in supply.

The secular decline incommodity pricesCommodity prices have declined relative tomanufactures’ prices during the past two de-cades. The real price indexes (nominal priceindexes divided by the manufactures unit valueindex) of both agriculture and metals andminerals fell by about 45 percent during 1980-98. Energy prices have fallen by 76 percent inreal terms since 1980 and in 1998 were lessthan half the level reached after the first oilprice rise in 1973–74. Even with the recoveryin 1999, the real energy price index likely willremain at only 55 percent of the 1974 level.Researchers have concluded that the declinein the ratio of the price of primary productsto manufactures is statistically significant(Bleaney and Greenaway 1993; Sapsford andBalasubramanyam 1994).

The view that the terms of trade of pri-mary commodity exporters show a seculardeterioration dates at least from the 1950s(Prebisch 1950; Singer 1950).4 The initialanalysis of this issue raised a host of method-ological and data questions (Spraos 1980), andlater researchers focused on compiling betterand longer time series (Grilli and Yang 1988)and on bringing more sophisticated statisticaltechniques to bear.5 Most of these writersfound that commodity prices followed closeto a random walk (where the best predictorof tomorrow’s value is today’s value), but withlarge, predominantly negative shocks thattended to persist. Thus, the decline in com-modity prices has been due to a series of ran-dom shocks that were more negative thanpositive, rather than a consistent, and there-fore predictable, trend.

The decline in commodity prices relativeto manufactures depends on the index usedfor manufactures as well as for commodities.Lipsey (1994) argues that the increase in manu-factures prices has been overestimated byroughly one percentage point per year becauseof the failure to adjust for changes in the qual-

ity and mix of manufactures over time. Since1980 real commodity prices have declined byabout 45 percent. Adjusting for qualitychanges in the index of manufactures, the de-cline in real commodity prices would fall to27 percent.

The recent boom and bust incommodity pricesCommodity prices appear to exhibit longcycles driven by technological improvements,as well as short-run booms or busts, whichhave occurred every ten years or so during thelast five decades (Varangis, Akiyama, andMitchell 1995). The most recent boom wasfrom 1994 to 1997, followed by a bust from1997 through early 1999. Energy and non-energy prices followed similar paths, withminor differences. Nonenergy commodityprices rose by 46 percent from the monthlylow in mid-1993 to mid-1997, before falling30 percent by late 1999. Energy prices rose74 percent from early 1994 to the end of 1996and then fell 56 percent by early 1999. Indi-vidual commodities were even more volatile.For example, the price of robusta coffee (ex-ported primarily from Africa and Asia) roseby 390 percent from 1992 to 1994, fell 66percent from 1995 to 1996, and then roseagain by 48 percent in 1997.

The decline in primary commodity pricessince 1997 was in part a response to an un-usually large increase in supply. The rate ofgrowth of world production of agriculturalcommodities rose from 1 percent per year in1990–94 to 2.6 percent in 1994–98, whereasworld production of metals and minerals wasflat in 1990–94, but increased by 3.5 percentper year in 1994–98.6 The acceleration of con-sumption of primary commodities between thetwo periods was much less. Increased produc-tion was in part a response to the high pricesthat prevailed during 1993–95. In the case ofagriculture, favorable weather conditions alsoled to higher yields, and global agriculturalproduction surged in 1996 and 1997. For ex-ample, world grain production increased 5percent per year and world soybean produc-

107

tion by 6.4 percent per year from 1995 to1997, compared with average annual produc-tion growth rates of 1.4 percent and 3.3 per-cent, respectively, for 1980 through 1994.

The supply increases were widespread andwere not confined to a few commodities or afew countries.7 The increases were most rapidin developing countries. The Food and Agri-culture Organization of the United Nation’s(FAO’s) index of the volume of agriculturalproduction rose by 3.8 percent per year forall developing countries from 1990 to 1997.The global supplies of metals and minerals alsoincreased rapidly following the sharp price in-creases of the mid-1990s and the investmentsmade in the late 1980s and early 1990s. Alu-minum production increased by 5 percent peryear during 1995–97, while consumption grewby only 3 percent per year. Copper produc-tion grew by 6 percent from 1995 to 1997,while consumption grew by 4 percent. Nickelproduction increased by more than 5 percentfrom 1995 to 1997, while consumption fellslightly.

The surge in commodity supplies was duein part to improved technology. Technologi-cal advances have taken many forms, includ-ing increased computerization in many areas,expanded use of and refinements to leachingand solvent extraction techniques in mining,horizontal drilling, three-dimensional com-puter seismology, progress in deepwater tech-nology in oil, and improved seeds inagriculture.

Policy reforms and increased privatizationhave also boosted production in some devel-oping countries. For example, in Argentinaderegulation of the ports and the maritimesector in 1992 led to heavy investment andreduced the cost of loading grain from $10/ton to $2/ton, according to the InternationalGrains Council. This made Argentina a morecompetitive exporter, and maize exports tripledin volume terms from 1993–94 to 1997–98.Sri Lanka privatized the management of teaestates in 1992, and the private sector wasallowed to take long-term leases in 1995,thereby contributing to higher investment and

better management. Tea yields increased 46percent from 1990–92 to 1996–98, and pro-duction increased 25 percent, despite a 15percent decline in growing area as unprofit-able fields were taken out of tea. As a result,exports increased by 25 percent. Many Afri-can countries have also undertaken importantliberalization efforts in the last decade, espe-cially in commodities that earn foreign ex-change, such as coffee, cocoa, and cotton. Forexample, Uganda reformed its coffee sectorfollowing the political turmoil of the 1970sand 1980s, and average annual export volumesin 1994–97 were 50 percent higher than in1980–93.

The Asian crisis contributed to the fall incommodity prices. Declines in incomes and thesteep currency devaluations significantly re-duced demand for commodities, with a sig-nificant impact on the prices of commoditieswhere East Asia had a large share of worldconsumption. For example, East Asia, includ-ing Japan, accounted for 21 percent of worldcrude oil consumption in 1997, and East Asianimports of oil dropped by 4 percent in 1998.Production continued to rise, stocks soared,and prices fell by 56 percent from November1997 to a low point in the first quarter of 1999.Furthermore, the East Asian economies hadaccounted for a significant share of the increasein world consumption of some commoditiesin recent years. Investments in production,which generally take a few years to comeonstream, are typically based on demand fore-casts, which prior to the crisis would haveincluded healthy increases in consumption inEast Asia. Thus, the crisis substantially reduceddemand below the levels that could be pro-duced given recent investments, further exac-erbating price declines. For example, the fivecrisis countries and Japan accounted for aboutone-fifth of the increase in world copper con-sumption from 1994 to 1996. When thesecountries’ copper consumption fell in 1998,copper consumption was some 500,000 tonsbelow expected levels, or about 4 percent ofworld demand. Because supply is highly in-elastic in the short term, production levels did


108


not fall as prices declined. The fall in demandfrom Asia thus contributed to the 44 percentdrop in copper prices from June 1997 to De-cember 1998. Other commodities hit by thedecline in East Asian demand included alumi-num, maize, sugar, and cotton.

The crisis also had important effects onthe supply side. Currency devaluations in-creased the competitiveness of the crisis coun-tries, thereby contributing to increases insupply in several commodities. For example,Indonesia, Malaysia, and Thailand account for70 percent of the global exports of naturalrubber, and prices for this commodity fell bynearly one-third in the two years followingthe start of the Asian crisis in July 1997 (al-though natural rubber prices had alreadynearly halved in the two years prior to July1997). The prices of logs from Malaysia (40percent of world exports) and rice from Thai-land (23 percent of world exports) alsodropped sharply after July 1997.

The results of simulations using a com-putable general equilibrium model indicatethat the crisis-induced fall in demand from EastAsia and the increased supply from the EastAsian exchange rate devaluations had a sub-stantial role in reducing commodity prices (seechapter 1 for an explanation of the method-ology used). The commodities exported by

East Asian countries were the hardest hit. Se-lected agricultural prices fell by 10 percent,mineral prices fell by 6 percent, and oil pricesfell by 8 percent (figure 4.2). The decline inaverage non-oil commodity export prices forother developing countries was just under 2percent.

Note, however, that the cyclical declinein commodity prices had already begun whenthe crisis hit. The World Bank’s food priceindex peaked in April 1996 and had declined12.7 percent by June 1997. The index of met-als and mineral prices peaked in August 1995and had declined 11 percent by June 1997.Petroleum prices had declined 24.1 percent byJune 1997 from their peak in December 1996.Beverage prices (coffee, cocoa, and tea) fellmore than 40 percent since their peak in May1997, due to large supply increases from SouthAmerica.

The decline in nonenergy commodityprices since May 1996 has now exceededprevious declines and has lasted slightlylonger. Following the 1980 and 1988 pricepeaks, commodity prices declined for anaverage of 35 months and by an average of25 percent before prices either stabilized orincreased (figure 4.3). Currently, nonenergycommodity prices have declined by 30 percentover 37 months and have since reboundedslightly.

The savings response tocommodity price cycles

Heavy dependence on highly volatile com-modity prices can impose significant

costs on an economy (World Bank 1994). Thepotential for massive changes in relativeprices, in real incomes, and in the level ofeconomic activity can increase uncertainty andcan have a negative impact on performanceand poverty. (The evidence is discussed fur-ther in chapter 2.)

Economists have long known that house-holds tend to smooth the impact of volatileprices on consumption (Harberger 1950). Therelationship between terms of trade and sav-

Figure 4.2 The decline in commodityprices due to the East Asian crisis

0

Percentage change due to crisis

–2

–4

–6

–8

–10

–12

Crude oilMiningAgriculture

Note: East Asia-5 includes Indonesia, Malaysia, the Philippines, the Republic of Korea, and Thailand.Source: Datastream.

East Asia-5

Other developing countries

109

ings depends on the expected duration of theterms of trade shock. If the rise in price isviewed as permanent, consumption increasesto the higher level of income, but if prices riseonly temporarily, savings rise to cushion thefall in income when the price declines (Sachs1981; Svensson and Razin 1983; Ostry andReinhart 1992). Both case study and econo-metric evidence indicate that private agents willtend to save substantial portions of tempo-rary commodity price windfalls.8

The rise in consumption expendituresfrom a temporary commodity price rise shouldbe extremely small if consumption is based onpermanent income. With a real rate of inter-est of 5 percent, the warranted increase inconsumption of a temporary boom is only 5percent of the present value of the windfallgain, assuming an infinite planning horizon(Cuddington 1988). For example, if oil priceswere expected to double in real terms (the larg-est percentage increase that occurred in a singleyear from 1960 to 1998) and then to returnto their former level after one year, a countrywhose oil exports equal about 40 percent ofGDP—for example, Saudi Arabia—should in-

crease consumption by only 2 percent of GDP,or one-twentieth of the initial increase in ex-port earnings.

Savings behavior should take into accountthe possibility that what may look like a tem-porary decline in prices actually represents amedium-term trend. If prices are on a declin-ing trend, then the rise in savings during boomsshould be higher (and the decline in savingsduring busts lower) than if prices are expectedto return to a long-term average. This asym-metric approach to managing commodity pricevolatility would minimize the cost over timeof adjusting to a secular decline in primarycommodity prices.

Savings decisions in countries that exportnonrenewable resources, such as oil or miner-als, should also take into account the poten-tial decline in these resources over time. Theexport of a nonrenewable resource is countedas an addition to GDP (and thus to savings)in national income accounts. However, thisexport actually represents the liquidation ofan asset rather than an increase in savings.Calculations of “genuine savings” reduce re-corded savings by the extent of natural re-


Figure 4.3 Current and previous price declines of nonenergy commodities

105

Nominal dollar indexes (peak =100)

100

95

90

85

80

75

70

65

60

0 12 24 36 48

May 1996

Months

June 1988

February 1980


110


source depletion, among other factors. Thereduction in savings in 1997 in the major oilexporting countries required to reflect thedepletion of energy resources ranged from 2percent of GDP to 44 percent (World Bank1999). Calculations of genuine savings in theMiddle East and North Africa are stronglynegative from 1980 to 1993 (Hamilton andClemens 1999), while savings recorded in thenational accounts averaged 23 percent duringthis period.9 Countries where proven oil re-serves are low or production is declining mayneed to achieve relatively high savings to sus-tain permanent income in the medium term.

Economies may not generate adequatelevels of savings to smooth consumption inthe face of volatile real incomes for severalreasons. Most important, distinguishing be-tween temporary and permanent shocks tocommodity prices can be extraordinarily dif-ficult. The swings in commodity prices can betoo large and uncertain to ascertain theircauses and nature (Deaton and Miller 1995).The degree of uncertainty about the durationof a price shock varies. For example, marketparticipants could see that the sharp jump incoffee prices caused by the Brazilian frost of1994 was likely to be reversed, assuming areturn to more normal weather. By contrast,most analysts assumed that the high oil pricesduring the mid-1970s and early 1980s wouldlast indefinitely.10

Second, revenues from some primary com-modity exports (particularly oil and metals andminerals) are channeled through the publicsector, and the public sector often capturesthese gains through taxation. Political andsocial pressures often lead governments toincrease expenditures during commoditybooms. For example, the commodity boom inthe 1970s resulted in increased public expen-ditures in many Sub-Saharan African coun-tries (Bevan, Collier, and Gunning 1990;Wetzel 1992; Alpine and Pickett 1993; andLittle and others 1993).11 In some cases, boomrevenues invested in external assets by respon-sible governments have been dissipated laterby less responsible ones. An alternative is to

transfer public sector revenues to the privatesector, although this can be difficult to achievein a transparent and efficient way (Stauffer1999).

Even responsible governments that at-tempt to capture the gains from commoditybooms and to channel this income to the pri-vate sector may create difficulties. Public sec-tor interventions may obscure the source ofthe rise in income. This makes it more diffi-cult for the private sector to determine whetherthe increase in their real incomes is tempo-rary or permanent. Collier, Gunning, and As-sociates (1999) found that savings rates outof positive shocks tended to be higher wheneconomic agents were the direct beneficiariesof increased prices than when the rise in in-comes was intermediated by the government,for example, by taxing exporters and reallo-cating funds to other groups in the economy.One important role for government is to en-sure that adequate information on the causes(and, if possible, the likely duration) ofprice booms and busts is disseminated widelyso that private agents can make appropriatedecisions about savings behavior and resourceallocation.

Finally, a country may be unable to smoothconsumption because of limited access to in-ternational financial markets (see below).

The savings response by oilexporting countries to the recentswing in oil prices

The major oil exporting countries, whichare among the most commodity-depen-

dent economies in the developing world, facedsubstantial difficulties in smoothing consump-tion over the commodity price cycle.12 Two-thirds of these countries receive more than 80percent of export revenues from fuels. Theavailability of large oil reserves in a few coun-tries and the huge difference between the av-erage production cost and the selling price haveencouraged specialization in oil. According tothe United Nations Conference on Trade andDevelopment, half of the countries reporting

111

(including developing and high-income coun-tries) that had export concentration ratios ofmore than 50 percent were oil exporters.13

Implications of the 1990s swingin energy pricesThe sharp swings in the price of fuels from1996 to 1998 had an enormous impact onexport revenues, economic activity, and realincome in the major oil exporting countries.14

During the 1996–97 price boom, exportrevenues for 11 major oil exporters rose by45 percent of 1993–95 imports, and duringthe 1998 collapse in oil prices, export revenuesfell by 14 percent of base-year imports. Onaverage, changes in the terms of trade increasedreal income in the oil exporting countries by4.6 percent of GDP per year in 1996–97 com-pared with 1993–95, and reduced real incomeby 5.4 percent in 1998 (relative to 1993–95)(table 4.1).15 The average data, however, maskconsiderable differences among countries.Countries where oil accounted for the largestshare of export revenues, and where the ex-ports were large relative to imports and GDP,experienced the sharpest swings in real in-comes. For example, in Angola and Nigeriafuels account for more than 90 percent ofexport revenues, and the 1996–97 real incomegains were 14 and 8 percent of GDP, respec-tively. In Oman, exports were 40 percent largerthan imports and almost half the size of GDP,and the 1996–97 real income gain was 8.5percent of GDP. The composition of imports

also had an important impact on the size ofreal income gains. Trinidad and Tobago’s im-port price index increased 10.5 percent in 1996because of the jump in raw materials prices,and real income rose by only 0.2 percent in1996–97.

If a substantial portion of the change inoil prices represented temporary deviationsfrom trend, basing consumption on permanentincome would have implied that savings shouldhave increased significantly during the boomin 1996–97 and then fallen during the 1998price decline.16 During the boom the averagesavings rate did increase by slightly more thanhalf the rise in real income, measured as apercentage of base-year GDP, and a few coun-tries actually increased their savings rates bymore than the increase in real income. Thegroup’s average was reduced by the decline insavings rates in Angola, the Islamic Republicof Iran, and Trinidad and Tobago because ofeconomic or political difficulties. However,during the bust savings rates fell by the fullamount of the decline in real incomes. Of the11 countries, 6 reduced their savings rates bymore than the fall in real incomes, and 3 oth-ers reduced their savings rates by more than70 percent of the decline in real incomes. Inother words, most of the major oil exportersacted as if the loss was almost entirely tempo-rary. This experience has disturbing implica-tions for the future if oil prices continue tofall, as countries would fail to adjust over thecourse of the cycle to the permanently lower


Table 4.1 Savings, investment, and real income changes, selected country groups, 1996–98(percentage of GDP)

Savings Investment Foreign savings Real income1996–97 1998 1996–97 1998 1996–97 1998 1996–97 1998

All oil exporters 2.5 –5.3 –2.5 –3.4 –5.1 1.9 4.6 –5.4Middle-income 2.2 –5.0 –2.4 –3.6 –4.6 1.4 4.2 –5.3 Debtors 2.1 –4.2 –3.9 –5.8 –6.0 –1.6 3.4 –3.5 Creditors 2.4 –6.6 0.5 0.5 –1.8 7.0 5.6 –8.6Low-income 6.5 –8.0 –4.5 –0.1 –11.0 7.9 9.3 –7.0

Note: Savings and investment refer to the average ratio to GDP during the period shown, minus the average ratio in 1993–95.Real income refers to the change in real income from 1993 to 1995 caused by changes in the terms of trade, as a share of 1993–95GDP. All group averages are weighted by base-period GDP.Source: World Bank staff calculations.

112


level of oil prices. An analysis of a longer pe-riod (1980–96) that takes other determinantsof savings behavior into account also indicatesthat oil exporting countries have tended totreat changes in the terms of trade as tempo-rary (see below).

Using historical data on savings and realincome to analyze the extent of adjustment tocommodity price cycles has limitations. Thewindfall gains and losses are measured bychanges in oil prices relative to trend, and de-termining the trend depends on arbitrary judg-ments, such as the period over which the trendis estimated. Thus, the windfall elements ofchanges in the terms of trade may not be ac-curately measured. Box 4.1 describes counter-factual scenarios of savings behavior, whichprovide an additional estimate of the extent

of savings out of windfall income. These sce-narios generally confirm the conclusionsreached earlier.

Investment and foreign savings duringthe swing in fuel pricesOil exporting countries’ investment fell rela-tive to output over the commodity price cycle.The decline in investment was actually greaterthan the decline in domestic savings, so thecurrent account deficit fell. During the boomin oil prices, the average ratio of investmentto GDP in the oil exporting countries was 2.5percentage points lower than during the 1993–95 base period, and during the bust invest-ment was 3.4 percentage points lower than inthe base period (table 4.1). Of the 11 coun-tries, 7 experienced a decline in investment

The scenario results provide one measurementof the deviation of actual savings behavior fromdesired savings behavior. When oil prices were high(in 1996–97), oil exporting countries on averagesaved only a portion of the windfall gain. In theIslamic Republic of Iran and Trinidad and Tobagoactual savings rates fell relative to the base yeardespite the windfall gain. By contrast, in Algeria,Nigeria, and Venezuela savings rates rose by morethan the hypothetical increase. In 1998 all of thecountries (except Algeria) dissaved by significantlymore than the windfall loss.

Box 4.1 Counterfactual scenarios

Simple counterfactual scenarios generally con-firm the conclusion that the savings of oil

exporting countries declined relative to permanentincome over the oil price cycle. The scenarios showa hypothetical measurement of the windfall gain(or loss) from oil based on the deviation of the oilprice from trend.17 If economic agents consider thewindfall as temporary, then savings should rise bythe full amount of the windfall gain during theboom, and fall by the amount of the loss duringthe bust.18 The table shows the change in the ac-tual and hypothetical savings rates from the baseyear in six of the oil exporters during the oil pricecycle of the late 1990s.19

Change in ratio of savings to GDP from base years, selected countries, 1996–98(percentage points)

1996–97 1998

Actual Hypothetical Actual Hypothetical

Algeria 5.8 4.0 5.6 –2.3Iran, Islamic Rep. of –2.9 2.9 –6.8 –1.6Nigeria 8.0 3.5 –7.9 –1.8Saudi Arabia 2.4 5.0 –6.3 –3.0Trinidad and Tobago –6.2 3.0 –13.1 –1.7Venezuela 7.7 4.6 –6.0 –2.8


113

rates during both base and boom periods. Thedecline in investment over the commodity pricecycle is likely to impede efforts to improve oilexporting countries’ disappointing growthperformance.

The decline in investment in the majoroil exporting countries appears to be concen-trated in the private sector, as the average ratioof private investment to GDP fell by 2 per-centage points during the boom, while theratio of public investment to GDP declinedby only 0.5 percentage points.20 However, thedata on private investment include state en-terprises. Therefore, the allocation of publicand private investment may reflect decisionsconcerning the transfer of resources betweenstate enterprises and the central government,rather than different investment behavior bythe public and private sectors. Also, in coun-tries with a substantial influence over price,investment in the state enterprise responsiblefor oil exploration may decline during boomsif the authorities anticipate that output reduc-tions will be required to establish more prof-itable price levels. Given the commanding roleof the public sector in many of the oil export-ing countries, even declines in private invest-ment rates may largely reflect public sectordecisions.

Increasing public investment substantiallyduring a commodity price boom may not beadvisable, because the return on public invest-ment can decline during booms (see box 4.2).Also, central government investment expen-ditures tend to fall more heavily on non-tradable capital goods, such as buildings, thanprivate investment, and the rise in demand fornontradable capital goods during a boom willtend to increase their price. By contrast, theprices of tradable capital goods are set in glo-bal markets and, therefore, will be relativelyunaffected by demand conditions in an indi-vidual country. In 12 of 14 case studies re-ported in Collier, Gunning, and Associates(1999), the relative price of nontradable capi-tal goods rose during the commodity priceboom and fell thereafter. Thus, it may makesense for the central government to delay a

portion of the investment of real income gainsuntil after the boom is over and the price ofnontradable capital goods has declined.

The decline in investment was largeenough to cause countries to reduce their reli-ance on foreign savings, which fell by a cu-mulative 12 percent of GDP during thetwo-year boom in oil prices. Thus, a signifi-cant portion of the 1996–97 boom in exportreceipts was allocated to reserves.21 Foreignsavings also increased somewhat during the1998 fall in oil prices.

Most of the oil exporting countries relyheavily on sales of external assets to adjust todeclines in real income because external fi-nance has often not been available to helpsmooth consumption in the face of decliningcommodity prices. Developing countries aresubject to credit constraints that generallybecome more binding in the face of adverseshocks, and capital flows are often procyclicalfor marginally creditworthy borrowers(Dadush and Dasgupta 1999). Most of themiddle-income debtors among the oil export-ing countries either have speculative gradecredit ratings (Bahrain, the Islamic Republicof Iran, Trinidad and Tobago, and Venezu-ela) or are not rated. Of the oil exporting de-veloping countries, only Oman and SaudiArabia have investment-grade ratings. In-creases in oil prices can help make the mar-ginally creditworthy countries eligible forloans from the international capital markets(thus facilitating spending the windfall or evenmore than the windfall), but sharp declines inoil prices can mean that access is reduced oreven shut off.

Although many considerations influencethe level of flows from private capital mar-kets, some evidence indicates that private lend-ing to the oil exporters has been positivelyrelated to changes in the oil price. In periodswhere the price of oil (relative to the averageprice of manufactures exports from industrialcountries) was falling, long-term gross dis-bursements from private creditors have gen-erally declined, while gross disbursementshave risen when real oil prices were on the


114


upswing (table 4.2). The real oil price wassignificantly and positively related to disburse-ments to the oil exporting countries during1972–97, although the oil price explains only22 percent of the variation in disbursements.22

However, the procyclical effect of capital flowshad only a limited impact during the mostrecent commodity price cycle. Gross disburse-ments to the net debtors increased by $3 bil-lion during the most recent oil price boom(table 4.2), which is consistent with someimprovement in access. Nevertheless, net dis-bursements averaged negative $1 billion peryear during this period (compared with anaverage of $700 million per year in 1990–95)because of the large repayments due on exist-ing debt. Thus, reliance on foreign savings fellduring the rise in oil prices, as indicated intable 4.1.

Mexico’s experience prior to the 1982debt crisis, when oil accounted for three-quar-ters of export revenues, provides a more dra-matic illustration of the procyclical nature ofinternational capital flows (figure 4.4).Mexico’s current account deficit to GDP ra-tio more than doubled following the oil pricerise of 1973–74 and remained high through1976. The ratio fell in 1977 as oil prices re-mained flat and then more than doubled againfollowing the 1979–80 price rise, despite thesharp increases in export revenues along withthe price of oil. Clearly the major factor driv-ing the deficit was increased borrowing in re-sponse to the improved access to internationalcapital markets. Mexico’s debt-to-GDP ratio

rose from 14 percent in 1973 to 53 percent in1982, when the country could no longer ser-vice its commercial bank debt.

Fiscal policy and adjustment to the oilprice swings. Oil revenues are usually chan-neled through the public sector so that publicsector policies have an important influenceon the patterns of adjustment observed dur-ing the oil price cycle. In the past, many oilexporting countries have greatly increased fis-cal expenditures in response to increases infuel prices. Some governments have used theirnew-found access to capital markets to bor-row, in effect spending their anticipated fu-ture wealth today. The spending of whatturned out to be temporary increases in fuelrevenues resulted in a need for sharp reduc-tions in expenditures once fuel prices declined.In several countries, the efficiency of theselarge increases in expenditures was question-able (see box 4.2).

The fiscal positions of governments in oilexporting countries deteriorated somewhatover the commodity price cycle. The averageratio of the current budget balance to GDPimproved by 3.6 percentage points during the1996–97 boom (compared with 1993–95),slightly less than the real income gain (table4.3).23 By contrast, private savings declined.24

Subsequently, public savings dropped by al-most 6 percent of GDP during the bust in1998. Furthermore, several countries hadhigh fiscal deficits during the base period. Forexample, in Angola the current budget bal-ance in 1993–95 averaged 17 percent of GDP,

Table 4.2 Capital flows to oil exporters and energy prices, 1970–97

1970–80 1980–86 1986–90 1990–95 1995–97

Change in gross disbursements (millions of U.S. dollars)a 9,658 –2,897 410 –3,249 2,956Percentage change in real oil priceb 962 –65 29 –37 23

a. Change in gross disbursements of long-term flows from international capital markets from beginning to end of period.b. Percentage change in oil price relative to manufactures unit value index.Note: The countries included are Algeria, Gabon, the Islamic Republic of Iran, Nigeria, Oman, Republic of the Congo, Trinidadand Tobago, and Venezuela.Source: World Bank staff calculations.

115

and Venezuela experienced a foreign ex-change crisis in 1995–96 as a result of thehigh government deficit and triple-digit in-flation (see box 4.3). Also, the rise in publicsavings in 1996–97 in seven of the ten coun-tries reflected a sharp improvement in 1996,followed by some deterioration in the fiscalbalance in 1997, largely due to a significantrise in expenditures.25 The deterioration inthe fiscal position in 1998 (relative to highdeficits in 1993–95), therefore, is a matterfor serious concern in some of the oil export-ing countries.


Figure 4.4 Mexico's current account, 1970–82

40

35

30

25

20

15

10

5

0

6

Percentage of GDP Dollars per barrel

5

4

3

2

1

0

Current account deficit/GDP(left axis)

Oil price(right axis)

1970 1972 1974 1976 1978 1980 1982


Table 4.3 Ratios of public and private savingsto GDP, 1996–98(change from 1993–95 average)

1996–97 1998Public Private Public Private

All countries 3.6 –1.1 –5.6 0.3Middle-income 3.6 –1.4 –5.2 0.1

Debtors 2.4 –0.3 –9.2 5.0Creditors 5.8 –3.5 2.4 –9.0

Low-income 3.7 2.7 –10.8 2.2


Long-term economic performanceThe rise in consumption and the fall in pri-vate investment during the oil price cycle ofthe 1990s has complicated efforts to reversethe weak economic performance of many oilexporting countries since the quadrupling ofoil prices in 1974. Although output growthwas rapid in the 1970s, growth slowed dur-ing 1980–97 to 2.1 percent per year, belowthe 2.8 percent annual growth in developingcountries as a group.26 Output and invest-ment performance in the oil exporting coun-tries was substantially poorer compared withother countries in the same region in LatinAmerica and the Middle East and North Af-rica, while performance was slightly betteramong the oil exporters in Sub-Saharan Af-rica (table 4.4).

Poor economic performance since 1980has been attributed to the sharp decline andhigh level of volatility of oil prices, as well asto a mixed record of policy reform. Some coun-tries have made substantial efforts to improveincentives for private sector activities. Never-theless, policy regimes in some oil exportingcountries have been characterized by inad-equate macroeconomic environments; poor

116


investment decisions; and the frequent use ofproducer and consumer subsidies, price con-trols, and trade restrictions (Gelb and Associ-ates 1988; McMahon 1997). The World Bank’scountry performance ratings for oil export-ing countries are 0.5 points below the aver-age for developing countries (on a scale of 1to 5). Ratings done by private services alsoindicate that some oil exporting countries areperceived as worse than the average of devel-oping economies in terms of corruption, tradepolicy, and the environment for foreign directinvestment.27

Inappropriate policy regimes may haveimpeded the diversification of production to-ward non-oil activities that could supportfaster growth in the wake of the secular de-cline in oil prices. For the most part, the ma-jor oil exporting countries have been slowerto diversify their exports (to either other pri-mary commodities or manufactures) thanother developing countries. Of the 12 coun-tries where oil accounted for more than 80percent of exports in 1980, only 2 had reducedtheir share of fuel exports below 80 percentby 1997 (table 4.5), despite a 65 percent de-

took up substantial portions of the boom revenuesand left a legacy of debt and losses in post-boomyears that contributed to fiscal deficits (McMahon1997). Capital output ratios in oil exporting devel-oping countries increased during the 1970s, in partreflecting heavy investments in long gestationprojects in infrastructure and human capital forma-tion, as well as capital-intensive hydrocarbonsinvestments. However, the reduction in efficiencyalso stemmed from ill-conceived investments,planned too hastily and subject to enormous con-straints on implementation (Ahamed 1984).

One popular use of oil windfalls involvedefforts to diversify into resource-based industries,for example mineral processing (iron ore into steel,bauxite into aluminum, and hydrocarbons intopetrochemicals). Ultimately these efforts had disap-pointing results, largely because the projects wereimplemented inefficiently. With few financial con-straints and a strong impetus toward diversifica-tion, feasibility studies greatly overestimated futuredemand. By the mid-1980s prices were betweentwo-thirds and half the levels projected, and poten-tial rents were minimal, even for effectively imple-mented resource-based industrial projects. Lowcapacity utilization combined with high levels ofdebt and rising interest rates to greatly reduce theprofitability of resource-based industrial projects(Auty 1988).

Box 4.2 Public sector expendituresduring the oil price boom

The low efficiency of public sector expendituresin several oil exporting countries has impaired

economic performance. In all these countries, thepublic sector is the main owner of fuels resourcesand accounts for a significant share of economicactivity. In 1996 public sector expenditures aver-aged about 30 percent of GDP in the major oilexporters, compared with less than 20 percent, onaverage, for the developing countries.

Particular problems have been evident in theefficiency of public investment undertaken duringbooms. During the past three decades many oilexporting countries have squandered a large por-tion of the income from oil revenue increases onlow-return public investment projects (Gelb andAssociates 1988). Huge investments in state-ownedenterprises and human skills unsuited to today’sglobal marketplace resulted in relatively low pro-ductive uses of oil revenues in several countries ofthe Middle East and North Africa (Page 1999).Investments during the oil price boom also fre-quently generated disappointing results because oftransport bottlenecks resulting from a flood ofinvestments and the inability to fund, when pricesretreated, the recurrent costs required for projectsuccess. Particularly unsuccessful examples of in-vestment include the establishment of state enter-prises in the manufacturing sector. For example, inNigeria and Trinidad and Tobago, such enterprises

117

cline in the price of oil relative to manufac-tures from 1980 to 1997. By contrast, of the27 countries where the share of nonfuel com-modities exceeded 80 percent in 1980, 8 coun-tries managed to reduce their nonfuelcommodity share below 80 percent. Thegreater success in diversification of nonfuelcommodities exporters is not comparable forless extreme export concentrations. Almost allthe countries where fuels accounted for morethan 50 percent, but below 80 percent, ofexports in 1980 had reduced this share below50 percent by the late 1990s.

The failure to diversify is likely to ham-per growth over the medium term. Oil-depen-dent economies remain subject to the depressed


of Bs1.2 billion ($2.35 billion) but ended the yearwith only Bs362 billion, about $600 million. Thus,by the end of 1998, the government had spent allof the extraordinary revenue from the rise in theprice of petroleum, and was left with a higherpublic sector wage bill and high external amortiza-tion payments.

By contrast, Saudi Arabia devoted a signifi-cant portion of the rise in oil revenues in the mid-1990s to retiring debt, which helped to ease itsadjustment to the oil price fall in 1998. In 1996the government paid SR22 billion ($5.9 billion) todomestic creditors, largely to cover arrears fromunpaid bills, such as the issuance of agriculturalcertificates to farmers in lieu of cash for the 1993–94 harvest. These payments increased total expen-ditures to 35 percent over budget. Essentially, thegovernment spent a portion of the unanticipatedrise in revenues from higher oil prices on clearingarrears, leaving government finances in a strongerposition to accommodate the 1998 fall in rev-enues. While expenditures were cut and govern-ment payments to contractors and suppliers werestretched to the 180-day limit, most public agen-cies maintained current payments, and the govern-ment also continued to redeem past arrears(Kemp 1998).

Box 4.3 Fiscal adjustment in Venezuela and Saudi Arabia

Venezuela’s experience illustrates the dangers ofincreasing expenditures as a result of a com-

modity boom. The government had adopted astabilization program in 1996 in response to infla-tion of 103 percent and a deficit of 7 percent ofGDP. Higher international oil prices, increasedtax revenues, increased fuel production because ofthe investment drive initiated earlier in the de-cade, and a cutback in civil service wages in realterms achieved a massive shift of the fiscal bal-ance to a surplus of 7.25 percent of GDP in1996.

However, the government’s balance deterio-rated to 1.5 percent of GDP in 1997 because ofwage increases that averaged about 90 percent,transfers to local governments and decentralizedpublic sector agencies totaling 2.75 percent ofGDP, and a real appreciation that reduced thecontribution of dollar-denominated oil receipts tothe budget. Despite efforts to restrain expendituresin the first half of 1998, the deficit of the nonfi-nancial public sector is estimated at 6 percent ofGDP. Almost the entire deficit was financed by theliquidation of assets, including throughprivatization, liquidation of external assets, and asharp decline in government cash balances. TheTesorería Nacional began in 1998 with a balance

Table 4.4 Economic performance of major oilexporters and other countries, 1980–97(percent per year)

GDP Investmentgrowth growth


Oil exporters 2.2 1.3Other countries 4.1 1.9

Sub-Saharan AfricaOil exporters 2.1 0.2Other countries 1.8 –0.9


Oil exporters 1.5 –1.1Other countries 2.2 0.7


118


level of oil prices relative to manufactures andto slow growth in demand. These economieswill continue to suffer from the adverse im-pact of commodity price volatility on invest-ment and welfare (discussed in chapter 2).Also, the existence of large government-con-trolled rents in economies dependent on oilexports encourages rent seeking behavior thattends to lead to inefficient expenditures (seeLane and Tornell 1995). The lack of diversifi-cation also means that economies fail to cap-ture the benefits of manufactures production.Manufacturing is characterized by positiveexternalities, that is, benefits to the economythat are not captured by the firm. For example,skills training provided either formally orthrough learning by doing to suppliers and em-ployees is readily transferable to other activi-ties (Matsuyama 1992; Mayer 1996). Also,increasing returns to scale may exist in manu-facturing or in the education and job trainingthat is appropriate for manufactures (Sachsand Warner 1999). Thus, increasing the pro-duction of manufactures may raise the totalproductivity of the economy.

Slow output growth coupled with highpopulation growth poses enormous challengesfor many oil exporters. Projections indicate thatthe labor force in oil exporting countries willincrease by 54 percent by 2010, compared with23 percent for developing countries as a group.In addition, unemployment is already high inseveral countries. For example, approximately

40 percent of the Saudi Arabian population isyounger than 14 years of age, and it is there-fore anticipated that new entrants to the laborforce will total almost 4 million people overthe next decade, or two-thirds of the currentlabor force. Given the high rate of labor forcegrowth, even employment growth of 5 percentper year from 2000 to 2015 would make onlylimited progress in reducing unemployment bythe end of the period. The unemployment prob-lem in oil exporting countries is exacerbatedbecause the bulk of the labor force is employedin the nontraded sector. The oil sector is anenclave that generates few jobs directly. Largeparts of the nontraded sector are exposed tolimited competition, often implying limitationson the growth of output and labor demand.

Improvements in policy regimes tostrengthen incentives for non-oil productionand to remove constraints on competition areessential in several countries to avoid grow-ing unemployment and deteriorating livingstandards. In particular, some governmentsneed to consider lowering taxes and remov-ing barriers to employment that restrict theflexibility of using labor. For example, in Al-geria firms pay 24 percent of the wage bill intaxes and face significant severance and ad-ministrative costs when laying off employees(Ruppert 1996). It should be noted that insome countries attitudes are shifting towardmore market-oriented policies, which provideshope for an acceleration of growth.

Table 4.5 Share of oil and non-oil commodities in merchandise exports, 1980 and most recent year(number of countries)

Between 50 percent and 80 percent More than 80 percentof merchandise exports of merchandise exports

1980 Most recent yeara 1980 Most recent yeara

Non-oil exporters 28 18 27 19b

Oil exporters 7 1 12 10c

a. Only covers those countries included in the column to the left. Figures for oil exporters are from 1997; figures for non-oilexporters vary from 1993 to 1997.b. Of the eight countries that reduced their share of non-oil exports below 80 percent from 1980, all remained with shares above50 percent in the most recent year.c. The two countries that reduced their share of oil exports below 80 percent from 1980 remained with export shares above 50percent in the most recent year.Note: Excludes countries where data were not available in either 1980 or the mid-1990s.Source: World Bank 1999; World Bank staff calculations.

119

The savings response by non-oilcommodity exporting countriesof Sub-Saharan Africa

Non-oil primary commodity prices underwent a pronounced cycle, similar to oil

prices, during the 1990s. However, the com-modity price cycle does not appear to haveadversely affected the prospects for growth inthe non-oil exporting countries of Sub-SaharanAfrica for two reasons. First, changes in theterms of trade and real incomes were gener-ally smaller than in the oil exporting coun-tries. Second, improvements in policies enabledcountries to achieve increases in savings andinvestment rates during both booms and bustsin commodity prices.

Next to the oil exporters, the non-oil ex-porters of Sub-Saharan Africa are the mostdependent on primary commodities.27 Non-oilprimary commodities accounted for 80 percentof merchandise exports from these countriesin 1997, which is about the same as in 1990(figure 4.5). By contrast, non-oil primary com-modities accounted for 15 percent of merchan-dise exports from East Asia and Pacific, 35percent from Latin America and the Caribbean,and 21 percent from South Asia.

The commodity price cycle andthe terms of tradeThe global market prices of non-oil primarycommodities have undergone sharp changessince 1993. Agricultural prices, weighted bythe share of exports from Sub-Saharan Africaduring 1987–89, increased by 52 percent in1994–95 before falling by 19 percent duringthe next three years (figure 4.6). However, the1994–95 increase in these countries’ exportdeflators was significantly less than the rise inthe global market prices of agricultural prod-ucts, and the average export deflator remainedbelow the commodity price index through1998. In part, this reflected the 23 percentshare of manufactures in these countries’ ex-ports (manufactures export prices rose by only4 percent in 1994).28 There is a considerablelag between changes in market prices andchanges in the prices actually received by ex-porters, owing to the existence of fixed pricecontracts. Also, the global commodity priceindexes are based on average prices quoted ininternational markets throughout the year,while export deflators reflect prices at the datesof sale, which are limited to certain times ofthe year for agricultural commodities. For


Figure 4.5 Shares of merchandise exports of Sub-Saharan Africa non-oilcommodity exporters, 1990–97

80

Percent

60

40

20

019971996199519941990

Note: The list of countries included in the data for this figure is given in footnote 28. Coverage of trade shares excluded because of the lack of data.Source: Comtrade.

Agriculture

Metals and minerals

Manufactures

120


example, one reason that the average exportprice deflator rose much less than the interna-tional market price in 1994 was that coffeeprices jumped in July when frost damaged theBrazilian crop, but coffee sales are usuallymade from November to March, and by thenprices had dropped significantly. Therefore, theaverage international market price for coffeein 1994 was much higher than the price thatwas actually received by most developingcountries.29

The average rise in the terms of tradeduring 1994–95 was even less than the increasein export prices, because these countries alsoimport substantial amounts of primary com-modities (about one-fifth of total imports,equal to one-third the size of primary com-modity exports). Higher prices for these coun-tries’ agricultural and fuel imports thuspartially offset their higher export prices. Con-versely, in 1998 export prices fell by less thanthe price of imported products, particularlyfuels, and the terms of trade rose by 8 per-cent. The gain in real income during boomsaveraged 1.5 percent of base period GDP forthe agricultural exporters, while real income

was roughly equal to the base period duringbusts. Boom and bust periods were chosen foreach country based on when agricultural pricesand export revenues were rising.30

While the terms of trade and real incomesrose only modestly during the commodity priceboom, export revenues soared, increasing byan average of 12 percent per year. A portionof the rise in export revenues may stem fromincreased reporting rather than actual increasesin exports, as the liberalization of trade andforeign exchange regimes reduced incentivesto evade tariffs. This strong export perfor-mance reflected somewhat higher prices andsharp increases in volumes as world trade in-creased. Several countries improved incentivesfor agricultural production by removing pricecontrols, by dismantling government-runboards that monopolized the purchase of keyexport crops, and by establishing market-basedexchange rates.

The metals and minerals exporters alsoexperienced sharp changes in the market pricesof the commodities they exported, but rela-tively small movement in the terms of trade.Global metals and minerals prices rose by 43

Figure 4.6 Agricultural exporters' trade prices, 1993–98

Index (1990=100)

150

140

130

120

110

100

90

80199819971996199519941993

Terms of trade

Export deflator

Agricultural commodity prices


121

percent during 1994–95, only somewhat lessthan the prices of agricultural goods (figure4.7). However, these countries’ export pricedeflator increased by only 2 percent in 1994and by 6 percent in 1995. 31 By 1998 the ex-port price deflator and the terms of trade werenot significantly different from their 1994 lev-els. During boom periods, the change in realincome caused by changes in traded goodsprices averaged 0.5 percent of base year GDP,and during bust periods, real incomes fell by0.6 percent.32 These countries’ export volumesincreased by an average of 3.4 percent per yearduring boom periods, compared with 7.7 per-cent for the agricultural exporters.

Policies and country performanceImproved policy performance enabled manyof the non-oil exporting countries of Sub-Saharan Africa to increase their savings andinvestment rates over the commodity pricecycle. Several countries in Sub-Saharan Africaadopted more prudent macroeconomic poli-cies, established market-determined exchangerates, reduced quantitative restrictions onimports, reduced tariff rates, introduced

greater private sector participation in theeconomy through privatization and the dis-mantling of marketing boards, removed pricecontrols and some other restrictions on pri-vate sector economic activity, and took stepsto improve the efficiency and soundness oftheir financial sectors.

These policy reforms usually increasedsavings through a number of channels. First,one goal of several programs was to raisepublic savings directly through increases inrevenues and through improved control overcurrent expenditures, in part to strengthenmacroeconomic management and in part tomake necessary increases in public investment.Second, reduced inflation and the establish-ment of more efficient financial systems in-creased the return on holding savingsdomestically, which may have had some im-pact on private savings behavior. Finally, policyreforms increased the expected return on in-vestment, particularly for exporters, becausea key element of reform programs involvedreducing biases against exports. Real invest-ment increased strongly in many of these coun-tries. Since they lack access to private capital


Figure 4.7 Metals and minerals exporters' trade prices, 1993–98

Index (1990=100)

110

100

90

80

70

199819971996199519941993

Terms of trade

Export deflator

Metals and minerals prices


122


markets and official flows have declined dur-ing the 1990s, the rise in investment had to befinanced by increased domestic savings.

The link between policies and perfor-mance can best be seen by grouping the sampleof non-oil commodity exporters in terms ofratings of their policies (table 4.6).33 Coun-tries with ratings of more than 3.5 (1 is worst,and 5 is best) increased their GDP by 5 per-cent per year during the boom, while coun-tries with ratings of 3.5 or less increased theirGDP by only 3 percent per year. Policy per-formance was by far a better predictor of sav-ings behavior than changes in real income. Thecountries with better policies increased sav-ings rates by more than those with relativelypoorer policies (almost 7 percent of GDP overthe base period versus 3 percent of GDP).However, the average rise in real income inthe countries with better policies was less thanin the countries with poorer policies (1 per-cent of GDP versus 2 percent).34

Note that this is a biased sample of Sub-Saharan non-oil commodity exporters. To ana-lyze macroeconomic adjustment over thecommodity price cycle, countries with severecivil conflicts and countries with inadequatedata were excluded. Both these groups of coun-tries tend to have lower performance ratingsthan countries with civil peace and more de-veloped statistical services. The average per-formance rating of the countries in the sampleis 3.4, while the average performance ratingof the non-oil commodity exporting countriesin Sub-Saharan Africa excluded from the

sample is only 2.7. Thus the improvements insavings, output growth, and investment for thesample are probably larger than for the Sub-Saharan African non-oil commodity export-ers as a whole.

Savings, real income changes, and thecommodity price cyclePolicy performance was the primary reasonfor differences in savings performance. Bycontrast, changes in real income caused bychanges in traded goods prices had a more lim-ited impact on savings during the commodityprice cycle. To analyze adjustment to real in-come changes during the commodity pricecycle, a useful approach is to group the sampleof the non-oil exporting countries by the per-centage change in export revenues and theterms of trade in the boom relative to the baseperiod for each country (table 4.7).

The two groups with more than one coun-try experienced substantial increases in sav-ings rates during both booms and busts, despitethe different magnitudes of changes in realincome (table 4.8). The group with flat termsof trade and relatively small changes in realincomes actually increased its savings rates bymore than the group with large terms of tradeand real income gains. Savings rates remainedabove base period levels during the decline incommodity prices. Both groups’ savings per-formance was much greater than that achievedby the oil exporting countries during the boomand bust of oil prices and was larger than theaverage savings out of windfall incomes (aboutone-half) recorded in other studies of savingsfrom commodity price windfalls (Collier, Gun-ning, and Associates 1999). An analysis of sav-ings behavior in Sub-Saharan African countriesover a longer period, from 1980 to 1996, thattakes other determinants of savings into ac-count also finds that savings were not closelytied to changes in real incomes (box 4.4).

The sharp increase in savings rates dur-ing boom periods enabled the Sub-Saharan Af-rican non-oil commodity exporters to increaseinvestment rates compared with the base pe-riod by 1.1 percentage points in the group with

Table 4.6 Policy performance and GDP,savings, and real income during boom periods(average annual percentage change from base period)

Average policy Realperformance ratinga GDP Savings income

Above 3.5 5.3 6.7 1.23.5 or below 3.2 3.1 1.9

a. Based on a survey of World Bank country economists.Note: Only boom periods are shown because none of thegood performers had bust periods.Source: World Bank staff calculations.

123

large terms-of-trade gains and by 3.8 percent-age points in the group with only small im-provements in the terms of trade. Savings andinvestment increased by much less during busts,but still rose by more than the negligible changein real income. This experience contrastssharply with that of the oil exporting coun-tries, where investment rates fell over the com-modity price cycle. The non-oil exporters ofSub-Saharan Africa also allocated a portionof the rise in domestic savings to reduce theirreliance on foreign savings, which fell by morethan 2 percent of GDP in both groups duringbooms and by 1 percent of GDP during busts.

The difference in savings performancebetween the groups is partially related tochanges in aid flows. Most of these countriesexperienced a decline in net concessional flowsduring boom periods (data are not yet avail-able on net flows during busts, which tookplace in 1998 for most countries), reflectingthe general decline in aid since the early 1990s.However, the countries in the first group (withthe smallest rise in savings rates despite thelargest rise in real income) saw a decline inthe ratio of aid to GDP of 5 percentage pointsin the boom compared with the base period.By contrast, the group of countries with the


Table 4.7 Exports and terms-of-trade changes, boom compared to base period

Large increase Little change Large decreasein terms of trade in terms of trade in terms of trade

Large increase in export revenues Benin Central African Togo Republic

Botswana Côte d’IvoireChad GhanaEthiopia GuineaGhana MadagascarKenya MaliMalawi SenegalTanzania ZimbabweUganda

Little change in export revenues Mauritania Niger Zambia

Note: The cutoff for large increases is plus or minus 5 percent for export revenues and 3 percent for the terms of trade (per year).The boom and bust periods were chosen for each country based on movements in export prices, the terms of trade, and exportrevenues during 1994–98, with the three years prior to the beginning of the boom as the base period. Most of the boom periodsare 1995–97, 1994–97, or 1994–98.Source: World Bank staff calculations.

Table 4.8 Changes in savings and real income relative to base periods(percentage of GDP)

Foreign RealCountry group Savings Investment savings income

During boomLarge increases in exports and in terms of trade 3.5 1.1 –2.4 2.3Large increases in exports, flat terms of trade 6.5 3.8 –2.7 0.6

During busta 1.5 0.4 –1.0 0.1

a. This includes only four countries, because several countries did not experience a decline in export revenues or real income, sothat the boom continues through 1998.Note: Savings indicates an average change in the ratio of savings to GDP in boom or bust periods relative to the base period. Realincome indicates average change in real income (as a percentage of the base period GDP) caused by changes in export and importprices. Base, boom, and bust periods differ by country depending on the evolution of export prices, revenues, and terms of trade.Group averages reflect the weight of GDP in the base period.Source: World Bank staff calculations.

124


longer period. First, the relationship between sav-ings and the terms of trade varies considerablyamong country groups. This relationship is posi-tive and significant for oil exporters in both up-swings and downswings and for metals andminerals producers during upswings (but notdownswings), but is not significant for agriculturalproducers. Note that the relationship betweensavings and the terms of trade is the same in theupswing as it is in the downswing during 1980–96.In the most recent cycle, however, oil exportingcountries tended to save less than the rise in realincome during the upswing and dissaved the fullamount of the fall in real income during the down-swing (see main text). It is unclear whether thisdifference stems from the longer time period anddifferent country coverage of the regressions, or ifit reflects a more accurate measurement of theimpact of terms-of-trade changes after accountingfor other determinants of savings. Second, whilechanges in the terms of trade are not significantlyrelated to savings by the Sub-Saharan Africanagricultural producers, a time trend is positive andsignificant. This result is at least consistent withthe view that policy improvements undertaken inseveral countries since the 1980s have had a moreimportant impact on savings than changes in theterms of trade.

Box 4.4 Savings and real incomes duringthe commodity price cycle

The main text shows that changes in real in-comes because of shifts in the terms of trade

have not played a major role in determining sav-ings behavior in the non-oil exporters in Sub-Sa-haran Africa. This box attempts to control forother determinants of savings when measuring theimpact of changes in real incomes. The economicsliterature cites several determinants of savings: (a)higher GDP growth is associated with higher sav-ings, (b) countries with higher per capita incomesnormally have higher savings rates, (c) high levelsof inflation discourage savings by reducing the realvalue of some financial assets, (d) countries with ahigher age dependency ratio (share of populationeither very young or very old) generally save less,and (e) increases in foreign savings and official aidare associated with reduced savings (Ghura andHadjimichael 1995; Ogaki, Ostry, and Reinhart1995; Dayal-Gulati and Thiman 1997; Loayza,Serven, and Schmidt-Hebbel 1999). The tableshows the estimated impact on savings of upswingsand downswings in the terms of trade during1980–96 after controlling for these other determi-nants of savings. 35

The regression results are roughly consistentwith two of the major findings presented in thischapter, even though the regressions account morefully for other determinants of savings and cover a

Terms-of-trade shocks and savings

Impact on savings rates Term-of-trade Terms-of-trade(number of observations) upswing downswing Time

Oil exporting countriesa (120) 0.143* 0.131*Metals and minerals producers

in Sub-Saharan Africab (78) 0.196** 0.110Agricultural producers in

Sub-Saharan Africac (207) 0.038 0.022 0.004*

a. Algeria, Equatorial Guinea, Gabon, the Islamic Republic of Iran, Nigeria, Republic of the Congo, Russian Federation,Trinidad and Tobago, Venezuela, and the Republic of Yemen.b. Botswana, Guinea, Mauritania, Niger, Togo, and Zambia.c. Central African Republic, Chad, Côte d’Ivoire, Ethiopia, Ghana, Kenya, Madagascar, Malawi, Mali, Senegal, Uganda, andZimbabwe.Note: Asterisks indicate significance at a 1 percent (*) or 10 percent (**) level. Estimation is based on panel data withcountry-fixed effects and instrumental variables estimators.Source: World Bank staff calculations.

125

largest increase in savings rates experienced adecline in the ratio of aid to GDP of only 1percentage point. The relationship betweensavings, real incomes, and aid flows on a coun-try basis is represented in figure 4.8, whichshows data on changes in real income andsavings (as a share of GDP) during boom pe-riods compared with base periods. All thecountries with real income gains that werelarger than average (1 percent of GDP), butwhich had increases in savings that weresmaller than average (5 percent of GDP), hadlarge declines in aid flows.

Both the private and public sectors con-tributed to the rise in savings rates during theboom.36 Stronger macroeconomic policieswere reflected in improvements in governmentcurrent balances for both groups. National ac-count data show that private savings also in-creased significantly during the boom, by 2.8percent of GDP in those countries with flatterms of trade and by 1.7 percent of GDP inthe other group (figure 4.9).

Improved policy performance led to anacceleration of GDP growth, which rosesharply during booms, averaging 3.8 and 4.6percent per year in the two groups (table 4.9).

The terms-of-trade effects account for only asmall fraction of the increase in GDP growth.Based on estimates by Deaton and Miller(1995), the income gains would have increasedGDP growth on average by 1.2 and 0.3 per-centage points during the boom periods forthe two groups.37 These results do not implythat the changes in real incomes did not havean impact on growth rates. Countries experi-encing a bust did see slower rates of GDPgrowth (by 2 to 3 percent per year) than coun-tries experiencing a boom in commodityprices.38

The commodity price cycle and theinternal distribution of incomeThe commodity price cycle had a significantimpact on the internal distribution of incomein many of the non-oil commodity exporters,including some countries that did not experi-ence a sharp change in aggregate real income.Table 4.10 decomposes the change in real in-come between changes that resulted frommovements in export prices and from move-ments of import prices. The countries withlittle change in the terms of trade experiencedan internal shift in real income when commod-


Figure 4.8 Real income, savings, and aid during boomsChange from base period

10

–5 5 10 15

5

0

0

–5

–10

Cha

nge

in r

eal i

ncom

e (p

erce

ntag

e of

GD

P)

Change in savings (percentage of GDP)

Aid flows greater than averageAid flows less than average


126


ity prices were high. Two factors lay behindthis shift. Firms and households that earn theirincome from exports saw a gain of up to 2percent of GDP because of higher prices, whilefirms and households that depend largely onimported goods lost as much as 1.4 percentof GDP.39 In many of these countries, higherexport prices increased incomes in rural ar-eas, where export crops are produced andwhere people tend to consume locally pro-duced food, but higher import prices reducedincomes in the urban areas that are dependent

43210

Private

Public

Flat exports and

terms of trade

High exports, flat

terms of trade

High exports, large

terms of trade gain

Figure 4.9 Change in savings rates during boom compared with base periodsPercentage of GDP


on imported food. Benin provides an interest-ing example of the internal distribution of realincome changes (box 4.5).

Conclusion. The recent swing in commod-ity prices posed a significant challenge to eco-nomic management in those developingcountries that are dependent on commodityexports, and the quality of economic policiesplayed an important role in determining coun-tries’ responses. Policies have been weak inseveral oil exporting countries, and many ofthe countries experienced a deterioration of

Table 4.9 Economic performance(average annual percentage change from base period)

Country group GDP Investment Export volumes

During boomLarge increases in exports and in terms of trade 4.6 5.3 6.4Large increases in exports, flat terms of trade 3.8 7.1 7.4

During busta 1.7 1.7 5.2

a. Country coverage differs substantially between boom and bust periods.Note: Data show change in real income caused by export and import prices. A negative sign indicates a fall in real income, eitherfrom lower export prices or higher import prices. See notes to table 4.8.Source: World Bank staff calculations.

127


savings and investment performance during thecommodity price cycle. By contrast, the pastfew years have seen substantial improvementsin policies in many of the non-oil exportingcountries of Sub-Saharan Africa, and thesecountries’ savings and investment rates in-creased.

Notes1. The impact of volatility on the poor is discussed

in chapter 2.2. Some measures of commodity price volatility

declined after the mid-1980s, but remained higher thanin the 1970s (Dehn and Gilbert 1999).

3. Dehn and Gilbert (1999) used a recursive fore-cast model to evaluate the degree of uncertainty or

Table 4.10 Decomposition of real income changes(percentage of base year GDP)

Country group Real income change caused by:Export prices Import prices

During boomLarge increases in exports and in terms of trade 2.7 –0.4Large increases in exports, flat terms of trade 2.0 –1.4

During busta –0.3 0.2

a. Country coverage differs substantially between boom and bust periods.Note: Data show change in real income caused by export and import prices. A negative sign indicates a fall in real income, eitherfrom lower export prices or higher import prices. See notes to table 4.8.Source: World Bank staff calculations.

was much greater than the gain from exports. Thenet loss to the economy was about 13 percent ofGDP during 1994–96 compared with 1992–93.

Rural areas saw an increase in incomes andlabor demand as the price rise encouraged greatercotton production (the area under cotton cultiva-tion almost doubled from 1993–94 to 1996–97).40

Urban workers, however, experienced a decline inliving standards following the devaluation as thecost of living increase exceeded the rise in wages. A46 percent rise in the salaire minimuminterprofessionel garanti was granted in May 1994,compared with a 54 percent increase in the con-sumer price index. Moreover, actual earnings didnot fully reflect the increase in wage rates as thegovernment was substantially in arrears payingwage increases. Finally, many workers, particularlythe more vulnerable, lower-income ones, wereemployed in the informal sector where incomesprobably did not keep pace with the rise in livingcosts.

Box 4.5 Real incomes in Benin duringthe commodity price cycle

Benin experienced sharp changes in the distribu-tion of income as a result of the commodity

price cycle and the CFA franc devaluation duringthe mid-1990s. During 1994–96 the U.S. dollarprice of cotton, which accounts for 80 percent ofBenin’s merchandise exports, rose by 48 percentfrom its 1992–93 level. Also in 1994, the 50 per-cent devaluation of the CFA franc greatly increasedthe prices of exports and imports in local currency.By 1996 the cotton producer price had risen totwice its 1992–93 level in CFA franc terms, and theboost to real income during 1994–96 from the risein export prices was 7 percent of 1992–93 GDP.Import prices also increased sharply, however,because of the devaluation and the rise in the dol-lar price of primary commodity imports. Importprices in CFA franc terms doubled during 1994–96compared with 1992–93, equivalent to a loss of 20percent of GDP. Imports were almost three timesthe size of exports (excluding re-exports) during1992–93, so the loss from the rise in import prices

128


unpredictability in commodity prices, where uncer-tainty was measured by the standard deviation of theforecast error.

4. Selected commodity prices (coffee, cotton, gold,and copper) have remained roughly flat relative to U.S.inflation since 1900. At a minimum, owners of a con-stant flow of one of these commodities would not haveseen much growth in real income during the twentiethcentury (Deaton 1999).

5. Statistical analyses of commodity price seriescan be found in Cuddington and Urzúa 1989;Cuddington 1992; Deaton and Laroque 1992; Reinhartand Wickham 1994; Léon and Soto 1997; and Cashin,Liang, and McDermott 1999.

6. These data are based on a limited sample ofcommodities. The agricultural products included arecocoa, coffee, tea, fats and oils, cotton, sugar, rubber,soybeans, maize, rice, and wheat. The metals and min-erals are aluminum, copper, gold, and steel.

7. However, these increases did not occur in thecountries of the former Soviet Union, where produc-tion continued to fall. According to the Food and Ag-riculture Organization of the United Nations, agricul-tural production in transition economies fell by 35percent from 1990 to 1998. This partially offset gainsin other countries.

8. Case studies are given in Collier, Gunning, andAssociates (1999), Ingham (1973), and Paxson (1992).Econometric studies include Dayal-Gulati andThimann (1997) and Loayza, Serven, and Schmidt-Hebbel (1999). Studies have also found thatconsumption behavior in some developing countriesis closely related to permanent income (Ostry andReinhart 1992; Ghosh and Ostry 1993; Borenszteinand others 1994).

9. The Middle East and North Africa region hasmany oil exporters. Note that these estimates of thesize of resource depletion are probably overstated.Estimates of depletion should reflect the price of theresource minus the marginal cost of extraction. In prac-tice, data on marginal extraction costs are generallynot available, so that average production costs are used.This practice overstates depletion, and hence under-states levels of genuine savings.

10. See, for example, the U.S. Department ofEnergy’s Annual Energy Outlook, various issues. TheWorld Bank’s forecasts also envisioned continued highenergy prices in the early 1980s.

11. However, many Sub-Saharan African coun-tries have liberalized marketing arrangements and re-duced export taxes since the 1970s, which has reducedthe government’s role in allocating commodity pricewindfalls.

12. This section covers 11 developing countrieswhere oil exceeded 50 percent of exports in 1997,namely: Algeria, Gabon, the Islamic Republic of Iran,Oman, Trinidad and Tobago, and Venezuela (middle-income debtors); Bahrain and Saudi Arabia (middle-in-come creditors); and Angola, Republic of Congo, andNigeria (low income). These countries account for 93percent of the total oil exports of the major oil export-ing developing countries. Equatorial Guinea, Iraq, Libya,the Republic of Yemen, and some of the transition econo-mies are excluded due to lack of sufficient data.

13. The concentration index ranges from zero toone, with one representing the most extreme concen-tration (UNCTAD 1994).

14. For the purpose of this analysis, the fuels priceboom is dated as 1996–97 (when prices were abovetrend) and the bust as 1998 (when prices fell belowtrend). The trend is calculated beginning in 1986 toexclude the last major collapse of oil prices.

15. The methodology used in the IMF (1998) wasused to measure the loss in real incomes due to termsof trade changes. The formula is [(PXt+1 – PXt)*Xt –(PMt+1 – PMt)Mt)]/GDPt, where PX and PM are priceindexes for exports and imports, X and M are exportand import volumes, and GDP is at current prices. Thiscalculation ignores changes in the volume of exportsand imports. To the extent that changes in volumes arein response to changes in prices, the formula fails toreflect the full impact of price changes on real incomes.

16. The flexibility of consumption levels in re-sponse to terms-of-trade changes is limited in somecountries because of the unstable security environment(military expenditures are typically counted as con-sumption in the national accounts).

17. Permanent income from oil was calculatedassuming that production was constant (the extent ofdepletion of oil reserves was not viewed as an impor-tant consideration over a short time period) and thatthe real oil price was equal to the trend. Permanentincome in the non-oil sector was set equal to the ac-tual level during the period.

18. This is an extreme assumption, as typicallyproducers will view a portion of the rise in price aspermanent.

19. The number of countries is limited based onthe availability of data for 1998 and on estimates fornon-oil GDP.

20. This calculation is based on a sample of sevencountries, determined by data availability. The datarefer to fixed investment, while the data on total in-vestment include changes in stocks.

21. There were substantial differences betweenthe magnitude of changes in foreign savings as shown

129

in national income accounts and in the current accountof the balance of payments, partly because of data dis-crepancies between the two sources and partly becauseof movements in factor income and other current ac-count items.

22. The regression equation is disbursements =6.6 + .66*oil price, where all variables are expressedin natural logarithms. The coefficient of the oil priceis significant at the 1 percent level.

23. The measurement of the change in real in-come refers to the economy as a whole, not to the gov-ernment. Determining the government’s share of thereal income gain is difficult. Typically, the government,or the public sector as a whole, directly captures thebulk of the rise in oil prices, but data to measure theimpact of changes in import prices on government in-comes are not available.

24. Private sector savings are calculated by sub-tracting the government’s current balance from theeconomy’s total savings. Thus, public enterprises areclassified under the private sector, which for somecountries reduces the accuracy of these data as indic-tors of private sector behavior. Also, discrepancies innational income accounts could distort the data onprivate sector savings.

25. Data on public savings in 1998 are not avail-able for the Republic of the Congo. Therefore, table4.3 covers only 10 countries (compared with 11 coun-tries for table 4.1).

26. Output growth is not a reliable indicator ofeconomic performance for some of the major oil ex-porting countries because it reflects, in part, changesin the volume of oil production in line with efforts tosupport prices. Also, the huge shifts in oil prices sincethe 1970s make interpreting long-term changes in con-stant price data difficult. Growth in non-oil output isa better indicator of performance, but sufficient his-torical data are available for only a limited number ofcountries.

27. International Country Risk Guide 1999.28. The countries discussed in this section include

the agricultural exporters Benin, Central African Re-public, Chad, Côte d’Ivoire, Ethiopia, The Gambia,Ghana, Kenya, Madagascar, Malawi, Mali, Senegal,Tanzania, Uganda, and Zimbabwe. Minerals export-ers are Botswana, Guinea, Mauritania, Niger, Togo,and Zambia. Burundi, Democratic Republic of theCongo, Liberia, Mozambique, Rwanda, Sierra Leone,Somalia, and Sudan are excluded because of civil strife.Burkina Faso, Cape Verde, Comoros, Djibouti, Eritrea,Guinea Bissau, Lesotho, Namibia, São Tomé andPrincipe, Seychelles, and Swaziland are excluded be-cause of the lack of data. The sample countries ac-count for approximately 75 percent of the total GDP

of the non-oil commodity exporters of Sub-SaharanAfrica.

29. However, a substantial portion of these coun-tries’ manufactured exports represents agriculturalgoods with limited processing, whose prices probablymove closely with primary commodities. The exportprice deflator roughly followed the average deflatorof the largest two commodities from each country.

30. The commodity price index also may differfrom export price deflators because the fixed weightsused in the former may not reflect the export compo-sition of the countries in our sample during the 1990s.For example, coffee and cocoa prices, which accountfor almost half of the Sub-Saharan Africa agricultureprice index using 1987–89 weights, increased morerapidly than total agricultural prices during 1990–97,but exports of coffee and cocoa from the principalproducers increased more slowly than the group’s to-tal exports. Therefore, in this case the weight of themore rapidly growing commodities among the exportprice deflators would have declined relative to the fixedweight index.

31. The sample of countries with bust periods issmaller than the sample with boom periods because insome countries export revenues and prices never fellsignificantly.

32. The price index for metals and minerals doesnot cover all the relevant goods exported by these coun-tries. For example, the average commodity price formetals and minerals does not include diamonds, whichis an important export of Botswana, because of thelack of data.

33. The country composition during boom andbust periods differs, because some countries never ex-perienced a bust.

34. The ratings are prepared by World Bank coun-try economists, based on their evaluation of macro-economic policies, structural policies, reduction ofinequality, and public sector management.

35. We discuss later how differences in savingsperformance depended in part on aid flows. Note thatchanges in aid flows do not explain the differences inperformance of the two groups, as they experiencedalmost the same decline in aid flows during boom pe-riods.

36. The dependent variable is savings divided byGDP. The terms of trade variables are constructed inthree steps. First, an index of the terms of trade is mul-tiplied by the ratio of exports to GDP and the growthrate is calculated. Second, dummy variables are con-structed to represent upswings and downswings in theterms of trade. A value of 1 or 0 is assigned to each yearand country, based on an analysis of the evolution ofprices and revenue of each country’s commodity exports.


130


Finally, the dummy variables are multiplied by thegrowth rate of the terms of trade variable.

37. Data are not yet available to distinguish be-tween public and private savings during bust periods,which occurred in 1998 for most countries.

38. Deaton and Miller (1995) estimate that a 1percent gain in GDP from terms of trade would addabout one half of 1 percent in GDP growth during theboom, in addition to the real income effects.

39. The country composition is substantiallydifferent between boom and bust periods. Thus theconclusion cannot be drawn that output in individualcountries necessarily grew more slowly during the bust.

40. Exporters may depend on imported inputs,and firms in the traded goods sector that depend onimported inputs may have benefited from increasedoutput prices. Thus, neither group experienced the fullamount of gain or loss represented by the change inexport and import prices.

41. However, producers’ prices rose by consider-ably less than what would be implied by the increasein international prices and the CFA franc devaluation.Therefore, the marketing board or the governmentreaped a large portion of the gain.

ReferencesAhamed, Liaquat. 1984. “Oil Exporting Developing

Countries: Macroeconomic Policies and Adjust-ment Issues.” Country Policy Department Dis-cussion Paper 1984–7. Washington, D.C.: WorldBank.

Alpine, Robin W., and James Pickett. 1993. Agricul-ture, Liberalisation and Economic Growth inGhana and Côte d’Ivoire 1960–90. Paris:Organisation for Economic Co-operation andDevelopment.

Auty, Richard M. 1988. “Oil-Exporters’ Disappoint-ing Diversification into Resource-Based Industry.”Energy Policy 16:230–42 (June).

Bevan, David, Paul Collier, and Jan Willem Gunning.1990. Controlled Open Economies. Oxford:Clarendon Press.

Bleaney, Michael F., and David Greenaway. 1993.“Long-Run Trends in the Relative Price of Pri-mary Commodities and in the Terms of Trade ofDeveloping Countries.” Oxford Economic Papers45(July): 349–63.

Borensztein, Eduardo, Mohsin S. Khan, Carmen M.Reinhart, and Peter Wickham. 1994. The Behav-ior of Non-oil Commodity Prices. Washington,DC: International Monetary Fund.

Cashin, Paul, Hong Liang, and C. John McDermott.1999. “How Persistent Are Shocks to World Com-

modity Prices?” Working Paper no. 99/80. Inter-national Monetary Fund, Washington, D.C.

Collier, Paul, and Jan Willem Gunning. 1996. “PolicyTowards Commodity Shocks in DevelopingCountries.” Working Paper no. 96/84. Interna-tional Monetary Fund, Washington, D.C.

Collier, Paul, Jan Willem Gunning, and Associates.1999. Trade Shocks in Developing Countries.Volume I: Africa and Volume II: Asia and LatinAmerica. Oxford: Oxford University Press.

Cuddington, John. 1988. “Fiscal Policy in Commod-ity-Exporting LDCs.” Policy, Planning and Re-search Working Paper no. 33. World Bank, Officeof the Vice President, Development Economics,Washington, D.C.

________. 1992. “Long-Run Trends in 26 PrimaryCommodity Prices.” Journal of DevelopmentEconomics 39 (October): 207–27.

Cuddington, John, and Carlos M. Urzúa. 1989. “Trendand Cycles in the Net Barter Terms of Trade: ANew Approach.” Economic Journal 99 (396):426–42.

Dadush, Uri, and Dipak Dasgupta. 1999. “The Ben-efits and Risks of Private Capital Flows to Devel-oping Countries.” World Bank, DevelopmentProspects Group, Washington, D.C.

Dayal-Gulati, Anuradha and Christian Thimann. 1997.“Saving in Southeast Asia and Latin AmericaCompared: Searching for Policy Lessons.” Work-ing Paper no. 97/110. International MonetaryFund, Asia and Pacific Dept., Washington, D.C.

Deaton, Angus S. 1999. “Commodity Prices andGrowth in Africa.” Journal of Economic Perspec-tives 13(3): 23–40.

Deaton, Angus S., and Guy Laroque. 1992. “On theBehavior of Commodity Prices.” Review of Eco-nomic Studies 59 (198): 1–24.

Deaton, Angus S., and Ronald I. Miller. 1995. “Inter-national Commodity Prices, Macroeconomic Per-formance, and Politics in Sub-Saharan Africa.”Studies in International Finance no. 79. PrincetonUniversity Press, Department of Economics, In-ternational Finance Section, Princeton, N.J.

Dehn, Jan, and Christopher L. Gilbert. 1999. “Com-modity Price Uncertainty and Economic Growthand Poverty.” Processed.

Gelb, Alan H., and Associates. 1988. Oil Windfalls:Blessing or Curse? New York: Oxford Univer-sity Press.

Ghosh, Atish, and Jonathan Ostry. 1993. “Do CapitalFlows Reflect Economic Fundamentals in Devel-oping Countries?” IMF Working Paper no. 93/34. International Monetary Fund, Washington,D.C.

131

Ghura, Dhaneshwar, and Michael T. Hadjimichael.1995. “Growth in Sub-Saharan Africa.” IMFStaff Working Paper no. 95/136. InternationalMonetary Fund, Washington, D.C.

Grilli, Enzo, and Maw Yang. 1988. “Primary Com-modity Prices, Manufactured Goods Prices andthe Terms of Trade of Developing Countries:What the Long Run Shows.” World Bank Eco-nomic Review 2 (January): 1–47.

Hamilton, Kirk, and Michael Clemens. 1999. “Genu-ine Savings Rates in Developing Countries.”World Bank Economic Review 13(2):333–56.

Harberger, Arnold C. 1950. “Currency Depreciation,Income, and the Balance of Trade.” Journal ofPolitical Economy 53(February): 47–60.

Ingham, Barbara M. 1973. “Ghana Cocoa Farmers—Income-Expenditure Relationships.” Journal ofDevelopment Studies 9(April): 365–77.

IMF (International Monetary Fund). 1998. World Eco-nomic Outlook. Washington, D.C.

International Country Risk Guide 1999. Syracuse,New York: PRS Group.

Kemp, Peter. 1998. “Wishing and Waiting for the OilPrice Recovery.” MEED Saudi Arabia SpecialReport. MEED East Business Weekly June 12:29–35.

Lane, Philip R., and Aaron Tornell. 1995. “PowerConcentration and Growth.” Discussion Paperno.1720. Cambridge: Harvard Institute of Eco-nomic Research.

Léon, Javier, and Raimundo Soto. 1997. “StructuralBreaks and Long-Run Trends in CommodityPrices.” Journal of International Development 9(May–June): 347–66.

Lipsey, Robert E. 1994. “Quality Changes and OtherInfluences on Measures of Export Prices of Manu-factured Goods.” Policy Research Working Pa-per no. 1348. World Bank, Washington, D.C.

Little, I. M. D., Richard N. Cooper, Max W. Corden,and Sarath Rajapatirana. 1993. Boom, Crisis andAdjustment. Oxford: Oxford University Press.

Loayza, Normal V., Luis Serven, and Klaus Schmidt-Hebbel. 1999. “What Drives Savings across theWorld?” In Klaus Schmidt-Hebbel and LuisServen, eds., The Economics of Savings andGrowth. Cambridge: Cambridge UniversityPress.

Maizels, Alfred. 1994. “The Continuing CommodityCrisis of Developing Countries.” World Devel-opment 22(11): 1685–95.

Matsuyama, K. 1992. “Agricultural Productivity,Comparative Advantage, and EconomicGrowth.” Journal of Economic Theory 58: 317.

Mayer, Jorg. 1996. “Learning Sequences and Struc-tural Diversification in Developing Countries.”Journal of Development Studies 33(2): 210–29.

McMahon, Gary. 1997. “The Natural Resource Curse:Myth or Reality?” World Bank, Washington, D.C.

Ogaki, Masao, Jonathan David Ostry, and Carmen M.Reinhart. 1995. “Saving Behavior in Low- andMiddle-Income Developing Countries: A Com-parison.” IMF Staff Working Paper no. 95/3. In-ternational Monetary Fund, Washington, D.C.

Ostry, Jonathan D., and Carmen M. Reinhart. 1992.“Private Savings and Terms of Trade Shocks:Evidence from Developing Countries.” Staff Pa-pers 39 (September): 495–517.

Page, John. 1999. “The Middle East and North Af-rica: An Overview.” World Bank, Washington,D.C.

Paxson, Christian H. 1992. “Consumption and IncomeSeasonality in Thailand.” Research Program inDevelopment Studies, Discussion Paper no. 150.Woodrow Wilson School of Public and Interna-tional Affairs, Princeton, N.J.

Prebisch, Raul. 1950. The Economic Development ofLatin America and its Principal Problems. NewYork: United Nations.

Reinhart, Carmen M., and Peter Wickham. 1994.“Commodity Prices: Cyclical Weakness or Secu-lar Decline?” International Monetary Fund StaffPapers 41(2): 175–213.

Ruppert, Elizabeth. 1996. “Unemployment Insurancein Algeria: Implications for a Labor Market inTransition.” Policy Research Working Paper no.1659. World Bank, Washington, D.C.

Sachs, Jeffrey D. 1981. “The Current Account andMacroeconomic Adjustment in the 1970s.”Brookings Papers on Economic Activity 1: 201–68.

Sachs, Jeffrey D., and Andrew M. Warner. 1999.“Natural Resource Intensity and EconomicGrowth.” In Jörg Mayer, Brian Chambers, andAyisha Farooq, eds., Development Policies inNatural Resource Economies. United NationsConference on Trade and Development.Cheltenham, U.K.; Northhampton, Mass.: Ed-ward Elgar in association with UNCTAD.

Sapsford, David, and V. N. Balasubramanyam. 1994.“The Long-Run Behavior or the Relative Price ofPrimary Commodities: Statistical Evidence andPolicy Implications.” World Development 22(11):1737–45.

Singer, H. W. 1950. “The Distribution of Gains be-tween Investing and Borrowing Countries.” AERPapers and Proceedings 40: 473–85.


132


Spraos, John. 1980. “The Statistical Debate on the NetBarter Terms of Trade between Primary Com-modities and Manufactures.” Economic Journal90 (357): 107–28.

Stauffer, Thomas. 1999. “Macroeconomic Manage-ment of Windfalls.” Lecture given at the WorldBank, Washington, D.C.

Svensson, Lars E. O., and Assaf Razin. 1983. “TheTerms of Trade and the Current Acount: TheHarberger-Laursen-Metzler Effect.” Journal ofPolitical Economy 91 (February): 97–125.

UNCTAD (United Nations Conference on Trade andDevelopment). 1994. Handbook of InternationalTrade and Development Statistics. N.Y.: UnitedNations.

Varangis, Panos, Takamasa Akiyama, and DonaldMitchell. 1995. “Managing Commodity Booms—and Busts.” In Directions in Development. Wash-ington, D.C.: World Bank.

Wetzel, Deborah L. 1992. “Temporary Trade Shocksin Developing Countries: A Case Study of theGhanaian Cocoa Boom.” Centre for the Study ofAfrican Economies, Oxford, U.K.

World Bank. 1994. Global Economic Prospects. Wash-ington, D.C.

World Bank. 1999. World Development Indicators.Washington, D.C.

133

East Asia and Pacific

Recent developments

GROWTH FOR MOST OF THE EAST ASIA AND

Pacific region in 1999, as anticipatedearlier this year, will revive from the

deep crisis-induced recessions of 1998. Theturnaround has been much stronger than ini-tially anticipated. Whereas the earlier forecastfor the five crisis economies1 saw growth re-bounding to 0.2 percent in 1999, comparedwith a 7.9 percent decline in 1998, this year’sperformance now looks to be in the vicinity

Appendix 1Regional Economic Prospects

of 4.5 percent (table A1.1). Growth has beenuneven across the five economies, with the Re-public of Korea noticeably leading the group.Under current projections, Korea and the Phil-ippines will return to their precrisis GDP lev-els this year, followed by Malaysia in 2000,and Thailand, just barely, in 2001. In contrast,Indonesia’s output in 2001 could still be some7 percent lower than precrisis levels.

Financial markets have improved mark-edly over the past year (figure A1.1). Curren-cies have appreciated between 10 and20 percent since the low point of late 1998,

Figure A1.1 Stock index, exchange rate, inflation, and interest rate for East Asia-4

Stock market (right axis)

Exchange rate (left axis)

Call rates (right axis)

Inflation (left axis)

100

Jan. 1997 May 1997 Sept. 1997 Jan. 1998 May 1998 Sept. 1998 Jan. 1999 May 1999 Sept. 1999

20

15

10

5

0

90

80

70

60

50

40

30

20

10

0

Note: East Asia-4 includes Malaysia, the Philippines, the Republic of Korea, and Thailand. Averages are calculated as simple arithmetic means for the four countries. Inflation is the 3-month moving average of the year-on-year inflation.Source: Datastream.

134


with the exception of Indonesia where theappreciation has been over 100 percent.2

Stronger currencies have been abetted by in-creasing reserve levels and declining levels ofshort-term debt. Prices have been subduedthroughout 1999, providing scope for reducedinterest rates and a related fall in public andprivate debt servicing.

China is confronting several economicproblems. The uncertainties surrounding re-structuring of state-owned enterprises (SOEs)and of employment prospects, coupled withalmost two years of monthly deflation, havedampened consumer demand. This has led toa significant increase in unsold goods and hasexacerbated the costs of corporate restructur-ing. Government efforts to spur demandthrough public investment programs werepartially successful as temporary measures,and will be pursued to maintain growth. Thegovernment has taken a variety of measuresto spur demand, but early evidence indicatesthat these measures have led to increased li-quidity in equity markets, without markedlyaffecting personal consumption. Thus, growthin China is expected to slow compared with1998, from 7.8 percent to 6.5 percent.

After lagging through the first half of 1999a combination of factors are buttressing Chi-nese exports, including renewed growth in EastAsia and continued import growth in theUnited States. China’s improving internationalcompetitiveness, spurred by strengthening EastAsian currencies, domestic deflation, and ex-port-linked tax rebates, has also been a fac-tor. Recorded imports have risen substantiallythis year, though this is mainly attributable toa crackdown on smuggling. The trade balanceis likely to narrow, but will remain positiveon aggregate.

Among the newly industrializing econo-mies (NIEs), both Singapore and Taiwan(China) have benefited from the improvingregional economy and the global electronicsboom. Because of its stronger integration withthe crisis economies, Singapore was moredeeply affected than Taiwan (China) by theregional downturn. Hong Kong’s (SAR, China)adjustment to the crisis has been more pain-ful. The currency peg has forced down assetprices and real wages as Hong Kong attemptsto remain competitive, and has depressed in-vestment through high interest rates. But re-tail sales are finally picking up, tourist arrivals

Table A1.1 East Asia and Pacific forecast summary(percent per year)

Baseline forecast

Growth rates/ratios 1989–98 1997 1998 1999 2000 2001 1999–2008

Real GDP growth 7.5 6.6 0.1 5.5 6.2 6.2 6.2 Consumption per capita 5.2 3.0 –4.0 4.2 4.3 4.7 5.1 GDP per capita 6.1 5.4 –1.1 4.3 5.1 5.1 5.2

Population 1.3 1.2 1.1 1.1 1.0 1.0 0.9Median inflationa 6.3 3.8 8.7 4.5 3.7 3.1 4.5Gross domestic investment/GDP 35.0 36.2 31.0 32.6 33.9 34.6 35.3Central government budget deficit/GDP –0.6 –0.3 –2.2 –2.6 –3.0 –2.8 –2.4Export volumeb 11.9 17.6 7.0 7.9 8.1 7.8 8.0Current account/GDP –0.5 0.2 5.6 3.5 2.8 2.1 1.1

Memo itemsGDP of region, excluding China 5.7 4.4 –7.6 4.3 5.3 5.1 5.2GDP of East Asia Crisis-5 countriesc 5.7 4.5 –7.9 4.4 5.3 5.1 5.2

a. GDP deflator.b. Goods and nonfactor services.c. Indonesia, Malaysia, the Philippines, the Republic of Korea, and Thailand.Source: World Bank Development Prospects Group, November 1999.

135

are increasing, and renewed Chinese exportswill revive trade-related activities.

Near-term outlookFor the crisis countries improving macroeco-nomic conditions abetted by fiscal measuresshould set the stage for a further consolida-tion of growth, averaging 5 percent or morein 2000 and 2001. Inflation, which had beenaveraging 4.5–6 percent before the crisis,peaked in most of the crisis countries at lessthan 11 percent (except for Indonesia). Theaverage year-on-year inflation rate is runningat less than 2 percent in the East Asia-4 (Ma-laysia, the Philippines, Korea, and Thailand)and is in single digits in Indonesia after peak-ing at 80 percent at the end of 1998. Interestrates have fallen continuously since reachinghighs in 1998. Overnight call rates in Korea,Malaysia, and Thailand are now well belowprecrisis levels.

The initial catalyst for growth was a re-vival in exports—led by electronics—drivenby better than anticipated import growth inthe United States and Japan, and intraregionalmultipliers. Low interest rates, rising assetprices, the end of the inventory cycle, and animproving near-term employment outlook,especially in Korea, have helped to broadengrowth. Current account balances are likelyto remain comfortably positive, though nar-rowing, as imports grow in tandem with con-sumption, production, and the end of theinventory cycle. Growth in the near term willbe buttressed by strengthening consumer de-mand and inventory replenishment, with in-vestment kicking in further down the road ascapacity utilization improves.

On the downside, there could be a slow-down in the electronics sector in the secondhalf of 1999 and perhaps into 2000. Equityprices have slumped more recently and theywill have to rely on renewed growth in corpo-rate profits rather than on liquidity. Interestrates have most likely reached a floor, andinvestment could continue to lag if banks’balance sheets fail to improve. Progress in thelatter area has been uneven to date, and non-

performing loan (NPL) levels remain near theirpeaks. In Indonesia NPLs are estimated torange from 60 to 80 percent of outstandingloans and are close to 50 percent in Thailand.In Korea and Malaysia, they are a much moremanageable 10–15 percent. Failure to achievemore substantial progress on restructuringcould lead to renewed financial instability andcould dampen near-term growth prospects (seechapter 3).

Expansionary fiscal policies are expectedto fade out, though the remaining off-budgetdeficit will take longer to eliminate, and couldaccumulate further. The rapidly growing publicdebt in East Asia could pose a threat to recov-ery down the road, but historically low debtlevels allow for a degree of flexibility. None-theless, Indonesia (with a history of high debtlevels, political instability, and corruption), andto a lesser extent Thailand (with a persistentlyhigh level of NPLs), are in significantly moreprecarious positions.

Long-term prospectsDespite the current rapid rate of growth inthe crisis countries long-term prospects forEast Asia have not been revised since fore-casts of one year ago. Lack of upward revi-sion reflects the ongoing difficulties inIndonesia and, importantly, prospective indus-trial restructuring in China. In particular, thecost of restructuring the Indonesian bankbalance sheets has risen dramatically, andtroubles with Chinese commercial banks andtheir links with ailing SOEs have emergedmore sharply this year. Inward investment inChina has also been negatively influenced bythe well-publicized collapse of several finan-cial institutions. Despite the weak perfor-mance of the domestic economy, China seemsdetermined to pursue its domestic reforms andto join the World Trade Organization (WTO).WTO membership is expected to benefitChina, notwithstanding intensified competi-tive pressures on protected sectors. Beyondthe one-time efficiency gains from reducingtrade barriers, China could anticipate a boostin foreign direct investment and improved

R E G I O N A L E C O N O M I C P R O S P E C T S

136


access to foreign markets. For the region asa whole, long-term growth is expected to av-erage 6.3 percent between 2002 and 2008,with the five crisis countries growing at5.3 percent.

Success with regard to human capitalgrowth will bolster longer term productivitygains in East Asia. Educational achievementsover the last decades have been impressive,with primary school enrollment nearly univer-sal and secondary school enrollment at highrates (for example, 50–60 percent in Thailandand Indonesia, and over 70 percent in China).Evidence from the hardest hit crisis countriessuggests that these achievements were not re-versed during the crisis. Governments madeefforts to maintain expenditures on education(relative to GDP and total expenditures), andhouseholds adjusted consumption patterns toensure continued school enrollment (see chap-ter 2). With the exception of Korea, which wasa high-income country before the crisis, coun-tries in East Asia—with per capita incomesranging from $800 to $4,500, and an averageclose to $1,000—still have significant poten-tial for moving toward NIE or industrial coun-try income levels.

Europe and Central Asia

Recent developments

Economic output contracted by 0.2 percentin 1998 for the Europe and Central Asia

(ECA) region as a whole, largely reflectingRussia’s liquidity crisis in August 1998 andsubsequent regional contagion. The ensuingeconomic downward spiral in Russia rever-berated throughout the Commonwealth of In-dependent States (CIS), as these economies inparticular retain strong trade links with theRussian market. Trade-finance and paymentssystems arrangements were interrupted andtrade declined sharply as Russian demandcollapsed (figure A1.2a) Central and EasternEuropean countries (CEECs) exports and bal-ance of payments deteriorated as well, as slug-gish growth in Western Europe translated intoweak external demand for the more diversi-fied regional exporters, dampening the inter-nal dynamics of the region (figure A1.2b).

The region’s weak growth performanceis likely to improve during 1999. The firmingof oil prices is improving prospects for thepetroleum exporting countries of Russia andthe Caspian Sea basin. While oil importers,

Figure A1.2a Russian imports and partner exportsPercent year-on-year of U.S. dollars trade

30

10

-10

-30

-40

-50

-60

20

0

-20

Mar. 1997 Sept. 1997 Mar. 1998 Sept. 1998 Mar. 1999

Note: Figures are three-month moving averages.Source: Datastream.

Russian imports

Turkish imports

Central and Eastern European imports

Other Commonwealth of Independent States imports

137

such as Turkey and the Central Europeancountries, are experiencing some terms oftrade losses, many of these countries are ben-efiting from the incipient strengthening ofdemand in Western Europe. The end of thewar in the Federal Republic of Yugoslavia(Serbia/Montenegro) has allowed reconstruc-tion efforts to begin and cross-border tradeto resume, albeit constrained by the need torebuild the transportation infrastructure.Russia has made some progress on debt re-structuring and on gaining another round offunding from the International MonetaryFund (IMF). Inflationary pressures have beenreduced significantly, such that monthly in-flation has slowed to 1.4 percent in Septem-ber from 11 percent in December 1998. Bondspreads throughout the region in secondarymarkets have generally narrowed, which is areflection of improving international finan-cial market sentiment. For example, Poland’sspread returned to early 1998 precrisis lev-els. The most striking exception to this trendis Ukraine, where the economy remains vul-nerable and Eurobond spreads are still el-evated at close to 8,300 basis points inmid-November 1999, up from about 3,000

points in June 1999. Spreads for Russia re-mained above 2,300 basis points in Novem-ber 1999, down from 5,200 basis points inApril 1999.

Estimates for growth in the ECA regionfor 1999 have been revised upward, from the1.5 percent decline expected in Global De-velopment Finance (GDF) 1999 to plus 0.3percent. Russian GDP is now forecast to in-crease by close to 1.0 percent (compared withthe 5 percent contraction anticipated in GDF1999), due largely to higher than expected oilprices and increased production for importsubstitution. While GDP and industrial pro-duction contracted in the first quarter of 1999,both posted gains in the second and thirdquarters. With positive spillover effects an-ticipated in neighboring countries, the fore-cast for growth in the Commonwealth ofIndependent States (CIS) has been revised fromminus 5.5 percent to a 0.7 percent advance.The most notable exception to the region’sgenerally lackluster output is Turkmenistan,where double-digit growth is expected as alarge natural gas project comes onstream. Incontrast to the CIS, expectations for growthin 1999 in the Central Eastern European coun-

Figure A1.2b German imports and Central and Eastern European exportsPercentage change year-on-year

15

10

5

0

–5

–10Mar. 1997 Sept. 1997 Mar. 1998 Sept. 1998 Mar. 1999

Note: Figures are three-month moving averages in U.S. dollars.Source: Datastream.

German imports

Central and Eastern European exports


138


tries have been revised downward, from 2.3percent to 1 percent, reflecting weaker thananticipated growth in Europe and higherprices for these countries’ energy imports.Also, the war in Kosovo has greatly reducedeconomic prospects in the Balkans. Not onlymust the region bear the direct consequencesof the destruction and disruption in the the-ater of conflict, but it must also absorb theimpact of the movement of refugees intoneighboring countries, the disruption of re-gional and transit trade, and a sharp fall intourism revenues.

Near-term outlookAccelerating world trade and stabilizing com-modity prices should contribute to a strongerrecovery in the ECA region in 2000, and realGDP is expected to increase by 2.5 percent. Arecent history of high real investment growth(in double digits for some countries since themid-1990s), boosted by strong FDI inflows(albeit tied in part to one-off privatizationsales), should provide additional impetus forfuture growth in the CEECs and the Balticcountries. Greater integration with the Euro-pean Union, as the CEECs progress at varying

stages toward EU membership, will also im-prove growth prospects. Average national pro-duction growth for the CEECs is forecast at3.2 percent in 2000. Output in most CIS coun-tries is projected to recover more gradually,owing to expected slow growth in Russia,where the aggressive implementation of struc-tural reforms is likely to await the results ofthe June 2000 presidential elections. Neverthe-less, the depreciation of the ruble and of otherCIS countries’ currencies is expected to boosttheir exports to a degree which, combined withmore positive external demand and sustainedhigher oil prices, should contribute to morepositive growth of 1.3 percent in 2000 (tableA1.2).

Long-term prospectsThe long-term growth forecast (2002–2008)for the ECA region has been reduced from 5percent per annum to 4 percent. This changeis due largely to a downward revision to theprojections for Russia, which continues tounderperform relative to its underlying poten-tial. Despite stabilizing more quickly than an-ticipated in 1999 after the August 1998financial collapse, the fundamentals required

Table A1.2 Europe and Central Asia forecast summary(percent per year)

Baseline forecast


Real GDP growth –3.5 2.7 –0.2 0.3 2.5 3.3 3.4Consumption per capita –2.1 3.4 1.0 1.4 3.0 3.1 3.3GDP per capita –3.8 2.6 –0.4 1.0 3.2 3.1 3.4

Population 0.3 0.1 0.1 –0.7 –0.7 0.2 0.1Median inflationa 37.0 50.0 12.3 8.4 7.4 7.5 8.5Gross domestic investment/GDP 27.2 24.4 23.0 22.4 22.5 22.9 23.5Central government budget deficit/GDP –7.3 –5.5 –5.3 –5.2 –4.3 –4.0 –4.2Export volumeb 0.5 11.5 6.1 –0.3 4.9 6.6 5.6Current account/GDP 0.7 –0.5 –1.1 –1.0 –1.5 –1.7 –1.5

Memo itemsGDP of middle-income Western Europe 4.2 7.5 2.9 –2.2 4.6 4.3 4.3GDP of Central and Eastern Europe –0.8 2.6 2.1 1.0 3.2 4.3 4.1GDP of CIS states –6.3 1.4 –2.7 0.7 1.3 2.3 2.6

a. GDP deflator.b. Goods and nonfactor services.Source: World Bank Development Prospects Group, November 1999.

139

to support longer term growth have deterio-rated. Savings and investment rates have fallen;the financial sector and state enterprises arein need of significant restructuring; the levelof FDI remains constrained; and capital out-flows continue. Population growth is projectedto continue to fall by 0.3 percent per year, withthe labor force contracting by 0.5 percent.While fiscal and monetary policies are likelyto improve in the coming years, the degree ofsupport for implementing far-reaching reformsto address the many imbalances in theeconomy is uncertain.

Risks to the forecast are mainly on thedownside and are primarily linked, within theregion, to achieving a sustainable recovery inRussia, where the situation remains fragile.A further downturn in Russia would havesignificant negative repercussions in the restof the CIS and in Turkey, in particular, asoccurred in the aftermath of the 1998 crisis.The economic situation in both Ukraine andRomania is tenuous, reflecting high debt-servicing payments. Turkey also faces signifi-cant economic challenges, as its economy issuffering from a sharp slowdown in growth,a growing budget deficit, and high short-termdebt turnover requirements. The disruption ofthe August earthquake (estimated to cost be-tween 2.5 percent and 5 percent of GDP) hascompounded the situation. Throughout the re-gion, slippage in reforms and failure to ad-dress significant financial and enterprise sectorproblems represent an ongoing risk. ECA’s re-lationships with the EU pose both opportuni-ties and challenges. A slowdown in Europecould weaken growth in ECA, especially inthe CEECs. In the longer term, however, fa-vorable prospects for several countries are tiedto EU enlargement.


Recent developments

The initial impact of the East Asian crisison Latin American countries was not

large. Most economies were at the peak of their

business cycles. Nonetheless, the adverse im-pact of the East Asian crisis on the global en-vironment—through falling world exportprices and volumes, and reduced capital flowsto developing countries—eventually took itstoll on Latin American countries. In 1998 theregion’s terms of trade fell by about 4 percent(a loss equivalent to 0.6 percent of GDP) andexport volume growth slowed from 11.5 per-cent in 1997 to 5.6 percent, widening the re-gion’s current account deficit by $22 billion.

Mexico, which benefited from strong U.S.import-demand growth and which had a flex-ible exchange rate, was the least affected bythe slowdown in world trade. ExcludingMexico, the region’s export volume growthslowed from 9.5 percent in 1997 to 3.3 per-cent in 1998, due to worsening competitive-ness. The countries with targeted exchangerates saw their real effective exchange ratesrise by an average 17 percent above their1990–96 levels (before the East Asian crisis),compared with a decline of 25 percent in theEast Asian crisis countries. Capital flows frominternational markets dried up in the wake ofthe Russian default of August 1998, causinggross new flows to fall 25 percent in 1998compared to 1997 levels. This precipitated amassive credit squeeze and a sharp reductionin current account deficits in several countries.The contagion effect from the Russian crisismade it impossible for countries to finance thepart of the trade shock that could be treatedas temporary. Even Chile, a country with solidcredit ratings, experienced speculative attacksand was eventually forced to float its currency.

The combination of a deterioration in theexternal environment, high initial debt levels,a large dependence on foreign savings, andtight monetary policies—aimed at preservingexisting exchange rate regimes during an elec-tion year for many large countries—causedregional GDP growth to slow from 5.4 per-cent in 1997 to 2.1 percent in 1998. By thefourth quarter of 1998 Argentina, Brazil,Chile, Colombia, Ecuador, Peru, and Venezu-ela were all experiencing recessions. In Brazil,Colombia, Ecuador, and Venezuela high do-


140


mestic interest rates in the second half of 1998added to government debt service and, along-side lower tax revenues in the wake of theeconomic slowdown, widened fiscal deficitsand put pressure on exchange rates. With theexception of Venezuela, these countries wereeventually forced to abandon targeting theirexchange rate, with the Brazilian and Ecua-dorian devaluations being the most acute, af-ter massive losses of reserves. In the case ofColombia (and Chile), the decision to floatcurrencies was not accompanied by large de-valuation because reserves were still at rea-sonable levels. As in other cases (Russia andTurkey), the East Asian crisis may not havebeen the fundamental cause of the downturnthat ensued, but it was a contributing factorand triggered a reaction to more deep-seatedproblems. By and large, countries where poli-cies are sounder (Chile and perhaps Mexico)avoided the worst.

Brazil’s devaluation, uncertainties in therun-up to elections, and tighter U.S. monetarypolicy helped spread the economic downturnin the region in 1999. The Brazilian devalua-tion worsened the external environment formany countries in the region, though to a muchlesser extent than the Russian crisis. Prices ofkey commodities exported by Brazil (coffee,soybeans, and sugar) fell sharply in the firsthalf of the year at the same time that Brazil-ian import demand collapsed, significantlyreducing export revenues in a number of coun-tries. Forthcoming elections in Argentina andChile, coupled with increasing civil resistanceto further fiscal tightening in some countries(Colombia, Ecuador, and Venezuela), led togreater uncertainty on the part of domesticinvestors. The tightening of U.S. monetarypolicy in mid-year, in conjunction with thedebate over increasing the burden-sharing ofinternational bond holders, contributed to areduction in the supply of private capital avail-able to the region from international markets.Ecuador’s prospective default on its Bradybonds and calls for debt restructuring in Ven-ezuela added to the uncertainty that privateinvestors faced. The net result was the spread

of the economic downturn within the region,with GDP in 1999 falling in nine countries,compared with only four countries in 1998.However, the expected economic outcome forBrazil will likely be much better than envi-sioned six months ago (no growth now ver-sus –3.9 percent then) and should keep thedecline in the region’s GDP to about –0.6 per-cent. The region’s current account deficit nar-rowed by about $33 billion in 1999, due to acontraction in imports of 2.6 percent.

Growth in per capita income of the Car-ibbean countries averaged 2.2 percent in 1999but economic performance varied widely, andgrowth prospects are expected to remain di-verse. Per capita GDP growth in 1999 aver-aged 3–4 percent in Barbados, DominicanRepublic, and Trinidad & Tobago; 2–3 per-cent in the Leeward and Windward islands;0–1 percent in Guyana and Haiti; but wasnegative in Jamaica for the third consecutiveyear. Movements in the terms of trade explainpart of this diversity in growth performance—the faster growing economies rely more ontourism and oil for export revenues while mostof the others depend heavily on exports ofbananas and sugar, whose prices on worldmarkets fell 12 and 30 percent, respectively.

Near-term outlookEconomic recovery in 2000–2001 is likely tobe gradual, as further fiscal tightening is nec-essary in many countries and reform fatigueis becoming widespread. Recovery should be-gin by the fourth quarter and into 2000 formany of the countries experiencing recessionin 1999 (figure A1.3). External factors behindthe rebound include: an acceleration of worldtrade; the stabilization of commodity pricesand the rise in some (for example, oil andmetals); a slow recovery of capital flows;greater exchange rate flexibility in many coun-tries; and less external debt amortization in2000. Domestically, the ending of destockingfacilitated by improved financial conditions(lower domestic interest rates, higher stockmarket valuations) should support a stabili-zation and consequent upturn in production.

141

However, there are a number of reasonswhy the economic recovery in 2000 is expectedto be modest compared with 1996 (post-Mexico crisis). First, fiscal tightening is re-quired in a number of countries to sustaininvestor confidence and help exchange ratestability (Argentina, Colombia, Ecuador, Ven-ezuela, and possibly Brazil). Implementationof further fiscal tightening could encounterresistance, especially since unemployment ratesare high. Second, political uncertainties arelikely to persist into 2000 as Mexico and Peruelect new presidents and many new adminis-trations lack clear majorities in their respec-tive congresses (Brazil, Colombia, andEcuador). This uncertainty is likely to resultin investor wariness. In addition, the expectedslowdown in the United States in 2000–2001could restrain Latin America’s export growth.The combination of these factors is likely tolead to a moderate economic recovery, withthe region’s GDP growing by 2.7 percent in2000 before accelerating toward 3.5 percentin 2001. The current account deficit duringthe next two years is expected to rise to a rangeof $60 to $70 billion from the 1999 level of$56 billion.

Near-term uncertainties. If private capi-tal inflows remain weak into 2000, or capitaloutflows increase sharply in response to po-litical developments, pressures on some cur-rencies could increase. Although the baselinescenario assumes that Argentina will imple-ment fiscal adjustment and obtain sufficientexternal support for financing its fiscal deficitand external payment obligations, risks remainthat the process will not be smooth. Tightermonetary conditions in the industrial coun-tries are likely to keep Argentina’s cost of capi-tal relatively high and the volume of capitalinflows modest. Venezuela has persisted inmaintaining a crawling exchange rate band,even though the country’s real effective ex-change rate is now 50 percent higher than its1990–96 average. Although Venezuela hasbeen able to fend off speculative exchange rateattacks in the past, and the rise in oil priceswill help with potential future episodes, un-certainty on the policy front could lead tocapital flight on a scale beyond the adminis-tration’s ability to counteract. Both Chileand Colombia, with a history of much bettermacroeconomic management were forced tomove toward an exchange rate float in 1999.


Figure A1.3 Latin America and the Caribbean quarterly GDP

Mexican devaluation Asian crisis

Russian crisis

Projection8

4

0

–4Q2

1992Q2

1993Q2

1994Q2

1995Q2

1996Q2

1997Q2

1998Q2

1999Q2

2000


Percent year-on-year

142


Long-term prospectsOutput growth for the region in the long term(2002-2008) is now projected to average 4.2percent (2.8 percent in per capita terms),which is a reduction of 0.2 percentage pointscompared to Global Economic Prospects1998/99 (table A1.3). The lower forecast isbased on several factors which, although char-acteristic of the region for many years, havebecome more evident during 1999. Nationalsaving rates in several large countries have notrisen in the 1990s and most continue to relyon foreign savings to accommodate increasesin investment. High external indebtedness hasincreased reliance on international capitalmarkets to finance debt repayments, and thereis increasing evidence of reform fatigue in anumber of countries. In Brazil the large in-crease in domestic debt over 1997–99 and acontinued low saving rate will require a strongand sustained fiscal adjustment, which hasbeen encountering public resistance. A simi-lar situation exists in Argentina, where sav-ings remain insufficient for sustainablegrowth. GDP per capita growth for the Car-ibbean countries is likely to average 2.5 per-cent a year reflecting the difficult structuraltransition that these countries will have to un-dergo in the face of increasing global compe-tition of commodity exports to traditionalmarkets.

Despite these uncertainties, longer termgrowth prospects—contrasted with the 1980sand 1990s—are favorable, as efficiency gainsfrom past reforms are realized. Growth in to-tal factor productivity, regionwide, is likely tocontinue its upward trend as Brazil, the last ofthe large economies in the region to embarkon liberalization, overcomes current difficul-ties. The privatization of large state enterprisesin the water, electricity, transportation, andtelecommunications sectors should begin tobear fruit into the first decade of the 2000s.Privatization, combined with the increasingmarket power of the Southern Cone countriesthrough Mercosur, should encourage FDI in-flows, and the nature of FDI should shift fromacquiring existing capital stock throughprivatization to investment in new capacity inthe services and manufacturing sectors.


Recent developments

The historically low oil prices of 1998 de-pressed growth in the Middle East and

North Africa (MENA) region (see chapter 4).Current account imbalances grew and govern-ment revenues came under severe pressure (fig-ure A1.4). The low absorption oil exporters(Bahrain, Oman, and Saudi Arabia) utilized

Table A1.3 Latin America and the Caribbean forecast summary(percent per year)

Baseline forecast


Real GDP growth 2.9 5.4 2.1 –0.6 2.7 3.5 3.5Consumption per capita 1.6 4.0 1.5 –3.6 0.2 1.4 1.3GDP per capita 1.1 3.7 0.5 –2.2 1.1 2.0 2.0

Population 1.8 1.7 1.6 1.6 1.6 1.5 1.4Median inflationa 17.0 8.5 7.9 8.3 7.9 6.8 6.3Gross domestic investment/GDP 22.1 24.2 24.0 22.8 23.5 24.1 24.6Central government budget deficit/GDP –2.8 –1.8 –3.7 –2.8 –1.6 –1.4 –1.3Export volumeb 8.0 8.9 5.6 3.8 6.9 6.0 6.5Current account/GDP –2.4 –3.3 –4.5 –3.1 –3.3 –3.5 –3.4


143

foreign reserves and external portfolios to fi-nance the deterioration of fiscal deficits andtrade balances in the short term. Other oil ex-porters, such as Algeria, the Islamic Republicof Iran, and the Republic of Yemen had simi-lar pressures but faced tighter financing con-straints. This led to a different adjustment path,including deeper expenditure cuts, exchangerate devaluation, rescheduling of external debt,and, in the case of the Islamic Republic of Iran,import compression and liquidity stress as for-eign reserves declined and import cover fell.The net oil importers benefited from lower oilprices, although some countries experienceda fall in workers’ remittance receipts from theoil exporting countries. The low level of re-gional integration into global capital marketsensured some protection against the worsteffects of the East Asian crisis.

GDP growth in the region averaged 2percent in 1999, 1.4 percentage points higherthan anticipated earlier in the year, due tostronger than expected oil prices and higherdomestic investment in the diversified export-ers. The terms of trade of oil exporters roseby 35 percent (equivalent to nearly 9 percentof GDP), boosting government revenues and

contributing to an easing of fiscal constraints.Nevertheless, the restriction of export volumesin order to support oil prices (OPEC cut pro-duction by 2 million barrels a day in April1999) limited the real output recovery in theoil exporting countries.

GDP growth in the more diversified ex-porters (the Arab Republic of Egypt, Jordan,Morocco, the Syrian Arab Republic, and Tu-nisia) slowed to 3 percent in 1999, from 3.5percent in 1998. The decline in activity in oil-exporting countries, as well as slower growthin Europe, put downward pressure on receiptsfrom worker remittances. Those countries withexchange rates pegged to the dollar experiencedslower export growth, in part as the value ofthe dollar increased and as the competitivenessof East Asian exporters in textiles, clothing,and machinery improved. Drought conditionsin several countries (Jordan, Morocco, andSyria) led to a decline in agricultural incomes,as well as to further pressures in urban labormarkets and higher food import bills.

Near-term outlookGDP growth in the region should accelerate to3.2 percent in 2000. The recovery in oil prices


Figure A1.4 Terms of trade, current account balance, and fiscal deficits in the Middle East and North Africa, 1990–98Percentage change

30

20

10

0

–10

–30

–20

Share of GDP

4

2

0

–10

–8

–6

–4

–2

1990 1991 1992 1993 1994 1995 1996 1997 1998

Source: World Bank.

Terms of trade (left axis)

Current account balance (right axis)

Fiscal deficit (right axis)

144


in mid-1999, following the implementation ofthe OPEC oil quotas and the high level of com-pliance thus far, points to a stabilization ofeconomic activity in oil exporting countries,and GDP growth for this group is expected torange between 2.5–3 percent in 2000–2001.Output growth will continue to be constrainedby quotas on oil production and the need tocontain fiscal deficits and public debt levels.Recent increases in sales tax and utility chargesand cuts in subsidies in the oil exporting coun-tries are expected to have modest inflationaryimpact. But central banks in the low absorp-tion oil exporters are likely to maintain tightmonetary policies to help keep inflation lowand maintain exchange rate stability. Inflationis likely to be more of a problem in the Repub-lic of Yemen and the Islamic Republic of Iran.The diversified exporters are likely to grow by4.7 percent in 2000, as growth in partner im-ports improves, the renewed drive for reformin several countries encourages domestic in-vestment, the effects of drought conditionsfade, and tourism receipts rise (table A1.4).

Long-term prospectsDespite an anticipated near-term revival ofactivity, forecasts for long-term output growth

(2002–2008) have not been revised from ear-lier projections at 3.7 percent. Oil prices arelikely to remain under downward pressure inthe longer term (see chapter 1), which will con-strain oil exporting countries’ fiscal and for-eign exchange revenues. But the diversifiedexporters are expected to maintain a growthrate of 4.6 percent, supported by continuedstructural reforms and the linkages that aredeveloping as a result of the Euro-Mediterra-nean agreements. Several North African coun-tries are implementing restructuring programsin line with their association agreements toboost efficiency in preparation for open com-petition from European firms. Privatizationprograms in Egypt, Jordan, Morocco, andTunisia, in conjunction with a deepening ofcapital markets, are attracting foreign invest-ment. Programs are in place to transfer publicenterprises to the private sector, and newprojects in minerals and infrastructure arebeing implemented under private contracts oron a build-own-operate-transfer basis. Largepublic enterprises such as utilities and telecom-munications are being sold or corporatizedprior to sale. A beneficial by-product of di-vestment of public enterprises will be the low-ering of fiscal deficits and public debt.

Table A1.4 Middle East and North Africa forecast summary(percent per year)

Baseline forecast


Real GDP growth 3.0 3.7 3.2 2.0 3.2 3.5 3.4 Consumption per capita –0.1 0.3 1.8 0.8 1.3 1.3 1.4 GDP per capita 0.6 1.5 1.2 –0.1 1.1 1.5 1.4

Population 2.4 2.1 2.1 2.1 2.1 2.1 2.0Median inflationa 8.7 4.9 3.3 5.3 4.9 4.9 4.5Gross domestic investment/GDP 22.8 23.9 23.7 23.8 23.9 24.1 24.7Central government budget deficit/GDP –5.2 –2.6 –4.5 –3.6 –3.3 –2.9 –2.2Export volumeb 4.5 5.0 2.6 1.4 3.4 4.0 3.9Current account/GDP –2.2 3.8 –1.5 4.2 4.5 3.6 2.2

Memo itemsGDP of oil dominant economies 3.6 2.9 1.6 1.1 2.2 3.0 2.9GDP of diversified exporters 3.7 3.1 3.5 3.0 4.7 4.4 4.4


145

Additionally, the peace process in the MiddleEast, which appears to be bearing fruit, maycontribute to a lowering of risk perceptions,particularly in the Mashreq countries, and mayprovide an opportunity for increased dyna-mism in trade and capital flows.

South Asia

Recent developments

In 1999, for a second consecutive year, SouthAsia was the fastest growing developing re-

gion, averaging 5.4 percent GDP growth.Negative external factors that slowed growthelsewhere had a smaller effect on the region.Among domestic factors, the most positive wasa favorable Indian monsoon, which led toimpressive agricultural sector performance,while among the smaller economies domesticfactors led to lower-than-trend growth. Theaftereffects of 1998 floods in Bangladesh andPakistan’s stabilization program dampenedgrowth in those two countries.

Relative independence from external in-fluences is a constant structural feature of theIndian economy, and India experienced onlymoderate spill-over effects from the series ofcrises in 1997–99. The country’s main exportmarkets are in Europe and the United States,so the crisis did not greatly reduce demand

for India’s exports. As well, financial liberal-ization in India has been limited, and the coun-try was not subject to capital account reversals.However, several other countries of the regionwere adversely affected. Bangladesh, Pakistan,and Sri Lanka were hit by weak external mar-kets and falling commodity prices, particularlyfor tea, rubber, and cotton.

Pakistan’s economy decelerated sharply in1998 and 1999 due to a wide range of causes.Many of these causes were related to deeperseated structural imbalances, especially tied tofiscal performance and shortfalls in privatesavings. The result was an emerging balanceof payments crisis, compounded by the sanc-tions that followed nuclear testing in 1998.Bangladesh suffered an economic downturnin 1998, due to extensive flooding which de-stroyed food crops, cut export volume, in-creased imports of food, and damagedinfrastructure. Expenditure throughout theeconomy was affected as rural incomes werecut. The East Asian crisis may have weakeneddemand for ready-to-wear garments manufac-tured in Bangladesh. GDP growth droppedfrom 5.7 percent in 1998 to 3.5 percent in1999. Sri Lanka, the most open of the SouthAsian economies, exports 37 percent of itsGDP. It depends on exports to provide a mar-ket for its manufactures and on strong worlddemand to sustain prices for tea, rubber, and


Table A1.5 South Asia forecast summary(percent per year)

Baseline forecast


Real GDP growth 5.5 4.7 5.1 5.4 5.5 5.3 5.2Consumption per capita 3.3 2.1 3.1 3.2 3.3 3.5 3.6GDP per capita 3.5 3.0 3.3 3.5 3.6 3.6 3.7

Population 1.9 1.8 1.8 1.9 1.8 1.7 1.5Median inflationa 9.5 9.1 7.6 7.2 5.8 5.2 5.3Gross domestic investment/GDP 22.0 22.9 23.2 23.3 23.5 23.6 23.5Central government budget deficit/GDP –6.3 –4.4 –3.1 –3.0 –2.9 –2.8 –3.1Export volumeb 9.8 2.6 1.9 6.0 7.7 7.8 7.7Current account/GDP –1.9 –1.2 –1.8 –2.0 –2.1 –2.3 –2.5


146


other materials. The slowdown in world de-mand was important in keeping tea and rub-ber prices low and reducing export receipts(table A1.5).

Near-term outlookLooking ahead, several positive factors willhelp India sustain growth in 2000 and 2001.These include a high level of private sectorbusiness confidence and a commitment in prin-ciple by both major political parties to a broadrange of economic reforms. But there are alsoforces that could restrain growth, includingtwin fiscal and balance of payments deficits,the deteriorating state of infrastructure, anda large public sector deficit. Substantially fastergrowth would require increased fixed capitalformation and saving, and foreign direct in-vestment would be a crucial element of financ-ing. Absent policy changes, prospects forBangladesh and Pakistan are not particularlyencouraging. An improvement in external con-ditions, including an acceleration of worldtrade growth and the stabilization of commod-ity prices, would be helpful in the near term,but the recent surge in petroleum prices willexact a toll on fragile import bills for bothcountries. The implications for economicgrowth of the recent coup in Pakistan are notyet clear. For Bangladesh, most of the flooddamage has been repaired, the water level hasfallen, and crops were replanted for the 2000harvest. Conditions should allow the economyto revive toward 5 percent growth in 2000.Sri Lanka is well situated to benefit from anacceleration in world trade growth in 2000and an increase in commodity prices over thesame period.

Long-term prospectsDespite cautious optimism concerning a near-term increase in economic activity in the re-gion, prospects covering 2002–2008 have beendowngraded from previous projections. Themain requirement for a strong long-term out-look is economic reform: in Pakistan, theimplementation of key reforms in banking,power, taxation, and public spending; and in

Bangladesh, the improvement of macroeco-nomic management, governance, and infra-structure. But critically, India’s long-termpotential output growth is linked to policy ac-tions, which must redress the current fiscaldeficit (over 6 percent of GDP), continued high(and increasing) tariff barriers, and increasesin nonperforming loans and other structuralproblems in the banking system.

Sub-Saharan Africa

Recent developments

Sub-Saharan Africa’s economic growthslowed progressively from 4.7 percent in

1996 to less than 2.5 percent in 1998–99. Sev-eral factors were responsible. A steep fall inoil prices in 1997–98 hurt countries that ex-port significant amounts of oil and other hy-drocarbons.3 Numerous conflicts—many ofwhich had appeared to be moving toward reso-lution—flared up, including in Angola, Demo-cratic Republic of Congo, Ethiopia and Eritrea,Guinea-Bissau, Lesotho, and Sierra Leone.Poor weather disrupted agriculture and tour-ism throughout eastern and southern Africaand heavy rains damaged infrastructure. Fi-nally, South Africa, the region’s largest andmost open economy, was caught in the turbu-lence of the East Asian crisis as a reversal ofcapital inflows in mid-1998 sent interest ratessharply higher and dampened interest-sensi-tive sectors through the first half of 1999.

Despite these adverse developments, pro-market reforms, trade liberalization, and bet-ter governance sustained growth elsewhere inthe region—in Mozambique, Uganda, andgenerally throughout the CFA zone, where thedevaluation of the CFA franc in 1994 triggereda sustained boom in investment and exports.Moreover, developments in 1999 augur wellfor near-term prospects. Especially encourag-ing have been the smooth political transitionin Nigeria; renewed prospects for peace inSierra Leone, Liberia, and the DemocraticRepublic of Congo, and a swift return of in-ternational investors to South Africa.

147

Near-term outlookGrowth in 1999 will be near the pace achievedin 1998—around 2.3 percent, or 3.2 percentexcluding South Africa and Nigeria. This fallsshort of what is needed to recover the groundlost in recent years, and per capita incomeswill decline modestly for a second successiveyear. Nevertheless, a gradual recovery fromthe slowdown appears to be underway.Growth is anticipated to rise to 3.3 percentover 2000–2001. Underpinning the strongerperformance will be a pick-up in exports overthe next 12–24 months, based on sharp gainsin oil prices, the unexpected strength of therecovery in East Asia, and the euro’s weak-ness against the U.S. dollar, which will fur-ther boost the competitiveness of CFA zonecountries. Following a decline of 0.3 percentin 1998, merchandise exports are expected togrow by 3.8 percent in 1999, 6.3 percent in2000, and 4.7 percent in 2001. On top of thebroader recovery, export performance in 2000will reflect the revival of Nigeria’s hydrocar-bon sector, and especially the commencementof liquid natural gas shipments from the newBonny Island facility. Excluding South Africaand Nigeria, Sub-Saharan export growth is

predicted to reach 4.3 percent in 1999 and4.6 percent in 2000–2001.

Import growth will not keep pace withexports, allowing the current account deficitto shrink from an unsustainable $14.2 billion(4.6 percent of GDP) in 1998. For the regionas a whole, the current account deficit is pro-jected to narrow to 2.8 percent of GDP in 1999and 2.6 percent in 2000. While falling shortof the level of performance in 1997–98, im-port growth will remain adequate to supportpurchases of capital and intermediate goodsfor production and investment needs. Mean-while, the forecast anticipates a continuationof policy reforms and a move away from theinward-looking trade regimes that contributedto Sub-Saharan Africa’s relative isolation overthe past 25 years. Overall, trade will continueto outpace GDP growth, resulting in highertrade shares and increasing openness (figureA1.5).

Long-term prospectsOutput for the region is expected to achievegrowth of 3.6 percent from 2002–2008 (fig-ure A1.5). That represents a 0.5 percentagepoint downward revision from last year’s Glo-

Figure A1.5 Real GDP growth by analytical groups for Sub-Saharan Africa

Percent per year

0

1

2

3

4

5

6

–2

–1

CFA franczone

Agriculturalexporters

Oilexporters

Mineralexporters

Undercivil strife

TotalSub-Saharan

AfricaSource: World Bank baseline forecasts.

19981999–20012002–2008


148


bal Economic Prospects and reflects greaterpessimism about prospects in commodity andfinancial markets, as well as an assessment thatSouth Africa will require a more extendedperiod to reach its long-run potential outputgrowth. A sustained 3.6 percent expansion ofGDP, together with a gradual slowing of theregion’s population growth (from current ratesof 2.6 toward 2.4 percent), will permit a mod-erate rise in per capita incomes of 1.3 percentper year over the long-term forecast period.The ability to encourage private investmentwill be the key to stronger performance, andthose countries with better policy environ-ments, effective export strategies, and morediversified economies will tend to do better(table A1.6).

Despite the downward revision, the out-look remains for a substantial improvementover the 2 percent average growth achievedduring the past two decades. Many of theregion’s intractable conflicts are likely to beresolved, and there has been a broad-basedtransition to democratic rule and more respon-sible governance. An accumulation of evidenceindicates that these developments are produc-ing results—for instance, they have contrib-

Table A1.6 Sub-Saharan Africa forecast summary(percent per year)

Baseline forecast


Real GDP growth 2.4 3.7 2.4 2.3 3.1 3.4 3.4 Consumption per capita –0.4 2.5 1.0 –0.8 –0.1 0.5 0.9 GDP per capita –0.3 0.8 –0.2 –0.2 0.6 0.9 1.0

Population 2.8 2.9 2.6 2.6 2.5 2.5 2.4Median inflationa 9.7 8.2 6.5 8.1 5.5 5.4 5.5Gross domestic investment/GDP 16.8 17.0 17.4 17.5 17.7 17.9 18.4Central government budget deficit/GDP –5.3 –3.3 –3.6 –4.1 –4.0 –4.1 –4.1Export volumeb 4.3 5.6 –0.3 4.1 6.1 4.8 4.9Current account/GDP –2.0 –2.8 –4.6 –2.8 –2.6 –2.6 –2.0

Memo itemsGDP of major oil exportersc 2.9 4.2 1.3 1.3 3.4 3.6 3.2GDP of region, excluding South Africa 3.1 4.6 4.0 3.3 3.6 3.9 3.9

and oil exporters

a. GDP deflator.b. Goods and nonfactor services.c. Angola, Gabon, and Nigeria.Source: World Bank Development Prospects Group, November 1999

uted to the rise in investment and exportsduring the 1994–97 growth cycle (see chap-ter 4). However, these factors will be offsetby other less favorable trends: commodityprices of critical importance to the region areexpected to show only weak advances, put-ting a damper on investment and exports; apoor outlook for foreign aid and negative for-eign investor sentiment will squeeze currentaccount balances and compress imports; fall-ing budget support is likely to cut into socialspending, reducing the growth rate of humancapital. Finally, the extent of the AIDS epi-demic is becoming clearer, and with it the in-evitability of a sustained, negative impact oneconomic performance. Medium-term pro-jected population growth in the worst affectedcountries has been lowered by as much as 1–2 percent annually compared to what wouldhave been expected without AIDS. Moreover,since the disease preponderantly affects bet-ter educated and more productive urban work-ers, output growth in these countries will slowby a substantially higher amount. For the re-gion as a whole, a reduction in per capitagrowth of 0.3 percent tied to this developmentseems likely.

149

Especially given the lackluster expecta-tions for commodity prices and the poor cli-mate for foreign aid, countries will need tocontinue diversifying and opening their econo-mies to sustain adequate growth. Ongoingstructural adjustment over the forecast periodshould raise domestic savings and investment,as well as make Sub-Saharan Africa more at-tractive to potential foreign investors. FDI isincreasingly being attracted not only to ex-tractive and resource-based sectors, but alsoto telecommunications, transportation, andutilities sectors as they are privatized and de-regulated. Meanwhile, though prospects forforeign aid are not promising, many Sub-Saharan countries stand to benefit from debtrelief under enhanced terms of the HeavilyIndebted Poor Countries Initiative (HIPC),improving the sustainability of current accountpositions. Of the 36 countries eligible under

the enhanced framework, 30 are in Sub-Saharan Africa, and the region stands to re-ceive over 80 percent of the relief worth $27billion in net present value terms. Not onlywill there be a transfer of real resources forpoverty reduction and other purposes, butstronger balance sheets and improved credit-worthiness will facilitate the integration ofrecipient countries into the world economy.

Notes1. Indonesia, Malaysia, the Philippines, the Re-

public of Korea, and Thailand.2. Since July the currencies have been drifting

somewhat downward for a number of reasons—re-structuring problems in Korea; corruption and con-flict in East Timor in Indonesia; and tensions betweenTaiwan (China) and China have affected the former.

3. Angola, Cameroon, Congo (Brazzaville), Equa-torial Guinea, Gabon, and Nigeria.


151

Appendix 2Global Economic Indicators

152


Table A2.1 Growth of real GDP, 1966–2008(GDP in 1987 prices and exchange rates—average annual percentage growth)

1998 GDP(current

billions of Estimate ForecastU.S. dollars) 1966–73 1974–90 1991–98 1998 1999 1999–2008

World 28,445 5.2 3.0 2.5 1.9 2.6 3.1

High-income economies 22,200 5.0 2.8 2.3 2.0 2.6 2.6 Industrial 21,505 5.0 2.7 2.2 2.0 2.6 2.6 G–7 18,425 5.0 2.8 2.2 1.8 2.6 2.5 United States 8,510 3.2 2.7 3.0 3.9 3.8 2.8 Japan 3,790 10.0 3.8 1.5 –2.9 1.3 1.7 G–4 Europe 6,130 4.4 2.3 1.7 2.2 1.6 2.6 Germanya 2,130 4.3 2.0 1.8 2.0 1.3 2.6 Euro Area 6,465 5.0 2.3 1.7 2.7 2.0 2.8 Other industrial 3,080 5.0 2.3 2.3 3.7 2.9 3.1 Other high-income 695 8.5 5.5 5.7 1.2 3.0 4.8 Asian NIEs 515 9.6 8.1 6.0 1.8 3.6 5.4

Low- and middle-income economies 6,245 6.2 3.8 3.2 1.6 2.7 4.5 Excluding Eastern Europe and CIS 5,390 6.0 3.7 5.3 2.1 3.0 4.7 Asia 2,360 5.8 6.5 7.6 1.6 5.4 5.9 East Asia and Pacific 1,800 7.8 7.5 8.5 0.1 5.5 6.2 China 960 8.4 8.7 11.4 7.8 6.5 … Korea, Rep. of 320 11.2 8.5 6.3 –5.8 8.0 … Indonesia 95 6.4 6.7 5.8 –13.2 0.0 … South Asia 560 3.6 5.0 5.9 5.2 5.4 5.2 India 430 3.8 4.9 6.1 5.1 6.0 … Latin America and the Caribbean 2,000 6.2 2.6 3.6 2.1 –0.6 3.5 Brazil 775 9.5 3.6 3.1 0.2 –0.4 … Mexico 425 6.3 3.2 2.6 4.8 3.2 … Argentina 340 4.3 0.4 5.2 3.9 –3.5 … Europe and Central Asia 1,040 6.2 4.2 –4.0 –0.2 0.3 3.4 Russian Federationb 337 6.6 5.2 –7.8 –4.6 1.0 … Turkey 190 1.9 4.2 4.2 2.9 –2.2 … Poland 145 7.3 0.3 4.2 4.8 3.5 … Middle East and North Africa 535 7.8 1.4 2.9 3.2 2.0 3.4 Saudi Arabia 130 8.5 0.9 1.6 1.6 –0.4 … Iran, Islamic Rep. 120 10.2 –0.3 3.6 2.1 1.2 … Egypt, Arab Rep. 83 3.7 7.1 4.1 4.9 4.9 … Sub-Saharan Africa 310 4.5 2.1 2.8 2.4 2.3 3.4 Republic of South Africa 115 4.9 2.1 1.5 0.1 1.0 … Nigeria 35 6.5 1.1 2.5 2.3 1.0 …

Note: Growth rates over intervals are computed using least squares method.a. Data prior to 1991 covers West Germany.b. Data prior to 1992 covers the former Soviet Union.Source: World Bank data and staff estimates.

Figure A2.1 Real GDP growth, 1989–2008Percent10

6

8

0

2

4

–6

–4

–2

High-incomeeconomies

Middle Eastand North

Africa

East Asiaand Pacific

South Asia Latin Americaand the

Caribbean

Europe andCentral Asia

Sub-SaharanAfrica

1989–98

1999–2008

153

Table A2.2 Growth of real per capita GDP, 1966–2008(GDP in 1987 prices and exchange rates—average annual percentage growth)

1998 GDP per capita Estimate Forecast(current

U.S. dollars) 1966–73 1974–90 1991–98 1998 1999 1999–2008

World 5,020 3.1 1.2 1.0 0.6 1.4 1.9

High-income economies 25,245 4.1 2.1 1.6 1.5 2.3 2.3 Industrial 25,720 4.1 2.1 1.6 1.6 2.3 2.3 G–7 26,929 4.0 2.2 1.5 1.3 2.2 2.2 United States 31,520 2.1 1.7 1.9 3.0 3.2 2.1 Japan 29,995 8.7 3.0 1.2 –3.1 1.2 1.6 G–4 Europe 23,795 3.8 2.1 1.3 2.0 1.5 2.6 Germanya 25,960 3.7 2.0 1.3 1.9 1.2 2.7 Euro Area 22,210 4.3 2.0 1.4 2.5 1.9 2.8 Other industrial 20,300 4.1 1.7 1.8 3.3 2.7 2.9 Other high-income 16,055 5.8 3.3 4.0 –0.5 1.6 3.7 Asian NIEs 16,165 7.2 6.3 4.7 0.4 2.7 4.7

Low- and middle-income economies 1,300 3.7 1.8 1.6 0.1 1.2 3.2 Excluding Eastern Europe and CIS 1,220 3.3 1.5 3.5 0.4 1.3 3.3 Asia 790 3.1 4.5 6.0 0.2 4.0 4.6

East Asia and Pacific 1,060 5.1 5.8 7.1 –1.1 4.3 5.2 China 775 5.6 7.1 10.2 6.8 5.5 … Korea, Rep. of 6,910 8.8 7.0 5.2 –6.7 7.2 … Indonesia 460 3.9 4.6 4.1 –14.6 –1.6 … South Asia 435 1.2 2.6 3.9 3.3 3.5 3.7 India 440 1.4 2.6 4.3 3.3 4.0 … Latin America and the Caribbean 4,115 3.5 0.4 1.8 0.5 –2.2 2.0 Brazil 4,680 6.8 1.4 1.6 –1.1 –1.7 … Mexico 4,410 2.9 0.8 0.8 3.0 1.6 … Argentina 9,350 2.7 –1.0 3.8 2.6 –4.7 … Europe and Central Asia 2,260 5.1 3.2 –4.2 –0.4 1.0 3.4 Russian Federationb 2,300 5.6 4.3 –7.9 –4.1 0.8 … Turkey 3,030 –0.7 1.8 2.7 1.3 –3.6 … Poland 3,775 6.5 –0.5 4.0 4.8 3.4 … Middle East and North Africa 1,950 4.9 –1.7 0.6 1.1 –0.1 1.4 Saudi Arabia 6,215 4.3 –4.3 –1.9 –1.7 –4.0 … Iran, Islamic Rep. 1,950 7.0 –3.6 1.9 0.5 –0.5 … Egypt, Arab Rep. 1,350 1.5 4.5 2.1 3.1 3.2 … Sub-Saharan Africa 510 1.8 –0.8 0.1 –0.2 –0.2 1.0 Republic of South Africa 2,825 2.7 –0.5 –0.5 –1.6 –0.6 … Nigeria 305 3.6 –1.9 –0.4 –0.5 –1.7 …

Note: Growth rates over intervals are computed using least squares method.a. Data prior to 1991 covers West Germany.b. Data prior to 1992 covers former Soviet Union.Source: World Bank data and staff estimates.

G L O B A L E C O N O M I C I N D I C A T O R S

Figure A2.2 Real per capita GDP growth, 1989–2008Percent

6

8

0

2

4

–6

–4

–2



Africa



Caribbean


Sub-SaharanAfrica

1989–98

1999–2008

154


Table A2.3 Inflation: GDP deflators, 1966–2008(percentage change a)

Estimate Forecast1966–73 1974–90 1991–98 1998 1999 1999–2008

World 5.4 7.8 4.2 2.5 2.5 2.7

High-income economies 5.5 7.0 2.3 1.1 1.2 1.9 Industrial 5.5 6.8 2.2 1.1 1.2 1.9 G–7 5.3 6.5 2.1 1.0 0.9 1.7 United States 4.8 6.3 2.2 1.0 1.3 1.9 Japan 5.8 3.6 0.3 0.4 –0.4 1.2 G–4 Europe 5.4 8.3 3.3 1.6 1.3 1.8 Germanyb 4.9 3.5 3.1 0.9 1.2 1.7 Euro Area 5.6 7.9 3.2 1.6 1.4 1.9 Other industrial 6.4 8.4 2.9 1.6 3.0 2.7 Other high-income 5.5 22.2 4.3 1.5 1.2 3.5 Asian NIEs 5.6 6.9 3.7 1.6 0.4 3.4

Low- and middle-income economies 4.7 11.1 11.3 7.6 7.3 5.4 Excluding Eastern Europe and CIS 4.8 11.4 10.1 6.5 7.2 5.3 Asia 6.7 9.2 7.2 8.2 5.6 4.4 East Asia and Pacific 6.7 8.2 6.1 8.7 4.5 3.8 China –2.0 3.8 9.6 –1.3 –0.9 … Korea, Rep. of 14.3 12.1 5.6 13.4 4.5 … Indonesia 65.0 13.3 12.3 73.1 38.1 … South Asia 6.9 10.6 9.4 7.6 7.2 5.3 India 7.6 8.1 8.7 7.6 9.8 … Latin America and the Caribbean 6.1 10.0 14.4 7.9 8.3 6.3 Brazil 23.2 145.0 348.0 3.8 11.0 … Mexico 6.4 48.0 19.5 15.9 16.8 … Argentina 24.0 203.0 9.8 –2.0 3.2 … Europe and Central Asia 2.0 6.5 41.4 12.3 8.9 6.3 Russian Federationc 0.9 1.2 250.0 11.6 56.0 … Turkey 7.3 44.5 79.3 69.2 60.0 … Middle East and North Africa 3.6 11.0 6.2 3.1 4.7 4.4 Saudi Arabia 11.4 4.0 1.4 –13.4 4.0 … Iran, Islamic Rep. 5.6 17.0 28.3 15.7 11.1 … Egypt, Arab Rep. 2.3 12.4 9.8 4.2 4.0 … Sub-Saharan Africa 4.2 10.6 10.5 6.5 7.9 5.6 Republic of South Africa 6.2 14.5 9.8 8.4 7.5 … Nigeria 10.7 14.5 38.8 10.0 13.1 …

Note: Deflators are in local currency units; 1987 = 100. Growth rates over intervals are computed using least squares method.a. High-income group inflation rates are GDP-weighted averages of local currency inflation. Low- and middle-income groups aremedians. World is GDP-weighted average of the two groups.b. Data prior to 1991 covers West Germany.c. Data prior to 1992 covers former Soviet Union.Source: World Bank data and staff estimates.

Figure A2.3 GDP inflation, 1989–2008Percent

20

25

5

10

15

0High-incomeeconomies


Africa



Caribbean


Sub-SaharanAfrica

1989–98

1999–2008

155


Table A2.4 Current account balances, 1970–2008(percentage of GDP)

1998 currentaccount

(billions of Estimate Forecast U.S. dollars) 1970–80 1981–90 1991–98 1998 1999 1999–2008

World –33.3 0.1 –0.4 –0.1 –0.1 –0.3 –0.5

High-income economies –0.8 0.1 –0.2 0.2 0.0 –0.4 –0.4 Industrial –24.7 0.0 –0.5 0.1 –0.1 –0.6 –0.5 G–7 –50.8 0.2 –0.4 0.0 –0.3 –0.8 –0.7 United States –232.5 0.0 –2.0 –1.5 –2.7 –3.5 –3.3 Japan 118.8 0.7 2.4 2.5 3.1 2.7 3.3 G–4 Europe 62.9 0.2 0.3 0.1 1.0 0.7 0.6 Germanya –4.9 0.5 2.6 –0.7 –0.2 –0.2 0.1 Euro Area 88.1 0.0 0.4 0.5 1.4 1.0 0.9 Other industrial 26.1 –0.9 –0.9 0.9 0.8 0.7 0.6 Other high-income 23.9 7.2 9.8 3.1 3.4 5.0 2.3 Asian NIEs 21.8 0.1 6.8 4.4 4.3 5.2 2.1

Low- and middle-income economies –32.5 0.0 –0.8 –1.3 –0.5 0.2 –1.0 Excluding Eastern Europe and CIS –19.3 –0.6 –2.1 –1.8 –0.4 0.4 –0.9 Asia 90.5 –1.1 –1.4 –0.8 3.8 2.2 0.3 East Asia and Pacific 100.7 –1.4 –1.1 –0.5 5.6 3.5 1.1 China 29.6 –0.4 0.1 1.4 3.1 1.6 … Korea, Rep. of 40.6 –6.1 0.7 –0.1 12.6 5.2 … Indonesia 4.0 –1.4 –3.1 –1.6 4.2 3.8 … South Asia –10.2 –0.5 –2.0 –1.7 –1.8 –2.0 –2.5 India –6.9 0.3 –1.7 –1.2 –1.6 –2.0 … Latin America and the Caribbean –89.4 –2.6 –1.8 –2.9 –4.5 –3.1 –3.4 Brazil –34.6 –4.1 –1.6 –1.7 –4.5 –4.3 … Mexico –16.3 –3.1 –0.7 –3.9 –3.9 –2.8 … Argentina –14.6 –0.3 –2.1 –2.7 –4.3 –3.5 … Europe and Central Asia –11.3 0.5 2.2 0.4 –1.1 –1.0 –1.5 Russian Federationb 1.7 2.0 3.6 2.1 0.5 4.3 … Turkey 1.9 –2.0 –1.3 –0.6 1.0 1.0 … Poland –7.5 –0.9 –1.4 –2.9 –5.1 –7.2 … Middle East and North Africa –8.1 6.6 –3.5 –2.3 –1.5 4.2 2.2 Saudi Arabia –6.7 19.8 –7.2 –7.7 –5.2 6.0 … Iran, Islamic Rep. –1.7 2.4 –0.4 –1.4 –1.4 2.2 … Egypt, Arab Rep. –2.6 –4.9 –3.4 2.2 –3.1 –2.7 … Sub-Saharan Africa –14.2 –1.9 –2.8 –2.3 –4.6 –2.8 –2.0 Republic of South Africa –2.3 –1.7 0.6 –0.3 –1.9 –2.0 … Nigeria –5.8 0.8 –0.7 –1.9 –15.7 –3.0 …

Note: Current account after official transfers. Shares over intervals are period averages.a. Data prior to 1991 covers West Germany.b. Data prior to 1992 covers former Soviet Union.Source: World Bank data and staff estimates.

Figure A2.4 Ratio of current account balance to GDP, 1989–2008Percent

2

3

–1

–2

–3

–4

0

1



Africa



Caribbean


Sub-SaharanAfrica

1989–98

1999–2008

156


Table A2.5 Exports of goods, 1997(percent)

Merchandise Average Effective Merchandise Average Effective Merchandise Average Effectiveexports annual market exports annual market exports annual market(US$ growth growth (US$ growth growth (US$ growth growth

millions) 1987–97 1987–97a millions) 1987–97 1987–97a millions) 1987–97 1987–97a

World 5,537,918 6.6 6.6 Latin America and Middle East and North the Caribbean (continued) Africa (continued)

All developing 1,369,615 7.7 6.9 Uruguay 2,726 2.8 9.0 Saudi Arabia 62,381 5.3 7.7 economies Venezuela 21,624 5.1 6.7 Syrian Arab 3,916 10.4 4.6

Rep.Asia 599,225 12.9 8.2 Europe and 255,721 2.2 4.5 Tunisia 5,559 8.1 5.2

Central Asia Yemen, Rep. 2,264 12.0 12.6East Asia and 546,173 13.2 8.4 Armenia 233 .. 4.0 Pacific Azerbaijan 781 .. 5.0 Sub-Saharan 85,598 2.8 6.2China 182,877 13.2 9.4 Belarus 7,301 .. 3.0 AfricaFiji 590 9.3 7.0 Bulgaria 4,950 –14.5 3.1 Angola 5,130 4.7 5.5Indonesia 53,443 14.1 8.3 Czech 22,746 .. 8.3 Botswana 2,842 –0.7 5.4Korea, Rep. 136,164 10.3 7.7 Republic Cameroon 1,860 2.4 5.9Malaysia 78,740 16.6 9.3 Estonia 2,924 .. 5.0 Côte 4,299 1.7 5.8Myanmar 866 12.0 11.4 Georgia 372 .. 3.0 d’IvoirePapua New 2,149 6.4 7.5 Hungary 18,374 1.3 3.5 Ethiopia 587 –4.6 4.5 Guinea Kazakhstan 6,497 .. 4.7 Gabon 2,941 6.6 6.3Philippines 25,088 12.6 7.5 Kyrgyz 531 .. 4.0 Ghana 1,637 7.5 5.6Thailand 57,388 14.4 7.7 Republic Kenya 2,054 7.3 4.5Vietnam 7,337 .. .. Latvia 1,672 .. 5.0 Madagascar 223 –6.3 6.7

Lithuania 3,860 .. 5.0 Nigeria 15,213 3.4 6.6South Asia 53,052 10.0 6.8 Moldova 805 .. 4.0 Senegal 933 –1.0 5.9Bangladesh 3,778 11.9 6.4 Poland 25,751 7.1 3.9 South Africa 31,027 2.3 6.1India 35,108 12.0 6.7 Romania 8,431 –6.2 3.9 Sudan 594 –0.5 8.1Nepal 402 9.3 5.7 Russian 88,326 .. 5.1 Zambia 915 0.4 7.7Pakistan 8,918 1.7 7.3 Federation Zimbabwe 2,424 3.9 7.4Sri Lanka 4,633 11.1 6.2 Slovak 8,254 .. 8.3

Republic High- 4,168,303 6.3 6.6Latin America 276,068 6.4 7.1 Tajikistan 770 .. 3.0 income economies and the Caribbean TFYR 1,147 .. 1.7Argentina 26,370 7.5 8.2 Macedonia Industrial 3,651,823 5.7 6.5Bolivia 1,167 10.3 10.6 Turkmenistan 1,628 .. 3.0Brazil 52,990 4.8 7.4 Turkey 26,245 5.8 5.6 G–7 2,659,715 5.2 6.7Chile 16,663 8.4 7.9 Ukraine 14,232 .. 4.5 France 290,151 4.6 5.9Colombia 11,522 12.6 6.4 Uzbekistan 3,781 .. 4.0 Germany 512,427 4.3 5.7Costa Rica 4,268 14.1 6.1 Italy 252,000 6.1 6.0Dominican 882 –6.0 5.9 Middle East 153,003 2.6 7.2 Japan 420,957 2.4 8.1 Republic and North Africa United 281,061 4.6 6.5Ecuador 5,264 6.8 6.9 Algeria 14,833 1.9 6.1 KingdomEl Salvador 1,359 10.4 6.7 Bahrain 4,384 4.6 8.3 United 688,697 7.5 8.2Guatemala 2,344 4.9 6.9 Egypt, Arab 3,921 –1.0 5.5 StatesJamaica 1,383 4.4 5.7 Rep.Mexico 110,431 12.7 6.5 Iran, Islamic 18,381 2.0 7.4 Other 992,108 7.1 6.1Panama 723 9.8 6.9 Rep. industrialParaguay 1,089 9.5 9.5 Iraq 8,179 –19.1 6.1 Australia 62,910 7.3 8.3Peru 6,841 9.0 7.2 Jordan 1,836 4.3 6.2 Austria 58,590 7.6 5.4Trinidad and 2,542 2.8 5.6 Morocco 7,032 3.6 6.3 Belgiumb 178,880 5.9 5.7 Tobago Oman 7,630 9.6 9.8 Denmark 47,715 5.4 5.9

157


Figure A2.5b Average annual growth rate of export volumes, 1987–97Percent

10

15

0

5

Industrialeconomies

World


Africa



Caribbean


Sub-SaharanAfrica

Figure A2.5a Merchandise exports as a share of GDP, 1997Percent

30

40

0

10

20

Industrialeconomies

World


Africa



Caribbean


Sub-SaharanAfrica

Table A2.5 Exports of goods, 1997 (continued)(percent)

Merchandise Average Effective Merchandise Average Effective Merchandise Average Effectiveexports annual market exports annual market exports annual market(US$ growth growth (US$ growth growth (US$ growth growth

millions) 1987–97 1987–97a millions) 1987–97 1987–97a millions) 1987–97 1987–97a

Other industrial (continued) Other industrial (continued) Other high-Finland 39,316 7.7 5.5 Sweden 82,802 7.5 5.9 income (continued)Greece 8,626 5.3 5.4 Switzerland 72,493 6.4 6.6 Kuwait 14,224 12.6 7.6Iceland 1,852 0.7 5.5 Qatar 4,377 2.7 7.4Ireland 53,512 11.6 5.8 Other high- 516,480 10.1 7.9 Singapore 124,985 14.5 9.0Netherlands 194,905 6.3 5.7 income Taiwan, 121,081 6.5 8.3New 14,207 4.8 7.7 Brunei 2,329 2.2 .. China Zealand Hong Kong, 188,059 12.8 8.1 United Arab 23,194 3.1 8.5Norway 48,542 6.0 5.7 China EmiratesSpain 104,359 11.0 5.9 Israel 22,503 8.1 6.6

.. Not available.a. Effective market growth is a weighted average of import volume growth in the country’s export markets.b. Includes LuxembourgNote: Merchandise exports are f.o.b. in current U.S. dollars. Growth rates are for export volumes. Growth rates over intervals are computed using leastsquares method.Source: see Technical Notes.

158


Table A2.6 Imports of goods, 1997(percent)

Merchandise Annual Merchan- Merchandise Annual Merchan- Merchandise Annual Merchan-imports average dise imports average dise imports average dise

(US$ growth imports/ (US$ growth imports/ (US$ growth imports/millions) 1987–96 GDP millions) 1987–96 GDP millions) 1987–96 GDP

World 5,541,179 6.4 19.2 Latin America and Middle East and the Caribbean (continued) North Africa (continued)

All developing 1,385,216 9.6 20.9 Uruguay 3,716 12.0 16.6 Jordan 4,102 1.6 61.2 economies Venezuela 14,606 2.4 11.3 Morocco 9,525 10.6 29.0

Oman 5,026 8.6 29.2Asia 600,510 12.0 24.2 Europe and 316,545 9.0 25.0 Saudi 28,742 1.0 19.8

Central Asia ArabiaEast Asia and 536,592 12.8 27.2 Armenia 892 .. 52.6 Syrian 4,028 10.6 30.1 Pacific Azerbaijan 794 .. 26.3 Arab Rep.China 142,189 11.1 15.5 Belarus 8,689 .. 30.7 Tunisia 7,914 7.0 40.7Fiji 965 6.1 46.8 Bulgaria 4,559 –15.1 46.6 Yemen, Rep. 2,407 1.0 40.6Indonesia 41,694 12.3 20.0 Czech 28,540 .. 58.5Korea, Rep. 144,616 12.3 31.6 Republic Sub-Saharan 76,253 4.8 22.3Malaysia 79,030 18.1 80.1 Estonia 4,429 .. 68.9 AfricaMyanmar 2,037 21.7 10.9 Georgia 686 .. 14.0 Angola 2,477 9.2 27.1Papua New 1,697 2.1 37.5 Hungary 20,668 4.4 34.7 Botswana 2,258 3.6 34.1 Guinea Kazakhstan 4,301 .. 19.1 Côte 2,741 –0.2 28.3Philippines 38,662 16.0 41.5 Kyrgyz 894 .. 50.7 d’IvoireThailand 62,854 14.4 49.2 Republic Cameroon 1,359 –3.6 13.5Vietnam 10,481 .. 39.8 Latvia 2,721 .. 42.0 Ethiopia 1,019 –0.8 15.7

Lithuania 5,644 .. 47.6 Gabon 1,034 3.5 18.8South Asia 63,918 7.3 12.6 Moldova 1,079 .. 57.6 Ghana 2,326 9.2 30.6Bangladesh 6,898 9.3 15.6 Poland 42,308 15.4 27.4 Kenya 3,279 0.3 27.9India 37,375 6.1 9.8 Romania 11,280 1.4 32.8 Madagascar 470 3.0 14.3Nepal 1,720 8.4 29.3 Russian 73,613 .. 15.4 Nigeria 10,330 8.0 17.4Pakistan 11,863 7.2 19.8 Federation Senegal 1,196 0.1 28.0Sri Lanka 5,851 10.1 36.1 Slovak 10,774 .. 58.7 South 32,998 6.2 20.4

Republic AfricaLatin America 270,280 12.3 13.4 Tajikistan 808 .. 73.5 Sudan 1,422 1.2 13.3 and the Caribbean TFYR 1,464 .. 66.5 Zambia 819 1.3 21.2Bolivia 1,851 9.3 20.5 Macedonia Zimbabwe 2,654 2.8 26.2Brazil 65,007 13.7 6.9 Turkmen- 1,173 .. 29.3Chile 19,662 10.9 23.1 istan High– 4,155,963 5.6 18.7Colombia 15,378 8.7 14.3 Turkey 48,585 9.9 23.0 income economiesCosta Rica 4,924 10.9 45.2 Ukraine 17,114 .. 37.5Dominican 4,821 7.7 27.4 Uzbekistan 4,712 .. 19.3 Industrial 3,630,228 5.0 16.9 RepublicEcuador 4,955 7.7 19.9 Middle East 121,628 2.6 23.2 G–7 2,669,104 5.0 14.4El Salvador 2,973 9.8 23.7 and North Canada 200,873 5.5 28.8Guatemala 3,852 12.5 17.7 Africa France 269,892 3.7 20.2Jamaica 3,132 6.9 71.7 Algeria 9,280 –0.3 19.3 Germany 445,616 5.0 22.0Mexico 114,846 16.3 23.2 Bahrain 4,026 1.1 77.6 Italy 208,364 2.9 18.1Panama 3,000 12.2 31.7 Egypt, 13,321 4.1 17.2 Japan 338,754 6.8 8.3Paraguay 3,403 22.4 33.5 Arab Rep. United 306,585 3.3 22.4Peru 10,264 9.6 14.8 Iran, 14,165 2.4 14.5 KingdomTrinidad and 2,990 4.8 36.4 Islamic Rep. United 899,020 6.1 10.5 Tobago Iraq 790 –27.6 1.3 States

159


Figure A2.6a Merchandise imports as a share of GDP, 1997Percent

20

30

0

10

Industrialeconomies


Africa



Caribbean


Sub-SaharanAfrica

World

Figure A2.6b Average annual growth rate of import volumes, 1987–97Percent

10

15

0

5

Industrialeconomies


Africa



Caribbean


Sub-SaharanAfrica

World

Table A2.6 Imports of goods, 1997 (continued)(percent)

Merchandise Annual Merchan- Merchandise Annual Merchan- Merchandise Annual Merchan-imports average dise imports average dise imports average dise

(US$ growth imports/ (US$ growth imports/ (US$ growth imports/millions) 1987–96 GDP millions) 1987–96 GDP millions) 1987–96 GDP

Other 961,124 5.1 32.7 Other industrial (continued) Other high-income (continued) industrial New 14,518 5.9 22.6 Hong Kong, 208,614 12.6 114.4Australia 65,892 5.9 16.1 Zealand ChinaAustria 64,766 5.1 32.7 Norway 35,709 3.9 23.2 Israel 30,781 7.9 32.2Belgiuma 166,639 4.3 66.9 Portugal 33,823 6.3 33.7 Kuwait 8,246 0.4 27.6Denmark 44,039 4.5 26.1 Spain 122,711 8.2 22.9 Qatar 3,322 10.1 30.8Finland 29,784 2.6 24.4 Sweden 65,020 6.2 29.4 Singapore 132,437 12.1 138.0Greece 27,799 8.8 22.8 Switzerland 71,064 –0.1 29.2 Taiwan, 113,924 10.2 35.8Iceland 1,992 –1.5 27.5 Other high- 525,735 10.8 71.4 ChinaIreland 39,238 8.3 46.5 income United 29,952 13.1 45.9Netherlands 178,130 5.6 49.7 Brunei 2,000 10.6 37.0 Arab Emirates

.. Not available.a. Includes LuxembourgNote: Merchandise imports are c.i.f. in current U.S. dollars. Growth rates are for import volumes. Growth rates over intervals are computed using leastsquares method.Source: see Technical Notes.

160


Table A2.7 Direction of merchandise trade, 1997a

(percentage of world trade)

High-income importers Low- and middle-income importers

LatinMiddle America All

Sub- East Europe East and low-Other All Other All Saha- Asia and and the and

United indust- indust- high- high- ran and South Central North Carib- middle-Source of exports States EU-15 Japan rial rial income income Africa Pacific Asia Asia Africa bean income World

High-income economies 8.4 30.1 5.5 8.2 52.3 7.3 59.7 0.9 6.2 0.7 2.7 1.9 3.7 16.1 75.8Industrial 6.8 28.5 3.6 7.7 46.6 5.3 52.0 0.9 4.0 0.5 2.5 1.4 3.6 13.0 64.9

United States … 2.8 2.1 3.5 8.4 2.2 10.7 0.3 1.4 0.2 0.2 0.2 2.5 4.8 15.5EU-15 2.5 20.8 1.2 3.4 27.9 1.5 29.4 0.5 1.1 0.2 2.1 0.8 0.7 5.4 34.8Japan 1.2 0.7 … 0.5 2.4 1.1 3.5 0.0 1.3 0.0 0.1 0.4 0.2 2.0 5.5Other 3.1 4.1 0.3 0.4 7.9 0.4 8.4 0.1 0.3 0.0 0.2 0.1 0.2 0.7 9.1industrial

Other high-incomeb 1.6 1.7 1.9 0.5 5.7 2.0 7.7 0.1 2.2 0.1 0.2 0.4 0.1 3.2 10.9

Low- and middle- 3.9 7.1 2.0 0.9 13.9 4.0 17.9 0.4 1.4 0.2 2.1 0.8 1.4 6.3 24.2income economies

Sub-Saharan Africa 0.1 0.5 0.1 0.0 0.8 0.2 0.9 0.2 0.1 0.0 0.0 0.1 0.0 0.4 1.3East Asia and Pacific 0.8 1.0 1.3 0.3 3.4 2.7 6.2 0.1 0.8 0.1 0.1 0.2 0.1 1.4 8.7South Asia 0.1 0.3 0.1 0.1 0.5 0.3 0.8 0.0 0.1 0.0 0.0 0.1 0.0 0.3 1.1

Europe and 0.2 3.3 0.1 0.2 3.8 0.3 4.0 0.0 0.1 0.0 1.8 0.1 0.1 2.2 6.2Central AsiaMiddle East 0.3 1.0 0.1 0.1 1.6 0.2 1.8 0.0 0.2 0.0 0.1 0.2 0.1 0.5 2.3and North Africa

Latin America 2.4 1.0 0.4 0.2 3.9 0.4 4.2 0.0 0.1 0.0 0.1 0.1 1.1 1.4 5.6and Caribbean

World 12.3 37.3 7.5 9.2 66.3 11.3 77.6 1.3 7.6 0.9 4.8 2.6 5.1 22.4 100.0

a. Expressed as a share (percent) of total world exports. World merchandise exports in 1997 amounted to some $5,500 billion.b. Other high-income group includes the Asian newly industrializing economies, several oil exporters of the Gulf region, and Israel.Source: IMF, Direction of Trade Statistics.

161


Table A2.8 Growth of current dollar merchandise trade, by direction, 1987-97(average annual percentage growth)

High-income importers Low- and middle-income importers

LatinMiddle America All

Sub- East Europe East and low-Other All Other All Saha- Asia and and the and

United indust- indust- high- high- ran and South Central North Carib- middle-Source of exports States EU-15 Japan rial rial income income Africa Pacific Asia Asia Africa bean income World

High-income economies 9.5 6.8 5.3 10.4 7.5 9.8 7.7 5.2 15.8 11.6 10.1 6.1 11.5 11.3 8.4Industrial 9.0 6.5 3.6 10.2 7.1 8.0 7.2 4.6 15.6 10.7 9.7 4.9 11.5 10.5 7.8United States … 5.9 3.4 9.2 6.3 6.6 6.4 7.2 18.5 13.0 12.0 4.2 13.9 13.5 8.1EU-15 8.4 5.8 4.5 12.3 6.6 10.0 6.7 3.8 16.3 11.0 9.6 3.6 6.3 8.3 7.0Japan 8.8 9.1 … 5.3 8.1 8.5 8.2 1.6 12.7 3.9 3.4 8.4 7.0 10.1 8.9

Other industrial 9.5 11.1 2.2 12.1 9.9 9.0 9.8 3.8 16.4 9.9 12.6 10.8 12.3 12.3 10.0Other high-income 12.2 12.3 9.6 12.2 11.3 17.3 12.5 16.5 16.1 16.2 16.5 11.7 13.7 15.4 13.3

Low- and middle- 13.1 9.6 9.3 8.0 10.3 17.5 11.5 12.1 15.2 9.6 7.4 7.5 13.6 10.4 11.2income economiesSub-Saharan Africa 7.0 2.6 -1.9 6.5 2.8 9.7 3.7 12.4 10.6 15.2 1.6 19.2 7.9 11.6 5.6East Asia and Pacific 16.4 13.9 14.3 11.5 14.3 20.3 16.5 13.7 21.3 20.0 9.5 27.0 15.1 18.7 16.9South Asia 7.7 4.7 0.3 9.1 4.8 12.8 7.0 23.9 4.6 12.4 -1.0 10.6 9.9 8.1 7.3Europe and 9.9 13.4 1.7 6.9 12.3 19.5 12.6 5.3 7.8 -0.2 8.6 1.2 3.9 7.4 10.4Central AsiaMiddle East and 6.9 4.7 0.9 4.3 4.6 7.0 4.9 13.0 7.5 6.6 -6.0 5.7 7.3 4.2 4.7North Africa

Latin America and 14.4 9.7 9.5 6.5 12.0 18.2 12.4 10.1 26.9 33.0 11.3 -1.4 15.8 14.5 12.9Caribbean

World 10.5 7.2 6.2 10.1 8.0 11.9 8.5 6.8 15.7 11.1 8.9 6.5 12.1 11.0 9.0

Note: Growth rates are compound averages.Source: IMF, Direction of Trade Statistics.

162


Table A2.9 Structure of long-term public and publicly guaranteed (PPG) debt, 1997(percentage of long-term PPG debt)

Non-concessional Non-concessional

Concessional Variable Fixed Concessional Variable Fixed

All developing 28.0 33.8 38.1 Europe and Central Asia (continued) economies Bulgaria 2.2 81.5 16.3

Czech Republic 0.7 32.7 66.6Asia 35.4 26.4 38.2 Estonia 16.8 49.4 33.8East Asia and Pacific 25.8 29.9 44.3 Georgia 53.6 11.1 35.3China 16.5 37.4 46.1 Hungary 3.0 21.7 75.3Indonesia 43.2 32.9 23.9 Kazakhstan 6.0 60.9 33.1Korea, Rep. 3.6 30.1 66.3 Kyrgyz Republic 63.5 31.8 4.7Malaysia 12.1 23.7 64.3 Latvia 24.1 58.8 17.1Myanmar 88.1 0.0 11.9 Lithuania 11.0 44.7 44.3Papua New Guinea 57.1 19.8 23.2 Moldova 24.8 41.9 33.3Philippines 38.5 26.9 34.5 Poland 22.5 61.0 16.5Thailand 29.6 17.6 52.8 Romania 6.3 30.1 63.6Vietnam 17.0 18.3 64.6 Russian Federation 26.2 56.6 17.1

Slovak Republic 5.7 30.3 63.9South Asia 58.0 18.4 23.6 Tajikistan 83.4 13.9 2.7Bangladesh 98.9 0.2 0.9 Turkmenistan 3.2 75.3 21.5India 48.4 19.4 32.2 Turkey 12.2 23.6 64.2Nepal 97.8 0.0 2.2 Ukraine 3.0 71.7 25.2Pakistan 62.4 27.6 10.0 Uzbekistan 10.1 64.2 25.7Sri Lanka 90.6 3.7 5.8

Middle East and 39.1 32.7 28.2Latin America and 14.0 40.5 45.4 North Africa the Caribbean Algeria 10.9 53.4 35.7Argentina 3.5 32.9 63.6 Egypt, Arab Rep. 84.4 4.5 11.1Bolivia 70.3 9.0 20.7 Jordan 50.2 23.4 26.4Brazil 1.8 59.7 38.6 Morocco 31.5 38.0 30.5Chile 8.0 74.3 17.7 Oman 21.9 45.0 33.1Colombia 5.9 42.3 51.8 Syrian Arab Rep. 92.2 0.0 7.8Costa Rica 25.7 24.9 49.4 Tunisia 27.9 23.3 48.8Dominican Republic 44.9 26.3 28.8 Yemen, Rep. 89.7 2.3 8.0Ecuador 15.0 50.4 34.6El Salvador 49.5 23.2 27.3 Sub-Saharan Africa 51.1 13.1 35.8Guatemala 44.8 19.3 35.9 Angola 21.1 14.1 64.8Jamaica 37.8 24.8 37.3 Botswana 55.7 12.2 32.1Mexico 1.7 34.6 63.7 Côte d’Ivoire 43.3 41.8 14.9Panama 8.0 52.8 39.2 Cameroon 54.2 11.5 34.3Paraguay 60.3 10.2 29.5 Ethiopia 91.6 0.3 8.1Peru 27.0 45.3 27.8 Gabon 24.4 12.5 63.2Trinidad and Tobago 0.9 52.6 46.5 Ghana 79.1 0.6 20.4Uruguay 5.1 48.5 46.4 Kenya 70.7 5.9 23.4Venezuela 0.3 58.1 41.6 Madagascar 67.8 6.4 25.8

Nigeria 5.9 20.2 73.9Europe and 16.8 48.9 34.3 Senegal 76.4 6.0 17.5 Central Asia Sudan 50.9 14.4 34.7Armenia 50.6 35.0 14.4 Zambia 65.7 12.3 22.0Azerbaijan 81.2 18.8 0.0 Zimbabwe 44.4 15.7 39.9Belarus 13.2 66.8 20.1

Note: Nonconcessional debt data are available only for countries which report to the World Bank’s Debtor Reporting System. Foraggregate figures, missing values are assumed to have the same average value as the available data.Source: World Bank data.

163


Figure A2.9c Top ten ratios of nonconcessional debt to GDP, 1997

Nicaragua

Angola

Congo, Dem. Rep.

Guinea-Bissau

Mozambique

Congo, Rep.

Bulgaria

Vietnam

Côte d'Ivoire

Guyana

0 20 40 60Percent

80 100 120

Figure A2.9a Structure of long-term PPG debt, by group, 1997Percent100

80

60

40

20

0Severely indebted

low-incomeModerately indebted

low-incomeSeverely indebted

middle-incomeModerately indebted

middle-incomeOther

Concessional

Variable rate

Fixed rate

Figure A2.9b Structure of long-term PPG debt, by region, 1997Percent100

80

60

40

20

0East Asia

and PacificEurope andCentral Asia

Latin Americaand the Caribbean

Middle East andNorth Africa

South Asia Sub-SaharanAfrica

Concessional

Variable rate

Fixed rate

164


Table A2.10 Long-term net resource flows to developing countries, 1997(millions of U.S. dollars)

Private Official

OfficialPercentage Net debt development

Total of GDP Total flows FDI Portfolio Total assistance Other

All developing countries 338,047 5.49 298,953 105,340 163,423 30,191 39,094 33,139 5,955

Asia 137,295 6.86 115,367 34,751 68,946 11,670 21,928 7,623 14,306

East Asia and Pacific 122,594 8.20 104,257 30,780 64,284 9,193 18,337 4,851 13,486China 65,370 7.28 60,828 8,134 44,236 8,457 4,542 771 3,771Indonesia 11,581 5.39 10,863 5,888 4,677 298 718 624 94Korea, Rep. 17,465 3.67 13,069 8,968 2,844 1,257 4,396 –254 4,650Malaysia 9,151 9.35 9,312 4,695 5,106 –489 –161 –256 95Myanmar 242 1.94 180 102 80 –2 62 63 –1Papua New Guinea 325 7.01 143 –57 200 0 183 215 –32Philippines 4,459 5.43 4,164 2,869 1,222 73 295 855 –560Thailand 9,452 6.34 3,444 7 3,745 –308 6,008 449 5,559Vietnam 2,621 9.94 1,994 287 1,800 –94 627 733 –105

South Asia 14,702 2.90 11,110 3,971 4,662 2,477 3,592 2,772 820Bangladesh 1,004 2.37 118 –28 135 11 886 892 –6India 8,506 2.23 8,307 2,839 3,351 2,116 199 –25 224Nepal 323 6.56 12 –11 23 0 311 311 0Pakistan 3,058 4.88 2,097 1,132 713 252 961 353 608Sri Lanka 984 6.52 575 47 430 98 410 400 10

Latin America and 115,987 5.74 118,918 47,399 61,573 9,947 –2,931 3,000 –5,931the CaribbeanArgentina 19,751 6.08 19,835 10,954 6,645 2,236 –84 –52 –32Bolivia 1,320 16.54 812 210 601 0 508 511 –3Brazil 41,620 5.07 43,377 19,890 19,652 3,835 –1,757 –229 –1,528Chile 9,296 12.06 9,637 3,734 5,417 486 –340 64 –405Colombia 9,780 10.21 10,151 4,053 5,982 116 –371 79 –449Costa Rica 85 0.89 104 47 57 0 –19 –40 21Dominican Republic 357 2.37 401 –4 405 0 –45 2 –47Ecuador 888 4.49 829 252 577 0 58 48 11El Salvador 330 2.93 61 50 11 0 269 118 151Guatemala 373 2.10 166 76 90 0 207 197 10Jamaica 166 4.02 377 240 137 0 –211 –64 –147Mexico 15,998 3.97 20,533 6,003 12,477 2,052 –4,534 –21 –4,513Panama 1,572 17.90 1,443 411 1,030 2 129 11 118Paraguay 480 5.02 273 23 250 0 207 139 68Peru 4,279 6.70 3,094 373 2,030 692 1,184 353 831Trinidad and Tobago 53 0.89 96 –244 340 0 –44 4 –48Uruguay 828 4.15 632 471 160 2 196 0 196Venezuela 6,039 6.90 6,282 765 5,087 429 –243 1 –244

Europe and Central Asia 59,052 4.66 49,874 22,753 22,314 4,808 9,178 5,713 3,465Armenia 185 11.33 51 0 51 0 134 114 20Azerbaijan 713 19.52 658 8 650 0 55 86 –31Belarus 238 1.05 169 –31 200 0 70 26 43Bulgaria 621 6.16 569 –59 498 130 52 123 –71Czech Republic 1,804 3.47 1,818 516 1,286 16 –14 89 –103Estonia 404 8.63 347 79 266 1 58 45 13Georgia 183 3.73 50 0 50 0 133 128 5Hungary 2,567 5.61 2,605 –1,284 2,079 1,810 –39 87 –125Kazakhstan 2,636 11.89 2,158 786 1,321 50 478 90 389Kyrgyz Republic 236 13.40 50 0 50 0 186 168 18Latvia 670 12.12 559 12 521 26 111 58 53Lithuania 765 7.98 637 283 355 0 128 69 59Moldova 311 16.59 257 196 60 0 54 42 12Poland 7,077 5.22 6,787 934 4,908 944 291 352 –61Romania 2,980 8.55 2,274 1,059 1,215 0 706 167 539Russian Federation 15,149 3.39 12,453 5,006 6,241 1,206 2,696 278 2,418Slovak Republic 1,343 6.90 1,074 862 165 48 269 195 74Tajikistan 101 9.16 20 0 20 0 81 70 11Turkmenistan 878 21.95 847 762 85 0 31 22 9Turkey 12,211 6.43 12,221 10,840 805 577 –10 –176 166Ukraine 1,496 3.01 1,419 796 623 0 77 81 –4Uzbekistan 476 1.95 435 150 285 0 41 128 –87

165


Table A2.10 Long-term net resource flows to developing countries, 1997 (continued)(millions of U.S. dollars)

Private Official

OfficialPercentage Net debt development

Total of GDP Total flows FDI Portfolio Total assistance Other

Middle East and 6,921 1.32 8,120 493 5,368 2,259 –1,199 3,669 –4,868 North AfricaAlgeria –391 –0.83 –543 –557 7 8 152 54 98Egypt, Arab Rep. 3,521 4.66 2,595 –109 891 1,813 926 1,067 –141Iran, Islamic Rep. –4,297 –3.82 –303 –353 50 0 –3,993 243 –4,236Jordan 620 8.84 61 –31 22 70 559 478 81Morocco 818 2.44 1,303 –141 1,200 243 –484 268 –752Oman 102 0.65 118 –9 90 38 –16 –10 –6Syrian Arab Rep. –108 2.12 69 –11 80 0 –176 104 –281Tunisia 1,141 6.03 903 587 316 0 239 92 146Yemen, Rep. 83 1.47 –138 0 –138 0 221 214 7

Sub-Saharan Africa 18,792 5.49 6,674 –56 5,222 1,507 12,118 13,135 –1,018Angola 247 3.28 –24 –374 350 0 271 318 –47Botswana 87 1.73 95 –5 100 0 –8 23 –31Côte d’Ivoire 64 0.62 –91 –436 327 18 154 382 –227Cameroon 171 1.87 16 –29 45 0 155 307 –153Ethiopia 530 8.30 28 23 5 0 501 466 35Gabon –65 –1.26 –105 –5 –100 0 40 115 –75Ghana 671 9.74 203 27 130 46 468 487 –19Kenya 73 0.69 –87 –119 20 12 160 300 –140Madagascar 792 22.33 13 –1 14 0 779 721 58Nigeria 1,032 2.59 1,285 –258 1,539 4 –253 39 –293Senegal 406 8.96 44 14 30 0 362 398 –36Sudan 140 1.39 0 0 0 0 140 135 5Zambia 395 10.05 79 9 70 0 316 389 –73Zimbabwe 285 3.32 32 –48 70 10 253 207 46

Source: World Bank data.

Figure A2.10b Change in share of private long-term flows, 1990–97Percent120

100

80

60

40

0

20






Figure A2.10a Ditribution of long-term net resource flows, 1997Percent100

80

60

40

0

20

–40

–20






Official

Private

166


1965 1970 1975 1976 1977 1978 1979 1980 1981 1982 1983

G–5 unit value index of 21.6 25.1 45.2 45.8 50.4 57.9 65.6 72.0 72.3 71.2 69.5manufacturesa

LIBORb 5.0 8.9 7.7 6.1 6.4 9.2 12.2 14.0 16.7 13.6 9.9

Commodity price indexes,current dollar terms

Weights(1990=100) (percent)Petroleum 6 5 46 51 55 57 135 161 155 143 130Nonfuel commodities 40 44 75 88 108 101 116 126 108 95 103

Agriculture 69.1 42 46 81 99 127 116 129 138 118 103 112Food 29.4 42 47 101 87 90 99 113 139 123 97 105Beverages 16.9 46 57 81 156 267 198 207 181 145 147 156Raw materials 22.8 37 36 55 72 72 76 93 105 90 80 88

Metals and minerals 28.1 38 41 53 61 66 68 85 95 83 75 82Fertilizers 2.7 39 30 158 76 75 73 100 129 122 105 98

Commodity prices,current dollars UnitsAgricultureCocoa cents/kg 37 67 125 204 379 340 329 260 208 174 212Coffee cents/kg 100 115 144 315 517 359 382 347 287 309 291Tea cents/kg 100 83 114 119 195 141 151 166 147 154 210Sugar cents/kg 5 8 45 26 18 17 21 63 37 19 19Banana $/mt 159 165 247 257 275 287 326 379 401 374 429Beef (cents/kg) 88 130 133 158 151 214 288 276 247 239 244Wheat $/mt 59 55 149 133 103 128 160 173 175 160 157Rice $/mt 119 126 341 235 252 346 313 411 459 272 257Maize $/mt 55 58 120 112 95 101 116 125 131 109 136Coconut oil $/mt 348 397 394 418 578 683 985 674 570 464 730Palm oil $/mt 273 260 434 407 530 600 654 584 571 445 501Soybean oil $/mt 270 286 563 438 580 607 662 598 507 447 527Soybeans $/mt 117 117 220 231 280 268 298 296 288 245 282Cotton cents/kg 63 63 116 169 155 157 169 205 185 160 185Rubber cents/kg 50 41 56 77 81 99 126 142 112 86 106

OtherLogs $/cm 35 43 68 92 93 97 170 196 155 146 138Sawnwood $/cm 157 175 223 264 265 272 366 396 349 339 328Urea $/mt .. 48 198 112 127 145 173 222 216 159 135

Metals and mineralsCopper $/mt 1,290 1,413 1,237 1,401 1,310 1,367 1,985 2,182 1,742 1,480 1,592Aluminum $/mt 474 556 797 896 1,050 1,088 1,230 1,456 1,263 992 1,439Nickel $/mt 1,735 2,846 4,570 4,974 5,203 4,610 5,986 6,519 5,953 4,838 4,673Gold ($/toz) 35 36 161 125 148 193 307 608 460 376 423Phosphate rock $/mt 13 11 67 36 31 29 33 47 50 42 37Steel products index

1990=100 25 31 52 54 53 68 76 79 82 71 67

EnergyCrude petroleum $/bbl 1.4 1.2 10.4 11.6 12.6 12.9 31.0 36.9 35.5 32.7 29.7Coal $/mt .. .. .. .. 33.4 39.6 35.4 43.1 56.5 52.2 44.5

Figure A2.11a Price indexes relative to manufacturers unit value index, 1985–99Index (1990=100)

300

250

200

150

100

50

01985 1986 1985 1988 1989 1991 1993 1995 1997 Jan.-Oct.

199919981996199419921990

Coffee

Copper

Wheat

Petroleum

a. Unit value index in U.S. dollar terms (1990=100) of manufactures exported from the G–5 countries (France, Germany,Japan, United Kingdom, and United States) weighted by the country’s exports to developing countries.

Table A2.11 Manufactures unit value, LIBOR, and commodity prices, selected years, 1965–99

167


Figure A2.11b Price indexes relative to manufacturers unit value index, 1985–99Index (1990=100)300

250

200

150

100

50

01985 1986 1985 1988 1989 1991 1993 1995 1997 Jan.-Oct.

199919981996199419921990

Beverages

Food

Raw MaterialsMetals and minerals

Jan.-Oct.1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

68.1 68.6 80.9 88.8 95.3 94.7 100.0 102.2 106.6 106.3 110.2 119.2 114.2 108.4 104.2 103.6

11.3 8.6 6.8 7.3 8.1 9.3 8.4 6.1 3.9 3.4 5.1 6.1 5.6 5.9 5.6 5.6

125 119 63 79 64 78 100 85 83 74 69 75 89 84 57 73104 91 92 93 111 107 100 95 92 91 112 122 115 118 99 88117 100 103 99 110 106 100 98 94 99 123 131 125 129 108 93107 86 77 84 107 108 100 99 100 99 107 117 124 116 105 89175 164 194 135 140 114 100 93 77 83 148 151 126 171 141 10787 71 70 90 91 97 100 99 98 110 126 135 127 114 87 8874 70 65 78 114 111 100 89 86 74 85 102 89 90 76 7398 89 89 94 109 106 100 102 96 84 93 104 120 120 122 116

240 225 207 199 158 124 127 120 110 112 140 143 146 162 168 118319 323 429 251 303 239 197 187 141 156 331 333 269 417 298 222274 175 166 165 158 182 206 167 160 161 149 149 166 206 205 18211 9 13 15 22 28 28 20 20 22 27 29 26 25 20 14

370 380 382 393 478 547 541 560 473 443 439 445 470 503 492 429227 215 209 239 252 257 256 266 245 262 233 191 179 186 173 182152 136 115 113 145 169 136 129 151 140 150 177 208 159 126 113232 197 186 215 277 299 271 293 268 235 268 321 339 303 304 252136 112 88 76 107 112 109 107 104 102 108 123 166 117 102 91

1,155 590 297 442 565 517 337 433 578 450 608 670 752 657 658 744729 501 257 343 437 350 290 339 394 378 528 628 531 546 671 451724 572 342 334 463 432 447 454 429 480 616 625 552 565 626 438282 224 208 216 304 275 247 240 236 255 252 259 305 295 243 202179 132 106 165 140 167 182 168 128 128 176 213 177 175 144 12196 76 81 98 118 97 86 83 86 83 113 158 139 102 72 61

157 122 139 202 201 191 177 191 210 390 308 256 252 238 162 185352 307 329 401 402 485 533 553 607 758 821 740 741 664 484 592171 127 80 96 132 108 131 151 123 94 131 194 187 128 103 78

1,377 1,417 1,374 1,783 2,602 2,848 2,662 2,339 2,281 1,913 2,307 2,936 2,295 2,277 1,654 1,5381,251 1,041 1,150 1,565 2,551 1,951 1,639 1,302 1,254 1,139 1,477 1,806 1,506 1,599 1,357 1,3314,752 4,899 3,881 4,872 13,778 13,308 8,864 8,156 7,001 5,293 6,340 8,228 7,501 6,927 4,630 5,610

360 318 364 446 437 381 383 362 344 360 384 384 388 331 294 27738 34 34 31 36 41 41 43 42 33 33 35 39 41 43 44

70 61 62 72 94 106 100 99 88 91 93 107 96 89 75 68

28.6 27.2 14.4 18.2 14.7 17.8 22.9 19.4 19.0 16.8 15.9 17.2 20.4 19.2 13.1 16.848.6 46.6 43.9 36.2 37.1 40.5 41.7 41.5 40.6 38.0 36.5 39.2 37.2 36.4 34.4 33.2

b. London interbank offer rate on six-month U.S. dollar deposits. For detailed descriptions of the price series, see the website http://www.worldbank.org/prospects/pinksheets.

Table A2.11 Manufactures unit value, LIBOR, and commodity prices, selected years, 1965–99 (continued)

168


Documents

Global Economic Prospects 2000