33799 Stagflation, Savings, and the State Perspectives on the Global Economy Edited by Deepak Lal and Martin Wolf Published for The World Bank Oxford University Press Oxford University Press NEW YORK OXFORD LONDON GLASGOW TORONTO MELBOURNE WELLINGTON HONG KONG TOKYO KUALA LUMPUR SINGAPORE JAKARTA DELHI BOMBAY CALCUTTA MADRAS KARACHI NAIROBI DAR ES SALAAM CAPE TOWN Copyright © 1986 by the International Bank for Reconstruction and Developmentffhe World Bank 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Manufactured in the United States of America First printing February 1986 The World Bank does not accept responsibility for the views expressed herein, which are those of the authors and should not be attributed to the World Bank or to its affiliated organizations. Library of Congress Cataloging in Publication Data Main entry under title: Stagflation, savings, and the state. (A World Bank research publication) "Published for the World Bank." Includes bibliographical references and index. 1. Unemployment-Effect of inflation on-Addresses, essays, lectures. 2. Debts, External-Developing Countries-Addresses, essays, lectures. 3. Developing countries-Economic conditions-Addresses, essays, lectures. I. Lal, Deepak. II. Wolf, Martin. III. International Bank for Reconstruction and Development. IV. Series. HG229.S74 1986 330.9172'4 85-31981 ISBN 0-19-520496-4 Foreword THIS BOOK contains the background papers prepared for the World Bank's World Development Report 1984 and distilled into the section of that report entitled "Recovery or Relapse in the World Economy?" As this title suggests, the report sought to put in historical and analytical perspective the most recent recession in industrial countries, which un like its predecessor was transmitted to developing countries. The report examined the causes of the stagflation that had plagued the industrial world in the 1970s, and its effects on developing countries, by asking whether the steady and sustained global growth of the 1950s and 1960s could be resumed. In answering this question, the report marked an extension in the scope of the Bank's annual analysis of the global economy. With the increasing integration of the world economy through trade and capital flows, the determinants and effects of the domestic policies of industrial economies were found to be increasingly relevant to the economic prospects of developing countries. By contrast, in the past the primary focus of the Bank's global analysis had been on the international economic policies of industrial countries, which impinge on the economic environment world wide, as well as on the domestic policies that determine growth in the developing countries. While continuing to emphasize the importance of domestic policies in the developing countries, the 1984 report argued that the apparent problems of the global economy since the oil shock of 1973-sporadic growth, incipient inflationary pressures, rising protec tionism, and burdensome international debt-are partly the result of a secular deterioration in the economic performance of industrial coun tries. The causes of this deterioration are similar to those that past World Development Reports have emphasized in explaining the relatively poor performance of many (but by no means all) developing countries. These are the rigidities in the workings ofthe price mechanism, which have been induced by public policy in many industrial and developing economies, and the macroeconomic imbalances reflected in the management of the public finances, which have increasingly led to a global problem of undesirable fiscal deficits. The second of these causes of global disorder is new; with the increasing integration of the world capital market in the v vi FOREWORD 1970s, the actions of national public sectors can affect the global market for savings and investment and the terms on which they are mediated. Because these underlying problems of the global economy have by no means been satisfactorily resolved, the subject of this volume-the crisis of public policy in both industrial and developing countries, which the editors believe to be continuing-remains of considerable current rele vance. A proper understanding of the processes at work and of their long-term consequences is important in changing the perceptions on which policies are based, and the papers in this book should therefore help to inform discussions of these continuing problems of the global economy. Many of the papers in the book summarize the historical record and the theoretical debates on the economic performance and policies of indus trial and developing countries; they should be of considerable general interest. Others are based on original research done at the World Bank and should interest the research community worldwide. Their integration into this volume reflects the two functions that research performs in the Bank, providing both an intermediate input into the Bank's own opera tions as a policy adviser and an important final output. Although these background papers are being published for the Bank, the views expressed and the methods explored should not necessarily be considered to repre sent the Bank's views or policies. Rather, they are offered as a modest contribution to the great discussion on the best means of promoting worldwide economic development. ANNE O. KRUEGER Vice President, Economics and Research The World Bank Contents Contributors xi Introduction Deepak Lal and Martin Wolf 3 The Overall Perspective 3 The Historical Perspective 5 Stagflation and Public Finance in Industrial Countries 5 The Transmission Mechanisms 9 Developing Countries in Crisis 10 The Global Balance 12 1. Developing Countries in the 1930s: Possible Lessons for the 1980s Angus Maddison 15 The Aggregate Dimensions of the Depression 16 Depression-Induced Changes in the World Economic Order 18 The Export Shock for Developing Countries 19 Service Payments 23 Perverse Capital Flows and Their Impact 25 The Income Effect of the Export Shocks 29 The Recovery Phase, 1932-37 30 The Impact of Policy in Developing Countries 31 Lessons from the 1930s 37 Statistical Appendix 39 Notes 44 References 44 2. The Slowdown of the World Economy and the Problem of Stagflation Gottfried Haber/er 48 The Cost of Stagflation and Disinflation 52 The Keynesian Position on Inflation and Stagflation 54 Monetarism and Rational Expectations 57 The Traditional Conservative, or Classical, Position 62 The Slowdown of Productivity and GNP Growth since 1970 70 The International Monetary System in the Postwar Period 74 vii viii CONTE""S Policy Implications 79 Summary 83 Notes 84 References 87 3. Systemic Risk and Developing Country Debt Robert Z. Lawrence 91 Objective Risk 93 Subjective Risk 96 Risk Aversion 97 Experience during the Crisis: A Cause for Reappraisal? 98 Future Innovations 99 Notes 102 References 102 4. A Description of Adjustment to External Shocks: Country Groups Pradeep K. Mitra 103 Methodology 104 External Shocks 104 Modes of Adjustment 106 Patterns of Adjustment 107 Conclusions 112 Appendix A. Data 112 Appendix B. Composition of Groups 113 Notes 113 References 114 5. The Oil Syndrome: Adjustment to Windfall Gains in Oil-Exporting Countries Alan Gelb 115 Dimensions of the Oil Windfall 116 Use of the Windfall: The Fiscal Response 117 Some Consequences of Increased Domestic Expenditures 119 After 1981: The End of the Oil Boom? 127 Conclusions 128 Notes 129 References 129 6. The Debt Crisis in Latin America Larry A. Sjaastad, Aquiles Almansi, and Carlos Hurtado 131 What Triggered the Crisis? 131 Latin American Public Finance during the 1970s 135 Money Creation and Government Finance 147 CONTENTS ix Gross versus Net Debt and Implicit versus Actual Amortization 155 Post-Debt Crisis Adjustment 165 Trading the Way Out 180 Notes 181 References 181 7. Government Deficits, the Real Interest Rate, and Developing Country Debt: On Global Crowding Out Deepak Lal and Sweder van Wijnbergen 182 Trends in OECD Real Wages, Social Expenditures, and Fiscal Deficits 183 The Financing of Social Expenditures, Budget Deficits, and Crowding Out 188 The Global Capital Market and Real Interest Rates 193 Third World Debt, Infrastructural Investments, and Fiscal Deficits 200 A Model of Global Crowding Out 206 A Summary of the Global Interactions 215 Appendix A. Rules for Optimal Foreign Borrowing 223 Appendix B. Data Description 226 Notes 232 References 235 8. Debt, Deficits, and Distortions Deepak Lal and Martin Wolf 239 Underlying Problems of the Global Economy 241 Recession in Retrospect 252 Prospects and Policies for Sustained Recovery 270 References 290 Appendix 1. DEeD Deficits, Debt, and Savings Structure and Trends, 1965-81: A Survey of the Evidence Leonardo Hakim and Christine Wallich 292 Trends in Government Revenues 292 Expenditures and Expenditure Trends 297 Controlling Future Growth 310 The Deficits 326 Trends in Savings 332 Uses of Savings 344 Deficits and Savings 350 Notes 357 References 357 x CONTENTS Appendix 2. rCORS, Savings Rates, and the Determinants of Public Expenditure in Developing Countries Constantino Lluch 361 ICORS and Savings Rates in Africa 361 ICORS and Savings Rates in Asia 364 ICORS and Savings Rates in Latin America 380 Public Expenditure and Its Composition 388 Notes 394 References 394 fud~ 3% Contributors AQUILES ALMANSI is an assistant professor of economics at the University of Michigan. ALAN GELB is an economist in the Productivity Division of the Develop ment Research Department at the World Bank. GOTIFRIED HABERLER is a professor emeritus at Harvard University and a resident scholar at the American Enterprise Institute in Washington, D.C. LEONARDO HAKIM is a senior economist at the Institute of International Finance in Washington, D.C. CARLOS HURTADO is a professor at the Instituto Tecnologico Automimo de Mejico. DEEPAK LAL is a professor of political economy at the University of London and the research administrator, economics and research, at the World Bank. ROBERT Z. LAWRENCE is a senior fellow at the Brookings Institution in Washington, D.C. CONSTANTINO LLucH is chief of the Labor Markets Division, Develop ment Research Department at the World Bank. ANGUS MADDISON is a professor of economics at the University of Gro ningen. PRADEEP K. MITRA is a senior economist in the Resource Mobilization and Public Management Division of the Country Policy Department at the World Bank. LARRY A. SJAASTAD is a professor of economics at the University of Chicago and the Graduate Institute of International Studies in Geneva. SWEDER VAN WIJNBERGEN is a senior economist in the Trade and Adjust ment Policy Division of the Country Policy Department at the World Bank and a research fellow at the Center for Economic Policy Research in London. xi xii CONTRIBUTORS CHRISTINE WALLICH is a senior economist in the Financial Policy and Analysis Department at the World Bank. MARTIN WOLF is the director of studies at the Trade Policy Research Center in London. Stagflation, Savings, and the State Introduction Deepak Lal and Martin Wolf IN JUNE 1984, as this introduction is being written, the world economy seems to be recovering from the deepest global recession in fifty years. Will the recovery be sustained? Can there be a renewed period of relatively stable and steady growth such as characterized the first two postwar decades, which are already wistfully being called a Golden Age? How can the world economy be made to recover its elan? Part I of World Development Report 1984 attempted to answer these questions by considering the underlying causes of the stagflationary crisis that has beset the industrial countries since the first oil shock and was transmitted in the late 1970s to the developing countries. The latter weathered the first emanation of this crisis fairly well after the 1973 oil price shock. Many middle-income oil importers were able to cope by means of countercyclical borrowing on international financial markets awash with newfound liquidity based on the financial surpluses of the oil producers. This commercial borrowing represented the reopening of international financial markets to developing countries after their defaults in the 1930s. This borrowing, however, has come to haunt many of the developing countries, as well as the international financial system, in the form of a global "debt crisis." The present book contains a set of background papers prepared for Part I of World Development Report 1984. Chapter 8, by Lal and Wolf, more fully states the arguments advanced in the Report. The rest of the back ground papers analytically and empirically validate these arguments. Each paper is relatively self-contained, although in content the chapters inevitably overlap somewhat. The Overall Perspective The 1960s were marked by confident predictions that technocratic economic advances had banished the business cycle. The 1970s have forcefully demonstrated that these predictions just reflected economists' hubris. The novel feature of the cycles in the late 1960s and 1970s has been 3 4 INTRODUCTION the persistence of inflation even during downturns and even with the progressive worsening of President Carter's "misery" index-the postu lated tradeoff between the rates of unemployment and inflation during each cycle. Furthermore, there has been a decline in productivity and net savings in industrial countries, a marked slowing of world trade, and a boom-bust cycle in commercial bank lending to developing countries. The most recent recession in industrial countries-which was very deep partly because of concerted attempts to control inflation-seems to have dampened the stagflationary pattern. Nevertheless, fears are being ex pressed that the world economy may as a result be falling into another great depression. Chapter 1 in this book, by Maddison, attempts to determine whether this is the case. The following chapters seek to identify the causes of the secular deterioration in economic performance in indus trial countries and the modes of transmission of the resulting stagflation to developing countries through the two major global channels of trade and capital flows. If we assume that business cycles will persist, the important questions for public policy in any country are: (a) whether there are feasible methods for smoothing the cycle and (b) what sort of economic environ ment should be established to allow economic agents to cope most efficiently with the volatile world economy in the sense of promoting the trend growth of productivity and output. Answers to the first question are at present the subject of fierce theoretical controversy. Chapter 2, by Haberler, assesses the various theoretical debates on macroeconomic policy. The macroeconomic poli cies adopted can have important effects on an economy's microeconomic flexibility and will therefore help condition answers to the second set of public policy questions. The trend growth rates of output and real in comes of economies are in large measure determined by the efficiency of their resource use and the incentives provided by their domestic eco nomic environment for productivity and thrift. The link between inappropriate macroeconomic policies and the mi croeconomic rigidities and distortions which were their inevitable by product is a major theme of chapter 8, by Lal and Wolf. It argues that the continuing crisis in the world economy has two deep-seated causes, of which the stagflation and rising protectionism in industrial countries, as well as the debt crisis, are dramatic symptoms. These causes are: (a) the microeconomic rigidities that were steadily built into many economies both developed and developing-partly as a result of inappropriate public policies, and (b) the management of the public finances, which has be come a global problem with the continuance of problematic fiscal deficits in many developing countries and the likely emergence of structural DEEPAK LAL AND MARTIN WOLF 5 budget deficits in the United States, and possibly in some other industrial countries, in the future. The role of protectionism and inappropriate public interventions that lead to rigidities in wages and prices has often been discussed in the context of the economic performance of both developed and developing countries. The role of public finances is a more novel and controversial topic, as it has only recently become apparent that, with the integration of international financial markets in the 1970s, there is now a global market for savings and investment. The actions of national public sectors in economies that bulk large in the world economy can affect the interna tional uses of world savings and the terms on which they are intermedi ated. In this global framework the public sectors of both developed and developing countries are increasingly competing with each other and with private agents for the available global savings. The problems of stagflation and Third World debt and the reactions to them can be seen, then, to be due in part to an emerging phenomenon: the "crowding out" of global savings by the public sectors in both developed and developing countries. Most of the chapters in the present volume are concerned with providing the empirical or analytical basis for this view. The Historical Perspective Chapter 1, by Maddison, sets the 1980-83 recession in historical per spective, comparing its effects with those of the Great Depression on some major developing countries (accounting for 70 percent of the population of the developing world). He shows that even though the two periods were marked by some similar phenomena-recession, a fall in world trade, the growing problem of debt, and a reverse flow of capital from developing to developed countries-their magnitude in recent years was far removed from that seen in the 1930s. Stagflation and Public Finance in Industrial Countries Chapter 2, by Haberler, and appendix 1, by Hakim and Wallich, are more directly concerned with the links between stagflation and public finance in developed countries. Haberler presents a survey of alternative theoretical perspectives on the problem of stagflation. He distinguishes between (a) the Keynesian view, which was dominant in the first two postwar decades and still continues to be influential; (b) a general mone tarist position, of which the so-called new macroeconomics based on 6 !NTRODUCTION rational expectations is a variant; and (c) a traditional conservative, or classical, position. The Keynesian position neglects inflationary expecta tions, being based on the notion of a Phillips curve that postulates a more or less permanent tradeoff between unemploy~ent and inflation. It emphasizes the importance of fiscal policy and the ability of governments to fine-tune the economy by skillful demand management. It deals with inflation by some form of incomes policy. This framework of thought is particularly unsuited, Haberler argues, to dealing with the problems of stagflation, being essentially "depression economics," or addressed to a time when wages and prices could be assumed to be constant. In an inflationary environment, incomes policies have not proved to be work able as a method for curbing wage inflation. The monetarist and rational expectations theorists, by contrast, em phasize the role of monetary policy, downgrade fiscal policy, reject fine tuning, incomes policy, and price controls, and emphasize the role of expectations. They reject the Phillips curve on the grounds that an ex tended period of inflation at a significant rate would ignite inflationary expectations, lead market participants to take anticipatory action, and undo the stimulatory effects of an inflationary macroeconomic policy on unemployment. The rational expectations view takes this approach to an extreme and argues that macroeconomic policies cannot affect the "real" economy even in the short run, as rational private agents will determine the consequences of any predictable government policy and will take anticipatory action. Only by acting in a surprising and unpredictable way can governments affect the real economy in the short run and even then not for long. Although this view seems extreme, argues Haberler, a variant identified by Fellner, called the "credibility approach," does have some merit. Fellner maintained that, if the government's policies-for instance, on disinflation-are credible, they can condition inflationary expectations and can thus mitigate the pain of disinflation. The monetarist position on fiscal policy is that loose fiscal policy and large deficits need not cause inflation if monetary policy is tight. Depending on the size of the deficits, however, interest rates could rise, which (a) could accelerate the velocity of circulation and lead to inflation even without an increase in money supply, and (b) could cause productive private investments to be crowded out by public expenditures at the cost of slowing down long run growth-hence the monetarist theorists' preference for combining monetary with fiscal restraint. The traditional, or classical, view, Haberler argues, agrees with the monetarist and rational expectations theories regarding the Phillips curve and regarding the role of monetary and fiscal policy and expectations, but stresses the importance of institutional rigidities-in particular, wage and DEEPAK LAL AND MARTIN WOLF 7 price rigidities, which are assumed away in the monetarist world of instantaneously or rapidly clearing markets. Haberler then cites the empirical evidence extending from the 1930s-when the New Deal period saw the first example of stagflation-to the present. This evidence sup ports the classical view as providing the best explanation for the stagfla tion of the 1970s and 1980s. Haberler also outlines the various explana tions provided for the productivity slowdown of the 1970s. Again, he finds that the basic causes are the increasing rigidities owing to inappropriate microeconomic interventions and the uncertainty caused by inappropri ate fiscal and monetary policies in the working of the price mechanism, rather than any fundamental underlying change in the technological bases of productivity growth. His survey of international money and exchange rates in the postwar period leads him to favor the existing floating-rate system and to counter various misconceptions about its workings as well as about the presumed current (1984) overvaluation of the dollar. The policy prescriptions that follow from Haberler's discussion are a tight monetary and fiscal policy and the removal of rigidities in the working of the price mechanism, particularly in the labor markets of industrial countries and especially those in Europe. This approach re quires both the promotion of free trade and the removal of various policy-induced distortions that have hardened the economic arteries of industrial economies. Appendix 1, by Hakim and Wallich, statistically summarizes on the basis of work done by the secretariat of the Organisation for Economic Co-operation and Development (OECO), the role and implications of the public sector's net expenditure policies in industrial countries. It docu ments the components of public expenditure and revenue that have led to the emergence oflarge actual (nominal) budget deficits in many industrial countries and asks to what extent they are structural in nature. Hakim and Wallich find that the growth of social expenditure on health and pensions is the major source of the dramatic growth of public expenditure in the OECD area in the 1960s and early 1970s. The increase was due largely to gains in the coverage and level of the health and pension benefits, with demographic factors (particularly the aging of the popula tion) playing a smaller role. By the end of the 1970s, however, benefit growth had slowed, as had coverage. After the turn of the century, demographic pressures could again make it hard for some countries to contain the growth of expenditures on health and pensions. At the same time, Hakim and WaUich show that it will be increasingly difficult to raise revenues in OECD countries, particularly by means of fiscal drag if tax systems continue to be indexed. Although the resulting budget deficits in the 1970s and 1980s had a controversially high cyclical element in OECD 8 INTRODUCTION estimates, there is a danger that structural deficits will emerge if the future growth of social expenditures is not contained, especially toward the end of the century. The appendix next summarizes the evidence on the decline in savings in the OECD area from about 1965. Though some of this decline is associ ated with the cyclical downturn, the major cause of the decline in savings is the increasing dissavings in the government sector (budget deficits), which have not been matched by corresponding increases in household or corporate savings. It is not clear, however, to what extent the decline in savings can be extrapolated into the future. If government budget deficits are reduced, inflation rates fall, and real interest rates remain positive, OECD savings may be expected to stabilize. It should be noted that there are certain omissions in the discussion of OECD savings, deficits, and inflation in this volume. The most important concerns the interaction of unindexed tax systems with variable inflation in altering the net returns to savers and borrowers on different physical and financial assets; the general effects of this interaction receive some discussion, however, in chapter 7. The second is the effect of inflation on the public finances themselves and the corresponding interpretation of the unadjusted budget deficits, as well as those of estimated deficits derived from various adjustments that are now increasingly being made by researchers. Some adjustment may need to be made to the nominal budget deficit; the analyst could have any of three different reasons for finding it of economic interest. For Keynesians, the relative size of the budget deficit is a measure of the net fiscal stance of the government, which needs to be judged by the conjunctural state of aggregate demand. If, as Haberler argues, this Keynesian framework is not very useful in formulating fiscal or monetary policies for these stagflationary times, adjustments to the nominal budget to judge the fiscal stance would not be very relevant. A second reason for looking at the budget deficit is to assess its impact on the public sector's net worth. A deficit, which implies an increase in the real public debt, will mean that the public sector's net worth (the differ ence between public assets and liabilities) will decline, a tendency that may be of concern in some contexts but not when we are trying to find a way out of stagflation. The third and for our purpose most relevant aspect of the budget deficit is its implications for financial policy and the effects of the changing government demand for private domestic savings: an increase in govern ment borrowing, if it occurs, has implications for interest rates, the money supply, and the division of financial flows between the private and public sectors. Acceptable inflation-adjusted national accounts suitable DEEPAK LAL AND MARTIN WOLF 9 for assessing the financial implications of deficits would be valuable but were not available for the OEeD countries as a whole. The Transmission Mechanisms Chapters 3 and 4, by Lawrence and Mitra, deal with the transmission mechanisms of shocks to developing countries that emanate from the policies and economic performance of industrial countries and are trans mitted through the trade and financial flows linking developed and de veloping countries. Lawrence is concerned with delineating the causes and nature of the current Third World debt crisis. He questions the popular view that this crisis is fundamentally a matter of liquidity, or solvency. The notion of insolvency is not applicable to nations, he argues, as they are likely to be unwilling to service their debts well before they are technically insol vent-in the sense that the present discounted value of the nation's assets is less than its liabilities. The resulting default decision will depend upon the costs and benefits of default rather than upon the country's net worth. The situation is, on the surface, clearly a liquidity crisis. Lawrence argues, however, that it is fundamentally due to an underestimation of the systemic risk associated with the conjunction of a number of events considered to have a low probability of occurring in the global economy between 1979 and 1983-the dramatic rise and fall in oil prices, unprec edentedly high and volatile real interest rates, large fluctuations in real exchange rates, the deepest global recession in fifty years, the largest slump in world trade in the postwar period, and the sharp decline in the global inflation rate. If these events had not conjoined to destabilize the world economy, studies suggest, developing countries could have sus tained and serviced their debt levels in the 1980s. Lawrence documents the various types of objective risks, particularly systemic risk (the possibility of a major series of synchronous losses for the lenders), that were underestimated. The error was not unreasonable, however, inasmuch as (a) developing countries seemed to have weath ered the 1974-75 recession well, maintaining and improving their export performance, which suggested that they might have been uncoupled from the global business cycle, and (b) the handling of the Herstatt and Franklin National bank failures suggested that the financial system could contain and absorb individual shocks. In addition, the subjective evalua tion of the risks of lending to developing countries was affected because of the expected response of policymakers. The praise that the interna tional banks received for recycling the financial surplus of the Organiza 10 INTRODUCTION tion of Petroleum Exporting Countries (OPEC) after the 1973 oil shock, and an assumption that they would be bailed out in the event of a default affecting the health of the global banking system, may have lowered their subjective estimates of the risk involved. Lawrence discusses various ways in which the systemic risk might be reduced, essentially through diversification. He notes that much of the debt was incurred to finance various public goods having no measurable returns and for which it is not possible to issue securities contingent upon income. He also argues that, given the relative success of cooperative solutions in the form of debt rescheduling that involves the banks, debtors, and international orga nizations, the past system based on bank rather than bond lending may have been relatively more stable, as banks are less concerned with transitory financing problems. Chapter 4, by Mitra, charts the trade links between developed and developing countries and in particular examines the question of efficient adjustment to external shocks. Mitra estimates the effects and responses to the external shocks suffered during the 1970s by developing countries in an open economy macroeconomic model for each of thirty-three countries over the 1963-81 period, assuming a structural break after 1973. The actual outcomes in dealing with exogenous shocks in the 1974-81 period for these countries are then compared with hypothetical outcomes that presuppose responses based on the countries' 1963-73 experience. The external shocks considered are (a) terms-of-trade effects, (b) the effects of global recession, and (c) net interest rate effects. The sum of those external shock effects during 1974--81 ranged from an unfavorable annual average of 7 percent to 9 percent of GNP to a favor able 10 percent. Of the thirty-three countries considered, twenty-four suffered adverse external shocks. Mitra then estimates the responses to the unfavorable shocks in terms of four modes of adjustment (and therefore of combinations): (a) trade adjustment, (b) domestic resource mobilization, (c) a slowing down of investment, and (d) additional external financing. Developing Countries in Crisis Chapter 5, by Gelb, and chapter 6, by Sjaastad and others, chart the nature of the economic crisis for two major sets of developing countries, namely the oil producers and the major Latin American debtors. Both chapters delineate the ways in which, for different reasons, the crisis essentially involves the public finances in these countries. It was brought about by unexpected shortfalls in public revenue (the case of the oil DEEPAK LAL AND MARTIN WOLF 11 exporters) and unexpected increases in expenditure on servicing the public external debt (the case of the Latin American debtors), which were in turn caused by the recession and its unexpectedly high level of interest rates as well as declining primary product prices. Gelb shows how the rise in oil revenues led to an unparalleled growth in the size and role of the public sector in virtually all oil-exporting countries. The windfall gains were channeled into industry---often large scale and capital-intensive industry-and into expanded programs of transfer payments and subsidies. The momentum of this vast increase in public expenditure proved difficult to control when oil revenues fell after 1981. The resulting need to curtail public expenditure led to a rapid deceleration in the non-oil part of these economies, to surplus domestic capacity, and to slack labor markets. The contraction was accentuated in some countries by private capital outflows. The real exchange rate appre ciation that had accompanied the earlier oil-based boom had led to disincentives to expansion of non-oil exports and had encouraged de pendence on imported intermediate and capital goods. When external circumstances worsened, it proved difficult to compress the previously expanded public expenditures-giving rise to fiscal deficits---or to reverse the relative price changes that were no longer appropriate. Many of the large-scale investments, for instance in steel, soured as the expectations of future global demand on which they were based worsened with the global recession. Microeconomic rigidities and unsustainable public finances were thus the fundamental causes of the economic distress of these countries in the global recession. Chapter 6 charts the paths to crisis and adjustment of five major indebted Latin American countries: Argentina, Brazil, Chile, Mexico, and Venezuela. The authors argue that the origins of Latin America's debt crisis precede the inability of the Mexican government to service its debt in August 1982. The implicated factors include a rapid and enormous growth of external debt with a shortening of its maturity structure and a sudden and large rise in dollar interest rates, accompanied by worldwide dollar deflation. As a result, the buildup of debt since 1973-in anticipa tion of oil price rises to finance ambitious public sector investments-and the change in the external environment in the early 1980s led to increases in debt service rates (interest plus amortization) that more than offset the earlier inflation-induced decline of the real value of their debt in the mid-1970s. The dollar prices of exports for many countries fell at the same time. The debt -service-to-debt ratio worsened (even though the level of debt in 1982 was not high by historical standards) without an accompany ing improvement in the ability of the countries to generate the requisite fiscal and trade surpluses to pay for debt service. 12 INTRODUCTION This adverse outcome, however, in turn reflected more deep-seated problems of overambitious public investment programs, which led to uncontrolled and growing fiscal deficits, financed most often by levying the inflation tax, and to trade policies with a bias against exports and against domestic savings. Low-cost foreign loans, obtainable when OPEC financial surpluses were being recycled after the first oil shock, helped to finance the actual and incipient fiscal deficits during the 1970s. As foreign loans slowed down, many of these countries resorted to the inflation tax once again. In coping with the crisis, governments have tried to cut back their fiscal deficits and have taken measures to restrict imports and to generate the fiscal and trade surpluses required to service their debt. The compression of imports has been an inefficient form of adjustment, as it has caused a fall in industrial output and employment, whereas exports have increased only slightly. Even if more efficient adjustment policies were followed under present capital market conditions, however, with nominal interest rates in excess of 10 percent and amortization rates in excess of 20 percent, the fiscal and trade adjustments that would be required are daunting. The normalization of world capital markets is therefore of importance if (given improved domestic policies) these countries are to find feasible ways of servicing their debts. The Global Balance Chapter 7, by Lal and van Wijnbergen, and appendix 2, by Lluch, are concerned with the global balance between the demand for and supply of savings. Lluch's appendix puts together data on incremental capital output ratios and savings rates for developing countries and considers how different regions have shifted from being net exporters or importers of capital in the postwar decades. For developing countries the major points that emerge are (a) the marked and sustained increase in domestic savings in the last three decades except in sub-Saharan Africa, where the upward trend was reversed in the 1970s, and (b) the extremely high variance and volatility of incremental capital-output ratios across coun tries and over time in the same country. For developed countries the data in appendix 1 show a declining trend in domestic savings and a change in the capital-exporting status of its major economy, that of the United States. Chapter 7 emphasizes how with the increased global financial integra tion there is a direct link between the actions and reactions of the public sectors in both developed and developing countries. In both sets there is DEEPAK LAL AND MARTIN WOLF 13 pressure for public expenditure increases. In developed countries the aging of the population will add to pressure for increased spending on health and social security. In developing countries the growth of popula tion intensifies pressure for increases in public expenditure on human and physical infrastructure. The financing of this public expenditure through taxation is posing problems in both sets of countries, and as a result incipient or actual fiscal deficits are being financed either through the inflation tax or through borrowing. The latter, in the increasingly inte grated world economy, puts upward pressure on world real interest rates and suggests the possibility of a global crowding out of expenditures private worldwide and public in developing countries-by the incipient structural deficits of developed countries. Lal and van Wijnbergen cali brate a three-region formal model of these global interactions with data for the 1970s. It seems to explain satisfactorily the terms-of-trade and real interest movements that have been observed in the last decade. The current debt crisis is seen as following essentially from a global fiscal crisis. The possibility then emerges that, unless "old age-related" public expenditures in developed countries can be contained, they will in creasingly crowd out at the margin the infrastructural developmental expenditures in developing countries that are required to raise the living standards of the latter's poor, young, and growing labor forces. Chapter 1 Developing Countries in the 1930s: Possible Lessons for the 1980s Angus Maddison LITERATURE on the world economic crisis of the 1930s is extensive and includes several respectable general surveys of the situation in the ad vanced countries (notably Arndt 1972; Haberler 1976; Hodson 1938; Kindleberger 1973; and Lewis 1949). Controversy is, of course, still lively about the causes of such a deep depression in the epicenter countries (Brunner 1981; Friedman and Schwartz 1963; and Temin 1976 on the United States; Balderston 1977; Falkus 1975; and Temin 1971 in Ger many), and none of the studies just mentioned has exploited the potential for comparative cross-country analysis that quantitative economic histo rians have made possible by reconstituting many of the basic modern cyclical indicators--GDP and its components, trade volumes, terms of trade, and payments flows. For the developing countries, the situation is much less favorable for sophisticated analysis. None of the older surveys that I have just cited gives adequate treatment to the situation in the Third World, and helpful recent attempts to fill the gap (Rothermund 1982; Thorp 1984) are constrained in cross-country analysis by lack of comparable data. The most ambitious cross-country study to use a standardized quantitative framework is that by Birnberg and Resnick (1975), which is only inciden tally concerned with the 1930s. Regional cross-country analysis has gone furthest in Latin America (Diaz Alejandro 1981, 1982), thanks in part to efforts to produce comparable quantitative data (EeLA 1976, 1978; Wilkie and Haber 1982). Although the data situation for the developing world is improving as research progresses, the basic weakness is the paucity and poor quality of the national accounts for prewar years. This chapter is therefore confined to the nine countries for which the basic data were either available or derivable. Fortunately these countries include almost 70 percent of the population of the developing world and may therefore be considered For a more detailed analysis of the issues discussed in this chapter, see Maddison (1985). 15 16 DEVELOPING COUNTRIES IN THE 1930s "representative," although they do not include any African countries. Even for this group, the available national account aggregates are based on GOP by industry of origin rather than by type of expenditure. Although inferential clues are available, it is difficult to provide hard or comparable evidence on movements in demand components except exports and imports.! It is also difficult to reach firm conclusions about the efficacy of policy, because fiscal, monetary, and balance-of-payments series that in princi ple can be reconstituted in comparable form are, for the present, often only sketchily available. Fortunately a growing number of scholarly country monographs, particularly from the Yale Growth Center, provide some basis for assessment. The Aggregate Dimensions of the Depression The world depression of 1929-32 was much bigger and more general than any earlier or later peacetime shock. Between 1929 and 1932, the aggregate GOP of the advanced countries fell 17.1 percent and world trade volume by 26.8 percent. In 1974-75, by contrast, the corresponding falls were only 0.6 percent and 5 percent, respectively. The 1981-82 recession in advanced countries was on the same order as that which occurred in 1974-75. The epicenters of depression were the United States and Germany, where the primary recession was exacerbated by a collapse of financial institutions. In the epicenter countries (and their immediate neighbors, that is, the United States, Canada, Germany, and Austria) the average peak-to-trough GOP decline was 24.6 percent (see table 1-1). The reces sion in the other advanced countries averaged only 7.5 percent. Because of the huge size of the United States, the weighted average decline was 17.1 percent as compared with a nonweighted average of 11.7 percent. In the nine developing countries in our sample, the average peak-to trough GDP decline was 12.2 percent within the 1929-34 period, with an average of 15.8 in Latin America and 4.9 percent in Asia. The weighted average peak-to-trough decline for the Third World was very small, at 3.6 percent, because of the huge weight of China, and because the fluctua tions in the Chinese economy differed in timing from those elsewhere. There were major changes in terms of trade in the 1929-32 period, which meant that the fall in real income in the Third World was invariably worse than the fall in GOP. For my sample of nine developing countries, the average impact of the 1929-32 worsening in terms of trade was a 4.5 percent fall in real income to be added to the fall in GDP. ANGUS MADDISON 17 Table 1-1. Amplitude of Recessions and GDP Growth Experience, 1929-38 Maximum peak- trough percentage Annual average fall in GDP compound growth Country (annual data) rate, 1929-38 Advanced countries Australia -8.2 1.9 Austria -22.S -0.3 Belgium -7.9 0.0 Canada -30.1 -0.2 Denmark -2.9 2.2 Finland -6.S 3.8 France -11.0 -0.4 Germany -16.1 3.8 Italy -6.1 1.4 Japan -7.2 4.7 Netherlands -9.1 0.3 Norway -8.3 3.1 Sweden -9.2 2.3 Switzerland -8.0 0.6 United Kingdom -S.O 1.9 United States -29.S -0.7 Unweighted average -11.7 1.5 Weighted average -17.1 1.0 Developing countries Argentina -13.8 1.3 Brazil -7.3 4.6 Chile -26.5 1.0 China -8.7 0.6 Colombia -2.4 3.S India -1.8 0.2 Indonesia -4.1 l.S Mexico -19.0 1.7 Peru -2S.8 2.S Unweighted average -12.2 1.9 Weighted average -3.6 0.8 Sources: Advanced countries from Maddison (1982); developing countries from statisti cal sources in References. 18 DEVELOPING COUNTRIES IN THE 1930s Depression-Induced Changes in the World Economic Order The recession was very deep and general because normal international transmission mechanisms were reinforced by the collapse of the liberal world economic order. This collapse had three (highly interrelated but analytically distinguishable) components. The international monetary system disintegrated. The surplus coun tries with big gold reserves (France and the United States) did not take conjunctural policy action to help nations in deficit. No country acted as lender of last resort, as the United Kingdom had done before 1913. No international institutions were available to provide liquidity as they do today. The demise of the gold standard was a messy and prolonged three-stage affair, involving sterling in 1931, the dollar in 1933, and the gold bloc (France and the Netherlands) in 1936. The breakdown itself was not a major tragedy. The economic damage was caused by the unneces sary deflation involved in adherence to the classical rules for debitors. A general decision to float all major currencies at an early stage would have been less deflationary. The second complication resulted from the quarrels of the big countries over war debts and reparations. In 1929 the total stock of private capital invested abroad was about $50 billion. 2 In addition there were govern ment war debts to the United States, the United Kingdom, and France that amounted to $20 billion and reparations due from Germany that had originally been fixed at $31.5 billion in 1921 and were twice scaled down, in 1924 and 1929. After a one-year moratorium in 1931, these govern ment-to-government claims became permanently delinquent in 1932. Acrimony among the major Western powers precluded mutual help with liquidity problems, and quarrels over intergovernment debt pre pared the way for massive default on private obligations and cessation of equilibrating international capital flows. In the 1930s the major capital exporters of the advanced world became capital importers. The trade problems of deficit countries, particularly in the Third World, were compounded by lack of credit. The third element of the international order that disintegrated was the liberal trade system. It is true that in 1929 tariffs were somewhat higher than they had been in 1913, but they were the only important trade restriction and were generally applied on a nondiscriminating basis. The United States gave an unfortunate lead to protectionism with the Haw ley-Smoot tariff of 1930 without having the excuse of the balance-of payments problems that drove other countries to such action. In 1930-32, the situation with regard to trade restrictions changed completely. Vir ANGUS MADDISON 19 tually all countries raised tariffs. New discriminatory trading blocs were created, and tariffs were reinforced by quantitative restrictions and ex change controls that were also applied in a discriminatory way. The most important of the new arrangements was the system of imperial preference devised in Ottawa in 1932 and supported by the sterling area payments system. Hitherto the British had not had tariff preferences in their colonies. The French had had colonial tariff preferences since the 1890s, and in 1931 these were reinforced and a new system of compensation funds and quotas was established to help Indochinese rice, tea, and corn as well as African coffee. In 1934-35 these tariff preferences were re newed and strengthened in the Imperial Economic Conference held in Paris. As a result France concentrated even more of its trade on the franc area. Japan, Germany, and Italy also built up new autarkic colonial and semicolonial trading blocs. Preferential systems favored some developing countries and damaged others, mainly in Latin America. The Export Shock for Developing Countries In 1929 developing countries were generally rather open economies. The exporting interests were strong politically in Latin America, where relatively high tariffs existed as much for revenue as for protection, and most African and Asian countries were colonies with low tariffs. Most of the latter were on the gold standard, and their initial response to pay ments difficulties was to play by the rules of the game and deflate. Their implicit long-term growth strategies involved growth diffusion via the liberal world order, without much governmental intervention. African and Asian countries were therefore rather vulnerable to external shock, and all the evidence suggests that the recessionary forces originated in the developed world. In China there was a very bad harvest in 1934 and war in 1937, but such autonomous cyclical causality was rare. Table 1-2 shows the relative size of the trade shock for my sample of developing countries and for the advanced countries on a peak-to-trough basis. Because the timing ofthe trough varied between 1931 and 1933, the peak-to-trough export volatility of table 1-2 is bigger than that for 1929 32 shown in table 1-9, but the same basic characteristics stand out. The faU in volume of commodity exports in both developing and advanced countries averaged somewhat more than a third, but in developing coun tries this drop was compounded by worsening terms of trade, which meant that the purchasing power of commodity exports over imports fell by more than half, as compared with just under a third in the developed world. The big European countries (France, Germany, the United King 20 DEVELOPING COUNTRIES IN THE 1930s Table 1-2. Fluctuations in Export Volume, Export Purchasing Power, and Import Volume, 1929-38 Purchasing Export power Import Country volume of exports volume Developing countries Argentina -35.8 -41.9 -53.2 Brazil -31.1 -45.6 -63.8 Chile -71.2 -84.5 -83.0 China -48.6 -64.8 -50.5 Colombia -12.5 -36.6 -63.1 India -30.6 -39.2 -30.3 Indonesia -15.2 -40.4 -47.5 Mexico -41.5 -64.8 -61.1 Peru -29.7 -57.1 -63.3 Average -35.1 -52.8 -57.3 Advanced countries Australia -6.0 -20.5 -48.0' Austria -45.5 -40.7 -45.4 Belgium -31.4 -28.9 -20.0 Canada -32.0 -41.2 -55.5 Denmark -20.9 -21.3 -25.8 Finland -15.7 -23.9 -46.7 France -46.9 -34.8 -28.0 Germany -50.1 -36.7 -36.2 Italy -67.9 -72.0 -53.1 Japan -8.4 -17.7 -21.9 Netherlands -33.4 -27.9 -24.8 Norway -12.0 -12.7 -21.7 Sweden -37.0 -40.6 -23.5 Switzerland -50.0 -41.5 -21.7 United Kingdom -37.6 -25.3 -13.0 United States -48.5 -38.3 -39.6 Average -34.0 -32.8 -32.8 Note: Fluctuations are given as maximum percentage peak-to-trough fall in annual data. a. 1929-31 period only. Sources: Latin American countries from ECLA (1976); China from Hsiao Liang-lin (1974:275); India from Birnberg and Resnick (1975); Indonesia from League of Nations, Review of World Trade, 1938; advanced countries from Maddison (1962, 1982), and sources cited in appendix F of Maddison (1982). ANGUS MADDISON 21 dom) and the United States all saw major improvements in their terms of trade. Because of the fall in capacity to import in the Third World, which was magnified by perverse capital flows, the payments problems of develop ing countries were very severe, and their import volume was cut on an average peak-to-trough basis by 57 percent. Thus the deflationary impact of these direct export losses and payments constraints was much bigger in developing countries than in advanced countries. The biggest loss of purchasing power occurred in Chile, where the spontaneous repercussions of exogenous recessionary forces from the outside world were compounded by specific U.S. protectionist action on copper imports, which led U.S. companies operating in both Chile and the United States to favor their U.S. products. Chile was also the Latm American country that clung most closely to deflationary gold standard rules in defending its exchange rate. (See Birnberg and Resnick 1975 on the impact of U.S. protection and Hirschman 1963 on Chilean monetary policy.) Colombia was the country most lightly affected by the recession in trade; its coffee export prices were sheltered because of Brazilian stock piling policy, whereas its exports of coffee were not restricted, as were those of BraziL Colombia also took internal policy measures to promote import substitution, hence its import volume fell a good deal more than its export purchasing power. (See Thorp 1984 for Colombian policy in the depression.) The size of the recessionary impact depended upon the luck of the commodity lottery rather than upon degree of export diversifica tion. China with its rather diversified exports suffered a bigger export decline than Brazil, which was much more highly specialized (see table 1-3). The traditional geographic distribution of exports (see table 1-4) also had relatively little effect on the degree of export decline of a particular country in the first stages of recession. Price falls for particular commod ities were felt worldwide, and in relatively free markets, export destina tions could be varied. In most cases, the initial fall in exports in 1929--32 was dominated by spontaneous repercussions of the deep decline in activity in the advanced countries and did not derive from protectionism, but the Hawley-Smoot tariff increases in the United States in 1930 sparked a great wave of tariff restrictions and quantitative export and exchange controls in 1931. These restrictions had their major effect on the recovery process rather than on the depth of the recession. The recovery of imports in advanced countries after 1932 was much smaller than it would otherwise have been, 22 DEVELOPING COUNTRIES IN THE 1930s Table 1-3. Percentage Composition of Commodity Exports in 1929 Country Exports Argentina wheat 29.2, maize 17.6, frozen, chilled, and tinned meat 12.8, linseed oil 12.6 Brazil coffee 71.0 Chile nitrates 42.1, copper 40.4 China vegetable oils and seeds 28.4, raw silk 16.3, hides and skins 4.5, tea 4.1, coal 3.0 Colombia coffee 60.6, petroleum 21.3 India jute and jute products 25.5, cotton 20.9, rice 10.1, oilseeds 8.5, tea 8.4 Indonesia sugar 21.6, rubber 16.0, petroleum 12.4, copra 6.7, tea 6.0, tobacco 5.8, coffee 4.8 Mexico silver 20.6, other minerals 47.0 Peru oil 29.7, copper 22.4, wool 21.1, sugar 11.5, lead 5.2 Source: Statistisches Reichsamt (1936). Table 1·4. Geographic Distribution of Exports in 1929 (percentage of total) United United Country France Germany Japan States Kingdom Argentina 7.1 10.0 0.0 32.2 9.8 Brazil 11.1 8.8 0.0 6.5 42.2 Chile 6.1 8.6 0.0 13.3 25.4 China 5.5 2.2 25.2 7.3 13.6 Colombia 0.5 2.1 0.0 4.7 75.2 India 5.4 8.6 10.4 21.4 11.7 Indonesia 0.0 2.6 3.3 8.9 11.4 Mexico 3.9 7.6 0.0 10.3 60.7 Peru 1.3 6.1 0.0 18.3 33.3 Source: Statistisches Reichsamt (1936). and the recovery process was highly import substitutive in the developing world. Specific studies of the impact of trade restrictions are rather scarce, but it has been rather conclusively demonstrated that, in the Chilean case, exports would have been a good deal higher without V.S. trade restrictions (see Birnberg and Resnick 1975). Similarly, V.S. sugar quotas favored V.S. Philippine and Puerto Rican producers at Cuba's expense. In general, Latin American countries were more exposed to the adverse impact of protection in the 19305 than were European colonies, which usually enjoyed some degree of discriminatory protection from their metropole. Germany did offer some discriminatory ANGUS MADDISON 23 Table 1-5. Impact of Trade Discrimination: Percentage Share of "Empire Trade" Imports Exports Country 1929 1938 1929 1938 United Kingdom with Commonwealth and colonies 30.2 41.9 44.4 49.9 France with colonies 12.0 27.1 18.8 27.5 Japan with Formosa, Korea, and Manchuria 13.9 39.0 19.3 46.6 Germany with SE Europe and Latin America 16.7 27.6 12.8 24.7 Netherlands with colonies 5.5 8.8 9.4 10.7 Source: League of Nations, Review of World Trade, 1938. market access privileges to Latin America in return for some reciprocity. The United States, with some exceptions just noted, held to the most favored-nation principle. Table 1·5 shows the impact of discriminatory blocs on trade patterns. Service Payments The information is poorer for service transactions than for commodity trade. Mexico, because of its proximity to the United States, was prob ably worst affected by the loss of tourist earnings. All developing coun tries enjoyed lower freights on shipping, supplied generally by the ad· vanced countries; the fall in shipping freights does not seem to have been as steep as the fall in commodity prices. Payments for profit remittances on direct foreign investment fell sharply because this income was cyclically very sensitive. In Latin Amer ica and Eastern Europe, such remittances were usually blocked or were limited by exchange control from 1931-32 onward, but even in colonies such as India and Indonesia, where such movements were not controlled, they fell substantially. With respect to payments of interest on bonded debt, the situation was much worse, because the fall in price levels made such payments very onerous at a time when there was little access to new funds of this character. The nature of the payments problem and the kinds of possible accom· modation are illustrated in table 1·6 for the Indian case. There was a Table 1-6. Items in India's Balance of Payments, 1929-38 (millions of rupees) Exports Imports Balance on Net foreign of of non- investment Merchandise precious Service Merchandise precious Service commercial and Year exports metals receipts imports metals payments transactions borrowing 1929-30 3,613.4 15.7 54.3 2,694.4 277.7 936.7 35.8 45.6 1930-31 2,578.5 24.3 40.9 1,892.6 269.1 790.5 -93.8 403.6 1931-32 1,819.0 629.1 29.5 1,403.5 74.1 706.0 166.2 -69.1 N 1932-33 1,538.6 679.2 29.7 1,459.1 31.2 744.2 178.8 45.2 "'" 1933-34 1,713.4 591.9 26.5 1,295.7 20.8 737.2 -368.0 112.7 1934-35 1,757.3 577.3 30.1 1,486.4 52.2 760.1 -227.3 4.9 1935-36 1,860.3 418.8 30.8 1,520.2 56.6 786.9 138.0 -51.7 1936-37 2,262.5 300.0 30.3 1,405.4 156.8 836.6 -106.0 -174.1 1937-38 2,254.7 183.2 37.2 1,775.2 40.2 870.6 -58.1 -37.1 1938-39 2,032.3 147.0 32.3 1,532.2 22.4 804.1 89.6 -64.5 Average, 1930-38 1,979.6 394.5 31.9 1,530.0 80.4 781.8 -127.8 18.9 Note: Figures are given for fiscal years beginning April 1. Source: Banerji (1963:27, 90, 137, and 195). ANGUS MADDISON 25 sharp 57 percent fall in export values from 1929 to 1932, and thereafter exports did not regain their 1929 value; on average, export values for 1930--38 were only 55 percent of export values for 1929. To meet this payments constraint, India also cut imports back sharply; for 1930--38 they averaged only 57 percent of their level in 1929. For developing countries during the period, service earnings were typically negligible, and service payments were considerable-equivalent to 35 percent of 1929 commodity imports. It was much more difficult to cut these miscellaneous service burdens than it was to cut imports, and 1930--38 service payments averaged 83 percent of those for 1929. Foreign borrowing for the period 1930--38 averaged only 41 percent of the 1929 level. A major balancing item was the reversal of the precious metals flow. India was a traditional importer of precious metals for indigenous savings hoards. In 1929 the net inflow was 262 million rupees. In 1930--38, however, the average annual outflow was 314 million rupees. Finally, we should note the drain stemming from noncommercial flows, which largely resulted from India's colonial status. These flows were substantial and negative in 1930--38. Perverse Capital Flows and Their Impact A feature of the 1930s situation that worsened the recessionary impact of the fall in export earnings was the reversal of the 1920s capital flow. This reversal is less well documented for developing countries than for developed countries because the annual prewar League of Nations figures for the balance of payments are rather weak. The overall situation of developed countries is rather clear from table 1-7, however; in the 1920s their net capital exports were more than $700 million a year, but in 1930--38 they had an average annual inflow of capital of about $540 million. The countries most affected by this reversal of capital flows were mostly in Latin America, where the bulk of the Third World's capital receipts had concentrated in the 1920s and where the debt was greatest (see table 1-8). When this capital dried up, it made the balance-of payments adjustment process more difficult, initially forcing deflationary policies to be more restrictive, or tariffs, quantitative restrictions, and exchange controls to be stricter. Except for Argentina, all the Latin American countries defaulted (but did not repudiate) their official debt obligations in the 1930s. In Asian countries that were colonies, debt default was not permitted, and official monetary and budgetary policy in India and Indonesia was tighter than in Latin America or in China, where debt default did occur. 26 DEVELOPING COUNTRIES IN THE 1930s Table 1-7. Net Foreign Investment Flows from and to Advanced Countries, 1924-38 (millions of U.S. dollars) Country 1924-29 1930-38 Australia 182 7 Canada -32 -14 Denmark 2 -7 Finland 6 -21 France -365 10 Germany 661 -77' Japan 54 -49b Netherlands -72 14 New Zealand 31 1 Norway 17 -3 Sweden -36 -12 Switzerland -68 42 United Kingdom 341 99 United States -762 552 Total -723 542 Note: Figures represent average annual flow in the years specified. A minus sign denotes flow from a country. a. 1930-35. b. 1930-36. Source: United Nations (1948). Table 1-8. Per Capita Net Foreign Capital Liabilities in 1938 (U .S. dollars) Country Amount Latin America Argentina 230 Brazil 51 Chile 258 Colombia 38 Mexico 93 Peru 48 Arithmetic average 120 Asia China 3 I[1dia 11 Indonesia 35 Arithmetic average 16 Sources: Lewis (1984) and Gurtoo (1961). ANGUS MADDISON 27 The transfer burden for bonded debt was probably most severe in Indonesia. There was (see Creutzberg 1975-86, vol. 2:82-83, vol. 5:70) a rise from 1.7 percent of income (net domestic product) in 1929 to 5.4 percent in 1934. A good deal of the rising burden stemmed from falling price levels, but some was due to debt redemption. Percentage of Indonesian Year net domestic product 1929 1.7 1930 1.9 1931 2.6 1932 4.6 1933 5.1 1934 5.4 1935 3.7 1936 3.5 1937 2.6 1938 2.7 In the 1930s, there was a very large capital outflow from both China and India in the form of precious metals, mainly gold from India and silver from China, amounting to $1 billion in each case. In general, the penalties incurred for debt default were rather small in the 1930s. The precedent had been set by the major powers in connection with war debts and reparations, and there was widespread domestic acceptance in those countries of the need to write down farm and mort gage debt, so the principle of temporary default (as distinct from repudia tion) was hard to refuse, and the number of defaulters was too large for sanctions to be workable. 3 Of the $5.3 billion Latin American securities outstanding in 1938, $3 billion were in default, as compared with $340 million in 1913 and about $1.4 billion in 1929. 4 The proportion of de faulted East European, Greek, and Yugoslav bonds seems to have been even higher. China and Turkey were also defaulters. Debt default was a very important item in alleviating the balance-of-payments problems of some countries. The saving in foreign remittances of interest was about $50 million annually for Brazil and $25 million for Mexico. In addition these countries economized on amortization.5 Much of the default was ultimately accepted by the creditors (mainly the United Kingdom and the United States) in wartime and postwar debt settlements. In the 1930s there was no real restoration of private international capital flows. There was, however, a considerable growth in government-guaranteed export credits tied to the goods of the creditor country. There was a big extension of such credits from Germany, the United Kingdom, and the U.S. Ex Table 1-9. Contribution of Change in Export Volume and Terms of Trade to Change in Real Income, 1929-32 Percentage Impact Direct real Real Percentage change Impact of income income 1929 change m of change effect effect ratio in purchasing fall in Percentage of 1929-32 of of volume power in commodity Percentage change change in other commodity of of commodity terms change in purchasing conjunc exports commodity commodity export of in real power of tural to exports, exports, volume, trade, GDP, income, commodity policy Country GDP' 1929-32 1929-32 1929-32 b 1929-32" 1929-32 1929-32 exports b differences Argentina 26.7 -12.6 -34.7 -3.4 5.9 13.8 -18.9 -9.3 9.6 I\.J Brazil 15.4 -19.2 -45.6 -3.0 -4.1 -4.9 -8.8 -7.1 -1.7 00 Chile 30.0 -71.2 -84.5 -21.4 -4.0 -26.5 -29.4 -25.4 -4.0 China (4.0) -41.1 -57.8 1.6 -0.7 (5.5) 4.8 -2.3 7.1 Colombia (21.0) -2.0 -28.2 -0.4 -5.5 4.0 -1.7 -5.9 4.2 India 7.8 -30.6 39.2 2.4 -0.7 -1.7 -2.4 -3.1 0.7 Indonesia 29.0 -6.5 -34.5 -1.9 -8.1 0.6 -7.5 10.0 2.5 Mexico 12.4 -41.5 -64.8 -5.1 -2.9 19.0 -21.3 -8.0 -13.3 Peru (30.0) -29.7 -57.1 -8.9 8.2 -25.8 -31.9 -17.1 -14.8 Average 19.1 -28.3 -49.6 -5.3 -4.5 -9.1 -13.0 -9.8 -3.2 Note: Figures in parentheses are estimates. a. At 1929 prices. b. As percentage of 1929 GDP. Source: Table 1-12 sources and tables 1-14-1-16. ANGUS MADDISON 29 port-Import Bank. Generally these credits were issued for a maximum term of five years. The Income Effect of the Export Shocks Table 1-9 shows the direct first-round impact of the export collapse on real income in 1929-32, weighting the impact by the ratio of exports to GDP. The decline in volume accounted for an average GDP fall of 5.3 percent for the sample, and worsening terms of trade meant an additional income loss equivalent to 4.5 percent of GDP. The total income loss was the GDP change multiplied by the terms-of-trade change, making a total average decline of 13 percent. The total income loss due to trade (that is, the change in export purchasing power) amounted to 9.8 percent of GDP for the sample. The last column of table 1-9 shows the residual GDP change in these years that is not "explained" by falling commodity exports. These residual, apparently endogenous items of conjuncture and policy accounted on average for an income fall of only 3.2 points, so for the group as a whole the direct export shocks "explain" the bulk of the recession. This situa tion contrasts with that in the developed world, where endogenous deflationary elements were bigger than those attributable to the export shocks. The "residual" effects in the developing countries should not in fact be interpreted as endogenous, because the export shocks and the sudden drying up of previous foreign capital flows produced further indirect deflationary effects, either spontaneously or as a result of government attempts to balance budgets and to curb payments deficits by expenditure cuts and tax increases. In view of the general lack of comparable quantitative information, it is not easy to assess the significance of the residuals on a country-by-country basis. In India and China, where international trade was a relatively small part of the economy, the secondary deflationary effects of the export shock would probably be more limited than in countries more exposed to trade. It is also clear that countries with relatively high income levels, such as Argentina and Chile, had higher levels of expenditure on cycli cally compressible items of demand-with respect to both investment and consumption-than did countries such as China and India, which had very large subsistance sectors. In the Latin American countries, invest ment levels were affected not only as a secondary effect of the export shock and its payments repercussions but also more directly by the reversal of the significant capital inflows of the 1920s and by the close 30 DEVELOPING COUNTRIES IN THE 1930s geographic, institutional, and psychological links with the extremely depressed U.S. markets. The impact of the export shock also depended on the nature of the export economy. In Chile, for example, the export sector was a geo graphic enclave with rather small employment, and a good part of the depressive impact was absorbed within the enclave by the profit remit tances of foreign companies. The secondary recessionary effects of the export shock would therefore tend to be smaller than in Argentina, where a larger part of the labor force was employed in production and in processing a given value of product. The residuals also provide some guide as to the efficacy of policy in withstanding the impact of recession. It would appear that Colombia and Brazil were most successful in this area and Peru, Mexico, and Argentina the least successful in Latin America. In Asia, by contrast, none of the three countries had negative residuals, and the situation was distinctly better than in Latin America. The Recovery Phase, 1932-37 In discussions of the 1930s, it is too readily assumed that the whole decade was one of depression, but table 1-10 shows that there was a vigorous recovery process under way from 1932 to 1937 in Latin America. In Asia the recovery process was much more modest. The recovery process was interrupted in 1938 by the sharp U.S. recession of that year, which did not affect other advanced countries but had an adverse effect Table 1-10. Annual Average Change in GDP in Three Cyclical Phases (compound rates) Country 1929-32 1932-37 1937-38 Argentina 4.8 5.2 1.2 Brazil 1.7 8.4 5.1 Chile -9.8 8.3 0.0 China 1.8 0.5 -2.5 Colombia 1.3 4.2 6.6 India -0.6 1.0 -0.3 Indonesia 0.2 3.3 -3.2 Mexico -6.8 7.2 1.8 Peru -9.5 11.5 -2.5 Average 3.3 5.5 0.7 Source: Table 1-12. ANGUS MADDISON 31 on several developing countries because of its effect on primary commod ity prices. In 1932-37 the volume of commodity exports increased in all countries except China, although in Argentina, Chile, and India commodity ex ports did not recover their 1929 levels. The only cases with very signifi cant terms-of-trade improvement were Argentina and Chile, and only in Argentina were 1937 terms of trade better than in 1929. As a result, trade expansion contributed only 6.6 percentage points of the average 34 percentage point increase in real income. The bulk of the recovery process was therefore explained by endogenous nontrade forces. The cyclical contrast between 1929-32 and 1932-37 was sharpest in Latin America. There the residual (nonexport shock) components in recession accounted for an average decline of 2.3 percent a year in real income in the three years 1929--32, and in the recovery phase, these endogenous expansionary elements were vigorous everywhere. The con trast was least marked in Colombia, which was most successful in dealing with recession in 1929--32 and therefore had less scope for sharp recovery in 1932-37. In Asia the situation was quite different. In the first place the recession phase involved no endogenous declines, but the recovery phase was so weak that it was indistinguishable from the recession phase. We could therefore claim that the 1930s were a decade of stagnation in Asia. The difference between the Latin American and Asian situations from 1932 onward is in fact traceable to policy differences. In Latin America most countries followed sharper policies of import substitution than in Asia because they were politically freer to impose trade and exchange controls and they were also freer politically to follow unorthodox budget and fiscal policies that achieved a fuller use of domestic resources and involved some degree of inflation. In addition they were free to follow policies of debt default. In contrasting Latin American and Asian policy, we must more or less exclude the Chinese case because of both the impact of war (the seizure of Manchuria in 1931, the Japanese invasion of 1937, and the continual civil wars of the 1930s) and the special currency experience, which helped put China's cyclical experience on a different time sched ule. The Impact of Policy in Developing Countries Although there is a distinct family resemblance in policy between the different Latin American countries in the recovery period from 1932 onward, policies differed in the first phase of recession. Argentina abandoned the gold standard at an early stage in 1929. This 32 DEVELOPING COUl\'TRIES IN THE 19305 action softened the deflationary impact of the fall in export earnings, helped to discourage imports, and meant that internal prices fell less, thus mitigating the secondary deflation caused in many countries by distressed debtors forced into liquidation. These policies were later followed by exchange controls and tariff increases. The recession created a loss of income and political power for the old oligarchy, which had favored free trade. The regime that took over in the 1930s was more willing to raise tariffs and to follow policies of import-substituting industrialization. In the case of Brazil, which did not leave the gold standard until 1930, the populist leader Vargas gave the state a major role in promoting import substitution through government enterprise. The new inward orientation of policy helped little with immediate cyclical problems but was of major long-term significance in reducing the trade ratio in these countries. Argentina's export dependence was twice as big as that of Brazil; Argentina was a more developed country, with higher ratios of the more volatile domestic GDP components, such as investment and consumer durables. For this reason and others, the recession was deeper in Argen tina. Furthermore, Argentina had strong bilateral trade ties with the United Kingdom, and the possibility of U.K. trade retaliation (articu lated in the negotiations for the Roca-Runciman treaty of 1933) led Argentina to honor its debt commitments, whereas Brazil was delinquent to a very significant extent and thereby eased its payments constraints. Brazilian domestic policy offset the exogenous shock to a significant degree because it happened to enter the recession with a well-established price support apparatus for stockpiling coffee. Unfortunately, lack of adequate planting restrictions or taxes on coffee production meant that Brazil continued to produce huge stocks of coffee that were eventually destroyed. Some of the apparent GDP buoyancy thus did not represent real income, though activity was maintained and the impact of recession was diffused away from the coffee sector. Cumulatively coffee's stock destruc tion over the period 1931-38 was equal to a loss of 17.7 percent of 1929 GDP. The major beneficiaries of Brazilian coffee policy were other produc ers, who enjoyed the price support benefit without the stockpiling costs. The most important of these producers were Colombia and Indonesia, which exported 3.8 million and 1.4 million bags respectively in 1934-38, as compared with Brazil's 14.6 million. In Chile, the impact of the depression via trade was extremely severe. Exports fell in volume by 70 percent from 1929 to 1932, and the terms of trade deteriorated nearly 40 percent. The purchasing power of imports fell more than 80 percent, and the substantial inflow of capital that had characterized the 1920s dried up. In the Chilean case the export loss was very severe partly because exports were heavily concentrated on miner als, the category of goods for which demand fell most, but also because ANGUS MADDISON 33 Chile was adversely affected by U.S. action in 1932 in imposing a specific tariff of four cents a pound on copper imports. The tariff was equivalent to a 70 percent ad valorem duty. It has been estimated (Birnberg and Resnick 1975:234-35) that Chile's exports in 1932 would have been twice as high had it not been for these U.S. export restrictions, which caused U.S. copper interests to reallocate production from their Chilean mines to their U.S. mines. Because Chile had an export dependence of about 30 percent of GDP in 1929, the fall in exports had a very large effect on GDP. This effect was exacerbated by the Chilean policy of defending the ex change rate by deflation until March 1932. The new central bank created on the advice of Kemmerer in 1925 pushed up the discount rate to 9 percent in a period of falling prices. Policy changed with the massive inflation of the brief socialist government in 1932, which paved the way for recovery (Hirschman 1963:179-83), and exchange controls forced profits to be reinvested domestically when they reemerged. In Mexico, the situation in 1929 was different from that elsewhere in Latin America in two contradictory respects. On the one hand, for political reasons there was a bigger endogenous policy element in Mex ican experience. The revolution and civil wars of 1910-20 had led to partial debt default, which meant a low capital inflow in the 1920s. The partial default also reduced oil exploration by foreign companies, which instead built up Venezuelan production, cutting into Mexican export markets. The Calles and Calles-dominated presidencies in the 1920s and 1930s followed conservative fiscal and monetary policy, both to payoff debt and to manage Mexico's rather complicated bimetallic currency. On the other hand, Mexico, being a closer neighbor of the United States with its epicenter economy, had a bigger across-the-board dependency rela tionship than several other Latin American countries. Its trade de pendency on the United States was high; it depended more on tourism and migrant remittances; its border areas were closely integrated with the U.S. economy; American enterprise and management were relatively important; and banking and credit links were strong. The first policy response of Mexico to the depression was as deflation ary as that in Chile. Montes de Oca, the minister of finance, followed a tight money policy and raised taxes to balance the budget. This policy changed sharply in 1932 when Pani became minister of finance, switched to a more expansionary budget and fiscal policy and to exchange depreciation, and tolerated more inflation. He also demone tized silver, switched to a paper currency, and sold some of the Mexican silver stocks in 1934-35, during the period when the U.S. government was supporting silver prices. From 1934 to 1940, under the Cardenas government, policy became more nationalist than it was elsewhere in Latin America, with greater use Table 1-11. Price Movements, 1929-38 (1929 = 100) Year Argentina Brazil· Chile China Colombia India' Indonesia" Mexico' Peru 1929 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 1930 101.0 98.5 98.9 112.4 79.2 80.7 96.5 102.4 95.8 1931 87.0 95.1 98.2 122.8 64.0 71.9 73.1 89.6 89.6 1932 78.0 95.1 104.4 114.8 53.6 68.7 57.S 80.0 85.6 1933 88.0 95.1 129.6 105.5 55.2 65.0 47.6 84.8 83.4 1934 78.0 101.9 129.7 97.2 76.8 66.4 45.4 87.2 S5.2 "'" ~ 1935 1936 82.7 89.7 107.8 123.3 132.4 143.6 97.1 108.9 80.0 84.8 66.9 67.1 45.8 44.4 88.S 96.8 86.1 90.4 1937 92.1 133.0 161.7 125.7 86.4 68.9 48.6 119.2 96.5 1938 91.5 138.3 221.2 97.3 69.3 49.5 125.6 Note: Dashes indicate that figures are not available. a. Brazil, India, Indonesia, and Mexico indexes are national income, or GDP, deflators. Sources: Brazil from Contador and Haddad (1975); India from Sivasubramonian (1965); Indonesia from Polak (1979); Mexico from Solis (1984:92). For other countries, the figures are generally cost-of-living indexes. Argentina and Colombia from Diaz Alejandro (1981:16). China from Feuerwerker (1977); 1929-36 is the average of the implicit Tientsin and Shanghai wholesale indexes for consumer goods given on pp. 46-47, and 1937 is from the Shanghai index, linked to the 1937-39 index ofretail prices in the main cities of unoccupied China. Chile and Peru from League of Nations, Statistical Yearbook (cost-of-Iiving indexes for Santiago and Lima). ANGUS MADDISON 35 of indigenous developmental and financial institutions to promote de velopment and outright nationalization of mineral resources. The nationalization included de facto expropriation of foreign oil interests in 1938, which provoked sharper sanctions against Mexico than the more diplomatic defaults farther south. This action also meant that oil explora tion in Mexico was delayed for several decades. In India and Indonesia, two big Asian countries that were still Euro pean colonies, the policy was very different in style from that in Latin America and paralleled action in the respective metropoles. India en tered the recession with tight monetary and fiscal policy to defend the rupee's parity with sterling and to defend sterling itself. From 1925 onward it was generally held that sterling was overvalued, and Indian nationalist economists argued that the rupee was overvalued relative to sterling (the rupee stood at a higher parity when India readopted the gold standard in 1927 than it had in 1913). In 1931 when sterling was allowed to float against other currencies, the rupee was pegged at the same rate relative to sterling. Virtually all Indian opinion (including that in the viceroy's executive council) opposed this policy, but Whitehall insisted. To back up the exchange rate and balance the budget, the government cut civil service salaries, reduced the military establishment, and by 1932 had cut capital expenditure on railways and civil works to a quarter of their 1929 level. The 1931 supplementary budget levied a 25 percent surcharge on income tax, excise taxes, and import duties. In the course of the 1930s, the government built up exchange reserves. It was able to do so because of the huge dishoarding of precious metals in the 19308. Between September 1931 and March 1939, the outflow was more than a billion dollars. Tariff autonomy had been regained in the 19208, and between 1930 and 1933 tariffs on non-British textiles were raised from 15 percent to 75 percent. Quotas were imposed on imports of Japanese piece goods, and the 1932 Ottawa agreements established imperial preference. Increased protection gave a boost to Indian manufacturing and to some import substitution in agriculture, for example in the case of sugar. The impact of the recession on Indian output was relatively mild for reasons that I have mentioned. The main impact of restrictive policy was to keep growth lower than it could have been in the 19308 had the Latin American options of easier monetary and fiscal policy and debt default been fol lowed. One of the most interesting differences between the Indian (and In donesian) situation and that in Latin America was the depth of fall in prices. The domestic price level fell 35 percent in India from 1929 to 1933, without much recovery before the war (see table 1-11). In Indonesia, the 36 DEVELOPING COUNTRIES IN THE 1930s price fall was even bigger, but only in Colombia did the price fall exceed that in India in the recession phase, and Colombian prices recovered substantially from 1932 onward. It is difficult to interpret the real impact of falling prices, because the intercountry variation in price experience bore no obvious relation to the depth of the recession, and it is difficult to judge the extent to which decisionmaking was affected by price changes and very high real interest rates. The price declines did not have as damaging an institutional impact in the Third World as they did on U.S. financial institutions, but they created many bankruptcy situations for debtors and windfall gains for moneylenders and landlords who were able to collect fixed money rents. They also raised the incidence of tax bur dens. It would appear that the price decline had an important distribu tional impact whose effects have not been very satisfactorily diagnosed. Governments made various attempts to mitigate the burden on debtors, and also to help credit institutions threatened by bad debt, but analysis of the macroeconomic impact of such policy is sadly lacking.6 In Indonesia, fiscal and monetary policy followed a deflationary style similar to that in India, with cuts in civil service salaries and in capital expenditure, increases in taxation, and fulfillment offoreign debt service commitments, with net redemptions in the 19305. The major difference was that the Netherlands guilder (the Indonesian currency unit) retained an unchanged gold parity until 1936, with the result that Indonesian prices had to fall further than prices in India to remain competitive on international markets. In response to foreign tariff increases and quotas, the Netherlands abandoned its traditional free trade policy, raised In donesian tariffs, imposed quotas on Japanese imports, and discriminated in favor of bilateral trade. The Netherlands also entered into restrictive commodity arrangements for sugar and rubber that it had previously eschewed. The creditor/debtor problem was probably worse in Indonesia than in India and led to liquidity problems for various kinds of banks catering to the indigenous population. The income cycle was worse for Indonesia than for India, because Indonesia had a much bigger traded sector. Indonesia, however, also had a bigger concentration of cyclically volatile income in the hands of expatriate and Chinese traders, so that the impact on the indigenous population was smaller than on the economy as a whole. The rather detailed Indonesian national accounts estimates by Polak (1979) for 1921-39 unfortunately relate to real income rather than to GDP. Our estimates for the latter are rather crude and partial, but the GDP volatility was much less than the real income movement. In Indone sia, the plantation labor of coolies was proportionately much more im portant than in India, and these coolies suffered very big pay cuts and ANGUS MADDISON 37 unemployment in the recession, for which some of them compensated in part by producing subsistance crops on small plots. Lessons from the 1930s Recent economic events in the Third World have been very different from those of the 1930s. There has not been a deep fall in output in the advanced countries, and the volume of their imports from non-oil Third World country exporters has been very well sustained. The terms-of-trade dichotomy was not between Third World primary producers and the advanced countries exporting manufactures but between oil producers and non-oil producers. The Third World had long cast aside deflationary gold standard economics as a technique for dealing with external shock and was equipped with a monetary-fiscal and exchange control armory that permitted more growth-oriented tradeoffs. Balance-of-payments accommodation to external shock was greatly facilitated after 1973 by the massive availability of credit, whereas the 1930s problem had been greatly complicated by lack of credit. The sharp decline in the pace of world inflation, the rise in U.S. interest rates after 1979, and the Mexican moratorium of August 1982, however, sparked off a debt crisis and checked the compensatory flow of capital. As a result, most of Latin America plunged into a recession in 1983. It is difficult to draw "lessons" from the experience of the 1930s for application to the 1980s because of the big difference between the two periods and also because the more obvious lessons of the 1930s, of the type that Arndt drew in 1944, have already been embodied in the new liberal international economic order, which has thus far proved rather robust. Another big difficulty is that the really substantial cyclical prob lems of the Third World arose rather recently (from 1982 onward), so that it is perhaps premature to push the conclusions too far. The "lessons" already learned have been several. L Articulate cooperation and consultation between the major ad vanced countries to avoid beggar-your-neighbor trade and payments measures. Developing countries have thus not faced the catastrophic trade volume collapse of the 1930s, and their terms of trade have not fared worse than those of developed countries. 2. Built-in stabilizers and more sensible discretionary action prevented developed countries from inducing a recession anywhere nearly as serious as that in 1929-32, but the advanced countries made a major mistake by pursuing a macropolicy (fiscal and monetary) that was too deflationary, with an increasing proportion of the labor force unemployed and other Table 1-12. 1929 GDP Levels at 1929 U.S. Relative Prices and Movement of Gross Domestic Product (1929 = 1(0) Year Argentina Brazil Chile China Colombia India Indonesia Mexico Peru Total 1929 level (millions of dollars) 5,190 2,842 1,759 40,729 798 23,741 5,470 1,339 491 82,359 1929 100.0 100.0 100.0 (llJO.O) 100.0 100.0 100.0 100.0 100.0 100.0 1930 95.9 95.6 95.9 (101.2) 99.1 100.0 100.4 93.2 88.8 99.9 1931 89.2 92.7 76.9 102.3 97.6 98.2 96.3 96.6 78.2 98.7 1932 86.2 95.1 73.5 105.5 104.0 99.6 100.6 81.0 74.2 100.6 "'" '00 1933 90.3 107.3 83.5 105.5 109.9 102.1 101.6 89.7 100.0 102.6 1934 97.4 116.9 94.7 96.3 107.6 100.7 102.5 95.6 116.9 98.9 1935 101.7 122.4 98.4 104.1 119.6 101.4 103.6 100.5 120.1 103.8 1936 103.0 137.1 101.6 110.7 126.0 106.4 112.5 110.9 125.4 110.1 1937 111.3 142.3 109.4 (108.0) 127.9 106.4 118.1 114.7 128.1 110.1 1938 112.6 149.6 109.4 (lOS. 3) 136.3 101.8 114.3 116.8 124.9 107.6 Note: Figures in parentheses are estimates. Sources: Argentina: EeLA (1978). Brazil: Zerkowski and de Gusmao Veloso (1982). Chile: EeLA (1978). China: 1931-36 from Yeh (1979); 1929-31 extrapolated from 1914/8-1933 growth rate in Perkins (1975:117); GDP assumed to drop 2.5 percent in both 1937 and 1938 because of the impact ofthe Sino-Japanese war. Colombia: EeLA (1978). India: Maddison (1971). Indonesia: derived from J. J. Polak in Creutzberg (1975-86, vol. 5:84); combination of food crop production and export crops. This crude indication should correspond more closely with the GDP movement than Polak's national income figures (p. 70), which are strongly affected by changing terms of trade. Mexico: EeLA (1978). Peru: Bolofia Behr (1981). ANGUS MADDISON 39 forms of underutilized growth potential. The reason was partly that the advanced countries exaggerated their payments problems with OPEC and partly that they were too ambitious in their aspirations to mitigate infla tion. The result was a rather weak recovery phase after the 1974-75 recession and also after the 1981-82 recession. 3. In the United States the recent recovery has been much more satisfactory than in Europe but has involved a policy mix that keeps the dollar and interest rates too high, greatly complicating debt problems. 4. During 1973-82, developing countries weathered external shocks rather well. Their macropolicy posture was generally much more expan sionary than that of developed countries, because of a willingness to trade off growth against higher inflation rates and to live with substantial payments deficits. The latter was possible to a much greater degree than in the 1930s, because of the willingness of the advanced countries both to provide the financing capital flow and to tolerate a dichotomous interna tional order such that developing countries can keep trade and payments restrictions while advanced countries remain relatively free of them. The debt crisis of 1982-83, however, demonstrated that Latin American coun tries had underestimated the risks involved and the potential costs of overborrowing. 5. Because credit was available on such liberal terms, important parts of the developing world based too much of their expansion on borrowing and not enough on export competitiveness. They would have done better to have grown somewhat more slowly and to have borrowed less. As in the 1930s, we see a contrast between Latin America and Asia. This time the colonial context has disappeared, and Asian countries have been better able to follow policies in their own interest. They seem, on the whole, to have judged their interests better than Latin America because they relied more on export competitiveness and less on borrowing to finance their expansion. Stabilization exercises (for example, in Korea), have acted more quickly and have been more successful than in Latin America. Statistical Appendix Economic data for Argentina, Brazil, Chile, China, Colombia, India, Indonesia, Mexico, and Peru in the period 1929-39 are presented in Tables 1-12 to 1-18. 40 DEVELOPING COUNTRIES IN THE 19305 Table 1-13. Relative Size of Economies in 1929 GDPat GDP 1929 U.S. per capita relative 1929 U.S. Exports Population prices relative Exports per capita Country (thousands) ($ million) prices ($ million) ($) Argentina 11,592 5,190 448 908 78 Brazil 32,894 2,842 87 462 14 Chile 4,333 1,759 406 283 65 China 490,382 40,729 83 650 1.3 Colombia 6,927 798 115 124 18 India 333,100 23,741 71 1,177 3.5 Indonesia 59,830 5,470 91 582 10 Mexico 16,337 1,339 82 285 17 Peru 5,860 491 84 117 20 Total 961,255 82,359 4,588 Other developing countries 437,000 3,611 World 2,100,000 33,024 Note: Dashes indicate that figures are not available. Sources: 1950 GDP levels in 1965 dollars adjusted for differences in purchasing power by the production method (except Chile and Indonesia) from Maddison (1983a); Chile from Maddison (1970); Indonesia (same method used) from Maddison (1983b). The 1950 esti mates are backcast to 1929 with the GDPindicators, givingGDPin 1929 at 1965 V.S. prices. To convert 10 1929 prices I adjusted all series by the ratio ofV.S, 1929 GDP in 1929 prices to V.S. GDP in 1929 at 1965 prices (from V.S. Department of Commerce 1977). Table 1-14. Export Volume (1929 100) Country 1930 1931 1932 1933 1934 1935 1936 1937 1938 Argentina 69.3 95.3 87.4 81.9 85.8 90.6 81.9 95.6 61.4 Brazil 109.6 117.3 80.8 100.0 111.5 128.9 142.3 128.8 155.8 Chile 65.0 60.0 28.8 41.3 66.3 67.5 67.5 95.0 88.8 China 87.7 91.3 58,9 54.1 51.4 54.9 54.4 n.a. n.a. Colombia 109.8 96.1 98.0 98.0 103.9 113.7 127.5 125.5 131.4 India 89.4 76.4 69.4 79.8 81.3 81.9 99.4 95.5 n.a. Indonesia 96.8 84.8 93.5 88.1 89.7 90.6 100.1 112.5 n.a. Mexico 81.1 82,1 58.5 62.3 84.9 86.8 95.3 112.3 50.0 Peru 91.9 81.1 70.3 86.5 102.7 108.1 113.5 129.7 105.4 n.a. = not available. Source: Statistical sources listed in references. ANGUS MADDISON 41 Table 1-15. Terms of Trade (1929 = 100) Country 1930 1931 1932 1933 1934 1935 1936 1937 1938 Argentina 95.9 71.4 74.5 70.4 85.7 85.7 105.1 120.4 110.2 Brazil 61.6 53.6 67.5 60.3 62.9 55.0 54.3 58.3 43.0 Chile 95.7 67.0 55.1 60.0 57.3 61.6 70.3 76.8 54.1 China 90.9 77.2 71.7 65.0 68.4 73.9 85.1 n.a. n.a. Colombia 73.6 83.6 72.9 63.6 81.4 64.3 64.3 67.1 60.0 India 92.4 85.3 87.6 87.1 88.7 94.6 94.1 87.1 n.a. Indonesia 76.4 72.7 70.0 67.6 69.0 69.9 71.3 67.8 58.1 Mexico 77.5 59.8 59.8 61.8 66.7 77.5 64.7 61.8 132.4 Peru 72.4 58.6 61.8 59.2 69.7 71.7 69.7 62.5 63.8 n.a. = not available. Source: Statistical sources listed in references. Table 1-16. Purchasing Power of Exports (1929 = 100) Country '1930 1931 1932 1933 1934 1935 1936 1937 1938 Argentina 66.9 68.5 65.3 58.1 74.2 78.2 86.3 115.3 67.7 Brazil 67.1 62.0 54.4 59.5 69.6 70.9 77.2 74.7 67.1 Chile 62.2 40.5 15.5 25.0 37.8 41.9 47.3 73.0 48.0 China 79.7 70.5 42.2 35.2 35.2 40.6 46.3 n.a, n.a. Colombia 81.7 80.3 71.8 63.4 84.5 73.2 83.1 84.5 78.9 India 82.6 65.2 60.8 69.5 72.1 77.5 93.5 83.2 n.a. Indonesia 74.0 61.6 65.5 59.6 61.9 63.3 71.4 76.3 n.a. Mexico 63.0 49.1 35.2 38.9 56.5 67.6 62.0 69.4 66.7 Peru 66.1 48.2 42.9 51.8 71.4 78.6 8004 82.1 67.9 n.a. = not available. Source: Statistical sources listed in references. 42 DEVELOPING COUNTRIES IN THE 1930s Table 1-17. Volume of imports (1929 100) Country 1930 1931 1932 1933 1934 1935 1936 1937 1938 Argentina 87.8 61.5 46.8 51.3 56.4 58.3 61.5 80.8 76.3 Brazil 59.4 39.1 36.2 50.7 55.1 62.3 63.8 78.3 72.5 Chile 92.0 48.0 17.0 19.0 25.0 38.0 43.0 48.0 44.0 China 93.7 92.8 72.8 62.3 54.4 53.4 49.5 n.a. n.a. Colombia 52.3 44.6 36.9 50.8 63.1 69.2 78.5 90.8 84.6 India 81.1 69.7 80.6 69.8 80.1 84.5 76.8 n.a. n.a. Indonesia 85.2 72.3 57.9 57.3 53.7 52.5 53.0 69.0 70.7 Mexico 74.1 48.1 38.9 44.4 55.6 57.4 66.7 85.2 70.4 Peru 73.3 50.0 36.7 40.0 70.0 80.0 83.3 90.0 90.0 n.a. not available. Source: Statistical sources listed in references. Table 1-18. Exchange Rates (U.s. cents per unit of national currency) Year Argentina Brazil Chile China Colombia India Indonesia Korea Mexico 1929 95.13 11.81 12.06 41.90 96.55 36.20 40.16 46.10 48.18 1930 83.51 10.71 12.08 29.92 96.49 36.07 40.23 49.39 47.13 1931 66.74 7.03 12.07 22.44 96.57 33.70 40.23 48.85 47.65 1932 58.44 7.12 7.91 21.74 95.28 26.35 40.39 28.11 31.85 1933 n.a. 7.96 7.68 28.60 81.70 31.82 51.72 25.65 28.11 ~ 1934 33.58 8.43 10.15 34.09 61.78 37.88 67.38 29.72 27.74 1935 32.66 8.30 5.08 36.57 56.01 36.96 67.71 28.71 27.78 1936 33.14 5.88' 5.12 29.75 57.08 37.52 64.48 29.02 27.76 1937 32.96 6.20' 5.17 29.61 56.73 37.33 55.04 28.79 27.75 1938 32.60 5.84 5.17 21.36 55.95 36.59 55.01 28.45 22.12 1939 30.85 5.13' 5.17 11.88 57.06 33.28 53.33 25.96 19.30 n.a. not available. a. Free rate. Source: Federal Reserve System, Banking and Monetary Statistics. 44 DEVELOPING COIJNTRIES IN THE 1930s Notes 1. In any case, even the GDP aggregates for these countries cannot be regarded as definitive. There are competing estimates for Brazil (Haddad 1978 versus Zerkowski and Veloso 1982), India (see Maddison 1984), and Mexico (Solis 1973 versus ECLA 1978). We had to make crude estimates for Indonesian GDP, as the Polak series refer to income rather than product. For China for 1929-38 we used trend estimates, as annual data were not available. 2. "Billion" means "thousand million." 3. A detailed description of the processes of debt "readjustment" is provided in Young (1971). China actually went into default in the 19208 when civil disturbance reduced the earmarked revenues below service obligations. Thereafter China played a prolonged cat and-mouse game with creditors that involved writing down and rollovers of debt, sweetened by occasional repayments. In the Chinese case a substantial part of debt also involved government claims arising from indemnities following the Sino-Japanese war and the Boxer Rebellion. 4. See C. Lewis (1948:42), and Royal Institute of International Affairs (1937:303). 5. The average annual amortization and interest receipts of the United States and the United Kingdom combined fell more than $800 million from 1925-28 to 1932-34. See Royal Institute of International Affairs (1937:283). 6. It is alleged by J. C. Scott (1976) that the increased burden of debt service and taxation caused by falling prices led to widespread falls in peasant living standards and to peasant rebellions in Asia (evidence is supplied mainly for Burma and French Indochina). A recent review of Southeast Asian evidence on prewar peasant living standards puts the Scott thesis in doubt; see Ian Brown (1983). References Advanced Countries GENERAL ANALYSIS Arndt, H. W. 1972. The economic lessons ofthe 1930s. London: Cass. Originally published 1944. Haberler, G. 1976. The world economy, money, and the Great Depression, 1919-1939. Washington, D.C.: American Enterprise Institute. Hodson, H. V. 1938. Slump and recovery, 1929-1937. London: Oxford University Press. Kindleberger, C. P. 1973. The world in depression, 1929-1939. London: Allen Lane. Lewis, W. A. 1949. Economic survey, 1919-1939. London: Allen and Unwin. Maddison, A. 1962. Growth and fluctuation in the world economy, 1870-1960. Banca Nazionale del Lavoro quarterly review, June. 1985. Two Crises: Latin America and Asia, 1929-38 and 1973-83. Paris: Organisa tion for Economic Co-operation and Development (OECD) Development Center. ANGUS MADDISON 45 GERMANY Balderston, T. 1977. The German business cycle in the 1920s: A comment. Economichi..vtory review. Falkus, M. E. 1975. The German business cycle in the 1920s. Economic history review. Temin, P. 1971. The beginning of the depression in Germany. Economic history review. UNITED STATES Brunner, K., ed. 1981. The Great Depression revisited. Boston: M. Nijhoff. Friedman, M., and A. J. Schwartz. 1963. A monetary history ofthe United States, 1867-1960. Princeton, N.J.: Princeton University Press. Temin, P. 1976. Did monetary forces cause the Great Depression? New York: Norton. Developing Countries GENERAL ANALYSIS Bimberg, T. 8., and S. A. Resnick. 1975. Colonial development: An econometric study. New Haven, Conn.: Yale University Press. Brown, Ian. 1983. Rural distress in South East Asia during the world depression of the 1930s. Manila; processed. Diaz Alejandro, C. F. 1981. Stories of the 1930s for the 1980s. New York: National Bureau of Economic Research; processed. - - - . 1982. Latin America in depression, 1929-39. In M. Gersovitz and others. The theory and experience of development. London: Allen and Unwin. Rothermund, D., ed. 1982. Die Peripherie in der Weltwirtschaftskrise. Schoningh: Pader born. Scott, J. C. 1976. The moral economy of the peasant. New Haven: Yale University Press. Thorp, R., ed. 1984. Latin America in the 1930s: The role of the periphery in world crisis. London: Macmillan. ARGENTINA Diaz Alejandro, C. F. 1970. Essays on the economic history of the Argentine Republic. New Haven: Yale University Press. BRAZIL Contador, C. R., and C. L. Haddad. 1975. Produto real, moeda, e prec;os: A experiencia brasileira no periodo 1861-1970. Revista brasileira de estatistica, July-September. Fishlow, A. 1972. Origins and consequences of import substitution in Brazil. In L. di Marco, ed. International economics and development: Essays in honor of Raul Prebisch. New York: Academic Press. Haddad, C. L. S. 1978. Crescimellto do produto real no Brasil, 1900-1947. Rio de Janeiro; Vargas Foundation. 46 DEVELOPING COUNTRIES IN THE 1930s Instituto Brasileiro de Geografia e Estatistica (IBOE). 1960. 0 Brasil en numeros. In Anuario estatistico do Brasil, 1960. Rio de Janeiro: !BOE. Zerkowski, R. M., and M. A. de Gusmao Veloso. 1982. Seis decadas de economia brasileira atraves do PIB. Revista brasileira de economia, July-September. CHILE Hirschman, A. O. 1963. Journeys toward progress. New York: Twentieth Century Fund. CHINA Feuerwerker, A. 1977. Economic trends in the Republic of China, 1912-1949. Ann Arbor: University of Michigan, Center for Chinese Studies. Hsiao Liang-lin. 1974. China's foreign trade statistics, 1864-1949. Cambridge, Mass.: Har vard University Press. Perkins, D. H., ed. 1975. China's modern economy in historical perspective. Stanford: Stanford University Press. Young, A. N. 1971. China's nation building effort, 1927-1937. Stanford: Stanford University Press. U.S. Congress. Joint Economic Committee. 1982. China under the four modernizations. INDIA Banerji, A. K. 1963. India's balance of payments. Bombay: Asia Publishing. Gurtoo, D. N. 1961. India's balance of payments (1920-1960). Delhi: Chand. Maddison, A. 1971. Class structure and economic growth: India and Pakistan since the Moghuls. London: Allen and Unwin. - - - . 1984. What did Heston do? Paper prepared for the Cambridge Conference on the Economic History of India. Cambridge, April. Sivasubramonian, S. 1965. National income of India, 1900-{)1 to 1946-47. Delhi: Delhi School of Economics; processed. INDONESIA Creutzberg, P., ed. 1975-86. The changing economy in Indonesia. 10 vols. The Hague: Martinus Nijhoff. Polak, J. J. 1979. The national income of the Netherlands Indies, 1921-1939. In P. Creutz berg, ed. National income. The Hague: Martinus Nijhoff. Originally published New York,I943. MEXICO Solis, L. 1984. La realidad economica mexicana. Mexico City: Siglo Veintiuno. PERU Bolofia Behr, C. A. 1981. Tariff policies in Peru, 1880-1980. Ph.D. dissertation, Oxford University. ANGVS MADDISON 47 Other Lewis, C. 1948. The United States and foreign investment problems. Washington, D.C.: Brookings Institution. Royal Institute of International Affairs. 1937. The problem of international investment. London: Oxford University Press. Statistical Sources for Tables 1-12 to 1-18 Federal Reserve System. Monthly issues, various years. Banking and monetary statistics. Washington, D.C. League of Nations, Various years. Statistical yearbook, World production andprices, Review of world trade, Balances of payments, Money and banking, Public finance. Geneva: League of Nations. Maddison, A. 1982. Phases of capitalist development. Oxford: Oxford University Press. - - - . 1983a. A comparison of levels of GDP per capita in developed and developing countries, 1700-1980. Journal of economic history, March. - - - . 1983b. Estimates of Indonesian population, GDP, export volume, and real product levels since 1820. Groningen: University of Groningen; processed. Mitchell, B. R. 1982. International historical statistics: Africa and Asia. London: Macmillan. Statistisches Reichsamt. 1936. Statistisches Handbuch der Weltwirtschaft. Berlin: Statistis ches Reichsamt. United Nations. 1948. International capital movements during the inter-war period. Lake Success: United Nations. - - - . Economic Commission for Latin America (EeLA). 1976. America Latina: Relacion de precios del intercambio: Cuadernos estadisticos de la CEPAL. Santiago: United Na tions. - - - . 1978. Series historicas del crecimiento de America Latina: Cuadernos estadisticos de la CEPAL. Santiago: United Nations. U.S. Department of Commerce. 1977. National income and product accounts ofthe United States, 1929-1974. Washington, D.C. Wilkie, J. W., and S. Haber, eds. 1982. Statistical Abstract of Latin America. Los Angeles: University of California, Latin American Center. Woytinsky, W. S., and E. S. Woytinsky. 1955. World commerce and governments. New York: Twentieth Century Fund. Chapter 2 The Slowdown of the World Economy and the Problem of Stagflation Gottfried Haberler THERE IS NOW general agreement that the first quarter century after World War II was a period of almost unprecedented growth and prosper ity for the advanced industrial countries as well as for the less developed countries as a group. In sharp contrast with the twenty years after World War I, the interwar period, 1919-39, was marked by the severe depres sion of 1920-21 and the Great Depression of the 1930s. The post-World War II period has seen recessions but no depression, if by "depression" we mean a decline in economic activity on the order of magnitude of the depressions of the interwar period and earlier. The climate of optimism or even euphoria that characterized the 1950s and 196Os, however, has given way to pessimism and gloom since the early 1970s. There is again talk of depression and crisis. Two related develop ments account for the changed outlook. First, the recent recessions, 1973-75 and 1980-82, have been more severe in the United States than the earlier ones and have been marked by the unsettling experience of stagflation, with prices rising even in periods of falling output and em ployment. Second, there has been a distinct slowdown of productivity growth in all advanced countries since about 1972. The talk about crisis and depression, however, and the widespread fear that the dismal experience of the 1930s is about to repeat itself seem to me greatly exaggerated. An extreme example of totally unwarranted pessi mism is provided by Raul Prebisch (1984:175). He refers to the recent recessions and slowdowns in productivity growth as "the second great crisis of capitalism." The first great crisis was the Great Depression of the 1930s. It certainly was a crisis, but as we shall see presently, it was due to avoidable policy mistakes. It is a fatal misinterpretation to call it a crisis of capitalism. This paper was written in winter 1983-84. It was possible to add, mainly in footnotes, references to later events and to new source materials, but there was no comprehensive updating of the text. 48 GOTIFRIED HABERLER 49 I will now try to put recent developments and the present malaise in historical perspective. During World War II and the immediate postwar period it was not generally realized that the world economy was on the verge of an ex tended vigorous expansion. On the contrary, it was widely believed that the dismal pattern of the interwar period would repeat itself. In fact, most Keynesian economists thought that deflation and unemployment, not inflation, would be the order of the day, and in each of the early reces sions many experts saw the beginning of the postwar depression. 1 The actual fate of the world economy during the twenty-five or even thirty years after World War II has been authoritatively described by Simon Kuznets: "Even in this recent twenty-five year period of greater strain and danger, the growth in peace-time product per capita in the United States was still at a high rate; and in the rest of the world, developed and less developed (but excepting the few countries and periods marked by internal conflicts and political breakdown), material returns have grown, per capita, at a rate higher than ever observed in the past (Kuznets 1977:14, emphasis added; also see Kuznets 1971 and 1976). In "Aspects of Post-World War II Growth in Less Developed Coun tries," Kuznets remarked: "For the LDCS as a group, the United Nations has estimated annual growth rates of total and per capita GDP (gross domestic product at constant factor prices) from 1950 to 1972. The growth rates of per capita product ... for the twenty-two years was 2.61 percent per year .... Such growth rates are quite high in the long-term historical perspective of both the LDCS and the current DCS." These high growth rates are largely a recent phenomenon. "While the historical data for LDCS rarely provide a firm basis for judging their long-term growth," it can be established indirectly that in earlier periods the growth rates must have been lower. Applying the recent growth rates to earlier periods "would have meant impossibly low levels of per capita product and consumption at the beginning of the preceding quarter of a century." Kuznets further observes that, for the current developed countries for which we have long-term growth rates, "the observed rates (for well over half a century of their modern growth) are generally well below those cited for the LDCS" (1976:40-41). (From 1960 to 1972 the average growth rate of per capita GNP of some sixty-seven developing countries with more than 1 million population each, and omitting major oil exporters, was 2.6 per year [Kuznets 1976:42].) Kuznets is, of course, fully aware of the danger of using broad aggre gate measures of growth for the developing countries as a group, given their great diversity. He discusses and carefully evaluates possible biases in the procedures. Still, after everything has been said and done, he 50 SLOWDOWN AND STAGFLATION confirms the basic soundness of his findings and expresses his bewilder ment that, despite the "impressively high" growth rates "in the per capita product of LDCS over almost a quarter of a century," the general senti ment in the developing countries is one of dissatisfaction and gloom that "seems to ignore the growth achievements." He conjectures, and gives ample reasons for the speculation, that "a rise in expectations has pro duced a negative reaction to economic attainments which otherwise migh t have elicited litanies of praise for economic miracles" (1976 :41 and passim). Kuznets's findings were confirmed in an important paper by Kravis and Lipsey (1984) based on statistical data that became available after Kuz nets wrote. The authors concluded that "the three decades 195~0 were unique in economic history in two important respects. First the industrial countries, which had enjoyed rapid economic growth in previous eras, experienced unprecedented rates of expansion. What is even more re markable is the diffusion of growth to, and rapid growth of, almost the entire world . . . including a large fraction of the people of the develop ing world." There are, of course, some exceptions, including a number of countries "concentrated in Africa between the Mahgreb and South Af rica and in the Indian subcontinent" (1984:134-35). When the widely expected depression again and again failed to mate rialize, pessimism about the future gave way to optimism and euphoria. Keynesians claimed credit for the good showing. I cite a typical example. In 1967 Sir Austin Robinson wrote: "In the year 1947-1948 we began to use in peacetime the principles that Maynard Keynes had worked out for war finance. We began to plan the use of national resources. If we are looking to the credit side I think we can honestly say that the world today is a different place from what it was in the 1930s in very large measure as a result of the economic thinking that began in this Faculty in Cambridge in those exciting years of the 1930s" (1967:43). Similar claims were made by American Keynesians. It was widely assumed that the business cycle had been all but eliminated by skillful demand management and fine tuning of the economy. The Keynesians proved to be mistaken. The business cycle is still with us, although it has indeed become much milder. No decline in economic activity even approaching the depression of the 1930s has occurred since 1945. The absence of such a decline is a great achievement, despite the recent slowdown and the threat of inflation and stagflation. In my opinion, monetarists and liberals-liberals in the classical nineteenth-century sense-have a better claim to credit for the improved performance than the Keynesians. For there can be no doubt that the avoidance of deep depressions during the postwar period is mainly due to GOTIFRIED HABERLER 51 the fact that there has been no serious deflation in the sense of a decline either in the supply of money or in nominal GNP. This general statement follows from what I regard as an established conclusion that all major cyclical downswings-depressions as distinguished from recessions, in modern terminology-have been strongly intensified, if not caused, by monetary contractions. 2 The monetary factor was especially pronounced in the case of the Great Depression of the 1930s. The exceptional severity of the Great Depression in the United States was due to the fact that the U.S. mone tary authorities, by acts of commission and omission, let the money supply shrink by 30 percent. (Similar mistakes have been made in other countries, for example in Germany.) This explanation is, of course, in line with the monetarist theory but also with the views of Joseph A. Schumpeter, whose theory of the business cycle was decidely non monetarist. Schumpeter emphasized that the Great Depression was not a regular cyclical downswing but was due to special "adventitious" cir cumstances. He wrote: "I do not see how it could be denied that it was the-avoidable-three bank epidemics [bank failures] that ... spread paralysis through all sectors of the business organism, turned retreat into rout and thus were the most important reasons ... for the prevailing distress and unemployment" (1951:214). What would have been a reces sion, perhaps a relatively severe one, was thus turned into a catastrophic depression. The monetarist-liberal interpretation of post-World War II events is strongly supported by the German "economic miracle," which started with the currency reform of 1948 and the simultaneous abrupt disman tling of all wartime controls by Ludwig Erhard. Restoration of sound money and continuing monetary discipline was essential for the spectacu lar growth of the German economy, which in turn had a galvanizing effect on Western Europe. The prospects of the German reform were com pletely misjudged by leading American and British Keynesians (Haberler 1980; Hutchison 1979, esp. p. 435 ff.).3 The first quarter century after World War II was by general agreement a period of almost unprecedented growth and prosperity for the whole Western world, including the developing countries. The last ten years have been marred by rising inflation in the modern vicious form of stagflation, by more severe recessions, and by a slowdown or even stagnation of productivity growth. In no country, however, has there been a decline in economic activity approaching the depression of the 1930s or earlier ones. Furthermore, the U.S. economy has staged an unexpectedly vigorous cyclical recovery that started in December 1982. It is still going strong and is stimulating the world economy. Inflation has 52 SLOWDOWN AND STAGFLATION sharply declined, but it is too early to say, as many experts (especially Keynesians) are inclined to do, that inflation has been licked and is no longer anything to worry about. My next five sections deal with different aspects of the problems of inflation, stagflation, and disinflation in the last ten years. I shall first discuss the nature and cost of rising inflation, stagflation, and disinflation. In the following three sections I shall present three alternative explana tions, the "Keynesian" position, the "monetarist-rational expectations" position, and the "traditional conservative," or "classical," position. I shall subsequently discuss the elusive problem of the productivity slow down in industrial countries since the early 1970s and the evolution of the international monetary system in the post-World War II period. Finally I shall describe the policy implications of my analysis. The Cost of Stagflation and Disinflation We have seen that two distinct but related unsettling developments account for the change from optimism and euphoria of the 1950s and 1960s to pessimism and gloom in the 1970s and 1980s-the slowdown of productivity growth, on the one hand, and, on the other hand, the relative severity and inflationary character of the 1973-75 and 1981-82 recessions. I shall address the second development first. The proximate cause of the severe recessions of 1973-75 and 1981-82 in the United States and the other industrial countries is not hard to find. Inflationary pressures had increased everywhere since about the middle 1960s and reached a high point in 1974, a recession year. Suffice it to say that, in the United States, inflation, measured by the consumer price index, topped 12 percent. Even in Switzerland it climbed to the unheard of level of 10 percent. Thus the industrial countries were forced to step on the monetary brake. In the United States the inflation rate was reduced to a little below 5 percent by the end of 1975, and Switzerland abruptly deflated to practically zero inflation; output contracted sharply, but un employment rose only slightly because of the buffer of foreign workers. In the United States, inflationary pressures again increased after 1976 when the Carter administration switched from fighting inflation to stimu lating the economy. In the late 1970s inflation reached again the two-digit level. It once again became necessary to take restrictive monetary mea sures, hence the famous "dollar rescue operation" of October 1979 after Paul Volcker had been appointed chairman of the Federal Reserve Board. We can say, then, that the proximate cause of the recession was the policy of disinflation, which in turn was forced on the authorities by the GOTTFRIED HABERLER 53 high rate of inflation, notwithstanding the fact that in the United States in the early as well the late 1970s the pronounced weakness of the dollar in the foreign exchange markets was the greatest worry of the monetary authorities and induced them to take restrictive measures. The weakness of the dollar, however, was the consequence of the fact that the high rate of inflation in the United States greatly exceeded that in a number of industrial countries, primarily Japan, West Germany, and Switzerland. Let me recall that the weakness of the dollar led to its devaluation in 1971, especially vis-a.-vis the Japanese yen, the German mark, and the Swiss franc, and later in early 1973 to the breakdown of the Bretton Woods regime of stable but adjustable exchange rates and the adoption of managed floating of all major currencies. The recessions, that is, were due to the fact that high rates of inflation forced countries to adopt restrictive monetary policies. It is practically impossible to wind down inflation without creating transitional unemployment-without, in other words, creating a recession. Two questions arise. What were the reasons for the high rate of inflation? Why did prices continue to rise even during the recession, as they had not done in earlier recessions and depressions? [n other words, why stagflation? It should be mentioned that in the 1930s an early case of stagflation, or cost-push inflation, as it was called at that time, was caused in the United States by the policies of the New Deal. Milton Friedman described the episode: The only example I know of in United States history when such a cost-push was important even temporarily for any substantial part of the economy was from 1933 to 1937, when the NlRA [National Indus trial Recovery Act], AAA [Agricultural Adjustment Administration], Wagner Labor Act, and associated growth of union strength unques tionably led to increasing market power of both industry and labor and thereby produced upward pressure on a wide range of wages and prices. This cost-push did not account for the concomitant rapid growth in nominal income at the average rate of 14 percent a year from 1933 to 1937. That reflected rather a rise in the quantity of money at the rate of 11 percent.... The cost-push does explain why so large a part of the growth in nominal income was absorbed by prices. Despite unprecedented levels of unemployed resources, wholesale prices rose nearly 50 percent from 1933 to 1937, and the cost of living rose by 13 percent. Similarly, the wage cost-push helps to explain why unemploy ment was still so high in 1937, when monetary restriction was followed by another severe contraction. [1966:22] The alarming rise in prices forced the Federal Reserve System to step on the monetary brake. The result was a short but very sharp depression. 54 SLOWDOWN AND STAGFLATION In thirteen months unemployment rose from 10 percent to 20 percent, real GNP dropped by 13.2 percent, and industrial production fell by 32.4 percent. An exogenous factor is often regarded as a major cause of inflation in the 1970s and 1980s-the two oil shocks, the quadrupling of crude oil prices by OPEC in 1973 and the doubling of oil prices in 1979. 4 These were indeed highly inflationary and disruptive events, but the direct impact of the oil price rise on inflation in the United States and other industrial countries has been greatly exaggerated. The first oil shock was preceded by a highly inflationary commodity boom, which in turn was superim posed on an inflationary groundswell that started in the middle 1960s (see IMF 1981). It can be argued, however, that the oil price had a significant indirect effect on inflation, through the manner in which the economy reacted to it. Let us assume, for example, that money wages and other incomes (or even real wages through widespread formal or informal indexation) are rigid downward, and let us also reflect that the oil price rise constitutes a sharp deterioration in the terms of trade, which implies a reduction in the standard of living. The consequence of the rise will be unemployment and inflation. I shall have more to say on this point later. The Keynesian Position on Inflation and Stagflation As we have seen, a sharp distinction must be made between Keynesian economics and the economics of Keynes himself. Keynes was concerned about inflation all his life except during the Great Depression of the 1930s when he wrote The General Theory. In 1937, a year later, however, he again became concerned about inflation. What he would have said about stagflation-the vicious combination of high or even rising unemploy ment and inflation-is a matter for speculation; he died before the problem became acute in the post-World War II period. Until very late in the game (see below), the views of Keynes's followers were characterized by emphasis on fiscal policy, by optimism that the economy can be fine-tuned by skillful demand management, and by unconcern about the rising danger of inflation as well as by neglect of inflationary expectations. This attitude of course sharply contrasts with monetarism and its offshoot, the modern theory of rational expectations, as I shall presently explain. The Keynesian unconcern about inflation and the neglect of inflation ary expectations is highlighted by the theory of the Phillips curve, which postulates a more or less permanent tradeoff between unemployment and inflation; lower levels of unemployment can be obtained by accepting GOTTFRIED HABERLER 55 higher rates of inflation. This theory for years played a leading or even dominating role in the Keynesian discussions of inflation. I cite two examples of prominent Keynesians. In 1960 Paul A. Samuelson and Robert M. Solow wrote the celebrated article "Analytical Aspects of Anti-Inflation Policy." It presented a modified Phillips curve for the United States, which the authors described as "the menu of choice[s] between different degrees of unemployment and price stability (1960:192). See figure 2-1. The authors mentioned Figure 2-1. Modified Phillips Curve for the United States Average annual price rise (percent) Unemployment rate (percent) Note: This shows the menu of choices between different degrees of unemployment and price stability, as roughly estimated from the last twenty-five years of U.S. data. Source: Samuelson and Solow (1960:192). 56 SLOWDOWN AND STAGFLATION specifically two "obtainable" choices: price stability with 5.5 percent unemployment and 3 percent unemployment with 4.5 percent inflation a year, marked as (A) and (B) on the graph. Now, the point I wish to make is that the authors do not say whether they regard other points on the curve as "obtainable" choices, for exam ple, the point that I mark (C), 1.10 percent unemployment with 10 percent inflation. As we see matters now, this point would not be "obtainable." With 10 percent inflation, the short-run Phillips curve would not stay put. Inflationary expectations leading to anticipatory action by market participants would shift the curve upward. Although the authors envisage shifts of the curve due to structural changes in the economy, they do not mention that inflationary expectations would shift the curve. It is true that there is a fleeting reference to inflationary expectations, but it comes earlier in the paper and is not made in connec tion with the Phillips curve. The authors say that inflationary expectations would be caused by "a period of high demand and rising prices." In flationary expectations would "bias the future in favor of further infla tion." The importance of the matter is immediately played down, how ever: "Unlike some other economists, we do not draw the firm conclusion that unless a firm stop is put, the rate of price increase must accelerate. We leave it as an open question: It may be that creeping inflation leads only to creeping inflation" (Samuelson and Solow 1960:185). Can we define "creeping"? Surely 10 percent inflation cannot be described as a creep. Inflation would accelerate, or if it is stopped the result would be recession. I think it is fair to say that the paper by Samuelson and Solow illustrates my point that Keynesian economics until recently was characterized by unconcern about the dangers of inflation and neglect of inflationary expectations. The paper, as I have noted, was written in 1960. The inflationary excesses of the last ten to fifteen years have sensitized in flationary expectations of market participants and have alerted econo mists to the dangers of inflation. The change in outlook has reached its climax in the theory of rational expectations, as I shall show. I now turn to my second example of the Keynesian blind spot concern ing inflation. As late as 1972, shortly before inflation in the United States and other industrial countries soared into the two-digit range, James Tobin extolled the virtues of inflation in adjusting "blindly, impartially and nonpolitically the inconsistent claims of different pressure groups on the national product" (1972:13). He later gracefully admitted that he had "been overoptimistic about the trade-off [between unemployment and inflation] and too skeptical of accelerationist warnings" (1973:622). Keynesian economics was ill equipped to deal with the problem of stagfla tion because it was essentially depression economics; wages and prices GOTTFRIED HABERLER 57 were assumed to be constant. It seems fair to say that this was accepted by Keynes's followers and his critics as the central message of The General Theory.s My point is illustrated by the following true story from Nazi Germany. When Hitler came to power, there was an influential Keynesian econo mist in the economics ministry named Wilhelm Lautenbach, who had advocated deficit spending for some time. In fact, there had been deficit spending even before Hitler, but under Hitler the deficit became much larger, though there was still great fear of inflation. Hitler called Lauten bach, who was not a Nazi, into his office and asked: "Isn't this a little dangerous, what we are doing? Are we not risking a serious inflation?" Lautenbach answered: "Mein Fuhrer, you are a very powerful man, but there is one thing you cannot do: you cannot make inflation with 30 percent unemployment!" What Hitler supposedly could not do-produce inflation in the midst of heavy unemployment-the New Deal accom plished in the United States. In the late 1960s and 1970s the dangers of inflation, despite substantial unemployment, could no longer be ignored. The first reaction of Keynesi ans was to recommend "income policies" to support macro policy. I shall return to the problem of incomes policy in the section on policy implica tions below. More important, there has been a gradual shift of Keynesian economics away from depression equilibrium with stable wages and prices of the General Theory to a neoclassical position. We might also describe this shift by saying that the Keynesians have been catching up with the master, who in 1937 already realized that the economic climate had changed. A good example of that shift is Paul A. Samuelson. In his sparkling contribution to the Keynes Centenary Conference in Cambridge in 1983 he stated that in the present world neither the "depression Keynesian model" of the General Theory nor "the market-clearing new classical theory model" works well anymore. He concludes: "If I had to choose between these two extreme archetypes, a ridiculous Hobson's choice, I fear that the one to jettison would have to be the Ur-Keynesian mode}"; and "people learn faster these days and the easy Keynesian victories are long behind us" (Samuelson 1983:212). I myself have expressed the same idea by saying that today the world is closer to the classical position than to the Keynesian one. Monetarism and Rational Expectations Monetarism and its offshoot the theory of rational expectations are characterized by reliance on monetary policy, the downgrading of fiscal 58 SLOWDOWN AND STAGFLATION policy, the rejection of fine-tuning, incomes policies, and controls, and emphasis on the role of expectations. The two schools are liberal in the classical nineteenth-century sense and are direct descendants of the quantity theory of money. Broadly speaking, the quantity theory of money is undoubtedly cor rect, namely in the sense that there has never been a significant inflation, with prices rising by, say, 4 percent or more per annum, for perhaps two, three, or more years without a significant increase in the quantity of money. Furthermore, a reduction of monetary growth is a necessary condition for reducing or eliminating inflation. Such statements are empirical. Exceptions are thinkable. If output fell sharply because of a natural catastrophe, a string of crop failures, or a war, for example, the price level would rise even with a constant quantity of money. War inflations have in fact been aggravated by a drop in output, but there can be no doubt that most wartime price rises were always due to inflationary methods of financing the war. The oil price rise imposed by OPEC is often mentioned as a nonmonetary cause of inflation. We have seen, however, that the direct impact of the oil shocks on the price level has been greatly exaggerated. There is furthermore the problem of a change in the velocity of circula tion of money. A sharp increase in velocity would spell inflation, and a decrease would spell deflation-even with a constant quantity of money. Does that statement not invalidate the quantity theory as formulated above? The answer is no for the following reason: it is true that the velocity does change over time. There are slow structural changes, and velocity has a cyclical pattern; it tends to rise in business cycle upswings and to decline in downswings. These changes are not large enough to invalidate the quantity theory as formulated above. Large changes that would invalidate the quantity theory, if they were autonomous, are clearly the consequence of inflation or deflation caused by changes in the quantity of money. The point is well illustrated in the case of high inflation. During the well-documented case of the German hyperinflation in 1922-23, for example, the velocity of circulation of money rose to fantastic heights, far outstripping the rise in the (nominal) quantity of money and in the price level. This inflation, however, was clearly the consequence of printing money.6 Policy conclusions are straightforward. To curb inflation, monetary growth must be reduced. It is impossible to stabilize a significant level of inflation, say, 6 percent per annum or more. Monetarists and rational expectations theorists reject the idea of a permanent tradeoff between inflation and unemployment as enshrined in the Phillips curve. An ex tended period of significant inflation cannot mark an equilibrium posi GOTTFRIED HABERLER 59 tion, because inflationary expectations would quickly develop and mar ket participants would take anticipatory measures. Interest rates would rise, labor unions would press for higher wages, and wholesale commod ity prices would rise-all of which would rob inflationary policies of their stimulating power, and unemployment would rise again. If an attempt was made to prevent a rise in unemployment by expansionary measures, the consequence would be accelerated inflation. Rational expectations theorists have pushed that argument to an ex treme. They say that macroeconomic policies of expansion or contraction have no effect whatsoever, not even in the short run, on the "real" economy, on output and employment. Such policies merely affect the price level and nominal interest rates. The reasoning behind this surpris ing conclusion is that market participants, "agents," by and large act rationally. They determine what the likely consequence of government policy will be, and they take anticipatory action. Limitations of space preclude my discussing the rational expectations theory, which has been developed, criticized, and defended in scores of papers, in greater detail. Suffice it to say that the "hard-line version" of the theory, to wit that macropolicies have no effect on the real economy, is widely regarded, also by monetarists, as an exaggeration that cannot be reconciled with the facts. There is widespread agreement, however, that the longer inflation lasts, even in the intermittent, stop-and-go form, the more firmly entrenched inflationary expectations become, and the harder it is to stop the inflation. This reasoning leads to the conclusion that it is impossible to stabilize a significant rate of inflation. The goal of the policy must be to bring down inflation by monetary restraint to a very low level, say, 2-4 percent. At this point I must mention the so-called credibility approach, which was developed by William Fellner in several important papers (1976, 1979a, 1980). Fellner was a sympathetic critic of the rational expectations theory. His theory tries to capture what he called "the valid core" of the rational expectations theory, namely that rational market participants are trying to determine what the government policy is likely to be. If the authorities make their policy of disinflation "credible," if they persuade the public that they will stick to it and will not give up, as they did so often in the past when unemployment rose a little bit, they will "condition" inflationary expectations. Labor unions and other pressure groups will be put on notice that they will price themselves out of the market if they push up wages and prices. As a result there will be a good chance that wage and price demands will moderate and thus the pains of disinflation will be reduced, although it is too optimistic to assume that disinflation will become entirely painless. 7 60 SLOWDOWN AND STAGFLATION Two important questions arise: what is the role, if any, of fiscal policy in the monetarist scheme of things, and is there any room at all for incomes policy in some sense ofthat ambiguous term? The importance of fiscal policy is highlighted by the current debates about the large budget deficits of the United States and other countries. Naturally the large current and projected deficits of the United States, whose economy looms so large in the world, have become the prime subject of the debate. Before discussing the question that interests us most at this point whether the deficit threatens to abort the current expansion and to reignite inflation-let me say a few words on the international aspects of the problem. The U.S. deficits have become a matter of great concern and anxiety abroad. U.S. fiscal policy has been strongly criticized by international officials and foreign governments: the budget deficits have raised U.S. interest rates, which attracts foreign capital, pushes up the dollar in the foreign exchange markets, and creates huge U.S. trade deficits. Foreign countries are thus forced to finance U.S. trade and budget deficits, are put under inflationary pressure, and are compelled to raise interest rates. There certainly is much truth in the criticism, but it has been somewhat defused by the vigorous recovery of the U.S. economy that has developed and is still going strong despite the drag of the large trade deficits and the appreciation of the dollar. It can, indeed, be argued that the trade deficit and the appreciation of the dollar have prolonged the expansion by slowing its pace to a more sustainable speed and by keeping inflation at a lower level. Moreover, the U. S. recovery is making a major contribution to the recovery of the world economy from the recession. I now come to the question of the role of fiscal policy and budget deficits in a closed economy as exemplified by the United States at the present time. There is widespread fear, which is probably shared by a great majority of professional economists of different schools, that the large budget deficits will reignite inflation and will jeopardize the recovery.8 How do fiscal policy in general and budget deficits in particular fit into the monetarist picture of the world? In every major inflation, government deficits have been blamed as a prime source of the trouble. Austerity programs prescribed by the International Monetary Fund (IMF) for trou bled countries that ask the Fund for help invariably include provisions for fiscal discipline. The reason is that troubled countries put the monetary authorities under irresistible pressure to "finance" the deficit by printing money. Monetarists would, however, insist that ifthe central bank stands firm and does not increase the money supply, even large budget deficits will not cause inflation. GOTIFRIED HABERLER 61 The statement is true, but monetarists would agree that it is not the end of the story. Large deficits push up interest rates, causing a doubly or triply adverse effect on the economy: higher interest rates could speed up the velocity of circulation of money and so could lead to inflation even without an increase in the money supply; more important, high interest rates could abort the cyclical recovery, and productive private invest ments would be crowded out by government profligacy, slowing long-run growth. We see, then, that loose fiscal policy and large budget deficits need not cause inflation if monetary policy is sufficiently tight. Depending on the magnitude of the deficits, however, the adverse side effects of loose fiscal policy can be very serious. For this reason monetarists often insist that monetary restraint brings best results when it is supported by fiscal restraint. Finally, I should say a few words on the Keynesian reaction to this picture. Let us recall that in the 1930s, during the Great Depression, when Keynes wrote his General Theory, the economic situation was entirely different from that which we see today: there was mass unem ployment, prices were falling, expectations were deflationary, and the general mood was one of despair and gloom. In that climate Keynes was quite right to call for deficit spending and monetary expansion; present day monetarists should agree. If my memory serves me right, I heard Milton Friedman himself on television say so. We may recall that Frank Knight, Jacob Viner, and Henry Simons, the teachers ofthe present-day monetarists in Chicago, recommended deficit spending and monetary expansion. 9 To be sure, if monetary expansion had been applied in time, it would have stopped deflation in its tracks, and after the deflationary spiral had gathered momentum, vigorous monetary expansion through open mar ket operations would eventually have pulled the economy out of the slump. The cure would have taken some time, however, and in the process a large pool of liquidity would have been created that would have caused inflationary troubles after the economy had turned upward again. It was therefore better to inject money directly into the income stream by deficit spending. The 1930s were the heroic age of Keynesianism. Today the situation is quite different; unemployment is low compared with its level in the 1930s, below 10 percent as against 25 percent. lO There is inflation, and expecta tions are inflationary. Keynes himself recognized, as we have seen, the change in climate already in 1937, one year after he wrote The General Theory, but some Keynesians still live in the heroic age. James Tobin (1983) repeatedly speaks of the "Volcker and Thatcher depression" (not 62 SLOWDOWN AND STAGFLATION recession), which suggests that the last cyclical decline was on the order of magnitude of the depression of the 1930s or earlier ones. The development of the British economy under the Thatcher govern ment was completely misjudged by the Keynesians, as The Economist observed: "When the government raised taxes and cut public borrowing in the 1981 budget, 364 academic economists" issued a manifesto predict ing dire consequences. "Shortly thereafter recovery began," and the economy is still growing at a healthy pace, despite the fact that the government has continued to shrink public sector borrowing (Economist 1984:13).11 The Traditional Conservative, or Classical, Position What I call the traditional conservative, or classical, view concurs with the basic tenets of the monetarist-rational expectations theory concern ing the role of monetary policy, fiscal policy, and expectations, including the rejection of a permanent tradeoff between unemployment and infla tion as outlined above. '2 The traditional conservative view stresses, however, the importance of growing wage and price rigidity caused by labor unions and other pressure groups, aided and abetted by govern ment policies-institutional rigidities-which are downplayed by the monetarist and rational expectations school. Members of the school (monetarists for short) usually use models that assume perfect competi tion-in other words, instantaneously or at least rapidly clearing mar kets. For many problems, especially those concerning long-run adjust ment, these models give an acceptable approximation. A satisfactory solution of the problems of unemployment and stagflation is not possible, however, if we presuppose perfectly competitive markets. In fact, persist ent unemployment and extended periods of stagflation are inconsistent with perfect competition. Unemployed workers would compete for jobs; thus wages would decline relative to prices, and full employment would be restored. An instructive contrast in policy that strikingly illustrates the impor tance of wage control is provided by the New Deal in the United States and Nazi Germany. Roosevelt and Hitler came to power at about the same time fifty-two years ago. Each man found his country in deep depression and immediately undertook expansionary measures. 13 Eco nomic recovery in the United States, although long and pronounced, was (as we have seen) marred by rising prices, which led to the interruption of the upswing, long before full employment had been reached, by the very vicious though brief depression of 1937-38. The German recovery, on the GOTIFRIED HABERLER 63 other hand, proceeded without interruption, and Germany reached sub stantially full employment within two or three years. Even more impor tant, the price level in Germany remained remarkably stable for several years. It would be tempting to attribute the rapidity of the German recovery as compared with the U.S. recovery to massive expenditures on arma ment. Heavy public spending there was indeed, but massive rearmament came only later. Possibly German public spending was comparatively larger than the American, but this difference would not explain the different price performances. The basic difference between the Amer ican and German recovery policies lies elsewhere: the New Deal com bined spending with deliberate price and wage boosting. As a conse quence, an exceptionally large part of the rising nominal GNP in the United States took the form of higher prices rather than larger output and employment. Full employment was reached only after the outbreak of the war in Europe when defense spending soared-a dismal failure indeed. In Germany, by contrast, money wage rates remained fairly stable, although average annual earnings of labor rose rapidly in mone tary and real terms, along with rising output and employment, because unemployment disappeared and the work week lengthened. We might object that price controls and rationing make real wage figures under the Nazi regime meaningless. True, there were wage and price controls right from the beginning of the Nazi dictatorship; and later-say from 1936 on, when the rearmament boom came into full stride-scarcities, unavailabilities, and the deterioration in quality of certain commodities made the official cost-of-living index increasingly unreliable. Fortunately we have the careful study by Bry and Boschan (1960). Bry makes adjustments in the cost-of-living index to take the controls into account. His figures show that, from about 1937 on, the official index understates the true rise in the cost of living. For the earlier years (1933-37), however, the corrections are minimal. There can, then, be no doubt about the great economic success of the Hitler regime. It is no exaggeration to speak of a German economic miracle. I< The present situation in all industrial countries is that the rise of powerful labor unions and other pressure groups, aided and abetted by government policies-minimum-wage laws, generous unemployment benefits, welfare payments-have made wages and many prices rigid downward. Before I go into detail, I must say a few words about the concept of unemployment and full employment. For two reasons full employment clearly does not mean that unemploy ment is zero or near zero. First, the measured unemployment always contains some spurious "voluntary" unemployment, that is, workers who 64 SLOWDOWN AND STAGFLATION do not care to work at the ruling wage for the type of work that they are qualified to do (although they may work at a higher wage) but who are somehow counted as unemployed and who receive unemployment bene fits. Voluntary unemployment has plainly become an important factor since the advent of generous unemployment benefits and welfare pay ments. Second, there is "frictional" unemployment, or workers "between jobs," in Keynes's felicitous phrase. Workers who have lost their jobs or wish to change their employment usually take their time to find a suitable new position. Again, generous support for unemployeds plainly leads to an increase in frictional unemployment because workers can afford to take more time to look for a new job. The same reasoning applies to the so-called structural unemployment, which can be described as fricitional unemployment writ large-disloca tion and layoffs caused by large structural changes in the economy, for example when horses are replaced by tractors or when cheap imports take the place of comparatively expensive home production. (I shall have more to say about structural unemployment below.) The difference between frictional and structural unemployment is one of degree, and the magnitude of voluntary, frictional, and structural unemployment can only be guessed. Our guesses take the form of state ments that at present in the United States 6 percent unemployment would be compatible with full employment. Measured unemployment minus voluntary, frictional, and structural unemployment is what Keynes called "involuntary unemployment. "15 The concept and Keynes's definition of it have given rise to much con troversy. The commonsense definition is this: a man is involuntarily unemployed if he is not working but is seeking work and is willing to work at the ruling wage and working condition for the type of work for which he is qualified. 16 Such a definition guides the compilers of unemployment statistics, who must be on their guard to keep voluntary unemployment out of their figures. The main causes of unemployment are fairly clear. Wages have always been sticky. They were sticky even before labor unions became as power ful as they are now and before the modern welfare state with generous unemployment benefits and welfare payments came into existence in the Great Depression of the 1930s. Labor markets have never been, and in the nature of the case cannot be, perfectly competitive auction markets. For this reason there was much unemployment in earlier depressions, even before unions reached their present level of power, even when the modern welfare state did not exist. There can be no doubt, however, that the stickiness of wages has greatly increased since the Great Depression. GOTIFRIED HABERLER 65 I shall first mention two striking examples of this trend. In the United States and some other industrial countries, legal minimum-wage laws have caused heavy unemployment among teenagers, as much as 40 percent among blacks. Second, in the U.S. steel and automobile indus tries, although unions have made some concessions, wages are still more than 50 percent higher than the average in U.S. manufacturing industries, despite the fact that unemployment in these two industries is very high because of foreign competition. The growing rigidity of wages and the resulting unresponsiveness of wages and of prices to the decline in economic activity during recessions have been widely discussed in the literature. I cite two papers that I found especially important, Cagan (1975) and Sachs (1980; this paper has exten sive references to the literature). See also Price (1982) and Sachs's reply (Sachs 1982). Phillip Cagan concludes that "wholesale prices show a smaller decline in the recessions after 1948-49 than formerly" and that "there has clearly been a gradual decline in price response to recessions over the postwar period, except mainly for raw materials prices" (1975:54-55). Sachs uses two methods to demonstrate the decreasing responsiveness of inflation to changes in aggregate demand, with special emphasis on wage behavior. The first approach follows Cagan's method and leads to the same conclusion for a longer period. A striking finding is that "for mild contractions, downward price flexibility seems to have ended with the pre-World War II period. For moderate and severe contractions, similarly, the response of wages and prices has fallen sig nificantly since 1950" (1980:81). The second approach is described as an econometric Phillips curve estimation. The result that the short-run Phillips curve has become steep er strongly supports the hypothesis of decreasing responsiveness of wages to declines in economic activity. The causes of this development are fairly obvious. The spread of unionization and the increased strength of labor unions, which is largely traceable to the New Deal legislation of the 1930s, is surely basic. Union wages are notoriously stickier than nonunion wages; their even greater stickiness on the upside of the cycle is far more than counterbalanced by the rapid lengthening of union contracts provid ing hefty wage increases for each year. The overlapping and leapfrogging of union wage contracts in the United States have greatly exacerbated the wage inflation caused by unions (see Taylor 1982 and the specialized literature that he quotes). In other industrial and industrializing coun tries, too, the power of labor unions has sharply increased, and so have wage push and what J. R. Hicks calls "real wage resistance." Naturally the pattern and force of this development vary from country to country, depending on the structure of the economy, the history of the labor 66 SLOWDOWN AND STAGFLATION movement, and labor's alliance in many countries with political, mainly socialist, parties. Without going into too great detail, I should mention that in most European countries a much larger percentage of the labor force is union ized than in the United States and that in some countries-Austria and Sweden, for example---central organizations represent labor as a whole and allow it to speak with one voice. Whether strong centralized organiza tions representing most or all of labor increase or decrease the dangers of inflationary wage push, compared with the U.S. decentralized system of essentially independent unions, is difficult to say. On the one hand cen tralized organizations increase the power of unions and their political clout. On the other hand, however, such organizations eliminate leap frogging, and the leaders of an all-embracing organization of labor may perhaps be assumed to be more responsible, more aware of the general good of the population as a whole, than the bosses of independent unions. International comparisons could throw some light on these questions. I must confine myself, however, to two remarks. International compari sons are complicated by the fact that union power depends also on the structure of the economy. It is well known, for example, that unions are more moderate in small countries, where competition from world mar kets is strong because foreign trade plays a much greater role than in large countries. Second, Japan is a special case that merits close attention. Wages of workers are much more flexible in Japan than in Europe and the United States because Japanese workers receive a large part of their earnings in th~ form of a bonus that varies in size with the level of profits. In recessions, when profits are low, wages and the cost of labor thus automatically decline. For this reason and others, unemployment is lower in Japan, and the Japanese economy rebounds more quickly from reces sions than do the economies of most other industrial countries. A few remarks should be made about recent developments in the United States as compared with the situation in Europe. In the last U.S. recession, which came to an end in November 1982, wages became more responsive to unemployment and slack than was expected. Partly for this reason, inflationary pressures until now have remained comparatively mild during the vigorous cyclical recovery that started in December 1982 and is still going strong. It is very doubtful, however, that there has been a basic change. The dilemma of stagflation has not disappeared, and infla tion will probably accelerate as the expansion continues. The contrast between the United States and Europe, however, is remarkable. The European economic picture is much less bright: the recovery from the recession started later and is slower than in the United States, unemployment is higher, and there is more inflation in most GOITFRIED HABERLER 67 European countries than in the United States. What accounts for this difference? The basic structural reasons have been very well described by Stephen Marris, the former chief economic adviser of the Organisation of Economic Co-operation and Development (OECO; Marris 1984). I quote some salient facts as presented by Marris: European economies are in important respects less flexible than the American economy .... European workers are generally better pro tected against economic misfortune than their American counterparts. Collective agreements and government regulations give them more job security. But this makes it more difficult and expensive for Euro pean employers to layoff workers when demand weakens. And, they are more reluctant to take on new workers when demand picks up, preferring instead to work overtime. Provisions for unemployment are also more generous in Europe. Laid-off workers have more time to look around for a new job. But, by the same token, this slows down the movement of labor from declining to expanding industries. Labor mobility is also inhibited in Europe by the greater rigidity of the relative wage structure between industries, occupations, and re gions. It is more difficult for employers in expanding industries to bid up wages to attract labor, or for laid-off workers in declining industries to bid down wages to get their jobs back .... The main culprit is the downward rigidity of real wages, coupled with the high taxes .... Europeans have been reluctant to swallow the rapid rise in taxation needed to finance the very rapid rise in public expenditure. Between 1960 and 1983 the ratio of general government expenditure to gross national product (GNP) in the European Com munity rose from 32 percent to 52 percent.... In America the overall burden of taxation is lower, and real incomes seem to have adjusted more flexibly to the shocks of the 1970s. 20 million new jobs have been created in America since 1973. . . . Against this, there was a net loss of around 2.5 million jobs in the European Community over the same period. Compared with Euro peans, Americans coming into the labor force have been more willing to accept whatever level of real wages was necessary to induce em ployers to hire them; in other words, to "price themselves" into jobs. [Marris 1984:14--15.] To Marris's list of European handicaps I would add the following: the U.S. economy enjoys the tremendous advantage of a large free trade area and of private competitive enterprise in the fields of transportation, communications, and electric power. The European Common Market is supposed to have established free trade among the members of the 68 SLOWDOWN AND STAGFLATION European Community (Ee). There still exist many impediments to the free movement of commodities, however. Customs formalities and in spection at the borders are still in place. Even more important, European countries are burdened in various degrees by the existence of national public monopolies in the areas of transportation, communications, and electric power. These public monopolies suffer to a varying extent from bureaucratic inefficiencies and are impervious to international competi tion. In addition, numerous nationalized industries suffer from the same handicaps. As we shall see presently, monopolies, private or public, also greatly strengthen the power of labor unions, making wages rigid and adding to the wage push. Another inflationary factor in all industrial countries, and one that derives its importance partly from the existence of powerful unions, is the rise of Keynesianism and the resulting emphasis on antidepression and antirecession policy. Keynesianism has had a double inflationary effect: it has reduced the price decline in the downswing of the business cycle and has stiffened the resistance of workers and their unions to wage cuts because they assume that unemployment caused by large wage boosts will not last long-in other words, that government policy will bail them out if they cause unemployment by excessive wage demands. I t is now widely recognized that wage rigidity, real wage resistance, and real wage push are the most serious impediments for regaining price stability at high levels of employment and higher growth rates. In recent years more and more economists have come to the conclusion that a decisive and lasting recovery from the world recession requires a reduc tion of the level of real wages and a substantial increase in profits to stimulate investment and growth. This solution implies a moderate de cline in the share of wages in GNP; "wages," of course, include salaries. Statements about the adverse effect of wage rigidity and real wage resistance apply also to other incomes or prices that have been made rigid by government action ranging from farm supports to social benefits of various kinds. I cite a few examples of the trend in thinking about wages. Two years ago, a group of prominent German economists, including several of monetarist persuasion, issued a statement urging a temporary wage freeze to let productivity growth catch up with the wage level and to permit a little inflation to bring down real wages. The plea was not heeded; wages continued to rise, and unemployment has reached the two-digit level. Herbert Giersch, an author with monetarist leanings, has argued in several important articles (1982, 1983) that all industrial coun tries suffer from excessively high real wages and too low profits. He thinks it will take several years to bring the necessary adjustment in the income distribution. GOTIFRIED HABERLER 69 The theme has been taken up by The Economist (London) in two important articles (1982a, 1982b). The Economist asks for a substantial cut in money wages to bring about an increase in profits for the purpose of stimulating investment, growth, and employment. Predictably, this call has shocked many of The Economist's Keynesian readers.17 The main argument against cutting money wages as a recovery measure is that it reduces total spending by reducing money income of labor and thus is a deflationary factor that intensifies the recession. This argument is fallacious, however, and rests on a misunderstanding of the aim of a wage cut. The purpose is not to reduce effective demand (nominal GNP); if such a reduction is necessary, it should be done by monetary-fiscal measures. The purpose of cutting money wages is to boost profits and stimulate investment, employment, and growth by making labor more competitive with robots and other machines-in other words, with capital. If hourly wage rates are cut by 10 percent, the wage bill and spending power of labor will not necessarily be reduced. If the elasticity of demand for labor is greater than unity (as it almost certainly is in the medium run), employment will rise by more than 10 percent, and sO will the wage bill and spending. True, if employment rises by less than 10 percent, labor incomes will decline, but total incomes and spending will not necessarily decline as well. A shift to profits will stimulate investment, employment, and growth. This tendency could be assisted by monetary expansion, for the reduction of unit labor cost would reduce the inflationary danger of easier money. These three statements of the problem share the assumption that, if a moderate cut in the wage level is achieved, and macroeconomic levers are set right, market forces will, in due course, bring about the necessary restructuring of the economy to achieve substantially full employment. The Economist (1982a) voices that assumption very clearly. It argues that entrepreneurs would find hundreds of ways of substituting labor for capital if labor costs were reduced, just as entrepreneurs found ways to substitute capital for labor when wages went up. The Economist's optimistic conclusion will be challenged by structural ists. In the 1930s it was widely believed that part of the unemployment problem was that labor-saving inventions had reduced the demand for labor or that the "structure of production" had been distorted in some other way. In other words, it was argued that a large part of unemploy ment was "technological" and "structural," requiring large-scale real location of factors of production, a time-consuming, painful process. There can be no doubt that subsequent developments were entirely at variance with the structuralist theory. Experience showed that as soon as deflation was stopped, the huge structural distortions that had been 70 SLOWDOWN AND STAGFLATION diagnosed by theorists during the depression shriveled as quickly as they had surfaced earlier. Monetary contraction, sometimes called "secon dary deflation," proved to be a much more important cause of high unemployment than structural distortions, which may have triggered the deflationary spiral. In other words, the great bulk of unemployment was "Keynesian," or monetarist, if you like, not structural, or "Hayekian." Extreme structuralist views can be heard again today. It has been said that robots and other "smart" machines have put human labor in the position that horses were in when tractors came into wide use. This analogy is, however, very misleading. Tractors replaced not only horse power but also manpower. Unlike horses, however, human labor could be shifted to producing tractors. I do not deny that technological progress may possibly require realloca tions of factors of production that may cause some structural unemploy ment until the transfer and retraining of labor have been carried out. As we have seen, a modest decline of the share of labor in the national product is probably required at the present time. It is most unlikely, however, that a large reduction of the marginal productivity of labor, an intolerable drop in real wages, and a massive decline of the share of labor (and salaries) would occur, as the analogy with the horses suggests. As far as we can tell, the share of labor in the national product has remained remarkably stable over the long pull-apart from cyclical fluctuations despite the tremendous technological changes, including mechanization and automation, that have occurred since the industrial revolution in England. Nowadays we hear gloomy forecasts of the disaster that will befall us unless radical reforms are undertaken that involve massive redistribution of income to spread work. I believe that these forecasts will prove to have been totally unfounded and will share the fate of earlier, similarly gloomy prophecies regularly made in periods of depression, beginning with those underlying the Luddite movement in the early nineteenth century and extending to the most famous one, Karl Marx's theory regarding the increasing misery of the working classes-prophecies that were com pletely disproved and discredited by subsequent developments. The Slowdown of Productivity and GNP Growth since 1970 We have seen that two developments account for the change in outlook from optimism and euphoria in the 1950s and 1960s to pessimism and gloom in the 1970s and 1980s, first, the severity and inflationary character of the last two recessions and, second, the slowdown of productivity growth in all industrial countries since 1970. GOTIFRIED HABERLER 71 It was suggested that the comparative severity of the recent recession can be attributed to the fact that all industrial countries were forced, after an extended period of inflationary abuse, to curb inflation by monetary restraint. It is simply impossible to stop an entrenched inflation without creating considerable transitional unemployment. The slowdown of pro ductivity growth is a much more elusive problem and indeed an ex ceedingly complex one. Many different explanations have been offered, and a combination of factors is doubtless responsible. 18 There are two connections between the two developments. The rising tide of world inflation, which started in the middle 1960s and led to the inflationary recessions of the 1970s and 1980s, was probably a major cause of the slowdown of productivity growth, which in turn reinforced inflation. Persistent inflation, especially the vicious form of stagflation, does not offer a healthy climate for the investment by innovating entre preneurs on which long-run productivity growth depends. Second, the "adjustment," or "supply-oriented," policies that have been recom mended by many experts to ease the pains of disinflation-that is, mea sures designed to bring the economy closer to the competitive ideal by attacking and eliminating monopolistic restrictions in labor and commod ity markets--are surely also required to speed up long-run productivity growth. In the extensive literature on the slowdown of the world economy, many other causal factors have been mentioned. A favorite among policymakers is the two oil shocks of 1973 and 1979, for which policymak ers cannot be held responsible. The oil price rise is supposed to have been largely responsible for inflation as well as for the slowdown. We have seen above, however, that the importance of the oil shock for inflation has been greatly exaggerated. True, the oil price rise was a burden on the importing countries, but it was not a very heavy one. For the United States, for example, the quadrupling of the oil price in 1973 caused an increase in the annual oil import bill of about $20 billion, or 1.2 percent of GNP. (For other countries that depend more heavily on imports the burden was, of course, greater, but for no country was it crushing.) It follows that a once-for-all reduction of real incomes by 1.2 percent would have eliminated the problem. In an ideal, fully competitive economy, a once-for-all decline in money wages (incomes) would solve the problem at a stable price level. Alternatively, if we assume wage rigidity, a once-for-all increase in the price level by 1.2 percentage points, not a matter of great importance in a period of two-digit inflation, would solve the problem. Only indirectly, via perverse reactions of the economy and of economic policy, is it possible to make the oil price rise responsible for continuing inflation, for example, if there is widespread formal or de facto indexa 72 SLOWDOWN AND STAGFLATION Hon of wages and other incomes. In other words, if labor unions and other pressure groups resist the reduction of their real income by pressing for money wage increases, there is bound to be trouble. If the monetary authorities stand firm, the consequence will be unemployment; if they try to prevent unemployment by expansionary ("Keynesian") policy, an inflationary spiral will be started. The oil price rise, or more broadly the rise in energy prices triggered by the Organization of Petroleum Exporting Countries (OPEC) is also widely regarded as one of the factors that have slowed down productivity growth. I believe that the argument put forward with respect to inflation also applies to productivity growth: only indirectly via adverse economic and economic policy reactions can the rise in energy prices cause a slowdown of productivity growth. If the economy adjusts promptly to the oil levy imposed by OPEC, that is to say, if real incomes are reduced by 1.2 percent, it is difficult to see why productivity growth should not resume, starting from the lower level. The proponents of the theory cite microeco nomic reaction as the cause of the slowdown of productivity growth, to wit, a shift from "energy intensive methods." There have been such shifts, of course; large, fuel-inefficient automobiles have been replaced by small, fuel-efficient ones, and producers everywhere have found hun dreds of ways to conserve energy. True, these shifts usually require investments. But I cannot see why the adjustment of economic agents to changes in relative prices, in a way that minimizes the impact of the OPEC oil levy on overall output and basic needs, should lead to a slowdown of productivity growth. Unfavorable indirect adverse effects on productiv ity growth through real wage resistance leading to unemployment and inflation, however, are possible and even likely. I believe the same analysis applies to several other factors that are usually mentioned as causes for the slowdown, for example: "Gov ernmental controls have required the diversion of a growing share of the labor and capital employed by business to pollution abatement and to the protection of employee safety and health. Also, rising crime has forced business to divert resources to crime prevention, and thefts of merchan dize have directly reduced measured output" (Denison 1979:75). The impact of these changes on output is negative, but if the economy adjusts promptly to the increased costs of regulatory compliance and crime prevention by accepting the unavoidable reduction of wages and other incomes, there is no reason why productivity and output growth should not resume from the lower level. Again, it is possible that the adjustment may be delayed and dragged out by real wage resistance, causing exten sive periods of unemployment and/or inflation. Clearly a string of inde pendent adverse effects could produce a period of stagnant output. It is GOTIFRIED HABERLER 73 generally assumed, however, that there is more to the productivity slow down than a series of exogenous shocks. Observers have offered several reasons for a permanent or at least long-lasting slowdown in productivity growth. To begin with, a decline in the number of new inventions and discoveries has been postulated; in other words, it has been suggested that technical progress has slowed down. The evidence cited is a decrease in the quantity of new patents granted by the patent offices of the United States and other countries. As Martin N. Baily (1983) notes, however, this evidence is far from convinc ing. Moreover, in the 1930s, we may recall, the theory of secular stagna tion was based partly on the assumption that technological progress had sharply slowed. Subsequent developments have emphatically disproved that theory. Much more plausible is the theory that the rate of savings and invest ment has been sharply reduced by government policies. The main culprit is tax disincentives for savings and investments. Specifically, high mar ginal tax rates blunt the incentives both to save and to invest. Inflation exacerbates the disincentive effects of the income tax by pushing tax payers into higher tax brackets. Indexation of the income tax would stop the process. Suffice it to say, quoting John Kendrick's excellent discus sion, that "the U.S. tax system is biased against capital formation-more so than is true of most other industrial countries, which have significantly higher ratios than the United States of gross saving and investment in GNP" (Kendrick 1979:53). The tax reform of 1980 should go a long way to correct the bias. There is a danger, however, that the fruits of the tax reform will be squandered by the huge budget deficits that have been allowed to develop. Another reason for a protracted slowdown of productivity growth is said to be an erosion of "work effort." Generous welfare payments, unemployment benefits, and other social welfare measures make it dif ficult for employers to enforce work disciplne, the argument goes; the threat of losing a job is no longer a strong spur to hard work. There is surely some truth in this statement, but it is hard to evaluate. The fact that labor unions have been weakened in the recent recession and that the wage push has abated much more than was expected would seem to weaken the force of the argument. In this connection Mancur Olson's well-known theory (Olson 1982) of the arteriosclerotic and arthritic afflictions of "aging" societies should be mentioned. The rise of what Olson calls "distributional coalitions" (vested interests and pressure groups) tends to make modern economies more and more rigid and inflexible. The process of rigidification can be haited, however, and is often reversed by defeat in war or by revolutions 76 SLOWDOWN AND STAGFLATION with the facts that in the 1970s floating shielded the strong currency countries from U.S. inflation and that the strong dollar in the 19805 has been a potent antiinflationary factor in the United States. Let us consider the alleged malfunctioning of floating, starting with the alleged overvaluation of the dollar. The rise of the dollar since 1980 has indeed been dramatic. From January 1981 to January 1985 the dollar has risen about 53 percent in trade weighted terms, 26 percent against the yen, 58 percent against the German mark, and 46 percent against the Swiss franc. This appreciation is said to have had most serious effects on the U.S. economy as well as on the economies of other countries. Quoting Amer ican sources The Economist writes: "American business has been priced out of markets both at home and abroad by the strong dollar; its costs are about 28 percent less competitive than they were in 1979. As a result, America's real gross national product was reduced by about 2Y4 percent between 1980 and early 1983 compared with what would have happened without an overvalued dollar, and more than one million jobs were lost (1983:9). This analysis is seriously flawed. True, the strong dollar and the large trade deficits have been depressive factors, but they are also potent antiinflationary factors, an aspect that the argument fails to take into account. If the dollar had not gone up, for example because investors had lacked confidence in the U.S. economic policy, inflation would have been much worse. If we assume, as we must, that inflation had to be brought down, the Federal Reserve Board would have had to tighten money. Thus the recession would have been about the same, although the impact on industries producing traded and nontraded goods would have been a little different. Alternatively, if the Federal Reserve Board had inter vened massively in the foreign exchange market to prevent the dollar from rising, a measure that some critics of floating urged, it would have undercut the antiinflation policy, a consequence that would obviously not have been acceptable. 20 The unexpectedly vigorous recovery of the U.S. economy since De cember 1983, despite large trade deficits and the strong dollar, has undermined the theory that the strong dollar produces a dismal effect. Although the speed of the expansion has been slowed down by the strong dollar to a more sustainable level, the length of the upswing has been increased, thus creating more jobs. It is true that the strong dollar has inflationary effects on other coun tries and forces them to tighten money and to raise interest rates. The dollar has remained strong longer than anybody expected. In fact, it has GOTTFRIED HABERLER 77 gone higher and higher, despite repeated attempts by central banks to hold it down by substantial interventions in the foreign exchange market. Since Stephen Marris (1983) raised the specter of an impending "col lapse" of the dollar in his article "Crisis Ahead for the Dollar," however, a chorus of voices has been heard predicting that the dollar will plunge from its present height, with disastrous consequences for the United States and the rest of the world. The argument has become familiar by now. The United States cannot go on indefinitely running large current account deficits and importing capital. Sooner or later "the nerves of foreign investors will crack," and there will be a stampede out of the dollar. To speak of an impending collapse of the dollar seems a bit melo dramatic. I agree, however, that sooner or later the dollar will decline and that a sharp reversal cannot be excluded if the huge budget deficits continue. I submit, however, that if the retreat became a rout, it is inconceivable that the IMF and central banks would stand idly by and let the dollar plunge. They would organize a dollar rescue operation, as they did in 1978. The markets probably understand this point, although the dooms ayers of the dollar do not. There is, therefore, a good chance that there will be no collapse of the dollar-in other words, that there will be a soft landing rather than a crash landing. This is not, however, the end of the story. Suppose that huge budget deficits continue, that capital imports and the trade deficit shrink, and that the dollar declines in a more or less orderly fashion. In such a case U.S. interest rates would go up and government borrowing would crowd out private investment. The consequence would he either a recession if the Federal Reserve Board stands firm or, if the monetary policy is lax, inflation followed later hy a more serious recession. It is, therefore, of the utmost importance for the budget deficits to be reduced gradually. If they were, more savings would be available for productive private investment, offsetting the decline of capital inflow from abroad. Thus a smooth transition is conceivable. In practice, of course, things may not work out so smoothly even if the macroeconomic levers, budget deficits, and monetary policy are set correctly. We cannot assume that fine tuning will be completely successful. After all, the business cycle is still alive. I would argue, however, that a mild recession is not a calamity. There is, I believe, fairly general agreement that various interrelated develop ments account for the strong dollar. The vigorous cyclical recovery of the U.S. economy and the huge budget deficits have pushed up interest rates, which pull in capital from abroad, thus financing the large trade and " RCt.", 80 SLOWDOWN AND STAGFLATION The policy conclusion is straightforward: for best results monetary policy should be supplemented and assisted by an appropriately tight fiscal policy, as monetarists themselves have often recommended. How about incomes policy? We have seen that Keynesians and some non-Keynesians often recommend that demand management should be supplemented by incomes policy to prevent labor monopolies (unions) and business monopolies from driving up wages and prices. The trouble is that "incomes policy" means different things to different people. It is often interpreted as denoting more or less comprehensive wage and price controls. In that form incomes policy must be rejected. Wage and price controls are too crude; they deal only with symptoms, distort the econ omy, and have never worked. This negative judgment also applies to TIP, which substitutes tax incentives and deterrents for absolute controls. TIP would be an administrative nightmare; for that reason it has never been tried. Similarly, wage and price freezes have been tried but never with success. It should be admitted, however, that proponents of incomes policy have a point when they say that the policy of disinflation will have a better chance of succeeding if excessive wage increases can be prevented. I myself have distinguished between Incomes Policy I, and Incomes Policy II. The former I define as more or less comprehensive wage and price controls (including TIP and wage and price freezes); the latter is a bundle of measures designed to move the economy closer to the competitive ideal. Before indicating the kind of measures that would bring about this situation, let me mention that the general approach sails under different flags. In Europe, for example, it is often called "supply-oriented eco nomic policy." This policy is largely the same as my Incomes Policy II, a bundle of measures designed to make the economy more flexible and efficient by breaking down impediments to competition in commodity and labor markets that would result in increasing supplies all around. Obviously the composition of the bundle of measures must be assumed to change over time and to vary from country to country, depending on the structural problems confronting each country. Supply-oriented policy is a much broader concept than, and must be distinguished from, supply-side economics, which flourishes in the United States but is hardly known in Europe. 22 Still another label for the same approach is "adjustment policies," a bundle of measures designed to speed up the economy's adjustment to changing conditions. Adjustment policies have been the theme of two important reports, one by the General Agreement on Tariffs and Trade (GAIT 1978) and the other by OECD (1983b). GOTIFRIED HABERLER 81 Adjustment policies and supply-oriented policies (or supply-side eco nomics, for that matter) should not be regarded as alternatives or antith eses to "demand-side economics" (demand management). The two are complements, not substitutes. Demand management, furthermore, should not be equated with Keynesianism. It is, after all, the task of monetary policy to keep aggregate demand on an even keel so as to avoid inflationary or deflationary spirals. (This statement is independent of the precise rule, monetarist or not, that monetary policy follows to achieve its goal.) I shall now cite concrete measures of adjustment or supply-oriented policies. This is, of course, a vast area. I can ~iscuss here only some basic facts and principles. To bring the economy closer to the competitive ideal, we must attack all forms of monopolid and restrictions of trade. It is convenient to consider business or industrial monopolies (including oligopolies and cartels), labor monopolies (labor unions), and the gov ernment separately. · Business and labor monopolies have rathdr different rules of conduct. Both sets of rules, however, restrict supply, eep prices and wages higher than they would be under competition, a d slow down productivity growth. In my opinion, labor unions in the resent-day world are much more powerful, present a greater danger for price stability and full employment, and are much more difficult to deal with than business monopolies. Compared with labor unions, private busi ess or industrial monopolies are little of a problem except in the area f public utilities. The most effective antimonopoly policy, which is at th same time easy to carry out from the economic and administrative poin of view (although it is not easy politically), is free trade. We have see enormous growth in world trade, especially in manufactures in the last f rty years, great advances in the technology of transportation, communic~tion, and information, and the emergence of new industries in scores or developed and developing countries. Because of this growth, few, if any, monopolies could survive in a free trade world outside the area of public utilities, where prices are under public control anyway. i Free trade policy must, of course, be definfd broadly. It would include not only the phasing out of tariffs, import q'!l0tas, and exchange control but also the elimination of administrative ~rotectionism, the so-called voluntary restrictions imposed on foreign exporters (often called "or derly marketing agreements," or OMAS) , the takeover of noncompetitive firms by the government, and their operation with great losses at the expense of the taxpayer. A policy along these lines would not require any new government bureaus or larger bureaucracies. On the contrary, it 82 SLOWDOWN AND STAGFLATION would reduce government activities, shrink the public sector, and lighten the tax burden. I now come to the problems of the labor market. To begin with, the liberalization of internal and international trade would go a long way to curb the monopoly power of labor unions. Unions know or quickly find out that striking against world markets is risky. For this reason labor unions in small countries where the international sector is a large fraction of the economy are usually much more reasonable and moderate than labor unions in large countries where the international sector is small. This point is strikingly illustrated by two recent developments, although they did not occur in the international arena: the deregulation in the United States ofthe trucking and airline industries. Until recently the two industries were tightly regulated by two huge federal bureaucracies. The dismantling of the controls is equivalent to the introduction of internal free trade. Deregulation-free trade---changed the structure of the two industries dramatically. New, largely nonunionized firms (regional airlines) with lower cost and dynamic management sprang up, providing better and much cheaper service to the public. In both cases the power of the unions was sharply reduced. The wage rates and wage costs of the new airlines are much lower than those of the old ones. It is not surprising that the unions are strongly opposed to deregulation. The unions of the airline pilots went so far as to threaten a general strike to force the government to restore tight regulations in order to protect the exorbitant salaries that they had been able to obtain under the earlier system. It is interesting to ask what would have happened if the competition for the established airlines had come from abroad rather than from domestic sources. Surely the resistance to deregulation would have been much stronger. 23 Finally I mention a few specific measures that would promote efficiency of the labor market, would reduce the monopoly power of labor unions, would increase the overall productivity of labor, and would thus lead to a rise in real wages. Consider minimum-wage laws, for example. They exist in many countries, including France and the United States. Legal mini mum wages serve no useful purpose. On the contrary, it has been shown conclusively that in the United States the legal minimum wage is largely responsible for the shockingly high unemployment (as much as 40 per cent) among teenagers, especially blacks. The minimum wage deprives young people of the on-the-job training that is so important to their future careers. Minimum-wage laws should be abolished (see Hashimoto 1981 and Rottenberg 1981). In the United States the Davis-Bacon Act and similar laws obligate the government to buy only from firms that pay highest union wages. Given GOlTFRIED HABERLER 83 the large size of the public sector and the huge volume of government purchases-ranging from paper and pencils to trucks and turbines-these laws add considerably to the monopoly power of unions and to the size of the government budget. Such laws, too, should be abrogated (see Gould and Bittlingmayer 1980). Now that I have discussed private monopolies of business and labor and what to do about them, I must say a few words about the role of the government. Actually, the government itself is the worst offender. The list of its misdeeds is very long indeed. To begin with, public policy is largely responsible for the power of private monopolies. We have seen that, without protection from imports, few, if any, private monopolies, oligopolies, or cartels would exist outside the public utility area. In fact, in most countries public utilities, postal services, railroads, and telephone and telegraph services are government monopolies. Prices in this area are notoriously rigid, and in many countries these public enterprises operate inefficiently and add substantially to the deficit of the government budget, crowding out productive private investment. We should further mention the enormous burdens of the welfare state and the overregula tion of industries that contributes to the tax burden, blunting incentives to work, save, and invest. When producers are too numerous to organize themselves to cut production and raise prices, the government steps in and does for them what unions do for their members. Farm price support is the most important example although not the only one. In the United States and even more so in the European Community (Common Market), farm price supports have become a heavy burden on the government budget and a source of inflation. There are thus infinite opportunities for adjustment policies and sup ply-oriented policies to improve the performance of the economy, to make it more efficient and competitive and to speed up the rate of productivity and GNP growth. Summary There should be general agreement that inflation, including stagfla tion, cannot be curbed without monetary restraint, or a reduction of monetary growth; that inflation must be stopped; that there is no perma nent tradeoff between unemployment and inflation; and that the longer inflation lasts, the more inflationary expectations become entrenched, the stronger the tendency of inflation to accelerate, and the harder it becomes to stop inflation. It is true that sufficiently tight money can stop 84 SLOWDOWN AND STAGFLATION inflation irrespective of large budget deficits and regardless of wage and price pressures exerted by powerful trade unions and other pressure groups. Monetarists should agree, however, that adverse side effects of large deficits, wage rigidity and wage pressure from powerful labor unions, and similar pressures exerted by other special interest groups will cause much unemployment that will make disinflation politically very difficult, if not impossible. Therefore, monetary tightness should be assisted by fiscal discipline and by what have become known as adjustment, or supply oriented, policies-that is, measures designed to bring the economy closer to the competitive ideal by attacking all kinds of monopolies and restrictions of trade. The most effective and administratively easiest antimonopoly policy is external and internal free trade. In a free trade world, there would be hardly any monopolies outside the public utility area. Free trade in commodities would also be a powerful restraint on union power. This point has been strikingly demonstrated by the deregulation of the airline and trucking industries in the United States. The deregulatory policy, which is equivalent to internal free trade in these industries, has dramati cally weakened the monopoly of the unions in these industries and has sharply reduced labor cost. Notes 1. As various writers have explained many times, a distinction must be made between Keynesian economics and the economics of Keynes himself. When he wrote The General Theory during the Great Depression, Keynes was concerned, it is true, with deflation, and he recommended deficit spending. Early in 1937, less than a year later, however, he argued in three famous articles published in The Times that it was time to switch from fighting deflation to curbing inflation, although at that time unemployment was still above 10 percent and inflation was low by post-World War II standards. Most of Keynes's followers, however, continued to preach expansionary policies. 2. The terminology-"depression" signifying severe cyclical declines and "recession" mild ones--is of modern origin. The words "depression" and "recession" were of course used in the older literature but in a different sense. The distinction between severe and mild cyclical downswings, however, was made in the earlier literature, for example by A. H. Hansen ("major" and "minor cycles") and by Milton Friedman ("mild" and "deep" depression cycles). The difference between depression and recession is one of degree, but it is a fact that most cyclical downswings can unequivocally be classified, with perhaps one or two borderline cases. 3. A distinction must again be made between Keynes and the Keynesians. It is true that Keynes had a nationalistic-protectionist-interventionistic period. During the war, however, when he worked on plans for postwar reconstruction, he returned to his early liberal beliefs. In a famous, posthumously published paper (Keynes 1946), he castigated the views of his GOITFRIED HABERLER 85 radical followers as "modernist stuff gone wrong and turned silly and sour" (1946:186) and urged that "the classical medicine" be allowed to do its work. 4. It has been argued that the oil price rise in the 1970s should not be regarded as "completely exogenous," that it resulted pardy from surging demand on the world oil market generated by rapid economic growth in the industrial countries and, in the later 1970s, in major developing countries as well (letter from Peter B. Clark to Gottfried Haberler). This reasoning is surely true, but it does not affect my argument in the text. 5. Again, we must distinguish between Keynesian economics and the economics of Keynes, but I think it is fair to say that in The General Theory wages and prices are assumed to be constant to the point offull employment, although Keynes hints at some qualifications. The assumption is clearly stated and repeated in the first part of The General Theory. It should be stressed, however, that in later chapters, especially in chapter 19, "Changes in Money Wages," important qualifications are made to what Samuelson called the "Ur Keynesian model" (see Samuelson 1983:212). 6. In the German case the stock of money in real terms (especially if it is measured by the foreign exchange rate) fell to a small fraction of its normal level. This fact was misinter preted by influential German economists at that time, who argued that it invalidated the quantity theory of money, according to which the increase in the money supply causes the price rise and the depreciation of the currency. For details see Haberler (1936:57-60). 7. There is evidence that in the current U.S. expansion there has indeed been a modera tion of wage demands as described by Fellner. See Cagan and Fellner (1984). 8. There is also a mixed group of dissenters: orthodox supply-siders and orthodox Keynesians. James Tobin (1983) recently hailed the large deficits as the proper "Keynesian" measure (adopted, of course, for the wrong reason) that lifted the economy out of the "depression." He would probably agree, however, that there may be too much of a good thing. 9. Milton Friedman writes: "There is clearly great similarity between the views expressed by Simons and by Keynes-as to the causes of the Great Depression, the impotence of monetary policy, and the need to rely extensively on fiscal policy. Both men placed great emphasis on the state of business expectations and assigned a critical role to the desire for liquidity [on the 1 'absolute' liquidity preference under conditions of deep de pression.... It was this that meant that changes in the quantity of money produced by the monetary authorities would simply be reflected in opposite movements in velocity and have no effect on income or employment" (1966:7). See also Stein (1969,1983) and Davis (1971). 10. Actually, the figures greatly understate the difference, because present-day unem ployment is quite different from that in the 19305. It contains much frictional and voluntary unemployment, which I shall discuss. 11. It has been argued that, if the figures of public sector borrowing quoted by The Economist are replaced by cyclically adjusted figures (oEeD 1983a:34), the fiscal restraint appears to have declined sharply in 1982 and to have given way to a mild stimulus in 1983. Actually, the fiscal restraint did not decrease in 1982, but there was indeed a mild fiscal stimulus in 1983. Thus it is still true that the policy of the Thatcher government was and is decidely "unKeynesian" and that the prediction of the 364 economists about the dire consequences of the Thatcher policy have turned out to be quite wrong. 12. I call this view traditional because it reflects, I believe, the outlook of the founders of the Chicago School. My favorite quotation comes from Frank H. Knight: "In a free market these differential changes would be temporary, but even then they might be serious, and with important markets so unfair as they actually are-and wages and prices as sticky ... -the result takes on the proportion of a soeial disaster" (Knight 1946:224). Knight wrote with the deflation of the 1930s in mind, but his comments apply also to the 86 SLOWDOWN AND STAGFLATION disinflation of our times, and surely the stickiness of wages and prices has greatly increased since he wrote. 13. It is true that the German economy had a head start, for it had turned the corner in August 1932, six months before Hitler carne to power. That fact does not, however, vitiate the comparison with the New Deal, because in the first six or seven months the German recovery was slow, and we cannot date the trough of the U.S. depression to 1933 with precision; the depression has been described as "double bottomed," the first bottom having occurred in 1932. 14. On Hitler's economic miracle, see Haffner (1979), described by the reviewer in The Economist as "a brief, incisive, adequate account of the monster which leaves no more, worth saying, to be said" (Economist 1979: 133). Another German miraclc is thc phenome nal rise of the German economy, starting in 1948, from the ashes of the Hitler regime. 15. Keynes expressly excluded frictional unemployment from his concept of involuntary unemployment (1936:15-16). He does not mention structural unemployment, but it seems to me that it is in the spirit of the Keynesian theory to treat frictional unemployment and structural unemployment in the same way. 16. Keynes's much-quoted definition is: "Men are involuntarily unemployed if in the event of a small rise in the price of wage-goods relative to the money wage, both the aggregate supply of labour willing to work for the current money wage and the aggregate demand for it at that wage would be greater than the existing volume of employment" (1936). This statement is, to my mind, unnecessarily complicated. I am encouraged by the fact that Richard Kahn, Keynes's closest collaborator and disciple, reached the same conclusion (Kahn 1975). 17. For samples of the dissenting letters, see The Economist for December 18, 1982. The Economist rightly argues that Keynes, if he were alive, would support its position and not that of its Keynesian critics. It is one thing to say, as Keynes did in the 1930s, that a deflationary spiral should be stopped by expansionary measures rather than by wage reduction; it is an entirely different thing to insist in a period of persistent, severe stagflation that the level of money wage rates must not be touched. As I mentioned earlier, one year after the publication of his General Theory, Keynes urged a shift in policy to fight inflation. We must distinguish between Keynesian economics and the economics of Keynes. 18. The literature on the subject is enormous. There are the few standard works by Edward Denison and John Kendrick. For reference see Kendrick (1979) and Denison (1979). See also the brief but very illuminating introduction by William Fellner, "The Declining Growth of American Productivity: An Introductory Note," and Mark Perlman, "One Man's Baedeker to Productivity Growth Discussions," both in Fellner (1979b); and Baily (1983). 19. Since this chapter was written in 1983 and 1984 an in-dcpth analysis of the interna tional monetary system by Morris Goldstein has appeared. See Goldstein (1984) and IMF Survey (1985). 20. Furthermore, we cannot counter the above-stated argument by saying that sterilized interventions would eliminate the danger of inflation, for there is general agreement that sterilized interventions are ineffective. 21. This situation could change overnight. The imposition of an import surcharge com bined with an "interest equalization tax" to restrain capital import, for example, would do the trick. The attraction of the surcharge is that it would deal with the trade and budget deficits at the same time and would be administratively easy to apply provided it was applied across the board. There are two hitches, however. There will surely be exceptions. Canada and some Latin American countries will be exempted. As a consequence, imports from these countries into the United States will soar, but so will imports into these countries from GOITFRIED HABERLER 87 Japan, Europe, and elsewhere. The result will be a distortion oftrade and a reduction in the effectiveness of the surcharge. The other hitch is that retaliation to a surcharge, especially from the European Common Market, will be swift. The interest equalization tax, a tax on interest payments to foreign investors, too, would have distorting and evasive reactions. Such a tax is discriminatory in two ways: it discriminates between interest income on financial assets on the one hand and income on other types ofinvestment on the other hand, such as income from real estate, speculative capital gains, and the like. There would thus be a strong inducement for foreign investors to switch from financial assets to other kinds of investment. The other area of discrimination is between foreign investors and American investors abroad. Foreign investors would be discouraged from exporting capital to the United States. American investors abroad, however, would be encouraged to repatriate some of their money to take advantage of the interest differential caused by the interest equalization tax. Nobody can forecast with certainty the comparative magnitude of these conflicting reactions, and the psychological effect of the whole operation, import surcharge, and interest equalization tax may be very powerful and disturbing, especially because it would highlight the unwillingness to solve the problem of the budget deficits. 22. Let me make my position clear. Supply-siders have a valid point when they say that high taxes, especially high marginal tax rates, have become a drag on incentives to work, to save, and to invest. It is dangerously overoptimistic to assume, however, as supply-siders do, that tax reductions all by themselves will turn the economy around almost overnight and that the resulting increase in output will take care of even large budget deficits. Supply oriented policy will include, as one element among many others, removal of tax impedi ments to work, save, and invest. 23. The large size of the U.S. market, which makes possible competition in industries that are subject to increasing returns to scale, is a great national asset. Few other countries enjoy that advantage. References Baily, Martin Neil. 1983. Will productivity growth recover? Discussion papers in economics. Washington, D.C.: Brookings Institution. Bry, Gerhard, with Charlotte Boschan. 1960. Wages in Germany, 1931-1945. Princeton, N.J.: National Bureau of Economic Research, Princeton University Press. Cagan, Phillip. 1975. Changes in the recession behavior of wholesale prices in the 1920's and post-World War II. In Explorations in economic research. Occasional papers, vol. 2, no. 1. Cambridge, Mass.: National Bureau of Economic Research. Cagan, Phillip, and William Fellner. 1984. The cost of disinflation, credibility, and the deceleration of wages, 1982-1983. In William Fellner, ed. Contemporary economic problems, 1983-1984. Washington, D.C.: American Enterprise Institute. Davis, J. Ronnie. 1971. The new economics and the old economists. Ames, Iowa: Iowa State University Press. Denison, Edward. 1979. Where has productivity gone? In William Fellner, ed. Contempo rary economic problems, 1979. Washington, D.C.: American Enterprise Institute. Economist. 1979. Review of "The Meaning of Hitler" by Sebastian Haffner. Economist (October 20):133. - - . 1982a. Work on a pay cut. Economist. November 27:11-12. 88 SLOWDOWN AND STAGFLATION - - - . 1982b. Wage cuts. Economist (December 18):14-15. - - - . 1983. The overblown dollar. Economist (December 17):9. - - - . 1984. UnKeynesian Britain. Economist (February 4). Fellner, William. 1976. Towards a reconstruction of macroeconomics: Problems of theory and policy. Washington, D.C.: American Enterprise Institute. - - - . 1979a. The credibility effect and rational expectations: Implications of the Gram lich study. Brookings papers on economic activity 1:167-78. - - - , ed. 1979b. Contemporary economic problems, 1979. Washington, D.C.: American Enterprise Institute. - - - . 1980. The valid core of rationality hypotheses in the theory of expectations. Journal of money, credit, and banking, pt. 2(November). Friedman, Milton. 1966. What price guideposts? In George P. Schultz and Robert Z. Aliber, eds. Guidelines, informal controls, and the market place: Policy choices in a full employment economy. Chicago: University of Chicago Press. General Agreement on Tariffs and Trade (GAIT). 1978. Adjustment, trade, and growth in developed and developing countries. Geneva: GAIT. Giersch, Herbert. 1982. Prospects for the world economy. Skandinaviska Enskilda Banken quarterly review: 104-10. - - - . 1983. Arbeit, Lohn, und Productivitiit. Weltwirtschaftliches Archiv 119(no. 1): 1-18. Goldstein, Morris. 1984. The exchange rate system: Lessons of the past and options for the fu ture. Occasional paper 30. Washington, D. C.: In terna tional Monetary Fund, Research Department. Gould, John P., and George Bittlingmayer. 1980. The economics of the Davis-Bacon Act: An analysis of prevailing wage laws. Washington, D.C.: American Enterprise Institute. Haberler, G. 1936. The theory of international trade. London and New York: Stonier and Benham. Reprint, New York: Augustus M. Kelly, 1968. - - - . 1980. The Great Depression of the 1930s-Can it happen again? In The business cycle and public policy, 1920-80, a compendium of papers submitted to the Joint Eco nomic Committee, 97th U.S. Congress, 2d session. Also available Washington, D.C.: American Enterprise Institute Reprint 118, January 1981. - - - . 1984. "The international monetary system in the world recession." In Essays in contemporary economic problems-Disinflation, 1983-1984, William Fellner, ed., Washington, D.C.: American Enterprise Institute. - - - . 1985a. "International issues raised by criticisms of the U.S. budget deficits." In Contemporary economic problems, 1985, Phillip Cagan, ed., Washington, D.C.: Amer ican Enterprise Institute. - - - . 1985b. The problem ofstagflation: Reflections on the microfoundation ofmacroeco nomic theory and policy. Washington, D.C.: American Enterprise Institute. Also to appear in Political business cycles and the political economy of stagflation, Thomas D. Willett, ed., San Francisco: The Pacific Institute for Public Policy. - - - . Forthcoming. "Liberal and illiberal development policy. 'Free trade like honesty is still the best policy. '" in Pioneers in development, vol. 2, Gerald M. Meier, ed., Washing ton, D.C.: The World Bank. Haffner, Sebastian. 1979. The meaning of Hitler. New York: Macmillan. Hashimoto, Masanori. 1981. Minimum wages and on-the-job training. Washington, D.C.: American Enterprise Institute. GOTIFRlED HABERLER 89 Hutchison, T. W. 1979. Notes on the effeets of eeonomic ideas on policy: The example of the German social market economy. In Currency and economic reform: West Germany after World War II: A symposium. Zeitschrift [iir die gesamte Staatswissenschaft 135(September):435ff. IMF Survey. 1985. New fund study sees floating rate system as a qualified success. February 18, p. 49. International Monetary Fund (IMF). 1981. International financial statistics. Supplement on price statistics, supp. ser. 2. Washington, D.C.: IMP. Kahn, Richard. 1975. Unemployment as seen by the Keynesians. In G. D. N. Worswick, ed. The concept and measurement of involuntary unemployment. London: Allen and Unwin. Kendrick, John. 1979. Productivity trends and the recent slowdown: Historical perspective, causal factors, and policy options. In William Fellner, ed. Contemporary economic problems, 1979. Washington, D.C.: American Enterprise Institute. 1984. Long-term economic projection: Stronger U.S. growth ahead. Presidential address before the Southern Economic Association. Reprinted in Southern economic journal (April). Keynes, J. M. 1936. The general theory of employment, interest, and money. London: Macmillan. - - - . 1946. The balance of payments in the United States. Economic journal 56. Knight, Frank H. 1946. The business cycle, interest, and money. In Frank H. Knight, ed. On the history and methods of economics. Chicago: University of Chicago Press. Kravis, Irving B., and Robert E. Lipsey. 1984. The diffusion of economic growth in the world economy, 1950-1980. In John Kendrick, ed. International Comparisons ofProduc tivity and Causes of the Slowdown. Cambridge, Mass.: Ballinger. Kuznets, Simon. 1971. Economic growth ofnations. Cambridge, Mass.: Harvard University Press. 1976. Aspects of post-World War II growth in less-developed countries. In A. M. Tang, E. M. Westfield, and James E. Worley, eds. Evolution, welfare, and time in economics: Essays in honor ofNicholas Georgescu-Roegen. Lexington, Mass.: Lexington Books. - - - . 1977. Two centuries of economic growth: Reflections on U.S. experience. Richard T. Ely lecture. American economic review 67, no. 1(February): 1-14. Marris, Stephen. 1983. Crisis ahead for the dollar. Fortune. December 26. - - - . 1984. Why Europe's recovery is lagging behind, with an unconventional view of what should be done about it. Europe (March-April). Olson, Mancur. 1982. The rise and decline of nations: Economic growth, stagflation, and social rigidities. New Haven: Yale University Press. Organisation for Economic Co-operation and Development (OECD). 1983a. Economic outlook (December):34. - - - . 1983b. Positive adjustment policies: Managing structural change. Paris: OECD. Prebisch, Raul. 1984. Pioneers in development. New York: Oxford University Press for the World Bank. Price, James E. 1982. The changing cyclical behavior of wages and prices, 1890-1976: Comment. American economic review 72(December). Robinson, E. A. G. 1967. Economic planning in the United Kingdom. Cambridge: Cam bridge University Press. 90 SLOWDOWN AND STAGFLATION Rottenberg, Simon, ed. 1981. The economics oflegal minimum wages. Washington, D.C.: American Enterprise Institute. Sachs, Jeffrey. 1980. The changing cyclical behavior of wages and prices: 1890-1976. American economic review 70(March):78-90. 1982. The changing cyclical behavior of wages and prices, 1890-1976: Reply. American economic review 72(December):1l91-93. Samuelson, Paul A. 1983. In David Worswick and James Trevithick, eds. Keynes in the modem world. Cambridge: Cambridge University Press. Samuelson, Paul A., and Robert M. Solow. 1960. Analytical aspects of anti-inflation policy. American economic review 50(May):I92. Schumpeter, Joseph A. 1951. The decade of the twenties. In Richard V. Clemence, ed. Essays of Joseph A. Schumpeter. Cambridge, Mass.: Addison-Wesley. Stein, Herbert. 1969. The fiscal revolution in America. Chicago: University of Chicago Press. - - - . 1983. Early memories of a Keynes I never met. AEI eCOl,omist (June). Taylor, John A. 1982. Union wage settlements during a disinflation. Working paper 985. Cambridge, Mass.: National Bureau of Economic Research. Tobin, James. 1972. Inflation and unemployment. American economic review 62(March):13. - - - . 1973. Comment of an academic scribbler. Journal of monetary economics 4:622. - - - . 1983. Keynes' policies in theory and practice. Challenge (November-December). Williamson, John. 1983. The exchange rate system. Washington, D.C.: Institute for Interna tional Economics. Chapter 3 Systemic Risk and Developing Country Debt Robert Z. Lawrence MUCH OF THE CURRENT DEBATE about the debt facing developing countries focuses on liquidity or solvency problems. For some purposes this debate is useful. If the present problems indicate simply insufficient finance, rescheduling and postponing of payments may be necessary. If, on the other hand, these problems are so large that a return to sustainable external financing would have to be postponed almost indefinitely, some of the special proposals for debt relief may merit consideration. By concentrating research on distinguishing between solvency and liquidity, however, we may miss some important considerations. The notion of solvency is misleading when applied to a nation. Tech nically a firm is insolvent when the present discounted value of its earn ings is less than its liabilities. Nations, however, are likely to default prior to reaching this point. The nation is a unit of consumption as well as production and will be unable and unwilling to service its debts before it is insolvent. For a nation, the relevant question, therefore, is when to default. The default decision is a function of the costs and benefits of default rather than of the difference between future income and liabili ties. If the current crisis is associated with major defaults or losses for lenders, it will have a substantial impact on future borrowing. Sim ulations indicate, however, that with appropriate adjustment by the developing countries and a return to a 3 percent growth path in the developed countries, the developing countries can restore their debt measures (for example, debt-export ratios) to their 1980 levels (see Cline 1983). The simulations imply a lack of liquidity in the world economy. With world growth, the characterization suggests, international lending can return to its former arrangement. These optimistic conclusions need qualification, however. Even if the current crisis reflects deficiencies only in debt-servicing capabilities rather than in ultimate net worth, its impact will not be transitory. If the debt indicators are restored to their 1980 levels, global financial market behavior is likely to undergo substantial 91 92 RISK AND DEBT structural change. Viewed ex post facto, the international system has severely underestimated the effects of systemic risk. Between 1979 and 1983, the global economy experienced an extraor dinary degree of turbulence. In only three years there was a dramatic rise and fall in oil prices, unprecedented levels and volatility in real interest rates, enormous fluctations in real exchange rates, the largest slump in world trade of the postwar era, and a global recession, all leading to great financial instability. Without these destabilizing influences, studies sug gest, developing countries could have sustained and serviced their 1980 debt levels (see, for example, Morgan Guaranty Trust 1983:6). With hindsight we may be surprised to note how vulnerable borrowers, lenders, and the international financial system were to the shocks that occurred in the 1980s. The urgency of borrowing changed the terms of typical lending arrangements. Much of the borrowing by developing countries took the form of credits financed by bank loans, with an increasing amount of borrowing denominated in dollars. Short-term debt became more common; even the long-term debt was typically negotiated at floating interest rates. Banks and developing countries appear to have neglected various types of risk. Developing countries ignored the risk of maturity mismatch, funding large long-term projects by short-term loans. They ignored the risks of income fluctuation by borrowing at fixed terms. They overlooked the impact of recessions on their ability to sustain payments, allowing the proportion of equity financing to decline significantly. They neglected the risks of exchange rate fluctuations, and in particular dollar appreciation, by borrowing in U.S. dollars. Thus, as calculated by the Federal Reserve Bank of New York, if developing countries had followed the practice of denominating their borrowing in proportion to their exports to developed countries, their burdens would have been approximately $30 billion lower (Solomon 1983:3). They ignored the risks of refinancing by increas ing the share of short-term debt that had to be renegotiated frequently. They assured the risk of import price fluctuations by committing their governments to maintaining the domestic prices of tradable goods such as food. In addition they overlooked the risk of interest rate fluctuations by funding at variable interest rates tied to the London interbank offered rate (LIBOR). The banks, which were the major creditors, took some steps to ensure against certain types of risks. They sought to avoid interest rate fluctua tions by linking both deposits and loans to LIBOR. The dramatic growth of the interbank market illustrates their concern about maturity mismatch, currency exposures, and so forth. In addition, ignoring the fallacy of composition, they individually sought to maintain their flexibility by ROBERT Z. LAWRENCE 93 financing with short-term loans. (This development might have lowered perceived risks individually but actually raised risks for the system.) Nonetheless, banks neglected the risk associated with possible default by major debtors among the developing countries. Banks were willing to expose amounts equal to 40 percent of their capital to individual develop ing country borrowers. They also ignored the dangers of creditor servic ing difficulties and the costs of immobilizing substantial portions of their balance sheets by participating in syndicated credits in nonmarketable instruments. Confidence in the creditworthiness of developing countries was not confined to a few banks. Capital markets in general gave borrow ing by developing countries low risk ratings. Banks are highly leveraged institutions with capital in major V.S. banks that is typically equal to about 5 percent of total assets. If these banks were threatened by interna tional default, their stock values would surely respond. There was no significant evidence, however, that they were initially penalized by the market for their developing country debt. Foreign branches in develop ing countries participated freely in V .S. and European interbank markets without incurring risk premiums. Spreads on developing countries' bor rowing as compared with developed countries' borrowing declined gently from about sixty-three basis points in 1976 to fifty-one basis points in 1979 (Sachs 1981:244). Apparently banking authorities viewed the rapid growth of this lending as fairly benign; they took little action to curtail it. They failed to consider the damage that a major default might have upon domestic banking systems. Viewed ex post facto, to disregard the implications of risk may seem negligent. Indeed many discussions of developing country borrowing assume so. Yet ex ante lenders may not have overlooked relevant market information. A risk-averse market equates the marginal benefits of risk reduction with its marginal cost. The benefit of risk aversion is in turn related to (a) the probability of risk-objective risk, (b) the perceived cost of that risk-subjective risk, and (c) the degree of risk aversion. Consider briefly the factors involved in evaluating sovereign nation risk and the possibilities for reducing it. Objective Risk The particular features of lending to sovereign nations change the nature and degree of risk involved and make evaluation more complex. Domestic lending to a private company allows creditors claims on the assets of the company. Creditors can also enter into legal contracts, for example covenants, which restrict the capacity of the borrower to under 94 RISK AND DEBT take other commitments. In contrast, the assets of sovereign borrowers are located in their own territories and cannot be liquidated. Further more, legal sanctions against sovereign borrowers are difficult to exer cise. To be sure, banks may attempt to seize the foreign assets of a defaulting nation; however, the borrower may take considerable precau tions to reduce such costs. International loans are subject to greater uncertainty than domestic loans. Loans to private institutions in develop ing countries involve the intrinsic risks of the project, the risks associated with international legal complications, and risks associated with profit repatriation. Yet other considerations suggest that sovereign lending may be less risky than private lending. Private liabilities may be restricted by bank ruptcy in the case of individuals and by limited liability provisions in corporate charters. Nations exist in perpetuity, however. Repudiation of debt may be particularly difficult if the nation (especially its present government) wishes one day to return to the international capital market. As in the case of the private market, default against bondholders may not forestall access to bonds for all times (Sachs 1982). Default against the relatively few banks in international lending, however, may deny the borrowing country access for longer periods of time. Behind a company's debts stand its limited assets, but behind the debts of a sovereign nation stands its taxing ability. For most nations, furthermore, international debt is below national income, not to mention national wealth. Thus outright debt repudiation is extremely rare. Countries in trouble will reschedule and will perhaps obtain relief from private and official credi tors. The international lending norms since World War II are very different from those for earlier periods. The recent norms are marked by coopera tive solutions that rarely entail default or forgiveness of principal. Instead lenders more typically extend maturity and renegotiate interest rates and fees. Official involvement through loan guarantees and so forth has led to renegotiation of government-to-government loans in the Paris Club. Private renegotiations have involved debtor countries and banks acting in a coordinated fashion with a committee. The International Monetary Fund (IMF) also plays a coordinating role. Thus the system has reduced the danger of outright default. Noting that market spreads have con tained little provision for risk, Guttentag and Herring (1985) suggest that banks expect these loans to be virtually risk free because write-offs can be delayed indefinitely. As my discussion indicates, the risk directly associ ated with loans to a particular borrowing country receives most of the attention. Yet considerations overlook some important factors about market risk. What matters in the market provision for risk is systemic risk. ROBERT Z. LAWRENCE 95 The evidence of relatively low market provision for risk that I adduced above need not indicate market failure. Even if assets have highly vari able returns, risk will command a low price, provided that asset returns are independent. Financial theory suggests that loan interest rates will reflect, not the total variance of an asset, but rather the covariance between the returns from that asset and from the rest of the portfolio (see Sharpe 1964). If an asset's return is independent of the rest of the portfolio, it is possible to diversify away from the risks associated with holding that asset. 1 The concern of individual banks for their own risk is quite understand able. To minimize the risk of a portfolio composed of assets with indepen dent returns, their shares should be inversely proportional to their variance. From the viewpoint of the market, however, the borrower's own risk is not important. Since no single asset has a significant share in the market portfolio, the dominant determinant of the price of risk will be the covariance of an asset's returns with those from the rest of the portfolio. Assets that have returns inversely correlated with the portfolio will have risk discounts; those whose returns show a positive correlation (positive betas) will require risk premiums. The loans to individual de veloping countries reflect risk only to the degree that they add to the nondiversifiable component oftotal market holdings of wealth ofall kinds. In addition, in an empirical investigation of the risks of international bank lending, Laurie Goodman concluded, "Country-specific risks appear to loom large relative to the common problems faced by non-OPEC [Organization of Petroleum Exporting Countries] LDCS" (see Goodman 1982:261).2 Thus the key issue from a systemic viewpoint was the prob ability of a major series of synchronous losses. This probability depends on: (a) the possibilities of major global disturbances and (b) the likeli hood that such disturbances would induce losses synchronous with those in the rest of the market. By a number of indicators, global shocks in 1981 and 1982 lay outside the range of relevant historical experience and were therefore beyond the range of outcomes that could reasonably be anticipated. 3 Judged by postwar norms, real interest rates, the decline of world trade, the volatil ity of real exchange rate fluctations, the downward movement in global average inflation rate, and so forth were extraordinary. In view of the experience in the 1974-75 recession, the dangers that a global recession would spread to the developing countries were in creasingly discounted in 1981-82. Indeed the ability of the developing countries as a group to weather the 1974-75 recession indicated the resilience of the system. Developing countries were able to sustain their growth and to avoid the stagnation that was occurring in developed countries. This trend sug 96 RISK AND DEBT gested that capital markets could decouple the business cycles of de veloped and developing countries. Borrowing by developing countries had been successfully used to finance investment rather than consump tion, which led to an effective export performance, indicating a capacity to service existing debts (Sachs 1981). Lending had proven profitable to lenders, with debt renegotiation clearly the exception rather than the rule. 4 Some countries had experienced difficulties (for example, Peru, Mexico, Turkey, Zaire, Sudan), but these were isolated instances and actually strengthened the view that contagion could be avoided. Overall, bankers frequently noted, loan loss experience with developing countries was superior to experience with domestic loans. Similarly, the limited impact of bank failures, for example of Herstatt and Franklin National, had also strengthened the view that the system could contain and absorb individual shocks. Increasing the share of these countries in portfolios thus seemed to be a way of reducing rather than increasing risk. Subjective Risk Even if risk is objectively high, lenders' perceptions of the costs of risk may be affected by anticipation of government assistance in the event of major shocks or by inaccurate or incomplete information. From a policy viewpoint, a key issue is the degree to which costs of risk were underesti mated in 1980-82 because of the responses that were expected from policymakers. Did banks believe that major defaults were unthinkable precisely because so much of the major banking system was involved? To what extent did they seek simply to match their competitors who securely believed that in the event of default, a lender of last resort and other public facilities would prevent major losses to shareholders? Banks playa major public role as a means of payment. Deposit insurance in the United States is one of the special privileges such institutions enjoy. In other nations, banks are government owned. Banks have serviced the interna tional system by recycling the OPEC surplus cash and have deservedly been praised by the policymakers. Did they perceive themselves acting in more than a private capacity, so that they were, in their eyes, due public aid in the face of trouble? Moreover, the use of official guarantees and of IMF money in cases of difficulty both indicated that public money would be available to support borrowers in the event of major difficulties. Yet bankers' expectations of policymakers should not be exaggerated. Bankers witnessed several reschedulings in which governments did not bail them out. Rescheduling avoided default, but the banks incurred the costs of involuntary lending and in some cases reductions in revenue. Did ROBERT Z. LAWRENCE 97 individual banks and other actors have too little information? Some actors undoubtedly made loans simply by following the herd. It seems clear that regional banks with little knowledge of the nations moved to lend in syndicated credits to countries and to foreign branches in the interbank market. Herd behavior, however, is not unknown in stock and bond markets. For markets to be efficient, complete information need not be available to all actors. As long as major actors can influence decisions at the margin, markets are efficient. There were undoubtedly lacunae in bank information, particularly about aggregate country bor rowing in the short run; it is difficult to make the case that positions built over a number of years were due solely to ignorance of the true state of affairs. Risk Aversion In public finance literature it is argued that a nation has superior risk-pooling abilities and therefore should ignore risk (Arrow and Lind 1970:364-78). Thus the riskless rate is the appropriate discount rate for social projects. The behavior of the developing countries did not appear to indicate much risk aversion. Their willingness to shoulder interest rate fluctation and exchange rate risk, to borrow short term, and to use debt financing all support this conclusion. On the other hand, even if risk reduction is desired, it may be seen as too costly. The very nature of country borrowing leads to the use of debt rather than equity. Nations often undertake projects to provide public goods; it is difficult and perhaps impossible to measure the benefits of these services precisely. Improvements in public health, education, transporta tion, population control, and regulation should all raise national income. To undertake such investments, however, the government must appropriate the returns indirectly by taxes instead of through service charges. Returns to these investments are implicit and hard to measure, and thus securities contingent upon income are unlikely to exist. Indeed there have been numerous proposals (for example, Kubarych 1982:245) for financing instruments that link returns to developing country export earnings or income growth, but the difficulty of measuring these accu rately (particularly when the country would have an incentive to under state them) is a major obstacle. Second, there are costs associated with institutional underdevelop ment. Private entrepreneurs could theoretically undertake many public investments, but entrepreneurial ability is often scarce in a developing country. If foreign entrepreneurs were used, then direct foreign invest 98 RISK AND DEBT ment in domestic equity would reduce the riskiness of investments. Institutional deficiencies, however, inhibit the development of large equity markets that allow portfolio investments to reduce risk by dis sociating ownership and control. Furthermore, political difficulties in hibit direct foreign investment because domestic corporations fear for eign control. Thus information and transactions costs have been a major impediment to risk reduction. Given the relatively low risk aversion exhibited by developing countries, they were unwilling to pay these costs. Experience during the Crisis: A Cause for Reappraisal? Experience during the past few years has confounded observers who argued that the system was too vulnerable to sustain major shocks. The pattern of cooperative solutions to debt problems that was set during the postwar period has been followed and extended. Instead of major de faults, solutions have entailed adjustment programs for the countries as well as rescheduling and increased (forced) lending by the banks and official institutions. These events and outcomes of the recent past, however, suggest that a reappraisal might be made of the risks and costs of international lending. Instead of avoiding the global business cycle, developing countries suf fered the most damage. International lending, once a means of decou pIing developing countries from the international business cycle, has become a source of pressure to magnify the cycle. Refinancing difficulties have driven countries to the International Monetary Fund for assistance. In return for loan rescheduling, developing countries have been forced to implement stringent adjustment programs. National authorities and policymakers in developing countries have spent considerable time deal ing with foreign creditors. Ultimately this activity has severely con strained their ability to pursue domestic policies. Even relatively credit worthy countries in Latin America, such as Colombia, have experienced financing difficulties as a result of systemic contagion. In the developed world, banks have had a difficult time. The stock market has been penalized. In the first quarter of 1982 the average price-earnings ratio of the four largest U.S. banks was about 70 percent of the market average; by tbe third quarter of 1983 it was 42 percent of the market average. Major portions of bank portfolios have been frozen. Overnight deposits in Mexican banks became long-term deposits. Pre vious commitments have required additional lending. Many bank execu tives who operated the programs have lost their jobs. Holders of develop ing country bonds have experienced substantial capital losses. The central banking authorities have been forced to monitor and at ROBERT Z. LAWRENCE 99 times to become actively involved in the crisis. As a result, international lending arrangements operate under tightened regulations and increased capital requirements as well as various other constraints, such as greater provisions in the event of loan loss. Future Innovations A system dependent on bank finance is highly vulnerable to default crisis but is probably less vulnerable to transitory refinancing crisis. Banks are easier to coordinate than bondholders. Banks have long-term rela tionships with borrowers. The relative success in the debt-restructuring experience suggests that the system may actually have been more stable than one with other institutional forms. The burden of debts that lenders currently shoulder, however, will make them wary. Furthermore, even if developing countries return to their previous levels of borrowing (that is, to previous debt-income ratios and debt-export ratios), they may not be of comparative creditworthiness, given the level of real interest rates. For the next few years, objective risks, subjective perceptions of those risks, and risk aversion are all likely to be high. I have argued in this chapter that what is important for the market is the likelihood of systemic loss. The recent short-run liquidity crisis has made this loss seem a possibility. Bank lending is likely to grow more sensitive to interest rate fluctua tions and recessions, thereby increasing the possibility that business cycle risks are transmitted internationally. Thus for the foreseeable future, the betas (systemic risk premiums) on developing country borrowing will be positive rather than close to zero or even negative. We can no longer assume that individual banks are able to protect themselves from immi nent crisis by holding short-term debt. The events of the recent past have also clarified subjective perceptions of policy response to crisis. Authorities have resisted efforts to bailout the banks. The authorities' responses to a crisis of this sort will now more easily be anticipated. Banks cannot reasonably expect to adjust their portfolios quickly in the face of future crisis. They will recognize the possibility of participating in future rescue packages by raising their exposure as dictated by the IMF and by work-out committees. The thought that the present crisis might recur will haunt financial markets for at least the next decade. Psychological studies suggest that individuals give rel atively higher weight to recent experience (Guttentag and Herring 1984). Thus actors throughout the system will be more willing to pay for risk reduction. Lending patterns will therefore change to reflect this prefer ence. Let us consider some of the forms these patterns could take. New types of instruments and contracts could evolve. Payment terms 100 RISK AND DEBT could be based on contingencies ratber than on fixed rates, perhaps involving equity or instruments that tie payments to major commodity prices and interest rates-for example, an instrument that automatically reschedules with fixed interest payments but variable maturity. Direct foreign investment could grow. In addition, the more developed of the developing countries could raise portfolio capital via equity market in vestment. Banks could move away from syndicated credits kept on their books toward more marketable instruments. They might continue to originate loans but seek greater flexibility in the disposition of loans. Hedging vehicles, such as futures markets for currencies, commodities, and in terest rates, could be used more frequently. Interest rate swaps that allow less creditworthy borrowers access to fixed interest rate financing via arbitrage between the fixed and variable interest rate markets are another possibility. What implications would such an evolution have for policymakers? The movement toward greater market risk diversification would make the overall system more resilient. Even the default of a major borrower such as Brazil could easily be handled by a portfolio fund with, say, Citibank's liabilities and assets, whereas Citibank as a leveraged bank could find that half its capi tal had been wiped out. The move toward more contingency-based payments forms would provide a greater buffer, thus reducing the danger of rescheduling. Improved hedging by developing country borrowers could likewise mitigate the potential adjustment costs from major global shocks, making the system better able to endure such shocks. On the other hand, movement too far in the direction of market based instruments and away from banks and official lending might lead to another set of instabilities, so that coordinated and cooperative solutions would become less likely. Observers who criticize banks for their herdlike behavior often overlook the fact that markets may be even more prone to violent swings in mood and behavior. Many policy measures should be instituted to deal with the emerging environment. First, policy should encourage officials in the developed countries to remove the regulations that prohibit major investment in stitutions from holding developing country loans, for example, pension funds. Second, countries such as Germany and Japan have at times had ambivalent attitudes toward the use of their domestic markets for inter national lending. Policy should encourage these countries to remove impediments to such lending, thereby diversifying the currency of de veloping country borrowing. Third, policy should encourage countries to build up the long-term and aid components of lending to developing countries, since market forces may tend toward short-term debt, which such countries have difficulty in servicing. ROBERT Z. LAWRENCE 101 The international organizations will also require changes. Since private international capital is less likely to provide the liquidity necessary to sustain programs in developing countries during recessions in developed countries, greater liquidity is required in the IMF. Programs to shoulder income fluctuation risks, such as the Compensatory Financing Facility, should be expanded. Countries with, for example, major exports of commodities traded on futures markets should be encouraged to use these to hedge against price fluctuations. The IMF could provide greater access to unconditional lending at market interest rates during recessions as a form of countercyclical stabilization policy. The Fund and the World Bank also need to monitor country risk exposure, planning not simply for one scenario of steady growth but also for potential recession and other major shocks. Continuous review of national risk exposure should be instituted. The World Bank should also be wary of contributing to the global cycle. The share of bank-funded projects provided by domestic finance could be tied to a global cyclical indicator. The World Bank and the IMF also need to provide increased expertise and to improve information. Banks enjoy economies of scale in informa tion processing. A more open and market-based system requires more information. In addition, programs to provide insurance for direct for eign investment could be expanded. The World Bank might consider emulating the Federal Home Loan Bank, which bundled mortgages for sale to private investors by bundling loans in developing country projects. Developing countries need to improve risk management. An impor tant component entails maturity matching of projects and finance. A second is avoiding commitments to unrealistic market prices and ex change rates that entail large subsidies and thus borrowing requirements in response to disturbances. Direct foreign investment has unfortunate political implications for many observers. The development of equity markets that allow for a dissociation of ownership and control needs to be encouraged. To create such markets, expertise as well as legal and regulatory infrastructures must be developed. The above-described view of greater systemic risk also suggests the dangers inherent in certain proposals to deal with the current debt crisis. The current rescheduling arrangements have unfortunately failed to deal with the interest rate risk faced hy developing countries. Rescheduled dehts have taken the form of fixed maturity and variable interest rates. A rise in global rates would leave these programs exposed. A superior method would have been to allow for fixed payment but variable amor tization schedules. Many of the proposals for long-term solutions require developed countries, the IMF, or the World Bank to provide guarantees that allow the banks to reduce their developing country debt holdings. These solutions, however, would remove from the banking system the 102 RISK AND DEBT systemic risk currently concentrated there and would concentrate it else where. It would be more desirable to diversify such risks by the use of nonguaranteed instruments that are widely held. Notes 1. Recall that the variance of a portfolio of independent assets is equal to the sum of the variances of the assets each multiplied by the square of their weights in the portfolio. Thus if two assets each have independent returns and a given variance V, a portfolio with equal shares of each will have a variance of O.5V; if four assets each have independent returns and a variance V, the portfolio with equal shares of each will have a variance of 0.25V. In other words, the investor's own risk can be reduced by holding a diversified portfolio. 2. Goodman (1982) found that for most countries the bulk of the risk was diversifiable. 3. Indeed Guttentag and Herring (1984) suggest that the recent experience leads to increased uncertainty rather than to risk because these were not regular events with a corresponding probability. 4. In fact it can be argued, using Mynsky's theory of financial crisis, that the very success of developing country adjustment earlier and the existence of high rates of return from long-term investment would induce a system with insufficient liquidity to deal with sudden disturbances. References Arrow, Kenneth, and Robert Lind. 1970. Uncertainty and the evaluation of public invest ment decisions. American economic review 6O(June):364-78. Cline, William. 1983. International debt and the stability ofthe world economy. Washington, D.C.: Institute for International Economics. Goodman, Laurie. 1982. Risk and international bank lending. In Paul Wachtel, ed. Crisis in the economic and financial structure. Lexington, Mass.: Lexington Books. Guttentag, Jack M., and Richard J. Herring. 1984. Credit rationing and financial disorder. Journal of finance 39(December): 1359-82. - - - . 1985. The current crises in international banking. Washington, D.C.: Brookings Institution. Kubarych, Roger. 1982. Discussion of Jeffrey D. Sachs, "we debt in the 1980's: Risk and reform." In Paul Wachtel, ed. Crisis in the economic and financial structure. Lexington, Mass.: Lexington Books. Morgan Guaranty Trust. 1983. Global debt: Assessment and prescriptions. In World financial markets, pp. 1-14. New York: Morgan Guaranty Trust Company of New York. Sachs, Jeffrey D. 1981. The current account and macroeconomic adjustment in the 1970's. Brookings papers on economic activity 1:201-82. - - - . 1982. Lne debt in the 1980's: Risk and reform. In Paul Wachtel, ed. Crises in the economic and financial structure. Lexington, Mass.: Lexington Books. Sharpe, William. 1964. Capital asset prices: A theory of market equilibrium under condi tions of risk. Journal of finance 19(3):425-42. Solomon, Anthony M. 1983. Toward a more resilient international financial system. Federal Reserve Bank of New York quarterly review 8(3):3. Chapter 4 A Description of Adjustment to External Shocks: Country Groups Pradeep K. Mitra THE 1970s will be remembered as a decade that witnessed serious convul sions in the world economy. · Petroleum prices quadrupled in 1973-74, fell by a sixth between 1974 and 1978, and then increased by 80 percent in real terms during 1979-80. · The industrial market economies went into a recession in 1974-75 and thereafter recovered strongly before plunging into another reces sion in 1979--80, from which a slow recovery is under way. Stagflation was born in the countries of the Organisation for Economic Co operation and Development (oEeD), with successive peaks of eco nomic activity occurring at ever higher levels of unemployment. · The end of the decade saw an interest rate shock following the use of restrictive monetary policies to combat inflation in the leading indus trial countries. In addition, in the early 1980s, the real prices of major primary products exported by developing countries-adjusted for rising prices of imported manufactures-fell to their lowest levels since World War II. Some perspective on the magnitude of these convulsions may be obtained from the following statistics. World trade in fuels increased from $29 billion in 1970 to $535 billion in 1980. 1 Paying for the 1970s' fuel price increases was equivalent to finding the money to buy all the exports of another United States or Federal Republic of Germany. The current account deficits of the oil-importing developing countries as a proportion of GNP doubled from about 2.5 percent in 1973 to 5 percent in 1980. Debt-servicing payments of all developing countries, deflated by their export unit values, rose nearly threefold between 1972 and 1979; interest The work reported in this paper has been undertaken in the context of World Bank research project no. 672-74, "Adjustment in Oil-Importing Countries." The author is deeply indebted to Hector Sierra for exceptional research assistance. 103 104 ADJUSTMENT TO EXTERNAL SHOCKS rates, deflated by export prices, rose from 10 percent in 1979 to 20 percent in 1981. This chapter is a first report on work on developing economies' adjust ment to external shocks in the 19705 and early 1980s. It articulates a framework to impose analytical order on descriptions of shock and adjustment; to construct comparators that can place individual perform ance in perspective; and to locate empirical regularities among growth performance, external shocks, modes of adjustment, and, when data permit, policy variables. The paper also develops the analytical framework with a number of examples. The analytical framework is used in the third section to classify thirty-three developing economies into five groups, according to certain features of their adjustment, which are then reviewed. The review pro vides a convenient backdrop against which individual country adjustment may be viewed, a task to which two companion papers (Mitra 1985a and 1985b) are devoted. The conclusion, in a fourth section, is followed by appendix A, which lists data sources. Appendix B classifies thirty-three economies into five broad groups. Methodology The following methodology underlies the comparative analysis. (A more formal account of the model and decomposition method appears in Mitra 1984, 1985a, and 1985b.) An open economy macroeconomic model is estimated for each country over the 1963-81 period, with an assumed structural break after 1973. The output of the model during the 1974-81 period is then compared with the output for the same period as if the 1963-73 parameters had prevailed and under certain assumptions about the course of variables exogenous to the model. I refer to the hypothetical development as the "counterfactual." The changes in the principal macroeconomic aggregates between the two scenarios are then decom posed into price and quantity changes. External Shocks External shocks comprise (a) international price effects, (b) recession induced effects, and (c) net interest rate effects. International price effects measure the balance-of-payments impact of changes in an econ omy's terms of trade relative to the counterfactual and are the sum of the export price effect and the import price effect. The export price effect PRADEEP K. MITRA 105 measures the net impact of a fall in the purchasing power of exports over manufactures exported by the OEeD countries relative to the counterfac tual. The import price effect measures the net impact of a rise in the purchasing cost of imports in terms of manufactures exported by the OEeD countries relative to the counterfactuaL Both export and import price effects may each be subdivided into two components. First, the relative increase (decrease) in import (export) prices exerts an unfavorable impact on the balance of payments-the direct effect. Second, the price effect impoverishes the economy and, with unchanged policies, restrains imports, thereby exerting a favorable impact on the balance of payments-the indirect effect. It can be shown that, if the economy's savings propensity is positive, the direct effect dominates the indirect effect, so that a relative increase (decrease) in import (export) prices always exerts a damaging effect on the balance of payments. When measured against a 1971-73 base as a percentage of GNP, interna tional price effects averaged on an annual basis over the 1974-81 period ranged from an extremely unfavorable 7.5 percent in Chile and 5.9 percent in Uruguay through a somewhat less unfavorable 3.5 percent in Malawi, 2.9 percent in the Philippines, and 2.6 percent in Taiwan to a moderately favorable 3.5 percent in Malaysia and 3.7 percent in Tunisia to an extremely favorable 9.8 percent in Nigeria and 14.2 percent in Indonesia. Although import price effects were unfavorable in all cases, the magnitude of export price effects was extremely unfavorable in Chile and Uruguay on the one hand and very favorable in Nigeria and Indone sia on the other. Recession-induced effects on the balance of payments are twofold. The export volume effect (a direct effect) is the shortfall in an economy's exports as a result of a slowdown in the rate of growth of GNP in principal trading partners. From this effect must be subtracted the import-saving effect (or indirect effect), that is, the restraint in the growth of imports, with unchanged policies, due to the slowdown in income growth induced by the export volume shortfall. Recession-induced effects were generally positive, ranging as a percentage of GNP from 0.1 percent in Spain and Uruguay through 1.4 percent in the Republic of Korea and 1.9 percent in Taiwan to 3.7 percent in Indonesia. Net interest rate effects are twofold. The payments effect measures the impact on the balance of payments of an increase in real interest rates (in terms of manufactures exported by OEeD countries) payable on a coun try's debt relative to the counterfactual. From this figure must be sub tracted the receipts effect, that is, the impact on the balance of payments 106 ADJUSTMENT TO EXTERNAL SHOCKS of an increase in real interest rates (in terms of manufactures exported by OECD countries) earned by a country's interest-bearing assets relative to the counterfactual. Net interest rate effects, when measured vis-a.-vis real interest rates prevailing in 1971-73, ranged as a percentage of GNP from -0.6 percent in Mali and - 0.2 percent in Kenya to 2 percent in Korea and 2.5 percent in Bolivia. Payments effects were particularly important in Bolivia, Korea, and Singapore; the payments effect on short-term debt was important in Singapore and, to a lesser extent, in Portugal. Modes of Adjustment Economies unfavorably affected by external shocks had four basic ways (and combinations thereof) of responding to external shocks: (a) trade adjustment, (b) domestic resource mobilization, (c) investment slowdown, and (d) additional external financing. To avoid unnecessary repetition the reader is asked to remember that, as with shocks, all modes of adjustment are measured as deviations from the counterfactual. The examples provided below have been drawn from a list of the economies that suffered rather than benefited from external shocks during the 1974-81 period. Trade adjustment is the sum of export expansion and import substitu tion. Export expansion is the increase in the responsiveness of exports to changes in GNP growth in principal trading partners. It has a twofold effect. The direct effect measures the favorable impact on the balance of payments of boosting exports. From this figure must be subtracted the indirect effect, that is, the boost in import growth due to the expansion in income growth induced by the direct effect. Of the thirty-three econo mies to which the analysis underlying this chapter has been applied, those in which export expansion played a prominent role include Singapore, Korea, the Philippines, Chile, and Thailand as well as Taiwan. Import substitution is the reduction in the responsiveness of the econ omy's import demand to income. The direct effect measures the bal ance-of-payments impact of restraining imports. From this figure must be subtracted the indirect effect, that is, the boost in import growth due to the expansion in income growth induced by the direct effect. Examples of adjustment through significant import substitution are Brazil, Yugosla via, and Malawi. Both export expansion and import substitution improve the trade balance and boost GNP growth. Domestic resource mobilization measures the import-restraining effect PRADEEP K. MITRA 107 of a slowdown in income growth induced by improved savings perform ance as defined below. It may be broken down into its private and public components. "Private resource mobilization" is the reduction in the responsiveness of private consumption to income. This was important in Honduras, Morocco, Singapore, Yugoslavia, Jamaica, and Korea. "Pub lic resource mobilization" has two parts, "public consumption restraint," or the reduction in the responsiveness of public consumption to income, and "tax intensification," or the increase in the responsiveness to income of indirect taxes less subsidies. This term therefore ignores any changes in the direct tax effort, an omission that may be justified on grounds of their relative unimportance in developing economies. El Salvador, Singapore, and Honduras favored this mode of adjustment. Investment slowdown measures the import-restraining effect of a slow down in income growth brought about through a reduction in the ratio of investment to income relative to the period 1971-73. 2 This was a domi nant mode of adjustment in Jamaica, Singapore, Mali, and Kenya. Net additional external financing measures changes in gross additional external financing (defined as capital flows, reserves, and transfers and services net of interest payments, deflated by a price index of manufac tures exported by OECD countries) less changes in net interest payments resulting from changes in real net debt relative to the counterfactuaJ.3 This measure played an important role in a large number of countries, for example, Mexico, El Salvador, Honduras, Morocco, Mali, Portugal, Spain, Guatemala, Turkey, the Philippines, Uruguay, and Kenya. Patterns of Adjustment An analysis of the experience of thirty-four developing economies over the period 1974-81 reveals that twenty-five of them suffered adverse external shocks. Their responses to these shocks varied considerably, a feature that is worth bearing in mind in the following discussion. To impose a measure of analytical order on the richness and diversity of experience, however, it is convenient to divide the economies into five groups, according to the sign ofexternal shocks and the degree of reliance on different modes of adjustment. 4 Group 1 (Chile, Costa Rica, Philip pines, Singapore, Korea, and Taiwan) adjusted principally through ex port expansion and public resource mobilization. Group 2 (Argentina, Brazil, Guatemala, Honduras, India, Kenya, Malawi, Mali, Thailand, Turkey, and Uruguay) relied on either export expansion or public re source mobilization, whereas Group 3 (Jamaica, Portugal, and Yugosla via) was characterized by import substitution and negative public resource 108 ADJUSTMENT TO EXTERNAL SHOCKS mobilization. Group 4 (EI Salvador, Mexico, Morocco, and Spain) resorted to financing without domestic adjustment. Finally, Group 5 (Benin, Bolivia, Colombia, Indonesia, Ivory Coast, Malaysia, Niger, Nigeria, and Tunisia) experienced favorable external shocks. The (un weighted) average shock adjustment figures for the 1974-81 period are shown for the five groups in table 4-l. Export expansion and public resource mobilization. The average shock was highest for Group 1 at 3.98 percent of GNP. International price effects accounted for roughly 60 percent of total shocks, with the recession induced and net interest rate effects contributing equally to the remain der. All economies of the group resorted heavily to export expansion, which exceeded external shocks by more than one-third and to public resource mobilization, of which the principal component was tax inten sification. Together, export expansion and public resource mobilization accounted for 154 percent of external shocks. Import substitution was significantly negative everywhere except in Costa Rica, especially during the later years of the period. Whereas Chile, the Philippines, and Taiwan relied on substantial additional external financing and stepped up their ratio of investment to GNP, Korea sustained an investment boom with comparatively limited recourse to additional external resources. In con trast, Singapore adopted a somewhat contractionary package, with a cut in the share of investment and real repayment of borrowed funds; Costa Rica had a similar adjustment profile as well. The ratio of external financing to external shocks was higher in 1974-81 than in 1974-78 but was nevertheless quite modest for this group in relation to the others. Export expansion or public resource mobilization. International price effects accounted for roughly 80 percent of external shocks for Group 2. This group occupies a position between Groups 1 and 3 in terms of adjustment characteristics. Three broad patterns of adjustment may be distinguished. First, Argentina, Guatemala, India, Mali, and Uruguay resorted to export expansion while exhibiting negative import substitu tion and negative public resource mobilization, which was significantly worse in the years 1979-81 than in 1974-78. Second, and in quite a contrast, Honduras and Kenya adjusted through a combination of import substitution and public resource mobilization, with export expansion turning negative. Third, the remaining countries-Brazil, Malawi, Thai land, and Turkey-relied on a combination of export expansion and import substitution, with negative public resource mobilization aggravat ing the balance-of-payments impact of disturbances from the inter national environment. For the group as a whole, negative public resource mobilization added 40 percent to external shocks. There was significant additional external financing, especially in countries such as Honduras, Table 4-1. Balance-of-Payments Effects of External Shocks and Modes of Adjustment, 1974-78 and 1974-81 Averages (percentage of local currency GNP) Group 1 Group 2 Group 3 Group 4 Group 5 Effect 1974-78 1974-81 1974-78 1974-81 1974-78 1974-81 1974-78 1974-81 1974-78 1974-8 External shocks 1. International price effects a. Export price effect i. Direct effect -1.97 -2.87 0.63 -0.45 -3.86 -3.24 3.16 -2.31 -7.59 -9.26 ii. Indirect effect -2.38 -3.05 -0.37 -0.37 -2.87 -2.34 1.75 -1.06 2.93 -3.57 Difference (:; i ii) 0.41 0.18 -0.27 -0.08 -0.99 -0.89 1.41 -1.25 -4.66 -5.69 b. Import price effect ' i. Direct effect 6.08 8.06 3.16 3.71 4.55 4.98 2.03 1.28 2.03 2.41 <::::> \Q ii. Indirect effect 4.81 5.80 1.44 1.72 3.20 3.47 1.03 0.53 0.90 1.24 Difference (= i - ii) 1.27 2.25 1.72 2.00 1.34 1.51 1.00 0.75 1.13 1.17 Sum (:; la+ Ib) 1.68 2.43 1.45 1.91 0.35 0.61 0.41 -0.50 3.53 -4.52 2. Recession-induced effect a. volume effect 1.97 2.04 0.60 0.69 1.18 1.30 1.22 1.46 0.73 1.27 b. Import saving effect 1.27 1.28 0.30 0.39 0.84 0.91 0.65 0.81 0.08 0.33 Difference (= 2a - 2b) 0.70 0.76 0.30 0.30 0.34 0.39 0.57 0.66 0.65 0.93 3. Net interest rate effect a. Payments effect i. Medium and long term 0.11 0.68 -0.09 0.18 0.05 0.72 0.06 0.45 0.10 0.75 ii. Short term -0.01 0.87 -0.01 0.16 0.00 0.40 -0.03 0.22 0.01 0.15 Sum (:; i + ii) 0.10 1.54 -0.10 0.34 0.04 1.12 0.03 0.68 0.09 0.90 b. Receipts effect 0.01 0.76 -0.01 0.09 -0.10 -0.15 0.00 0.06 0.00 0.28 Difference ( 3a 3b) 0.10 0.78 -0.09 0.25 0.14 1.27 0.04 0.62 0.09 0.63 Total shock (= 1 + 2 + 3) 2.48 3.98 1.66 2.47 0.83 2.27 0.20 0.77 -2.79 -2.96 (Table continues on the following page Table 4-1. (continued) Group 1 Group 2 Group 3 Group 4 Group 5 Effect 1974-78 1974-81 1974-78 1974-81 1974-78 1974-81 1974-78 1974-81 1974-78 1974-81 Modes of adjustment 1. Trade adjustment a. Export expansion i. Direct effect 12.79 17.05 0.75 1.66 -7.60 -7.31 0.63 -0.02 -0.D2 0.25 ii. Import augmenting effect 9.09 11.60 0.18 0.55 -5.41 -5.23 0.32 -0.13 -0.91 -0.58 Difference (= i - ii) 3.70 5.45 0.57 1.11 -2.19 -2.08 0.31 0.15 0.89 0.83 b. Import substitution i. Direct effect 0.97 -4.20 0.87 0.85 4.68 4.43 -3.32 -3.28 -3.88 -5.04 ....... ....... ii. Indirect effect 1.45 -2.59 0.36 0.38 3.38 3.13 -1.55 -1.28 -0.17 -0.36 c Difference (= i - ii) -0.48 -1.61 0.50 0.48 1.31 1.30 -1.77 -2.00 -3.71 -4.68 Sum (= 1a + 1b) 3.22 3.84 1.07 1.59 -0.88 -0.78 -1.46 -1.86 -2.82 -3.85 2. Resource mobilization a. Private 1.08 0.54 -0.61 -0.44 -1.53 -0.96 0.72 0.65 0.98 1.27 b. Public i. Publicconsump. restraint -0.09 0.19 -0.69 -0.88 -2.93 -4.04 -0.61 -0.87 0.25 0.16 ii. Tax intensification 0.49 0.49 -0.10 -0.12 0.28 0.39 -0.25 -0.24 -0.86 -1.14 Sum (= i + ii) 0.40 0.68 -0.79 -1.00 -2.65 -3.64 -0.86 -1.11 -0.61 -0.98 Sum (= 2a + 2b) 1.48 1.22 -1.39 -1.44 -4.18 -4.61 -0.14 -0.46 -0.37 0.29 3. Investment slowdown -1.13 -1.91 -0.46 -0.69 2.48 2.78 -1.60 -0.84 -1.31 -1.74 4. Net additional ext. financing -1.09 0.83 2.45 3.01 3.41 4.88 3.39 3.93 0.97 2.34 Total (= 1 + 2 + 3 + 4) 2.48 3.98 1.66 2.47 0.83 2.27 0.20 0.77 -2.79 -2.96 Note: Definitions of Groups 1-5 are in Appendix B. Source: See text and Appendixes A. PRADEEP K. MITRA 111 Mali, Guatemala, Turkey, Kenya, and Thailand, with this mode of adjustment exceeding external shocks by more than 20 percent for the group as a whole. There was some increase in the share of investment in GNP in all countries except Mali, Kenya, and Malawi. Import substitution and negative public resource mobilization. Although the shocks experienced by Group 3 were less unfavorable than those affecting Groups 1 and 2, their composition was rather different. International price effects accounted for less than 30 percent of external shocks, whereas net interest rate effects exceeded 55 percent of shocks, largely because of their relative importance in Jamaica. The adverse balance-of-payments impact of negative public resource mobilization was more than one and one-half times as large as that of external shocks in this group, with the effect being extremely strong in Jamaica. Import substitu tion played a dominant role in all of the countries in Group 3; export expansion was significantly negative. External financing was much more important than in Groups 1 and 2 but much less so in the later years of the period. The average, however, conceals marked intercountry differ ences: although it played a prominent role in Portugal, it was much less important in Jamaica and was virtually negligible in Yugoslavia. Financing without adjustment. External shocks averaged 0.77 percent of GNP for Group 4. Recession-induced and net interest rate effects accounted for one and two-thirds times this figure, principally because of their overwhelming importance relative to external shocks in EI Salvador and Mexico. Table 4-1 clearly indicates the virtual lack of domestic adjustment across the board. Export expansion was negative except in Mexico (because of petroleum) and especially in Morocco and EI Salva dor. A major import and investment boom was under way in Morocco and, in relation to external shocks, in EI Salvador. Public resource mobilization was positive in EI Salvador but was more than offset by worsening performance in the other countries, especially Morocco and Spain. Additional net external financing was extremely important in all countries and was more than five times as important as external shocks for the group as a whole. Favorably affected countries. The countries of Group 5 experienced favorable shocks usually because they had been exporters of petroleum or of other primary commodities, so that the boom in prices in the mid-1970s allowed them to benefit over the period as a whole. Inter national price effects alone exceeded total shocks by more than one-half in absolute terms. Export price effects, as a proportion of shocks, were extremely favorable in the nonfuel primary producers (the Ivory Coast, Bolivia, Tunisia, and Malaysia), followed by petroleum exporters (Ni geria and Indonesia), which were in turn succeeded by Colombia. s Im 112 ADJUSTMENT TO EXTERNAL SHOCKS port price effects, though significant in the Ivory Coast, were distinctly less important. Differences in the relative price movements of primary commodities during the 1970s accounted for variations in the pattern and timing of adjustment among members of the group. On average, how ever, adjustment to favorable shocks took the form of an import boom that intensified in 1979-81 as compared with 1974-78, a stepping up of the share of investment in GNP, a slackening of public resource mobilization efforts, and substantial additional external financing at the end of the period under review. With respect to particular countries, there was an import boom in Bolivia, Colombia, Malaysia, and Tunisia and, to a somewhat lesser extent, in Indonesia and Nigeria. It was accompanied by an investment boom, which was particularly marked in the Ivory Coast and Benin. There was a slackening of public resource mobilization efforts in the Ivory Coast and less of one in Malaysia and Tunisia. Net real additional financing was important in Tunisia, Colombia, and Bolivia, was negligible in Indonesia and Nigeria, and was negative in Malaysia and the Ivory Coast. Conclusions The framework developed in this chapter serves to impose a measure of analytical order on the richness and diversity of individual experience. It has been applied to thirty-three economies, and the results have been aggregated to describe the broad contours of group adjustment, both as an end in itself and with a view to placing individual performance in perspective. It is against this background that the experience of individual countries is discussed in two companion papers (Mitra 1985a and 1985b). Appendix A. Data Data on national accounts, price deflators, and exchange rates are taken from the World Bank's World Tables. The index of international inflation is the unit value index of manufactured exports f.o.b. from developed countries and is taken from various issues of the United Nations Monthly Bulletin of Statistics. Export and import trade weights are taken from the International Monetary Fund's Direction of Trade Statistics. The calculations distinguish public and publicly guaranteed medium and long-term debt from short-term debt. The latter has a maturity ofless than one year. Outstanding medium- and long-term disbursed debt be PRADEEP K. MITRA 113 longs to different vintages and carries different interest rates. Data on interest payments therefore reflect such terms and conditions. In the absence of a detailed breakdown, the nominal interest rate on medium and long-term debt has been calculated as interest payments outstanding and disbursed debt Both numerator and denominator are taken from the World Bank's Debtor Reporting System, which, however, reports only public and pub licly guaranteed medium- and long-term debt. It is assumed that the rate payable on short-term debt as well as that earned by the country's interest-bearing assets, equals the London inter bank offered rate (UBOR). This rate has been understood to correspond to six months' maturity (source: Salomon Brothers until 1978 and the Inter national Financial Statistics [IFS] of the International Monetary Fund thereafter). Short-term debt data is derived from the Bank for Interna tional Settlements' Maturity Distribution of International Bank Lending. Interest-bearing assets are defined as follows: Total Reserves minus Gold (line ll.d. in the IFS) less Use of Fund Credit (line 2 e.s. in the IFS), expressed in dollars. Gold has not been included as part of reserves. Appendix B. Composition of Groups Group 1: Chile, Costa Rica, Philippines, Republic of Korea, Taiwan, Singapore. Group 2: Argentina, Brazil, Guatemala, Honduras, India, Kenya, Malawi, Mali, Thailand, Turkey, Uruguay. Group 3: Jamaica, Portugal, Yugoslavia. Group 4: EI Salvador, Mexico, Morocco, Spain. Group 5: Benin, Bolivia, Colombia, Indonesia, Ivory Coast, Malaysia, Niger, Nigeria, Tunisia. Notes 1. "Billion" means "thousand million." 2. This measure could be broken down, data permitting, into its private and public investment components. 3. See equation (A.26) in Annex 1 of Mitra 1984 or Mitra 1985a for an algebraic statement. 114 ADJUSTMENT TO EXTERNAL SHOCKS 4. The members of each group are listed for easy reference in appendix 2. 5. Export price effects were extremely favorable in Niger as well, but here external shocks were positive in 1974-78, with the terms of trade improving sufficiently thereafter to yield negative shocks for the period 1974-81 as a whole. References Mitra, Pradeep. 1983. World Bank research on adjustment to external shocks. World Bank research news 4:3 (Fal1lWinter). - - - . 1984. A description of adjustment to external shocks: Country groups. Develop ment Research Department discussion paper 85. Washington, D.C.: World Bank. - - - . 1985a. Adjustment to external shocks in selected semi-industrial countries. In G. Szego, ed. Studies in banking andfinance. Amsterdam: North Holland. Forthcoming. - - - . 1985b. Adjustment to external shocks in selected less developed countries. Washington, D.C.: World Bank, Country Policy Department. Processed. World Bank. 1981. World development report 1981. New York: Oxford University Press, chap. 6. Chapter 5 The Oil Syndrome: Adjustment to Windfall Gains in Oil-Exporting Countries Alan Gelb DEVELOPING COUNTRIES with only a limited range of exports-typically primary products-have greater fluctuations in their terms of trade than more diversified advanced economies. Mineral exports tend to be among the most volatile, and because highly specialized exporting countries tend to have high ratios of exports and imports to GDP, mineral exporters are prone to exceptionally large fluctuations in national income. Because a high proportion of natural rent on rich mineral deposits usually accrues to producer governments, the conduct of fiscal policy is often central in determining the use of resources from favorable but temporary move ments in terms of trade and their ultimate benefit to producing economies. l The external shocks experienced by oil producers during the past decade have been exceptional even by the standards of monoexporters. World oil prices quadrupled in 1973-74, then decreased slightly in nomi nal dollar terms in 1975-78. They then redoubled in 1979-80, peaking at about thirty-five dollars per barrel. As the world economy moved into recession, and conservation measures in the major consuming countries began to affect the demand for energy (particularly petroleum-based energy), prices fell by six to eight dollars per barrel. New sources of supply, notably the North Sea, came on stream. Output increased rapidly in Mexico, and energy sales from the Soviet Union to Europe rose. These developments placed additional stress on traditional exporters, who saw their sales contract in 1980-83, in some cases to little over half of their peak levels. A small group of producers-the capital surplus exporters such as Saudi Arabia and Kuwait-have exceptionally large oil reserves, with low recovery costs, small populations, and underdeveloped non-oil econ omies. With very limited absorptive capacity, such countries face, in the 115 116 THE OIL SYNDROME first instance, a portfolio-choice problem: whether to store their major asset in the ground or to deplete reserves more rapidly and accumulate assets abroad. 2 The capital surplus exporters are not considered in this chapter, which focuses on a sample of countries-Algeria, Ecuador, Indonesia, Nigeria, Trinidad and Tobago, and Venezuela-with smaller reserves and projected oil incomes insufficient to defray the costs of development for more than perhaps two decades, a short period in historical perspective. The main questions addressed concern (a) the magnitude of the windfalls from oil over the past decade, (b) how they have been used, and (c) the impact on non-oil producer economies. Has oil laid a basis for self-sustaining growth at a higher rate than would otherwise have been possible? Or have the difficulties of economic man agement through fluctuating income severely reduced the benefits of oil windfalls and resulted in increased oil dependence of producing coun tries? Dimensions of the Oil Windfall Gains from higher oil prices can be measured in a number of ways. My approach is to estimate the increase in domestic income resulting from an enlarged oil sector in current-value terms relative to the non-oil econ omy. So that the data remain comparable, such computations are better performed with the economy partitioned into mining and nonmining segments rather than into oil and non-oil segments, a change that has only a minor effect on the results. Figure 5-1 indicates the (unweighted) average time profile for 1973-81 of the windfall for the above-mentioned six countries expressed in each year relative to their nonmining economies. 3 In 1974 the average windfall peaked at 33 percent of nonmin ing income, but by 1978 this figure had contracted to 15 percent. The time and country average, for 1974-78, was 22 percent. The second oil price increase raised the windfall to 27 percent of non mining output in 1980 and maintained it at 23 percent of nonmining GDP for 1979-81. The main factor reducing the impact of the second oil price rise was the reduced size of oil sectors in constant prices relative to the rest of producer economies. For the period 1979-81, constant-price mining sectors were on average 8.5 percent smaller relative to the non mining economies than they had been in 1970--72. The second oil price shock thus impacted on a relatively smaller oil sector than had the first. Slumping prices and contracting sales for 1982-83 in combination appear to have halved the windfall gain. ALAN GELB 117 Figure 5-1. The Oil Win'd!all and Its Use, 1970-72 to 1981 Percentage of nonrnining GDP 0.5 OA 0.3 Oil windfall 0.2 0.1 o ~~~ ____L _ _ ~ ____~ _ _~ ____~ _ _~ ____~ _ _~ 1970-72 1975 1978 1981 Note: Unweighted average for Algeria, Ecuador, Indonesia, Nigeria, Trinidad and Tobago, and Venezuela. Source: World Bank data. Use of the Windfall: The Fiscal Response With little direct linkage between the oil sector and the rest of a developing economy, and upward adjustment of tax and royalty rates to reduce the share of rent accruing to the oil multinationals, the above mentioned fluctuations were mainly reflected in fiscal revenues, as shown in table 5-1. For the six countries mentioned above, central government revenues jumped from 20 percent of nonmining GDP to 37 percent with the first oil price increase.' The implication is that on average about four-fifths of the windfall as previously measured accrued to producer governments. Iran, for which data after 1977 are limited, experienced a particularly large windfall, 36.7 percent of non-oil GDP for 1974-77, which was reflected in fiscal revenues. Although there were significant differ ences in non-oil fiscal performance between countries, almost all the increase in the ratios of fiscal revenues to nonmining income is attribut 118 THE OIL SYNDROME Table 5-1. Central Government Revenues as a Percentage of Nonmining GDP Country 1970-72 1974-78 1979-81 Algeria 32.6 59.9 57.4 Ecuador 14.2 12.9 14.2 Indonesia 15.6 23.1 30.9 Iran 31.7 71.1' Nigeria 12.3 27.7 Trinidad and Tobago 19.9 55.9 57.2 Venezuela 25.2 42.1 36.3 Mean: six countries (excluding Iran) 19.9 36.9 Mean: five countries (excluding Nigeria) 21.4 38.8 39.2 Note: Dashes indicate years for which data were judged unreliable or estimates that were not compatible with series for earlier years. a. 1974-77. Sources: IMF 1983; World Bank data; and national authorities. able to inereased taxes and royalties on oil, except in Algeria, where non-oil taxes were unusually high and buoyant. Levels of development and income per head vary significantly between the above-mentioned producers. So do the economic role of government and the weight of the public sector in the economy. All countries had extensive and growing public involvement in the hydrocarbon industry during the 1970s, with virtually total nationalization in Algeria and Venezuela. At the one extreme, however, central government and public enterprises are estimated to account for about 90 percent of domestic Algerian investment, whereas at the other, the role of the Ecuadorian public sector beyond the traditional functions of administration, defense, and the provision of physical and human infrastructure has been quite limited. These differences reflect both the varying ideological tendencies of successive governments and historical accident. The extensive involve ment in agriculture of Algeria's socialist government, for example, and (conservative) Indonesia's considerable public sector holdings in timber and plantation crops both stemmed from the departure at independence of colonial proprietors, French and Dutch, respectively. Nevertheless, all of the oil exporters saw unparalleled growth in the size of the public sector after 1973, and most experienced a considerable extension of its role, toward direct participation in industrial production. Although most governments expanded their activities in virtually all directions, there were considerable differences of emphasis between countries. Algeria ALAN GELD 119 and Trinidad and Tobago placed high priority on natural gas, the former for sale in primary form and the latter through gas-based industrializa tion. Venezuela emphasized the development of metals industries, and Nigeria the expansion of road networks and school (and later university) enrollment. Ecuador's public programs were directed largely toward promoting private industry, wnereas Indonesia pursued a strategy rel atively balanced between physical infrastructure, education, agricultural promotion, and capital-intensive industrial ventures. 5 For 1970--72, current account deficits in the six countries had averaged 5.1 percent of nonmining GDP. In these years, deficits were especially high in Trinidad and Tobago (14.5 percent) and Ecuador (7.1 percent) be cause of large investment expenditures to finance oil extraction and because of spending in anticipation of higher revenue flows. As indicated in figure 5-1, during 1974-78 about one-quarter of the windfall was saved abroad through reductions in trade (and current) deficits and one-quarter was consumed. Slightly more than half of the increase in consumption relative to nonmining value added was public, slightly less than half private. The remainder was used for domestic investment. Although non-oil private investment boomed in certain countries, notably Vene zuela in 1976-78, increased investment outlays were overwhelmingly those of the public sector. The pattern for 1979-81 was similar except that private consumption increased its share of the windfall at the expense of domestic investment, which accounted for only one-third of the second windfall. Central government investment expenditures and net lending thus expanded particularly rapidly. On average they grew more than twice as fast as the respective non-oil economies after the first oil price increase, but their growth slowed after 1978. Recurrent expenditures also in creased but less rapidly. Wage and salary expenditures of central govern ment grew, on average, at about 110 percent of the rate of increase in non-oil incomes. An exception to the pattern of moderate growth in central government current expenditure was the category of subsidies and transfers, which I shall discuss further below. Some Consequences of Increased Domestic Expenditures As would be predicted from a standard Salter-Swan model (see, for example, Bruno 1975; Corden and Neary 1982; and Gelb 1981), the rapid increase in domestic absorption of goods and services was reflected in a tendency for real exchange rates to appreciate. The degree of real appre ciation is shown in table 5-2. Relative to their average levels in 1970--72, 120 TIlE OIL SYNDROME Table 5-2. Real Exchange Rate Movements, 1974-83 Nonmining output deflator Trade-weighted relative to M UV of imports real exchange rate' of developing countries Country 1974-78 1979-81 1982-83 1974-78 1979-81 1982-83 Algeria 90.8 103.3 121.7 88.3 92.1 Ecuador 106.4 112.7 120.0 91.8 101.0 Indonesia 133.8 129.5 140.2 115.6 103.2 Iran 100.4 119.2 151.1 Nigeria 131.0 170.0 209.1 98.9 108.0 Trinidad and Tobago 101.4 107.7 138.8 88.0 104.2 Venezuela 97.6 103.2 124.7 81.6 93.4 Mean b 110.2 121.1 139.1 94.0 100.3 Memo: United States 92.3 93.4 105.2 82.1' 82.9' 98.2' Note: MUV manufacturers' unit value. Base 1970-72. Dashes indicate that data are not available. a. Averages. b. Excluding Iran. c. Wholesale price index relative to MUV deflator. Sources: International Monetary Fund, International Financial Statistics; World Bank data; and United Nations, Monthly Bulletin of Statistics. trade-weighted real exchange rates (here defined as the ratio of the domestic price level of the oil exporter to those of its trading partners, converted at average market exchange rates) were 10 percent higher in 1974-78,21 percent higher in 1979-81, and almost 40 percent higher in 1982-83. 6 It should, however, be noted that during the 1970s the unit value of manufactures imported by developing countries (manufactures' unit value, or MUV, index) rose markedly relative to the price levels of most countries largely because of the oil and other primary intermediate price shocks. This relative price shift limited the tendency for consump tion and investment prices to decrease relative to the cost of domestically produced non-oil goods, as would be normal with real currency apprecia tion. Purchasers in producing countries were thus cushioned from in creased import price shocks by their own real exchange appreciation. As noted above, in addition to expanding their traditional functions, governments typically channeled windfall gains into industry, especially petrochemicals and heavy metals. They also invested heavily in physical infrastructure, notably to develop their transport and communications systems. Public projects tended to be large and complex and were fre quently highly capital intensive. In fact, among a sample of the top ALAN GELD 121 nineteen developing countries with investments in projects exceeding $100 million each, all but five were found to be oil exporters (Murphy 1983). The dimensions of the part of the investment programs of the above-mentioned set of countries that consisted of such large projects may be seen from table 5-3, which is mostly based on a sample of some 1,600 large projects in the developing countries for the period 1970-79. Iran, which ranked an overall second after Saudi Arabia, included in its investment program 108 projects averaging more than $1 billion dollars each. The total capital cost was equivalent to more than one and one-half times its 1977 GNP, or ten times its 1977 oil windfall as previously com puted. Venezuela's investment program, which in contrast to Iran's was largely directed toward metals (notably steel and aluminum), repre sented five times its 1980 oil windfall, or half its 1980 GNP. The large projects identified in table 5-3 represent, on average, very roughly four and one-half years' average oil revenue for 1974-81. Indonesia and Nigeria, the poorest producers, were somewhat less inclined to mortgage oil for large projects, but their investments of this type were still consider able. In addition to such large investment commitments, most countries created or expanded programs of subsidies and transfer payments directed toward holding down the rate of inflation and supporting loss making firms. Between 1970-72 and 1974-78, fiscal subsidies and trans fers expanded, on average, twice as rapidly as nonmining GDP. Between 1974-78 and 1980-81 they rose about 1.6 times as rapidly. In addition, producer governments were reluctant to raise domestic oil prices, choos ing to pass part of the windfall on to domestic consumers in the form of lower prices. In a number of cases, these were set at roughly the cost of production, so that government derived no revenue from the part of oil output that was consumed at home. Several producers, notably Ecuador and Indonesia, raised domestic prices of oil derivatives in the early 1980s, but they still typically remained below world levels, and as domestic consumption grew more rapidly than non-oil economies, the implicit fiscal burden on the state increased. 7 Energy subsidies in 1980 were estimated to be equivalent to almost 10 percent of household income in Ecuador, while fiscal subsidies rose sharply in Trinidad and Tobago, to about 7 percent of GDP or 11 percent of nonmining GDP by 1981. These subsidies were directed primarily to keeping down prices of consumer goods but were also extended to support unprofitable industries. Such estimates do not include the subsidies implicit in loans made to loss-making (and frequently public) firms, nominally for investment, and in guarantees permitting them access to commercial sources of finance. It is difficult to estimate such subsidies (since many such firms would prob Table 5-3. Macroprojects in Oil-Exporting Countries Rank among Project sector (percentage) Number Average devel of Cost cost Cost I Costl oping Other Infra- projects (US$ bil (US$ mil 1980 1980 oil coun Hydro indus struc Country included lions) lions) GNP Windfall tries carbons Metals try ture irati 108 119.6 1.107 1.57' 10.2' 2 30 7 9 54 Algeria 69 38.7 561 1.07 4.2 5 36 7 33 23 ' Venezuela 27 27.4 1,015 0.51 5.4 10 33 41 7 19 ~ Mexico 59 26.0 441 0.18 5.1 2 46 17 12 25 Nigeria 19 14.4 758 0.17 0.9 15 26 11 16 47 Indonesia 44 14.4 327 0.23 1.1 16 41 18 16 25 Trinidad and Tobago b 7 6.9 983 1.35 4.5 -' 61 29 Od 0" Note: Data are for projects in which costs exceeded US$I00 million. a. 1977 GNP and oil windfall. b. Gas-based industrial projects only. Includes Tenneco-Midcon liquefied natural gas project proposed for 1988. c. No data. d. Zero by definition, because gas-based industrial projects alone have been included. Sources: Murphy (1983), table 2.5; Auty (1984). ALAN GELB 123 ably have been unable to borrow from commercial sources at any price without support), but they appear to have been considerable and to have been accorded to some extremely unprofitable firms. By 1983, for exam ple, it was estimated that the production costs of Caroni Sugar in Trinidad were five times those of efficient world-scale producers, despite the fact that some of the latter, notably in Australia, had unit labor costs several times higher. In addition to supporting firms, some oil-producing governments stimulated employment directly through public works programs. The INPRES (Instruksi Presiden) programs in Indonesia and the Special Works programs in Trinidad and Tobago employed some 2.5 percent of the two countries' respective labor forces. The impact of such programs depends on their administration and on the extent to which labor is a major constraint to production, particularly in agriculture. While the impact of INPRES appears to have been beneficial in labor-surplus Java, the Trinidad programs (which offered pay at least twice as high as the rate prevailing in agriculture) contributed to the acceleration of a rapid movement off the land that led to a drop in agricultural output. Per capita food production and agricultural production as a whole were both reduced by about 20 percent in the period 1969-71 to 1982, during which time population expanded by only 16 percent. The main loser was sugar, which saw its output fall by 62 percent. The larger public projects had a greater tendency to overrun initial estimates both in terms of cost and time, as shown in table 5-4. One-third of the largest projects in the sample on which the table is based experi- Table 5-4. Cost and Time Overruns in Macroprojects Project size (US$ millions) Item 100-249 250-499 500-999 1,000+ Percentage of total projects with cost escalations, completion delays, or postponements I suspensions 21 28 38 47 Average cost escalation (percent) 30 70 106 109 Percentage of total projects with cost escalation 10 18 28 34 with completion delay 11 14 16 16 with postponement/suspension 7 10 13 20 Source: Reprinted by permission of Westview Press from Macroproject Development in the Third World: An Analysis of Transnational Partnerships by Kathleen J. Murphy. Copyright © 1982 Westview Press, Boulder, Colorado. 124 THE OIL SYNDROME enced cost overruns that averaged 109 percent. Overruns on the smaller projects were less frequent and more modest at 30 percent. Delays of between one and two years plagued half the troubled projects; a further 25 percent experienced delays of three to four years. These estimates greatly understate the true extent of cost and time overruns, because many projects were not completed by 1980; many are still under construc tion, and a number may never reach completion. The tendency to overrun initial estimates and the poor operating performance of many plants once installed reflect a variety of factors, none specific to oil-exporting countries, but all accentuated by the scope and pace of their investment growth. B First, projects were, in many cases, inadequately prepared and assessed. In no country does there appear to have been systematic assessment of relative costs and benefits across a spectrum of potential projects. Second, larger projects tended to be more complex, both technologically and in terms of the organization necessary to integrate the project with its necessary infrastructure. Some involved state-of-the-art technology, which, in certain cases, was installed without the involvement of an experienced expatriate operating company. With little detailed knowledge of the industry or plant in question, public financing agencies were also sometimes slow to detect and correct emerg ing problems in construction, startup, and operation. Third, about half of the purchasing power of oil relative to domestic construction costs was eroded by increases in the latter for 1973-78. Increased construction costs were a major factor in real appreciation of the exchange rate. As noted above, international inflation in traded manufactured goods was also high. Finally, certain industrial investments of the oil producers were severely affected by the global recession in the 1980s, as described below. The momentum of accelerated public investment (some of which im plied large future recurrent obligations) and growing subsidies proved hard to curb when oil revenues fell, as they did in 1978 and after 1981. Central government deficits averaged 4.1 percent of nonmining GDP in 1978, and excluding Trinidad and Tobago, where expenditures acceler ated more slowly, current account deficits averaged 11.8 percent of nonmining GDP. A number of exporters moved to slow domestic absorp tion of goods and services. Indonesia devalued by 50 percent in Novem ber 1978, seeking to restore the domestic purchasing power of oil rev enues and to promote non-oil exports. These contractionary moves were interrupted by the second oil price increase, which resulted in a current balance surplus of $11.8 billion in 1980 for the six countries (excluding Iran). As current-dollar commodity exports contracted by 21.6 percent during 1980-82 and imports rose by 22.3 percent, this surplus shifted to a current deficit of $19.6 billion by ALAN GELB 125 1982. Of the current account deterioration between 1980 and 1982, 58 percent may be attributed to decreases in merchandise export revenues and 36.8 percent to increased imports of goods. These swings in the exporters' current accounts often mirrored de velopments in their respective public sectors. Ecuador's public sector, for example, ran surpluses of about 2 percent of GOP in 1973-74, but these turned into deficits of 5 percent of GOP in 1977-78. With the second oil price increase, the deficit declined, but with contracting revenues and mushrooming subsidies and interest payments it rose to about 8 percent of GOP in 1982. Economic management through the fluctuation in oil prices was ren dered more difficult by the fact that access to international capital tended to vary with the level of oil prices, which affected future price and revenue expectations rather directly. Algeria was able to boost the ex penditure impact of increased oil revenues by 50 percent during 1974-78 through borrowing abroad, largely to finance a transition from an oil to a natural gas-based hydrocarbon sector. Mexico augmented its compara tively small oil windfall (3.5 percent of nonmining GOP in 1979-81) by two-thirds by financing a growing deficit on goods and nonfactor services. In addition, Venezuela and Mexico were able to cushion the impact of growing private capital outflows by large public borrowings until the outlook in world oil markets deteriorated. The impact of expanded investment on growth has been, at first sight, disappointing, as shown in table 5-5. Excluding Iran, where data are limited, only Ecuador proved able to accelerate the growth rate of its nonmining economy for 1972-81 significantly relative to performance for 1967-72. On average, non-oil economies (excluding Iran) were 4.1 per cent smaller in 1979-81 than they would have been had they maintained their 1967-72 growth trajectories. On closer examination, the growth record is less adverse than appears from historical trends. Algeria, Indonesia, and Nigeria had all previously been in recovery phases, two from internal disturbances and one from a protracted war of independence, while Ecuador and Trinidad had been stimulated by oil development and the prospect of growing export rev enues. The non-oil growth performance of the sample had therefore been exceptional in 1967-72 at 7.3 percent, some 1.5 percent higher than the average growth of GOP in middle-income developing countries. Although the higher growth initially stimulated by spending from the first oil price increase was not sustained, the non-oil economies still grew 0.9 percent more rapidly than they had through the favorable period of the 1960s. Much of this growth was, however, demand led rather than supply generated in the sense that non-oil growth responded to increased Table 5-5. Growth Trends in the Oil Exporters, 1967-81 (percent) Goods and non/actor service Nonmining Domestic GDP investment Exports Imports Country 1967-72 1972-81 1967-72 1972-81 1967-72 1972-81 1967-72 1972-81 Algeria 9.5 8.6 16.7 10.8 5.7 -1.0 11.6 10.8 Ecuador 4.7 7.6 3.2 10.2 15.9 6.0 6.0 9.7 Indonesia 8.5 8.2 24.3 13.0 15.7 4.3 16.7 19.1 .... tv Iran' 10.1 13.3 10.2 21.1 12.9 -0.3 17.7 23.7 0\ Nigeria 9.2 5.3 8.7 - b 4.2 _b 15.3 Trinidad and Tobago 5.3 5.4 6.1 9.3 2.5 -6.5 6.6 8.4 Venezuela 6.5 5.1 11.9 3.5 1.3 -8.7 7.7 12.8 Unweighted mean (excluding Iran) 7.3 6.7 12.4 9.3 7.7 -1.7 9.7 12.7 Memo: middle-income oil importers 5.8' 5.1' 8.2 5.6 6.7 4.0 7.4 1.5 a. 1967-72 and 1972-77. b. Deflated data unreliable before 1970. c. 1960-70 and 1970-82 GDP. Source: World Bank data. ALAN GELB 127 absorption after 1974 but slowed after 1978 despite the expectation that the large investment undertaken in 1975-78 would begin to contribute to output growth. Although growth was only moderate, it was often poorly balanced. Construction had been the leading growth sector over the 1970s, followed by services and protected import-competing industry, with agriculture and non-oil export industry lagging in most cases. Only Ecuador, Indone sia, and Venezuela managed to raise domestic food and agricultural supply per head during the 1970s, the latter from an extremely small base (in 1970-72 Venezuelan agriculture represented only 8 percent of non mining GDP). Despite a policy objective common to all governments, that of reducing dependence on oil, the volume of non-oil exports contracted in all countries except in Ecuador (which saw a considerable shift toward processed products and manufactures) and in Indonesia, which main tained a fairly strong non-oil export performance across a wide range of traditional and nontraditional commodities. Overall export volumes con tracted, on average, by 1.7 percent annually in the period 1972-81. After 1981: The End of the Oil Boom? The downturn in world oil markets after 1981 revealed the fragility of the development patterns of the oil exporters. Shifts in the allocation of resources toward the nontraded sectors that had cumulated over the 1970s could not be rapidly reversed, and reluctance to devalue (plus competitive devaluations of trading partners) caused real exchange rates to remain at an appreciated level in 1982-83, as shown in table 5-2. The massive infrastructural and educational investments that had been under taken since 1974, whatever their implications for future productivity, did not represent an autonomous source of income to replace oil incomes. More seriously, the global outlook changed for a number of sectors notably steel, aluminum, and natural gas-which had figured promi nently in the investment programs. In 1980, for example, the Organisa tion for Economic Co-operation and Development (oEeD) was forecast ing a doubling of global steel demand to 1,400 million tons by the year 2000. More recent forecasts project a 20 percent rise to only 900 million tons. The difference has serious implications, particularly for countries with domestic markets too small to absorb full-capacity output of large capital-intensive plants that had gone forward without foreign partners to assure marketing outlets. Such countries would need to be competitive with the globally most efficient (or most highly subsidized) exporters to overcome trade and transport margins and a preference for domestic 128 THE OIL SYNDROME supply in major markets. In the case of steel, a producer such as the Iron and Steel Company of Trinidad and Tobago (Iscorr) was required to undercut U.S. minimills by 15 percent, although its production costs were some 50 percent higher than their estimated level of $270/tonne in 1982. As the pressure of demand slackened, the transient boom of the mid-1970s was followed by deceleration in non-oil growth, by surplus capacity, and by slackening labor markets. A further factor decelerating demand was the tendency for private capital to flow abroad, particularly in oil-exporting countries with open capital markets, such as Venezuela and Indonesia. In 1978-81 the total cumulative current balance deficits of the six countries, at $5.6 billion, accounted for only 31 percent of the deterioration in their net foreign assets, where "net foreign assets" is defined as the change in external debt less that in currency reserves. Venezuela may have experienced an outflow equivalent to almost 10 percent of GDP in 1982, impelled by a stagnant economy, interest rate ceilings, and reluctance to adjust the exchange rate in line with perceived trends in world oil prices. In 1979-82 its non-oil economy virtually stagnated despite massive investments and considerable increases in the labor force, which should have assured growth of some 4 percent per annum even in the absence of any productivity improvements. Conclusions Oil-exporting countries entered the mid-1970s with high expectations that access to seemingly unlimited quantities of foreign exchange would accelerate development and would lead to the creation of a modern, productive, self-sustaining non-oil economy. Toward this end, they allo cated the bulk of their increased oil income to domestic investments, which were mainly large scale and were overwhelmingly carried out by the public sector. Multiplier effects of investment expenditures, cost overruns, subsidy growth, and the recurrent spending needs of much past investment all resulted in a tendency to overshoot available revenues when the latter fell. The result has been a pronounced "stop-go" rhythm that has made economic management difficult. It is not yet possible to assess the impact on producer economies, because many domestic investments, notably in transportation and education, would be expected to have long gestation lags. Overall, however, the yield on much domestic investment has probably fallen well short of that which could have been obtained abroad, and its supply-side growth impact has been moderate. With hindsight, the oil exporters would probably have enjoyed a larger benefit from their windfalls had ALAN GELB 129 they saved a higher proportion abroad and limited domestic investments through applying market criteria more rigorously. This conclusion ab stracts, of course, from the impact of such a strategy on the global recycling problem. Notes 1. Government may also absorb fluctuations in nonmineral export revenues. 2. Although the Hotelling rule predicts that unit natural resource rents should rise at the rate of interest, the medium-run fluctuations about any such long-run relationship have major fiscal consequences. An extra 250,000 barrels/day sold through 1981, with the proceeds invested in U.S. governmenttreasury bills, would, by February 1984, have yielded approximately $4 billion ("billion" means "thousand million") against an estimated value of $2.6 billion for the same volume of oil valued at February prices. The capital surplus countries are discussed in Hablutzel (1981). 3. The windfall is expressed as the difference between the ratio of mining sector value added to value added in the nonmining economy and the value of this ratio in the base period 1970-72. Consumption and investment effects are similarly expressed. For details, see Gelb (1984). 4. The decrease for Ecuador is explained by (a) the fact that some oil revenues accrue to special funds outside government as defined here and (b) the fact that certain non-oil taxes eased after 1974. 5. These choices and the reasons behind them are discussed more extensively in Gelb (1984). 6. Algeria, the most notable exception to the pattern of real exchange appreciation, is analyzed in Conway and Gelb (1984). 7. Petroleum subsidies conceded by producer governments are usually implicit rather than fiscal, because revenues forgone through selling oil for domestic use at levels below world prices are not included in fiscal accounts. 8. As witness the $8 billion cost of the trans-Alaskan oil pipeline versus its $900 million original budget, cost overruns can be large in developed countries also. Their peculiar significance for the oil exporters is due to the weight of large projects relative to the size of their economies. References Auty, R. 1984. The deployment of oil rents in a small parliamentary democracy: The case of Trinidad and Tobago. Washington, D.C.: World Bank; processed. Bruno. M. 1975. The two sector open economy and the real exchange rate. Jerusalem: Hebrew University of Jerusalem, Falk Institute. Conway, P. J., and A. H. Gelb. 1984. Oil rents in a controlled economy: A case study of Algeria. Development Research Department discussion paper. Washington, D.C.: World Bank. Corden, W. M., and J. P. Neary. 1982. Booming sector and de-industrialization in a small open economy. Economic journal 11:2(February):119-41. 130 THE OIL SYNDROME Garcia Araujo, M. 1982. The impact of petrodollars on the economy and the public sector of Venezuela. Paper delivered at the tenth national meeting of the Latin American Studies Association, Washington, D.C., March 4. Gelb, A. H. 1981. Capital importing oil exporters: Adjustment issues and policy choices. World Bank staff working paper 475. Washington, D.C.: World Bank. 1984. Adjustment to windfall gains: A comparative analysis of capital-importing oil exporters. Developmental Research Department discussion paper. Washington, D.C.: World Bank. Hablutzel, R. 1981. Development prospects of the capital surplus oil exporting countries. World Bank staff working paper 483. Washington, D.C.: World Bank. International Monetary Fund (IMF). 1983. Government financial statistics. Washington, D.C.: IMF. Morgan, D. R. 1979. Fiscal policies in oil importing countries. International Monetary Fund Staff Papers 26. Murphy, K. 1983. Macroprojects in developing countries. Boulder: Westview Press. Nankani, G. 1979. Development problems of mineral exporting countries. Staff working paper 354. Washington, D.C.: World Bank. Chapter 6 The Debt Crisis in Latin America Larry A. Sjaastad, Aquiles Almansi, and Carlos Hurtado THIS CHAPTER discusses the various aspects of the current external debt problem in five Latin American countries: Argentina, Brazil, Chile, Mexico, and Venezuela. These countries were chosen because they account for a very large portion of total Latin American external debt and because they have all experienced severe payments difficulties since 1982. Other countries could have been included. Costa Rica, for example, was the first Latin American country to experience an external debt crisis in recent times, but Costa Rica has a very small external debt (at least in a comparative sense), and its problems do not conform to the global debt syndrome. Similarly, Panama could have been included on the ground that its debt is one of the largest in the world relative to its GDP. The cases that we have considered, however, permit us to cover a great deal of the problem and involve countries with highly different characteristics and histories. What Triggered the Crisis? The debt crisis was sprung on the world in August 1982, when the Mexican government became unable to continue debt service, and quickly spread to a number of other Latin American countries. We shall argue, however, that the crisis formed much earlier, with the coincidence of a number of quite independent phenomena: the rapid and enormous growth of external debt (again, particularly in Latin America), the shortening of maturities as more and more of that debt was owed to the world capital market rather than to international financial institutions, the sudden and spectacular rise in dollar interest rates, worldwide dollar deflation, and the political crisis in Latin America. All of these events were primary ingredients. From 1971 to 1982, the total external debt of the developing countries grew, in nominal terms, by 600 percent, and debt service by 1,100 131 132 DEBT IN LATIN AMERICA percent. (Estimates based on the 1983 World Debt Tables of the World Bank.) Even if we exclude dollar inflation during that period, growth was at an unsustainable rate, as it far exceeded the rate of growth of real output in the creditor nations. Nevertheless, by historical standards, the level of debt in 1982 was not extraordinarily high when measured against variables such as gross domestic product and exports. Debt service rela tive to debt, however, had doubled, with no commensurate increase in the ability of the countries to generate the requisite fiscal and trade surpluses to pay that service. Moreover the countries borrowing most heavily were also those most heavily in debt; their debt service was enormous but still less than their annual borrowings. Once the ability to borrow began to wane, the heavy borrowers had to convert fiscal and trade deficits into surpluses virtually overnight, a trick that none of them could turn. The result was that country after country had to go hat in hand to the International Monetary Fund (IMP), seeking the blessing that would prevent the lines of credit from drying up entirely. The great buildup of private international lending occurred during the 1970s and early 1980s and was closely related to, if not a consequence of, the oil price increase(s), which produced a virtual explosion of liquidity in the international commercial banks. Perhaps because of unanticipated inflation, and perhaps in part due to the surpluses of the Organization of Petroleum Exporting Countries (OPEC) following the oil price increase of 1973-74, real interest rates on dollar-denominated external debt were very low and indeed frequently negative, giving the developing countries a rather strong incentive to incur that debt. When real interest rates are negative (and are expected to remain so), it is clearly impossible to have "too much" external debt. To be sure, there was concern by economists and government officials, but it focused on the more narrowly defined recycling problem rather than on the ability of the developing countries to meet the interest service on their external debts. The recycling issue arose from the fact that the commercial banks were absorbing short-term liabilities (from a relatively concentrated group of depositors) while acquiring assets that were, by their very nature, long term (despite the particular contractual terms of any given loan), as the bulk of the loans were going to a limited number of the developing countries. By the late 1970s and early 1980s, less and less was heard about the recycling problem. There were basically two reasons. First, the insatiable appetite of at least certain developing countries for foreign capital was clearly matched by the willingness of the larger commercial banks to add to their international exposure, and second, growing fiscal deficits in LARRY A. SJAASTAD AND OTHERS 133 several of the industrialized countries were absorbing an ever larger portion of the OPEC surplus. So strong were these forces that the second oil shock, one much larger than the first in terms of the transfers involved, failed to reestablish any significant degree of concern with the recycling problem. Three nearly simultaneous events in the early 1980s served to trigger the current international debt "crisis," which is, of course, precisely the doomsday scenario envisaged by the more pessimistic participants in the recycling debate. The first was the sharp recovery of the dollar beginning in late 1980, followed by the extraordinary strength of the dollar during much of 1982. Associated with the recovery were two important (and related) developments: dollar interest rates rose sharply, and dollar prices of many traded goods (particularly commodities that figured heavily in the exports of the debtor countries) fell abruptly, especially during the spectacular rise of the dollar from October 1980 through February 1981. As most of the debt of the developing countries was denominated in dollars, and much of it at the floating London interbank offered rate (LIBOR), the appreciation of the dollar and the rise in dollar interest rates implied an equivalent rise in debt service. Not only did the nominal interest service increase because of the rise in interest rates, but real interest service rose even more because of the decline in the dollar prices of many tradables. The latter effect comes about because countries obviously service their foreign debt by importing less or exporting more (in the final analysis); a decline in the dollar price of tradables implies an increase in the real burden of servicing foreign debt. In addition, there is some evidence that the appreciation of the dollar, coupled with the subsequent world recession, resulted in an adverse turn in the terms of trade facing the debtor countries. This latter effect came about because the appreciation of the dollar had, initially at least, a stronger downward effect on the prices of homogeneous commodities than on those of manufactures, and the former are very important exportables of most debtor countries. Whereas appropriately defined real rates of interest on external debt had been negative during much of the 1970s for a number of developing countries, short-term real rates of interest on that debt rose abruptly to the 15-20 percent range at the end of 1980 and have remained very high ever since. Although it is difficult to have too much debt (in the short run, at least) when real rates of interest are negative, a debt service problem quickly emerges when real rates increase as they did. Chilean data readily illustrate the point. The second column of table 6-1 indicates the annual rate of change of unit (dollar) values of Chilean imports and exports, and the third column indicates the behavior of 134 DEBT IN LATIN AMERICA Table 6-1. External Prices of Chilean Tradables and Interest Rates (percent) Annual rate Real of change- Interest interest Period prices' rate" rate' 1977 5.1 7.0 1.8 1978 6.1 6.4 0.3 1979 26.8 11.2 -12.3 1980 16.6 13.9 -2.3 1981 -5.5 15.7 22.4 1982 -8.6 13.3 24.0 --- a. Rate of change in a simple average of unit values of Chilean imports and exports as calculated by the United Nations Economic Commission for Latin America. b. Annual averages of six-month LIBOR rates, based on U.S. dollars. c. Defined against the prices of Chilean tradables. Source: Gil-Diaz 1983. dollar nominal interest rates. The fourth column combines the two into a real rate of interest on Chilean foreign debt. Note that that rate was very low or negative from 1977 through 1980 and then rose into the 20-25 percent range. Although nominal dollar interest rates rose somewhat, the main source of the increase in the real rate is the dollar deflation-the prices of Chilean tradables actually fell during 1981 and 1982. The same phenomenon occurred in most other debtor countries in 1981, and it was probably the most important single element in the making of the crisis. The dollar deflation was a direct consequence of the recovery of the V.S. dollar vis it vis other major currencies. The second development was an intensification of the tendency toward greater fiscal deficits in the industrialized countries, a development that was exacerbated by the rapid decline in the OPEC current account sur pluses that began in early 1981. Clearly the excess supply of funds available to the developing countries was shrinking; nevertheless, they were still able to finance their debt service (and then some) by rollovers of the principal and further borrowing that more than covered interest payments. The change in real interest rates that began in late 1980 did not immediately provoke the crisis, but competition for international funds was plainly making it more difficult for the developing countries to maintain their level of foreign borrowing. Even without further develop ments, it was but a matter of time until the debt crisis. The third development exposed the underlying rot for all to see and precipitated the crisis. That development was the South Atlantic conflict of May-June 1982. As is well known, several major V.S. banks had an LARRY A. SJAASTAD AND OTHERS 135 exposure in Latin America well beyond their capital and reserves; such an exposure could be considered prudent only if the United States could reasonably be expected to come to the aid of any Latin American country with large debts to U.S. banks when that country encountered a pay ments difficulty. The decision by the United States to support Britain against Argentina in the South Atlantic conflict, however, exploded the credibility of that assumption. The bankers, unnerved by the event, began immediately to restrict the flow of loans to Latin America. Mexico, whose reserve position was very fragile after the February 1982 devalua tion, was the first country forced to suspend payment, but Argentina, Chile, Brazil, and even Venezuela suffered the same fate. Thus the debt crisis resulted from a number of factors no single one of which would have been sufficient alone. The crisis has been particularly intense owing to the fact that much of the borrowing helped, directly or indirectly, to sustain fiscal deficits; as that borrowing declined, the fiscal deficits had to be turned into surpluses to generate the local currency required to buy the dollars to service the debt. Countries such as Argen tina and Brazil have found it extremely difficult to undertake the fiscal reforms that are required of them and have turned more and more to inflation as a source of finance. In the process, control of international capital flows has been tightened (as the country tries to avoid runs on its own currency), with the effect that governments have no access what soever to the very substantial foreign currency earnings accruing to their citizens on their foreign investments. Governments increasingly find themselves being abruptly cut off from such funds, and this aspect makes the current debt crisis particularly dangerous. Latin American Public Finance during the 1970s In this section we look at the revenues, expenditures, and deficits of Latin American governments during the 1970s and, when data are avail able from our source, the Government Finance Statistics of the IMF, during the early 1980s. 1 We also look at the main expenditure items in order to detect the sources of growth in the aggregate. Because we are primarily interested in the connection between public expenditure and foreign indebtedness, we have taken the rather unusual step of presenting these figures in current U.S. dollars. To do so we used the market exchange rates as they are recorded in the International Financial Statistics (IFS) of the IMF. Since we are dealing with flow data, we use the average market exchange rate in each period. The presenta tion of data in current U.S. dollars has two implications: the U.S. infla 136 DEBT IN LATIN AMERICA tion imparts a positive trend to all series, and the figures reflect the volatility of Latin American exchange rates. The reason for looking at the data in current U.S. dollars rather than in units of constant domestic purchasing power, as is customary, is that, with free access to the world capital market, it is the current dollar value of expenditures, and not its size in terms of units of constant domestic purchasing power, that affects the extent of foreign borrowing. The interested reader could easily relate these series to foreign debt data in other sections of this book. Finally, we should note that the four main expenditure items that we have considered here-general public services, defense, capital expendi ture, and our aggregate "social, economic, others"-do not add up, in general, to the figure we present under the heading "total expenditure and net lending." In consequence, the reader should not expect to see the average rate of growth of the latter equal to a weighted average of the average rates of growth of our four expenditure items. General Patterns Revenues have been growing faster, on average, than expenditures in four out of five countries surveyed, Venezuela being the exception, with both items growing at the same pace. The ranking of different expenditure items in terms of growth rates varies widely across countries. Some patterns emerge clearly, however, from table 6-2. First, the aggregate "social, economic, others" is a very fast-growing item. It ranks first in Argentina and Brazil and second in Chile and Mexico. Capital expenditure, on the other hand, is a slow growing item in the southern cone countries. It ranks fourth of four items in Argentina and Brazil and third in Chile. Third, defense has been a fast-growing item in Chile and Argentina only. Two striking features of these ran kings are their similarity in the cases of Argentina and Brazil and their perfect disparity between Chile and Venezuela. Argentina Time series in Argentina reflect the highly unstable nature of its economic behavior during the 1970s and early 1980s. Except in 1975, when a series of "maxi" devaluations by former economic ministers Gomez Morales and Rodrigo reduced the dollar value of revenues about 49 percent, they have been growing all along the period (see table 6-3). Expenditures, on the other hand, suffered two contractions, the first in 1975, for the reason explained above, and the second in 1977, when the LARRY A. SJAASTAD AND OTHERS 137 Table 6-2. Growth of Revenues and Expenditures in Selected Latin American Countries, 1973-80 Item Argentina Brazil Chile Mexico Venezuela Total revenue and grants (percent) 36.7 21.4 23.3 27.5 32 Total expenditure and net lending (percent) 34.7 20.9 15.1 26.4 32 General public services (percent) 36.2 19.5 2.5 0.4 30.4 Rank' 3 2 4 4 1 Defense (percent) 37.9 11.3 15 17.5 13.5 Rank' 2 3 1 3 4 Social, economic, other (percent)h 43.4 23.7 11.4 24.4 22 Rank' 1 1 2 2 3 Capital expenditure (percent) 33.5 10.4 5.5 29.2 23.3 Rank' 4 4 3 2 a. According to growth rate. b. Education, health, social security and welfare, housing and community amenities, other community and social services, economic services, other purposes. Source: Tables 6-3 through 6-13. Table 6-3. Total Revenues and Grants, Total Expenditures and Net Lending, and Deficit in Argentina, 1973-80 Total revenue Total expenditure and grants and net lending Deficit US$ Percentage US$ Percentage (US$ Year billions change billions change billions) 1973 4.68 42.16 6.50 56.51 1.81 1974 8.54 82.43 11.46 76.61 2.92 1975 4.37 48.81 8.44 -26.33 4.07 1976 6.63 51.63 10.51 24.54 3.89 1977 7.26 9.56 8.77 16.63 1.42 1978 10.55 45.23 12.66 44.38 2.11 1979 17.88 69.50 20.70 63.49 2.82 1980 26.64 49.00 32.14 55.34 5.51 Average 36.70 34.70 Sources: International Monetary Fund, Government Financial Statistics and International Financial Statistics, various issues. Table 6-4. Expenditure in Argentina, by Function, 1973-80 General public Social, economic, Capital services Defense other' expenditure US$ Percentage US$ Percentage US$ Percentage US$ Percentage Year billions change billions change billions change billions change 1973 0.64 74.47 0.64 74.47 4.89 103.8 1.06 24.6 1974 0.90 40.82 0.79 23.22 8.65 76.9 1.69 58.4 ...... 1975 0.55 -39.21 0.60 -23.58 5.09 -41.2 1.04 -38.4 ~ 35.94 69.93 8.50 67.0 2.68 1976 0.74 1.02 158.0 1977 0.75 0.4 0.93 -8.73 6.91 -18.7 2.23 -16.7 1978 1.06 42.54 1.59 70.77 10.71 55.0 2.64 18.5 1979 1.80 69.63 2.64 66.01 15.96 49.0 3.60 36.1 1980 2.98 65.10 3.48 31.49 24.81 55.5 4.59 27.7 Average 36.20 37.90 43.4 33.5 a. Education, health, social security and welfare, housing and community amenities, other community and social services, economic service other purposes. Sources: International Monetary Fund, Government Financial Statistics and International Financial Statistics, various issues. LARRY A. SJAASTAD AND OTHERS 139 recently installed military government tried, with some initial success, to reduce the size of public bureaucracy. The fastest-growing item during the period was the aggregate "social, economic, other." Within this aggregate, the leading item was what the IMF Government Finance Statistics presents as "other purposes," with a growth rate of 107.3 percent. The defense average growth rate ranks second among the four aggre gates displayed in table 6-4. It is also the more volatile item, reflecting both the timing of the limits conflict with Chile and the changing political influence of the Argentine military. Finally, Argentina shares with Brazil and Chile, the other southern cone countries in the group, a pattern of slow-growing capital expendi tures. Brazil Brazil shows the smoother set of time series within the group of countries surveyed in this study. Figures for the central government are displayed in tables 6-5 and 6-6 and figures for state and local governments in table 6-7. The central government's revenues presented a slightly larger average rate of growth than expenditures. The opposite was true of state governments. An interesting feature of the central government's aggregate figures is the negative trend in growth rates for both revenues and expenditures. No definite trend was observed in the rates of growth of both items at the state government level. With respect to the components of central gov ernment expenditure, there are two fast-growing items, general public services and "social, economic, other," and two slow-growing ones, defense and capital expenditure. "Social, economic, other" ranks first, in accordance with the leading role this aggregate plays in explaining ex penditure growth in our group of countries. Capital expenditure shows the same slow growth pattern of the other two southern cone countries, Argentina and Brazil, and within this group, Brazil is the only exception with respect to high defense spending growth. It is interesting to note that the only item that clearly follows the trend of decrease observed in aggregate expenditure growth is "social, eco nomic, other." The other three items show more volatile behavior. State governments' expenditures also show a decreasing growth rate. Chile In Chile, much as in the other countries, central government revenues have grown at faster rates than expenditures over the 1973--80 period. 140 DEBT IN LATIN AMERICA Table 6-5. Total Revenues and Grants, Total Expenditures and Net Lending, and Deficit of Brazil's Central Government, 1973-79 Total revenue Total expenditure and grants and net lending Deficit US$ Percentage US$ Percentage (US$ Year billions change billions change billions) 1972 11.12 23.3 11.36 20A 0.24 1973 15,44 38.8 15.18 33.6 -0.26 1974 20,41 32.2 19.3 26.0 -1.28 1975 24.20 18.6 24.76 29,4 0.55 1976 30.14 24.5 30,45 23.0 0.31 1977 33.86 12.3 35.18 15.5 1.32 1978 39.05 15.3 41.08 16.8 2.04 1979 41.39 6.0 41.92 2.0 0.53 Average 21.4 20.9 Sources: International Monetary Fund, Governmenl Financial Statistics and International Financial Statistics, various issues. Data relating to the Chilean government revenues and expenditures appear in table 6-8. This tendency results in surpluses for 1979 and 1980 of about $1 billion and $1.5 billion. 2 It must be considered throughout that central government data give a narrow definition of the public sector. Although these data include several official agencies, they leave out important public enterprises such as Corporacion Nacional del Cobre (copper) and Linea Aerea Nacional (airlines). Therefore these figures are significant more in relation to the evolution of the government finances over time than in relation to the absolute level of the aggregates. In the 1975-78 period the central government registered only minor surpluses and deficits. Nevertheless, in 1979-80 it achieved significant surpluses, probably with the intention of paying off part of its domestic outstanding debt. The composition of Chilean government expenditure over the 1973-80 period was rather peculiar (see table 6-9). The relative average growth of expenditures for social and economic purposes and capital was the lowest among the countries under consideration (aggre gate expenditure was also relatively low). General public services spend ing was specially reduced during 1974--75 as a consequence of the govern ment's efforts to reduce the size of the public sector. Important decreases in other expenditure items in 1975 are also explained (at least in part) by the huge devaluation experienced in that year. In the 1976-77 period, all expenditure grew significantly, especially social and economic spending in 1977. By 1978 public spending in general had slowed down. Social and economic spending, however, kept rising rapidly. Table 6-6. Expenditure in Brazil's Central Government, by Function, 1973-79 General public Social, economic, Capital services Defense other" expenditure US$ Percentage US$ Percentage US$ Percentage US$ Percentage Year billions change billions change billions change billions change 1972 1.37 4.6 0.84 11.4 8.22 33.0 1.38 12.4 1973 1.40 2.8 1.09 29.8 10.92 32.8 1.52 9.9 ' 1974 2.14 52.1 1.28 17.2 14.32 31.1 2.15 41.7 ..... ~ 1975 2.62 22.7 1.38 7.6 19.22 34.2 3.06 42.5 1976 4.04 54.1 1.78 29.2 21.63 12.5 4.13 34.9 1977 4.49 11.2 1.83 2.9 25.6 18.4 4.01 -3.0 1978 5.63 25.4 2.04 11.5 31.2 21.9 3.90 -2.8 1979 4.68 16.9 1.65 -19.3 32.92 5.5 1.87 -52.1 Average 19.5 11.3 23.7 10.4 a. Education, health, social security and welfare, housing and community amenities, other community and social services, economic service other purposes. Sources: International Monetary Fund, Government Financial Statistics and International Financial Statistics, various issues. 142 DEBT IN LATIN AMERICA Table 6-7. Revenues, Expenditures, and Deficits of Brazil's State and Local Governments, 1973-79 State governments' State governments' Local total revenue total expenditure govern and grants and net lending ments' Deficit deficit US$ Percentage US$ Percentage (US$ (US$ Year billions change billions change billions) billions) 1972 5.61 17.8 5.80 22.1 0.19 0.05 1973 7.62 35.8 7.87 35.7 0.24 0.08 1974 9.87 29.4 10.28 30.7 0.41 0.15 1975 11.04 11.9 12.30 19.6 1.26 0.22 1976 11.37 3.0 13.13 6.8 1.76 0.25 1977 13.28 16.9 13.97 6.4 0.69 0.23 1978 12.11 -8.8 17.31 23.9 1.72 0.29 1979 17.05 40.8 18.38 6.2 1.33 0.07 Average 18.3 18.9 Sources: International Monetary Fund, Government Financial Statistics and International Financial Statistics, various issues. Table 6-8. Revenues and Expenditures of Chile's Central Government, 1973-80 Total expen- Total revenue diture and and grants net lending Deficit US$ Per- US$ Per (US$ Year billions centage billions centage billions) 1973 2.94 3.70 0.76 1974 3.13 6.5 3.73 0.8 0.6 1975 2.54 -19.0 2.53 -32.2 -0.01 1976 3.14 23.9 3.01 19.0 -0.13 1977 4.25 35.2 4.40 46.2 0.15 1978 5.97 16.9 4.99 13.4 0.02 1979 7.07 66.4 6.07 21.8 -1.0 1980 9.41 33.1 7.87 29.5 -1.55 Average 23.3 15.1 Sources: International Monetary Fund, Government Financial Statistics and International Financial Statistics, various issues. Table 6-9. Expenditure in Chile's Central Government, by Function, 1973-80 General Social, public economic, services Defense other" Capital US$ Per US$ Per US$ Per US$ Per Year billions centage billions centage billions centage billions centage 1973 0.6 0.4 2.65 0.78 ..... 1974 0.4 -36.8 0.5 39.8 2.82 6.4 1.13 44.9 ~ 1975 0.3 -19.0 0.3 -37.3 1.92 -31.9 0.49 -56.6 1976 0.4 28.1 0.4 16.8 2.01 4.69 0.43 -11.7 1977 0.5 35.7 0.5 35.5 3.34 66.2 0.50 16.1 1978 0.6 4.3 0.6 20.0 3.73 66.7 0.61 22.7 1979 0.71 16.7 1980 0.76 6.6 Average 2.5 15.0 11.42 5.53 Note: Dashes indicate that data are unavailable. a. Education, health, social security and welfare, housing, other community and social services, economic services, and other purposes. Sources: International Monetary Fund, Government Financial Statistics and International Financial Statistics, various issues. 144 DEBT IN LATIN AMERICA It is noteworthy that, in growth, official capital expenditures are always below aggregate spending except in 1978. This tendency may well be misleading, however, because, as noted above, several public enterprises are excluded from these figures. Finally it is also noteworthy that both expenditures and revenues show the lowest average growth in the group of countries considered. Mexico Although it is also observed in the Mexican case that the relative increase in revenues is slightly higher, on average, than that of expendi tures, the deficit of the central government shows substantial increases in the 1972-80 period. Data relating to the central government finances appear in table 6-10. This behavior is explained by the changes in eco nomic policies in the 1970s. A very expansionist policy was clearly pur sued until 1975, which concluded with a significant devaluation at the end of 1976. The low deficits registered in 1977 and 1978, then, reflect the implementation of the adjustment program undertaken after the de valuation. In 1978 another fiscal expansion began, probably one of a higher magnitude, and continued until 1981. The reduction of all figures Table 6-10. Revenues and Expenditures of Mexico's Central Government, 1972-80 Total expen- Total revenue diture and and grants net lending Deficit US$ Per- US$ Per- (US$ Year billions centage billions centage billions) 1972 4.68 6.04 1.4 1973 5.59 19.5 7.8 29.0 2.2 1974 7.64 36.8 10.4 33.3 2.7 1975 10.70 40.0 15.0 44.0 4.3 1976 10.96 2.4 15.1 0.8 4.2 1977 10.68 -0.2 13.4 -11.4 2.7 1978 14.19 32.9 16.9 26.6 2.8 1979 19.23 35.7 23.7 40.0 4.5 1980 29.42 52.8 35.2 48.6 5.8 Average 27.5 26.4 Note: Data for 1972 percentages are not available. Sources: International Monetary Fund, Government Financial Statistics and International Financial Statistics, various issues. LARRY A. SJAASTAD AND OTHERS 145 in table 6-7 during 1977 had the monetary causes implied by the devalua tion, and a similar effect was probably observed in 1981. Until 1975 the most important expenditure item of the government, in levels and rates of growth, was social and economic spending (see table 6-11). General public services and capital expenditure also showed sig nificant increases, increases in the latter being of a higher level. Although defense spending shows a high growth rate on average, its share in total spending is not significant by comparison with the figures for the other countries. The effects of the devaluation and the adjustment program im plemented in 1976 were felt most in government public service and capital expenditures. Social and economic spending fell insignificantly in 1977. Nevertheless, the spending level in general public services kept falling until 1979, and that of capital rose after 1977 at the highest level. Even though this definition of government excludes many huge public enterprises--for example, PEMEX (petroleum), CFE (electricity), twenty one steel enterprises, and seventeen chemical companies--the govern ment's capital expenditures have, on average during 1973-79, the highest growth rate, whereas spending for general services has the lowest. Venezuela The central government in Venezuela registered important deficits, especially in 1977-78. Data relating to Venezuelan government finances appear in table 6-12. The same qualification expressed about the defini tion of government used for the other countries applies here: the Vene zuelan data exclude many important publicly owned enterprises--such as Petr6leos de Venezuela (petroleum) and Petroquimica de Venezuela (petrochemical plants). Government expenditure consistently increased over the 1972-81 period although at very different rates. The highest sustained increase seemed to take place starting in 1980. Notice that monetary factors could not affect these figures as much as in the other cases, because the exchange rate remained fixed during practically the whole period. It is noteworthy that the benefits from the world oil shock to Venezuela were immediately reflected in the 1974 increase in spending (and revenue) figures. Regarding the composition of government expenditure (table 6-13), the highest growth was registered by general public services, whereas defense had the lowest. Furthermore, the latter's level was relatively unimportant. Expenditure on social and economic activities showed rather significant increases until 1977 and slowed down later in 1978-79. The general slowdown in government spending, however, was most Table 6-11. Expenditure in Mexico's Central Government, by Function, 1972-80 General Social, public economic, services Defense other Capital US$ Per US$ Per US$ Per US$ Per Year billions centage billions centage billions centage billions centage 1972 0.5 0.2 0.9 1.61 1973 0.6 15.8 0.3 24.1 5.9 37.2 2.06 28.0 ..... 1974 0.7 20.0 0.4 35.4 8.3 40.7 2.52 22.3 ~ 1975 1.1 50.3 0.5 23.8 11.1 33.7 3.0 19.0 1976 1.2 11.7 0.5 5.3 11.2 1.0 3.43 14.4 1977 0.5 -55.0 0.4 -17.7 10.1 -9.8 2.82 17.8 1978 0.4 -32.1 0.5 18.5 11.2 10.9 4.03 42.7 1979 0.3 7.3 0.6 31.5 14.7 31.3 6.83 69.6 1980 0.8 19.2 22.1 50.3 10.61 55.4 Average 0.36 17.5 24.4 29.2 Note: Dashes indicate that data are unavailable. a. Education, health, social security and welfare, housing. other community services, and economic services. Sources: International Monetary Fund, Government Financial Statistics and International Financial Statistics, various issues. LARRY A. SJAASTAD AND OTHERS 147 Table 6-12. Revenues and Expenditures of Venezuela's Central Government, 1972-81 Total expen- Total revenue diture and and grants net lending Deficit US$ Per- US$ Per (US$ Year billions centage billions centage billions) 1972 3.0 3.1 0.13 1973 3.95 31.4 3.7 17.8 -0.3 1974 10.4 162.0 9.8 164.0 -0.6 1975 9.9 -4.3 9.8 0.5 -0.1 1976 9.3 -6.0 10.6 8.5 1.3 1977 10.0 7.5 12.3 15.9 2.3 1978 to.O -0.2 12.7 3.3 2.7 1979 12.0 19.8 11.8 -7.8 -0.2 1980 15.7 31.1 16.1 37.1 0.4 1981 22.8 46.6 24.3 48.2 1.5 Average 32.0 32.0 Note: Data for 1972 percentages are not available. Sources: International Monetary Fund, Government Financial Statistics and International Financial Statistics, various issues. important in capital expenditure: negative rates of growth were reg istered in 1979 and 1980, whereas a positive but low growth took place in 1978. Until that time capital expenditure had shown the most rapid increases. There are thus two major characteristics of government spending in Venezuela during the 1972-80 period. Until 1977, total expenditure increases were sustained (though at varying rates), and capital accumula tion appeared to be the major concern of government fiscal expansion. In addition, social and economic activities seemed to be important targets during the period. After 1977 there was a generalized slowdown in total expenditure-which seems to have ended in 1981-and a change in the major concerns of government policy, away from capital spending and toward current expenditure, mainly in general public services. Money Creation and Government Finance The manner in which money creation is linked to public sector deficits, and the degree to which it is, are complex questions. Not all central bank expansion takes the form of credit to the government, nor does all credit Table 6-13. Expenditures in Venezuela's Central Government, by Function, 1972-81 General Social, public economic, services Defense other' Capital US$ Per- US$ Per- US$ Per- US$ Per- Year billions centage billions centage billions centage billions centage 1972 0.2 0.3 2.4 0.74 1973 0.2 16.5 0.3 6.7 2.73 13.8 0.90 21.7 ..... 1974 0.4 54.1 0.5 43.2 4.57 67.4 1.70 89.3 & 1975 0.6 50.3 0.5 18.1 5.54 21.2 1.90 12.0 1976 0.6 12.8 0.4 16.6 6.1 10.1 2.95 55.1 1977 0.8 33.2 0.6 26.2 8.4 37.7 4.17 41.3 1978 1.0 17.1 0.6 10.4 8.9 6.0 4.49 7.8 1979 1.0 1.1 0.7 12.0 8.2 -7.9 2.75 -38.7 1980 1.6 58.3 0.8 8.3 10.5 28.0 2.69 -2.2 Average 30.4 13.5 22.0 23.3 Note: Data for 1972 percentages are not available. a. Education, health, social security and welfare, housing, other community services, economic services, and other purposes. Sources: International Monetary Fund, Government Financial Statistics and International Financial Statistics, various issues. LARRY A. SJAASTAD AND OTHERS 149 to the government come from the central bank. Much money is created, in some countries, to finance off-budget expenditures such as interest subsidies (for example, in Brazil) or to aid ailing private banks (for example, in Chile and Argentina). Finally, the relation between deficits and money creation is subject to great and frequent change, making it difficult to formulate any generalizations. To estimate the amount of resources provided to the public sectors of Argentina, Brazil, Chile, Mexico, and Venezuela by their monetary authorities, we combined direct and indirect financing. Direct financing consists of the change in central bank claims on government (line 12a of the IFS). Indirect financing comes about because central banks lend to commercial banks, which in turn lend to public sectors. To obtain an estimate of this indirect finance, we constructed the net claims of the commercial banks on the public sector (commercial bank claims on government, line 22a, plus commercial bank claims on the rest of general government, line 22b, minus government deposits in commercial banks, line 26d), and compared it with central bank credit to commercial banks (line 26g). When changes in commercial bank claims on the public sector exceeded the change in central bank claims on commercial banks, we used the latter as the amount of indirect finance, and vice versa. It is unlikely, then, that either private saving or central bank lending to the private sector is included in our measure of indirect finance. The sum of the direct and indirect provision of resources appears in tables 6-14 and 6-15 under the heading aCG. Tables 6-14 and 6-15 also show the change in reserve money (labeled aRM, line 14 in the IFS) and the ratio of aCG to aRM. As can be seen, the ratios fluctuate enormously, indicating that contemporaneous monetary expansion is not highly correlated with cen tral bank lending to the public sector. In Argentina, claims on the public sector have increased every year during the 1972-82 period except for 1977, the average ratio being about two-thirds. In Brazil, the pattern has been highly erratic, and the average ratio is slightly negative; only in 1982 did direct plus indirect central bank claims on government expand by a greater amount than reserve money. The Brazilian case is obviously a special one in that central bank lending is largely to the Banco do Brazil, which then extends subsidized credit to both the public and the private sector. In Chile there was a sharp contraction of lending to the public sector in 1980 and 1981 but also a sharp decline in the stock of reserve money in 1981 and 1982. During an earlier period (1973-75), lending to the govern ment far exceeded money creation, indicating that credit to the private sector was severely squeezed. During the 1977-80 period, central bank claims on government grew by only one-third as much as reserve money; Table 6-14. Changes in Central Bank Credit to Government and Flows of Reserve Money: Argentina, Brazil, and Chile Argentina' Brazil" Chile" Year ,iCG ,iRM ,iCG/,iRM ,iCG ,iRM ,iCG/,iRM ,iCG ,iRM ,iCG/,iRM 1972 7 6 1.17 0.4 5.2 0.08 53 39 1.36 1973 24 108 0.22 3.5 15.0 0.23 605 251 2.41 1974 30 75 0.40 13.0 14.8 -0.88 2,917 692 4.22 1975 160 349 0.46 -9.9 17.4 0.57 14,394 2,565 5.61 ..... 1976 468 1,960 0.24 5.5 40.0 0.14 23,788 10,254 2.32 v. c 1977 -57 1,103 0.05 33.3 56.1 0.59 41,252 15,391 2.68 1978 1,669 3,504 0.48 25.2 73.2 0.34 12,860 16,637 0.77 1979 1,705 6,095 0.28 74.6 206.4 0.36 22,296 19,471 1.15 1980 15,060 10,386 1.45 64.6 261.0 0.25 -48,640 25,359 -1.92 1981 54,478 27,794 1.96 -163.8 502.0 -0.33 -62,647 -8,381 7.47 1982 112,297 381,847 0.29 1,504.2 1,004.3 1.50 16,466 -21,006 -0.78 Note: I:!.CG = sum of the direct and indircct provision of resources (see text). I:!.RM = change in reserve money. a. In billions of current pesos. b. In billions of current cruzeiros. c. In millions of current pesos. Sources: International Monetary Fund, Government Financial Statistics, various issues. LARRY A. SJAASTAD AND OTHERS 151 Table 6-15. Changes in Central Bank Credit to Government and Flows of Reserve Money: Mexico and Venezuela, 1972-82 Mexico' Venezuela b Year /lCG /lRM /lCG//lRM /lCG /lRM /lCG//lRM 1972 23.1 24.2 0.95 30 619 0.05 1973 22.7 17.6 1.29 65 1,363 0.05 1974 34.8 30.1 1.15 -1,279 2,722 -0.47 1975 32.5 35.5 0.92 -872 4,009 -0.22 1976 24.3 -9.9 -2.45 -1,743 2,558 -0.68 1977 148.4 165.0 0.90 -201 3,517 -0.06 1978 58.2 84.9 0.69 -3,529 2,078 -1.70 1979 109.0 132.7 0.82 1,148 2,791 0.41 1980 159.0 208.2 0.76 -1,538 1,613 -0.95 1981 284.8 323.3 0.88 372 4,630 0.08 1982 1,459.5 1,023.6 1.43 3,165 5,203 0.61 Note: /leG = sum of the direct and indirect provision of resources (see text). /lRM = change in reserve money. a. In billions of current pesos. b. In billions of bolivars. Sources: International Monetary Fund, Government Financial Statislics, various issues. this was the period of accumulation of international reserves in that country, with 1982 being a year of major decumulation. The Mexican case is much more stable (until 1982). In most years, the growth in claims on government was very close to the growth in reserve money, the major exception being 1976, when the devaluation caused a flight from the peso and hence an actual decline in reserve money. With the onset of the debt crisis in 1982, however, the government began to rely more heavily on the central bank for financing. Venezuela presents a picture similar to that of Brazil (but for different reasons). Reserve money has grown steadily and systematically, but very little of that growth has been captured by the government. This situation sharply contrasts with that of Mexico, where in most years the bulk ofthe money created in the central bank was passed directly to the fiscal authorities. The generalizations that the data permit are quite limited. All five countries except Venezuela have made heavy use of the central bank for fiscal purposes. This practice seems to have accelerated in 1982 with the onset of the debt crisis; central bank direct and indirect lending to governments increased very sharply in all five countries, including Vene zuela. Second, new central bank credits to governments exceeded reserve money creation in 1982 in all five countries except Argentina and Vene Table 6-16. Ratios of Money Creation to GDP and of Inflation Tax to GDP and Total Revenue: Argentina, Brazil, and Chile (percent) Argentina Brazil Chile ~RM TAX TAX flRM TAX TAX ~RM TAX TAX Year GDP GDP TAX + REV GDP GDP TAX+GDP GDP GDP TAX + GD 1972 2.6 3.5 23.7 1.2 0.8 8.5 9.8 9.1 32.1 1973 29.2 7.0 38.6 2.6 0.8 8.1 12.2 14.2 41.2 ....... 2.0 2.0 16.7 1974 13.0 9.4 39.5 6.6 9.3 26.7 ~ 74.0 1.6 1.7 16.1 6.2 1975 19.8 31.2 7.4 21.5 1976 25.8 25.5 67.8 2.5 2.2 17.8 7.8 5.7 16.6 1977 4.7 13.3 50.0 2.3 2.1 17.7 5.5 3.6 10.5 1978 5.8 8.8 38.2 2.1 1.9 16.3 3.9 2.3 6.2 1979 3.7 5.3 27.7 3.4 3.1 26.6 3.1 2.9 6.7 1980 3.0 3.3 19.0 2.2 3.0 22.2 2.8 2.3 5.7 1981 4.0 4.3 2.1 2,8 23.8 -0.7 0.7 1982 24.9 13,1 2.1 2,5 -1.8 1.2 Note: aRM = changc in reserve money. GDP gross domestic product. TAX inflation tax. REV = revenue. Some data were unavailable, indicated by the dashes. Sources: International Monetary Fund, Government Financial Statistics, various issues. LARRY A. SJAASTAD AND OTHERS 153 zuela (in Argentina much of the enormous acceleration in money crea tion stemmed from the ruinous Carvallo plan for bailing out bankrupt financial institutions; much the same sequence of events was to unfold in Chile in 1983). Prior to the debt crisis, however, reliance on central bank financing was erratic in all countries except Mexico. The above data indicate only qualitative behavior, not the magnitude of the financing, nor do they indicate the consequences for inflation. Indeed, reliance on central bank financing need not be inflationary if the resources transferred to the government do not exceed the normal growth in the demand for reserve money. Indeed, if the central bank does not allocate that growth to the government, the result will be a buildup for international reserves, which is essentially lending abroad, a phe nomenon that occurred in all five countries during the second half of the 1970s and in 1980. Tables 6-16 and 6-17 give a better idea of the magnitude of central bank finance, and the consequences for inflation, by presenting the growth of reserve money as a fraction of GDP, distinguishing between growth in that money and the "inflation tax." Reserve money is always a very small fraction of GDP, so that, in the absence ofinflation, the growth in that stock relative to the level of GDP will be correspondingly small. Tables 6-16 and 6-17 indicate, however, that some countries have made very heavy use of reserve money growth. During the 1973-76 period in Argentina, growth in reserve money averaged about 20 percent of GDP-a level that cannot be sustained without igniting a very high rate of inflation-and returned to that level in 1982. The growth ofreserve money in Brazil and Vene zuela has been stable in the neighborhood of somewhat more than 2 percent of GDP. In Chile it has fluctuated widely, as it has in Mexico, in the latter country reaching more than 10 percent of GDP in 1982. The growth in reserve money indicates the potential (or sometimes the consequences) for central bank finance of fiscal deficits (or the acquisition of international reserves). This growth includes both the increased de mand for reserve money because of normal economic growth and the replacement demand that comes about because of inflation. Inflation depreciates the real value of the stock of reserve money, so money holders must demand more of it simply to sustain their holdings in real terms. The latter source of demand is identified as TAX in tables 6-16 and 6-17; the ratio of TAX to G D P is simply the inflation rate (as measured by the various CPIS) multiplied by the beginning-of-year stock of reserve money, the product being expressed as a fraction of GDP. The third column in each country panel of table 6-10 presents TAX as a fraction of normal revenues (REV) plus the inflation tax. The excess of the column labeled IlRMIGDP and that labeled TAXIGDP is the change in reserve money in real terms expressed as a fraction of GDP. Table 6-17. Ratios of Money Creation to GDP and of Inflation Tax to GDP and Total Revenue: Mexico and Venezuela Mexico Venezuela tJ.RM TAX TAX tJ.RM TAX TAX - Year GDP GDP TAX+REV GDP GDP TAX + REV 1972 4.0 0.4 4.0 0.9 0.2 0.8 1973 2.5 1.8 15.3 1.6 0.4 1.6 ' u, 1974 3.4 1.9 15.1 2.7 0.9 1.6 .e:.. 1975 3.3 1.2 9.0 3.2 0.7 1.6 1976 0.7 2.4 16.1 1.8 0.7 1.9 1977 9.4 2.2 14.1 2.2 0.9 2.4 1978 3.9 2.3 13.6 1.2 0.8 2.4 1979 4.6 2.8 15.8 1.3 2.1 6.1 1980 5.4 4.1 19.4 0.7 1.9 4.9 1981 5.6 3.8 1.6 1.0 1982 10.4 10.5 1.5 0.7 Note: tlRM = change in reserve money. TAX inflation tax. REV revenue. Revenue data are unavailable for 1981 and 1982. Sources: International Monetary Fund, Government Financial Statistics, various issues. LARRY A. SJAASTAD AND OTHERS 155 Tables 6-16 and 6-17 indicate that all countries except Venezuela have relied heavily on the inflation tax, particularly Argentina (and also Chile during the first half of the 1970s). During 1975 and 1976, Argentine collections of the inflation tax accounted for more than 25 percent of its GDP (and half or more of total resources available to the government)! The tax has been important in both Brazil and Mexico since the early 1970s; again, Venezuela is the exception. In Argentina and Mexico, the inflation tax increased sharply in 1982 (unfortunately, revenue data are not available for 1981 and 1982 for those countries); in the other three countries, the change is modest. The data give some support to the idea that the debt crisis has forced governments to rely more heavily on the inflation tax as a source of finance, but we must be careful in evaluating this finding. Clearly the governments of Argentina, Brazil, and Chile were using the inflation tax as an important source of revenue long before the debt crisis. Indeed, it appears that in some countries (with the exception of Chile and possibly Venezuela), the low level of reliance on the inflation tax was a direct consequence of the ready availability of funds on the international capital market. That a drying up of those funds would cause the governments to revert to their old ways should surprise no one. Gross versus Net Debt and Implicit versus Actual Amortization In evaluating the implications of the enormous increase in Latin Amer ican external debt-particularly in the countries under consideration over the past decade, we must take into account the world inflation and the special circumstances of each of the countries involved. Dept service is always at the expense of domestic consumption (and investment) of tradables, even if the debt service is covered entirely or in part by new borrowing. Consequently, it is the price of tradables that is relevant for converting nominal external debt into real debt. Moreover, as most of that debt is denominated in dollars, it is the dollar price of tradables that is relevant for this purpose. The dollar prices of tradables have risen a great deal over the past decade for all countries but particularly for Mexico and Brazil. Part of the reason is the general world inflation, but the prominence of oil in the tradables of Mexico and Brazil has magnified the price increase for these countries. The two countries are in quite different circumstances, of course-Brazil is an oil importer, whereas Mexico is a major oil exporter. Mexico can take double advantage by exporting more, whereas Brazil can benefit only by contracting its imports. Nevertheless, by this means and others Brazil will, in the final analysis, service its external debt. 156 DEBT IN LATIN AMERICA When the nominal magnitudes of external debt are deflated by coun try-specific price indexes for tradables, the reduction is very great indeed. The results of the exercise are contained in table 6-18. The growth rate of foreign debt is greatly reduced, and service of it begins to appear much more feasible. Another factor, however, offsets the first. The world inflation, among other things, has resulted in much higher than normal interest rates; indeed, much of the interest payments are actually amor tization. Worse still, the debt is more and more owed to the private capital market (as opposed to international financial institutions), where interest rates are normally higher and maturities shorter. The conse quence, reported in table 6-19 below, is that debt service rates-interest plus amortization-have increased by an amount that more than offsets the inflation-induced decline in real external debt. The two effects are, of course, opposite sides of the same coin. The outcome is that external debt service is so enormous that none of the five countries can be expected to meet that service from trade sur pluses alone (with the possible exception of Venezuela and the even more remotely possible exception of Mexico). This issue is taken up in further detail in a later section, but it seems most unlikely that any of the countries in question can trade their way out of the debt problem. In the remainder of this section, we shall analyze each country's situation individually. For the most part, the results speak for themselves. In table 6-19, adjustments for interest and amortization rates are made; the column labeled "difference" is the increase in debt service owing to the rise in interest rates and shortening of maturities. The column labeled "current account adjustment" is the part of interest payments that is really amortization and that should be shifted from the current to the capital account of the balance of payments. In table 6-18, the debt and amortization figures are adjusted for inflation. The results of table 6-18 are self-explanatory. Brazil One of the key elements of Brazil's external debt problem is the evolution of its interest payments. The average interest rate on new commitments has risen from 7.2 percent in 1972 to 14.6 percent in 1981 (according to the World Bank's World Debt Tables). If we consider areal interest rate of 3 percent and assume that this is a reasonable estimate of the perceived real rate at the time when the major increase in indebted ness began (in the second half of the 1970s), we can calculate the interest payments that would have been made if this rate had been realized. The actual nominal interest rate increased as a result of world inflation, Table 6-18. Adjustments of Debt and Amortization t dollars) Total govern- Govern- Total Total Nominal Real ment ment nominal real borrowing' borrowing" nominal De real amorti amorti Year debt' jlator' debt zation zation Gross Net Gross Net Argentina 1976 3,150 212 1,486 774 351 1,883 1,279 887 536 1977 4,429 219 2,022 907 415 1,236 604 948 533 1978 5,033 197 2,555 1,976 999 3,330 1,715 664 335 1979 6,748 304 2,220 1,261 415 2,704 1,809 788 373 1980 8,557 330 2,593 1,730 524 2,776 1,630 785 261 1981 10,128 357 2,854 1,844 517 1,411 319 1,986 1,419 ' v. " 1976 23,080 228 10,101 3,547 Brazil 1,522 8,217 5,696 1,156 -396 1977 28,776 296 9,705 4,833 1,630 10,007 6,343 4,081 2,451 1978 35,119 289 12,156 7,265 2,515 16,539 11,347 4,478 1,963 1979 46,466 329 14,119 9,902 3,009 11,524 5,016 2.343 -667 1980 51,482 383 13,452 11.580 3,026 11,067 4,274 3,787 761 1981 55,756 392 14,213 13,196 3,364 14,948 7,857 6,110 2.746 Chile 1976 4,498 175 2,570 753 430 509 -134 53 -377 1977 4,364 199 2,193 942 473 1,120 291 704 231 1978 4,655 192 2,424 1,331 693 2,357 1,270 707 14 1979 5,925 243 2,438 1,722 709 2,940 1,623 770 61 1980 7,548 302 2,499 2,154 713 3,326 1,865 1,405 692 1981 9,413 295 3,191 2,939 996 4.945 3,148 2,770 1,774 (Table continues on the following pag Table 6-18 (continued) Total go vern- Govern- Total Total Nominal Real ment ment nominal real borrowing' borrowingd nominal De real amorti amorti Year debt' flator" debt zation zation Gross Net Gross Net Mexico 1976 11,580 199 5,819 1,892 951 5,503 4,350 1,581 630 1977 15,930 247 6,449 3,073 1,244 7,066 4,828 3,939 2,695 1978 20,758 227 9,144 5,603 2,468 9,265 4,857 439 -2,029 1979 25,615 360 7,115 9,199 2,555 10,739 3,672 134 - 2,421 1980 29,242 623 4,694 6,991 1,122 8,375 4,349 389 -733 1981 33,591 848 3,961 7,474 881 12,907 9,125 -569 1,450 ..... v, 00 Venezuela 1976 2,192 416 527 341 82 1,054 769 192 110 1977 2,%1 465 637 739 159 2,071 1,466 474 315 1978 4,427 465 952 616 132 2,822 2,466 274 142 1979 6,893 630 1,094 1,342 213 3,802 2,912 68 -145 1980 9,805 1,033 949 2,670 259 2,803 1,068 218 -41 1981 10,873 1,198 908 2,722 227 1,831 479 267 40 a. Figures for Argentina and Venezuela refer to public sector debt, whereas for the other countries they refer to total outstanding debt. Figur correspond to the beginning-of-year stock. b. The deflators correspond to the simple average of export and import unit values as they appear in the "world tables," World Bank (1980), un 1978. Afterward the deflators were updated with the available information about export and import unit values as they appear in the Internation Financial Statistics, IMF. c. From the first column and total (true) amortization. d. From the third and fifth columns. Sources: World Bank (1980); International Monetary Fund, International Financial Statistics, various issues. LARRY A. SJAASTAD AND OTHERS 159 however, and possibly also because of rises in the real rate. Therefore the explicit interest payments that result from the increase in the nominal interest alone, above the 3 percent level, should be regarded as implicit principal repayments that, had real interest rates stayed constant, would have been avoided. This implicit amortization amounted to more than $6 billion in 1981, six times the level of 1976 (see table 6-19). The above-mentioned component of interest payments actually repre sents a capital account item rather than a current account item. If the capital and current accounts are corrected by this factor, we find that Brazil's capital surplus and current account deficit is significantly re duced. This correction is, of course, most significant in the later years. Both accounts, for example, are cut in half, when corrected, in 1981 (see table 6-19). In assessing the external position of borrower countries, we must recognize that the stock of foreign debt in real terms has grown much less rapidly than in nominal terms. If the stock in nominal terms is deflated by an average of import and export unit values, the corresponding real debt service is an approximation of the sacrifice that the country has to make in terms of traded goods in order to enjoy the benefits of the borrowed capital. By using an index of this kind, we can see that, although nominal indebtedness in Brazil as of 1981 was more than ten times that of 1972, the real indebtedness has only slightly more than doubled over the same period (see table 6-18). On the other hand, during most of the years between 1976 and 1981, there was positive net new borrowing, and in all years the gross real new borrowing was positive (see table 6-18). The Brazilian foreign borrowing, in other words, not only covered interest payments on previous debt but also more than offset the negative effect that international inflation had on the real value of Brazil's outstanding debt. In addition to the amortization implicit in the increase in the nominal interest rates, principal repayments were also increased by a reduction of the loan maturities. It is reasonable to assume that long-term borrowing implies annual amortization of about 10 percent of the outstanding debt per year. The difference between the actual debt service and a hypotheti cal debt service, if we assume both a (real) interest rate of 3 percent and amortization of 10 percent of the stock of debt, indicates about how much extra was paid in contrast to the amount that would have been paid if the interest rate had not increased and the maturities had not shortened. In the Brazilian case, this calculation indicates that by 1981 the debt service actually paid was about twice the hypothetical figure (see table 6-19). The implication is that the combined effect of increasing nominal interest rates and shortening maturities was responsible for a doubling of Brazil's Table 6-19. Adjustments for Interest Rates and Amortization Periods (millions of current dollars) Hypothetical Debt service debt service Current account Foreign Inter- Amoni- Inter- Amorti- Differ- adjust- Year debt' est zation Total est b zation' Total ence ment Argentina 1976 3,150 264 604 869 95 315 410 459 170 1977 4,429 318 722 1,040 133 423 556 484 185 1978 5,033 503 1,615 2,117 151 503 654 1,463 352 ..... 0 1979 6,748 568 895 1,463 202 675 877 586 366 <:::> 1980 8,557 841 1,146 1,987 257 856 1,113 874 584 1981 10,187 1,058 1,092 2.150 306 1,019 1,325 825 752 1982 10,506 315 1,051 1,366 Brazil 1976 23,080 1,718 2,521 4,239 692 2,308 3,000 1,239 1,026 1977 28,776 2,032 3,664 5,696 863 2,878 3,741 1,955 1,169 1978 35,119 3,127 5,192 8,319 1.054 3,512 4,566 3,753 2,073 1979 46,466 4,750 6,508 11,258 1,394 4,647 6,041 5,217 3,394 1980 51,482 6,331 6,793 13,124 1,544 5,148 6,692 6,432 4,787 1981 55,756 7,778 7,091 14,869 1,673 5,576 7,249 7,620 6,105 1982 63,613 1,908 6,361 8,269 Chile 1976 4,498 245 643 888 135 450 585 303 110 1977 4,364 244 829 1,074 131 436 567 507 113 1978 4,655 384 1,087 1,471 140 466 606 865 244 1979 5,925 583 1,317 1,899 178 593 771 1,128 405 1980 7,548 919 1,461 2,380 226 755 981 1,399 693 1981 9,413 1,424 1,797 3,221 282 941 1,223 1,998 1,142 1982 12,561 377 1,256 1,633 Mexico 1976 11,580 1,086 1,153 2,239 342 1,158 1,505 734 739 1977 15,930 1,313 2,238 3,552 478 1,593 2,071 1,481 835 1978 20,758 1,818 4,408 6,226 623 2,076 2,699 3,527 1,195 1979 25,615 2,855 7,112 9,966 768 2,562 3,336 6,630 2,087 1980 29,242 3,842 4,026 7,868 877 2,924 3,801 4,607 2,965 1981 33,591 4,670 3,782 8,482 1,008 3,359 4,367 4,115 3,692 1982 42,716 1,281 4,272 5,553 " Venezuela 0 " 1976 2,192 122 285 407 66 219 285 122 56 1977 2,961 223 605 827 89 296 385 442 134 1978 4,427 393 356 750 133 443 576 174 260 1979 6,893 659 890 1,548 207 689 896 652 452 1980 9,805 1,229 1,735 2,964 294 981 1,275 1,689 935 1981 10,873 1,696 1,352 3,049 326 1,087 1,413 1,636 1,370 1982 11,352 341 1,135 1,476 Note: Dashes indicate that data are unavailable. a. Beginning-of-year stock of public sector debt for Argentina and Venezuela and beginning-of-year total debt for other countries. b. Assumes an interest rate of 3 percent text). c. Assumes an amortization rate of 10 percent of outstanding debt (see text). Sources: World Bank (1980); International Monetary Fund, International Financial Statistics, various issues. 162 DEBT IN LATIN AMERICA debt service. Therefore the trade surplus that must be generated in order to avoid default was doubled by factors that may be regarded as external to Brazil. Mexico High interest payments on external debt also played a dramatic role in the evolution of the Mexican foreign debt problem. Between 1972 and 1981, the average interest rate on new commitments increased from 6.9 percent to 15.1 percent, and the interest charged by private creditors went from 7 percent to 16.1 percent. The debt service due to higher nominal interest rates was increased further by the fact that the participa tion of private creditors on total loans to Mexico grew from 64 percent in 1972 to 88 percent in 1981. (A similar change also occurred in Brazil.) Assuming once again that 3 percent is a reasonable figure for the long-run real interest rate, the result is that, by 1980-81, the implicit amortization due to nominal interest increases accounted for about 10 percent of the total stock of debt outstanding, and in 1981 it was nearly ten times as large as it had been in 1974. If the capital account surplus and the current account deficit are adjusted by this implicit amortization, they decrease from $19.3 and $12.8 billion to $15.6 and $9.1 billion, respectively, in 1981 (see table 6-19). In Mexico the stock of foreign debt grew most rapidly between 1977 and 1981. This growth was matched, however, by the oil boom, which consisted of both a tremendous increase in the proven reserves and a doubling of the international price. When we assess the international debt position, we must consider two points: first, some of the debt was contracted only in order to anticipate higher consumption that would be made possible by the newly found wealth. Second, the stock of real debt was significantly reduced (with respect to the nominal stock) because of the oil price rise and oil's growing participation in the Mexican exports. Indeed, net real new borrowing may actually have been negative after 1978, although gross real borrowing was positive for most of the period (see table 6-18). As in the Brazilian case, borrowing was sufficient to cover at least all real total amortization. The change in the average maturity of the loans to Mexico also plays an important role in the debt problem. The average maturity from all creditors on new public debt commitments went from 13.7 years in 1972 to 7.8 years in 1981. Actual amortization was about twice what it would have been if the interest rate had been 3 percent and amortization 10 percent (see table 6-19). As in the Brazilian case, the increase in debt service implies that a LARRY A. SJAASTAD AND O1l1ERS 163 much larger trade surplus must be generated to achieve equilibrium in the balance of payments, if no further lending is forthcoming. The two cases show an important difference, however: the increase in the prices of Mexican traded goods comes from an increase in the price of its export ables, whereas the main increase in the prices of Brazilian traded goods comes from its imports-oil being probably the most significant price change. Therefore, with no more foreign borrowing available, the adjust ment will probably imply an increase in exports in the first case and a reduction of imports in the latter. Argentina Although the average interest rate on new public debt commitments has not risen as much as for the other countries, the rate of interest on Argentine borrowing was already higher than for the other countries in 1972. In addition, interest on loans from private creditors (and the share of these in total indebtedness) has increased from 8.6 percent in 1972 to 13 percent in 1981 (and that share increased from 69 percent in 1972 to 82 percent in 1981). Therefore, when the implicit amortization is calculated using an assumed 3 percent long-run interest rate, the amount proves to be significant, reaching nearly 10 percent of total debt in 1980-81. For the same period, when the current account is corrected by the implicit principal repayments, the deficit is substantially reduced (see table 6-19). The price index of traded goods in Argentina, in contrast with Brazil, Mexico, and Venezuela, has increased little. It grew significantly from 1978 to 1981 (about 80 percent) but declined sharply in 1981. Therefore, the stock of Argentina's debt in real terms has not been affected by inflation as much as the stock of debt in Brazil, Mexico, and Venezuela. In the second half of the 1970s, net real borrowing was negative only in 1978, whereas the gross equivalent was always positive, so that, in gen eral, foreign indebtedness was enough to cover the real (implicit and actual) amortization (see table 6-18). The maturities on new public debt commitments have lengthened somewhat during the period 1972-81 in Argentina, in contrast to the rest of the countries, but they have been quite variable. Amortization (actual and implicit) has increased mainly because of the effect of higher interest rates. Actual amortization has been below the hypothetical figure implied by the assumed 10 percent annual rate. The result is that the service of Argentine debt, relative to the hypothetical figure (assuming 3 percent interest and 10 percent amortization), amounted to about 10 percent in 1980-81, a lower percentage than in the countries discussed above (see table 6-19). 164 DEBT IN LATIN AMERICA The implication is that, although Argentina has also been hurt by high rates of interest on its debt, the corresponding maturity terms have been relatively favorable. The price of its exports, however, has not grown to match the growth in the debt service; should this situation continue, the required adjustment (in case no further indebtedness is made possible) would imply an enormous reduction in imports. Venezuela The total foreign indebtedness of Venezuela has not reached the absolute levels that we saw in the cases of Brazil and Mexico but is important in relation to its GNP-about 17 percent as of 1981. In addition, the participation of private creditors in the total debt is extremely high: 97 percent in 1981 (versus 71 percent in 1972). The effect of higher interest rates on the debt service payments is therefore especially important; indeed, the average interest rate on new commitments from private creditors nearly tripled between 1972 and 1981. The above-mentioned factors result in an implicit amortization (again, if we assume a 3 percent long-term interest rate) of about 13 percent of the outstanding stock of foreign debt in 1981. The Venezuelan capital and current accounts, unlike those of Brazil and Mexico, are not regularly in surplus and deficit, respectively; in fact, the (unadjusted) capital account shows net outflows in some years, and when it is adjusted by the implicit amortization, it is mainly negative. As in the Mexican case, the important foreign debt contracting was matched by an oil boom that began in 1974. Because of that boom, real debt has grown far less than its nominal counterpart; indeed, it may even be argued that real debt has stayed at a more or less constant level since 1972 (see table 6-18). The greatest increase in real foreign debt occurred in 1975-77, but it is mainly a result of the reduction in real terms during 1973-74. In the 1978-81 period, the gross real borrowing was about equal to (actual plus implicit) amortization (see tables 6-18 and 6-19). In addition to the implicit amortization coming from the increase in interest rates, the average maturity of the loans to Venezuela also de clined, as in the other countries. These two developments have made the service of the debt double with respect to the hypothetical service, which assumes a 3 percent interest rate and an amortization rate of 10 percent (see table 6-19). If Venezuela can borrow no additional funds, its pros pect of running a trade surplus that offsets the debt service may be brighter than in the cases of Brazil and Mexico. The reason is Venezuela's tremendous oil export potential and stock of debt, which in absolute terms is not as high as that of the other two countries. LARRY A. SJAASTAD AND OTHERS 165 Chile The implicit amortization due to high interest rates on Chile's external debt also provoked an important increase in its debt service. This in crease was fueled not only by a general interest rate rise but also by a dramatic change in the participation of private debt in the total stock, which went from 13 percent in 1972 to about 65 percent in 1981. Private debt comes from private sources, whose share in public debt also in creased significantly in the 1977-81 period. These factors help explain why in 1977 the implicit amortization was only about 5 percent of the total outstanding debt in 1978, whereas it reached 12 percent in 1981. This situation is also reflected in a reduction of the capital account surplus and the current account deficit of about 25 percent each in 1981 once both amounts have been adjusted for the implicit amortization in interest payments (see table 6-19). As with Argentina, the real stock of debt is far less affected by inflation of Chilean tradables than it is in the case of the other countries. The only negative real borrowing takes place in 1976 and is due to a reduction in nominal debt; most of the increase in external indebtedness took place in 1980-81 (see table 6-18). As with Brazil and Mexico, gross borrowing was more than enough to offset debt service. The average maturity from all creditors on loans to Chile declined from 13.5 years in 1972 to 10.7 in 1981. When the change of maturities is taken into account, the actual debt service was nearly twice as great as the hypothetical service, in contrast with the situation in Mexico, Brazil, and Venezuela (see table 6-19). Debt service in the case of Chile is a more serious problem than in the other four countries because Chile's foreign debt is nearly 85 percent of GDP-roughly double the relative debt of Mexico and Argentina. In addition, since January 1983 nearly all of the debt has become a liability of the government. The fiscal problem in Chile is therefore clearly a very intense one. Chile's ability to service its foreign debt plainly does not lie exclusively in trade; obviously more lending will be required. Post-Debt Crisis Adjustment The debt crisis burst on the world in August 1982 with the Mexican announcement that a suspension of payments was unavoidable. It quickly spread to other nations, particularly Latin American countries. The crisis was brought on by a number of factors, but the immediate cause was a 166 DEBT IN LATIN AMERICA growing reluctance on the part of the banking community to extend new credits, credits with which the debtor countries had been paying both interest and amortization. No immediate adjustment could be made to the sudden unavailability of funds in the international capital market. Rescheduling was the only possibility in the short run. Rescue packages were forthcoming from the Bank of International Settlements (BIS), the U.S. Treasury, and the International Monetary Fund. The World Bank has since instituted its structural adjustment loan program. None of these short-run measures, however, can substitute for adjustment at the coun try level. In this section, we examine the adjustments that have been made by Argentina, Brazil, Chile, Mexico, and Venezuela. In all cases, we find that a very considerable amount of progress has been made, but it has taken the form of accelerating the inflation rate (to collect more inflation tax), curtailing imports (at the expense of domestic industrial activity and employment), and appealing to the IMF for assistance. We conclude that these adjustments are insufficient if major lending does not continue; indeed, in most cases, no conceivable amount of domestic adjustment will replace the need for more borrowing. Unless some miracle reduces the ratio of debt service to debt, major defaults will be avoided only by writing down existing loans or by lending more funds. Argentina, 1981-83 During the last three years, Argentina has gone through a very com plex process of political change, including the inauguration of a new constitutional government last December. These developments have had an impact on its ability and willingness to impose the fiscal and monetary discipline required to adjust the economy to the current world capital market situation. The Argentine case is noteworthy in that most of the difficulty seems domestic in nature. As the country is nearly self-sufficient in oil, it was not seriously affected by the two oil shocks of the last decade. The need for capital inflows to foster economic development is far less acute than in other countries of the region (for example, Brazil); mere avoidance of recurrent politically motivated capital outflows would ensure sufficient development. At the end of 1980, the approaching change of command in the military government led to a major crisis from which the country has yet to emerge. Between 1978 and 1980, the country attracted foreign resources, channeled through the international banking system, in order to finance an ever-growing public sector. This capital inflow led to the sharp appre LARRY A. SJAASTAD AND OTHERS 167 ciation of the peso and growing trade deficits, as shown in table 6-20. The appreciation of the peso generated not only a change in domestic relative prices but also an increase in real interest rates for both the exporting and import-competing sectors of the economy, which resulted in widespread bankruptcy. The economic policy changes introduced since 1981 have been more closely related to these domestic developments than to the state of the world capital market. During the second and third quarters of 1981, the trade balance began to recover, and since the first quarter of 1982, it has consistently been in surplus. Such adjustment, however, has not been made on the basis of fiscal austerity. As table 6-21 shows, public expenditure has been rising since 1980. Although the new civilian government has announced its intention to reduce spending, significant measures are yet to be taken. The adjustment that has been realized to date essentially consists of a massive transfer of resources to the public sector, implemented by means of private credit rationing and a huge increase in inflation tax collections. The proportion of domestic credit directed to the government gives a clear picture of the increasing pressure of the public sector on the domes- Table 6-20. Trade Balance for Argentina, 1980-83 (millions of U.S. dollars) Year and Exports Imports quarter (f·o.b.) (eif) Balance 1980 1 2059.6 2282.2 -226.6 2 1927.3 2290.7 - 363.4 3 2035.5 2782.1 -746.6 4 2002.2 3190.4 -1188.2 1981 1 1989.9 2614.0 -624.1 2 2848.2 2622.0 226.2 3 2719.2 2196.0 523.2 4 1585.7 1999.0 -413.3 1982 1 2170.2 1484.0 686.2 2 2346.1 1333.0 1013.1 3 1622.6 1217.0 405.6 4 1483.7 1306.0 177.7 1983 1 1933.7 977.0 956.7 2 2106.9 1184.6 922.3 Note: LO.b. free on board. d.f. cost, insurance, freight. Sources: International Monetary Fund, International Financial Statistics, various issues. 168 DEBT IN LATIN AMERICA Table 6-2l. Basic Data for Argentina, 1980-83 Government Consumer credit I prices Public total credit Real (twelve- Year expenditure (percent) GDP month and (1982 (1980Q1 percentage quarter US$ billions) Stock Flow = 100) change) 1980 1 1.114 14.0 100.0 2 1.224 11.8 0.9 100.83 3 1.326 13.0 19.4 104.78 4 1.288 15.3 26.3 106.43 1981 1 1.1% 14.8 12.9 100.28 2 1.363 16.7 21.3 100.83 89.3 3 1.652 21.1 36.1 93.67 112.8 4 1.812 26.9 42.3 92.19 122.7 1982 1 1.090 29.0 38.6 92.38 147.2 2 1.171 28.4 25.9 90.54 129.7 3 1.343 24.3 17.5 91.00 156.4 4 1.808 25.4 28.9 92.29 202.9 1983 1 1.598 32.4 48.2 92.93 244.7 2 34.3 39.4 93.21 313.5 3 338.7 Note: Dashes indicate that data were not available at the time of writing. Sources: International Monetary Fund, International Financial Statistics and Government Financial Statistics, various issues. tic capital market. This development is dearly revealed in table 6-21, which shows the ratio of the stock credit to the government to total domestic credit, the ratio between the respective flows, and the time pattern of the inflation rate, measured by the CPI. As should be expected, this type of adjustment has had devastating effects on economic activity. Table 6-21 also shows the declining trend of the GDP since 1981. A direct consequence of this process has been the sharp fall of imports, which accounts for most of the observed adjustment in the trade balance. This adjustment is still well below the level necessi tated by the present debt service schedule. Owing to the military govern ment's political inability to carry out the IMF-sponsored austerity program after the South Atlantic conflict, and the unwillingness of the new civilian government to accept that program, the Argentine case is one of the most troublesome spots in the map of the debt crisis. Moreover, the sort of LARRY A. SJAASTAD AND OTHERS 169 adjustment realized so far (particularly the lack of fiscal austerity) will not be politically viable for much longer. Brazil, 1981-83 An understanding of the difficulties currently faced by Brazil in its current adjustment to the international capital market developments since mid-1982 requires a brief account of the initial conditions. Before the 1973 oil shock, Brazil constituted one of the major success stories in the developing world. The average rate of growth of real GoP was about 10 percent per annum and that of exports was 20 percent. The maximum figures for both variables since 1970 occurred in 1973: 14 percent for GOP and 55 percent for exports. At that time, the bright economic perfor mance, the small size of the foreign debt ($12.5 billion), and a high level of international liquidity made it possible temporarily to avoid the un avoidable: adjustment to a higher oil price. Despite stagnation in the world economy, Brazil was able to sustain an average GoP growth rate of 8.6 percent and an export growth rate of 24.9 percent during the 1973-77 period. In 1979 a new oil shock coupled with rising international real interest rates put an end to this process. After a $5 billion loss of international reserves between 1978 and 1980, a rather drastic revision of monetary policy was introduced in 1981 (see table 6-22). Interest rates were freed, and under the assumption of unrestricted (although expensive) access to the international capital mar ket, additional measures were taken with the explicit objective of increas ing domestic real interest rates in order "to stimulate demand for foreign resources (in the short run), and in the medium and long run to increase domestic savings" (Langoni 1981). At the expense of a nearly 10 percent Table 6-22. Basic Data for Brazil, 1980-83 Item 1980 1981 1982 1983 Exports (U3$ billions) 20,132 23,293 20,175 21,899 Imports (U3$ billions) 22,955 22,091 19,395 15,408 Oil imports (U3$ billions) 9,405 10,600 9,566 7,800 Trade balance (U3$ billions) -2,823 1,202 780 6,491 Inflation (percent) 86.4 100.0 97.9 172.9 Devaluation (percent) 95.8 286.2 Monetary growth (percent) 86.8 89.1 Note: Dashes indicate that data were not available. Sources: International Monetary Fund, International Financial Statistics and Government Financial Statistics, various issues. 170 DEBT IN LATIN AMERICA decline in industrial output, the $2.8 billion trade deficit was turned into $1.2 billion and $0.8 billion surpluses in 1981 and 1982, respectively. With this strategy of medium- and long-term reduction of the growth in foreign indebtedness Brazil confronted the disruption of the world capital market in 1982. Since the debt-restructuring agreement signed with the IMF on Febru ary 25,1983, a substantial deepening of the adjustment process has taken place. The cruzeiro was devalued by 30 percent at the end of that month and the accumulated devaluation for 1983 was 286 percent, compared with a 211 percent increase in the general price level. The implied "real" devaluation is nearly 30 percent. A $6.5 billion trade surplus was achieved in 1983, although the monthly trade balance data indicate a decay in that balance since midyear. Given the present conditions in the international capital market, the adjustment still seems to be below the required level in magnitude; consequently a $9 billion trade surplus has been targeted for 1984. During the first eleven months of 1983, public expenditure fell by 5.5 percent, and revenues increased by 1.9 percent. Accordingly, the mone tary base grew by only 98 percent (implying a decline, in real terms, of the monetary base by 40 percent), as far less revenue was collected by money creation. According to Conjuntura economica (Funda~ao Getulio Vargas 1984), the nominal deficit target on which Brazil and the IMF had agreed for 1983 had been met. An important aspect of the adjustment in Brazil was the decline in imports, which accounted for 61 percent of the trade balance improve ment. Much of the decline was in crude petroleum, the imports of which fell by 18.5 percent. Associated with this improvement on the trade front was a decline in industrial employment of 3.7 percent from January to August of 1983. As that sector had been seriously depressed since early 1981, this development cast some doubts on the political viability of the current adjustment process in Brazil. Chile, 1981--83 The Chilean economic adjustment was quite different from that in the other Latin American economies considered here. The reason lies in the roots of the problem, which are peculiar to Chile. One key difference is that the crisis in Chile was not associated with a huge government deficit; indeed, the government ran a surplus in 1980--81 and only a small deficit in 1982. Nevertheless, the current account deficit as a proportion of GDP reached 14 percent in 1981 and 10 percent in 1982. LARRY A. SJAASTAD AND OTHERS 171 The current account and trade deficits of 1980 and 1981 (table 6-23) can be explained (aside from the rise in interest rates and the fall in the price of copper) by an excess of domestic private absorption over income. Indeed, if we look at the financial flows, we see that only about half of the growth in credit to the private sector in these two years came from an increase in quasi-money, whereas the remainder matched the rise in the net foreign liabilities of the financial system. During 1982 the current account deficit was cut in half and the trade balance became positive, and the tendency continued in 1983. At least two factors help explain the 1982 turning point in the trade balance. First, the devaluation in June 1982, when the peg to the U.S. dollar was abandoned and the basket peg began. Second, the severe recession of 1981-83, which can be linked to several factors, including the incredibly high levels of the real interest rates since early 1981. In a related development, the unemployment rate-one of the major concerns of the adjustment program-reached a peak of 25 percent (in Santiago) by mid-1982. Although it may be thought that the government deficit of 1982 (the first since 1978) responded to the recession, this deficit clearly reflected the reduction in tax collections and in copper revenues. In fact, total government expenditure (measured in U.S. dollars) was the same as in 1981. On the other hand, in 1982 there was an important increase of credit to the private sector from both the central and the commercial banks. It is not clear how this increase in credit was financed, as the flow of quasi-money was about equal to that of the previous year. "Net foreign" and "other net" liabilities show the largest increases of the financial system liabilities. An obvious conclusion from the discussion thus far is that the Chilean financial problems arose more from excessive expenditure in the private sector than from fiscal imbalances. The low rate of internal saving made necessary the use of external saving, which is reflected in the current and trade accounts of 1980-82. As external debt service also rose because of general shortening of maturities and higher interest rates, larger domestic savings will be needed to generate a trade balance high enough to cover future debt service and imports. From this point of view, the Chilean adjustment problem is no different from that of other Latin American countries. During 1983, the adjustment of the economy in Chile seems to be well under way, even though it is occurring at a high cost in terms of domestic welfare. Until October 1983, a trade surplus of nearly $1 billion had been achieved, mainly at the expense of imports (which fell 26 percent on top of the 44 percent decline in 1982). The flow of credit from the domestic Table 6-23. Selected Economic Indicators, Chile, 1980-83 Item 1980 1981 1982 1983 Range' " ;:j Trade balance (US$ billion) 764 -2.677 63 986 (to Oct.) Percent change In exports 22.7 -18.5 -3.4 0.8 (to Oct.) In imports 30.5 19.1 -44.1 -26.1 (to Oct.) Current account (US$ billion) 1.971 -4.733 -2.304 Capital account (US$ billion) 3.165 4.698 1.215 Change, int!. reserves (US$ billion) 1.244 0.067 1.165 Percentage change In monetary base 37.9 -8.7 -25.5 -5.5 (to Sept.) In money, M1 77.1 -8.6 -7.9 9.6 (to Sept.) Fiscal def./GDP (percent) 5.5 1.1 2.7 Currency account deL/GDP (percent) 7.1 14.1 7.9 Copper price (cents /pound) 99.2 79.0 67.1 73.9 (to Nov.) Inflation (CPI) Average 35.1 19.7 9.9 27.3 December-December 31.2 9.5 20.7 23.1 Exchange rate (pesos! dollar) 39.0 39.0 50.9 78.8 Percentage change in GDP 7.8 5.7 14.3 -7.0 (to June) Unemployment rate (Santiago, Oct.-Dec.) 10.1 11.0 21.9 17.7 (to Aug.-Oct.) Real interest rate b 5.4 29.1 23.9 3.7 (to Flow of financial system credit To public sector (billion pesos) 5.1 -36.0 84.8 -18.1 (to Sept.) To private sector (billion pesos) 207.2 204.3 354.2 -45.9 (to Sept.) Change in net foreign liabilities, excluding intI. reserves (US$ billion) 79.7 89.8 158.4 -105.5 (to Sept.) Central bank intI. reserves (US$ billion)' 4.074 3.775 2.578 1.998 (to Nov.) Financial system flow of quasi-money (billion pesos) 120.8 98.8 94.0 -16.9 (to Sept.) Other net liabilities (billion pesos) -34.9 -29.1 163.7 97.2 (to Sept.) Note: Dashes appear in the column for 1983 where the range extended from December to December. Blank cells in the last column indicate th ' data were not available. t:;:l a. 1983 only. b. Short-term deposit rates (monthly) corrected by CPl inflation and annualized. c. Excludes the use of credit from the IMF. Source: Banco Central de Chile, Boletin mensual, November and December 1983. 174 DEBT IN LATIN AMERICA financial system to both private and public sectors was negative in 1983. As of November 1983, international reserves (excluding IMF credit, which was reactivated in August) stood at $1.998 billion, or nearly $600 million below the level of December 1982; however, they had been increasing since April 1983. An external debt renegotiation agreement reached in July 1983 in cluded a new loan for $1.3 billion, a medium- and short-term debt rescheduling of $3.4 billion, and a short-term rollover (to December 1984) of $1.8 billion (Brau, Williams, and others 1983). If we assume that the current account deficit was in fact cut by one-half in 1983, relative to 1982, this agreement provides Chile-at least in the short run-with sufficient external liquidity . As in other cases in Latin America, however, the question remains as to how long the low level of consumption can be sustained. The need to generate large trade surpluses in the next three years is evident from the terms of the renegotiation agreement. Finally, important external obligations must be met during 1987-88 (owing to the new loan and the medium- and short-term debt reschedulement). Mexico, 1982-83 The financial crisis that had been building up since 1980 culminated in 1982 (see table 6-24). The balance-of-payments situation in 1981 gives a clear picture of the Mexican financial crisis at that time. The current account deficit reached a record of $14 billion, with the "errors and omissions" deficit at a record $10 billion. These two outflows were nearly totally offset by an extraordinary inflow of capital, so that there was but a minor decline of international reserves. The enormous deficit in the current account has been attributed to a variety of causes, including the high government deficit, the fall of the international oil price, and the increase in debt service payments. On the other hand, the deficit in the "errors and omissions" account reflects private capital outflows due to the perceived increase in exchange and political risk. The exchange risk is also evident from the increase in savings channeled through dollar denominated assets in Mexican banks. Although the nominal exchange rate was "sliding" down (slowly depreciating) in a controlled fashion, the perspectives concerning the balance-of-payments situation, and the "confidence crisis," made the devaluation rate appear insufficient, so that capital flight went on. In 1982, the major devaluation of February (from twenty-seven to forty-five pesos per dollar) marked the beginning of a series of confusing and contradictory policy developments, which resulted in an acceleration of the capital flight, despite the devaluation. There was, however, an LARRY A. SJAASTAD AND OTHERS 175 Table 6-24. Mexico: Selected Economic Indicators Item 1980 1981 1982 1983 Range Imports, Lo.b. (US$ billion) 18.896 24.037 14.489 6.485 (to Oct.) Trade balance (US$ billion) -2.830 -4.099 6.885 10.766 (to Oct.) Service account (US$ billion) 5.607 10.089 -10.110 -7.645 (to Oct.)' Errors and omissions (US$ billion) -3.933 8.840 6.157 0.289 (to Sept.) Change, intI. reserves (US$ billion) -0.749 -1.106 3.011 2.556 (to Sept.) Exchange rate (pesos! dollar) Preferential 22.95 24.52 54.99 119.80 (to Dec.) Free 22.95 24.52 61.52 150.79 (to Dec.) Percentage change b In monetary base 40.5 44.8 98.0 45.5 (to Nov.) In central bank claims on govt. 33.1 47.2 137.5 77.3 (to June) In claims on govt.!mon. base 85 86 103 101 (to June) In money, Ml 32.2 33.1 64.9 29.0 (to Nov.) In quasi-money 39.2 59.0 73.5 67.6 (to June) In industrial production 9.8 8.8 1.9 -7.0 (to June) In real GDP 8.3 8.0 -0.2 Inflation (CPI) 26.4 27.9 59.0 101.9 (to Dec.) Fiscal def.!GDP (percent)' 7.7 14.8 18.6 Current account def.! GDJ>" 4.0 5.9 1.7 Short-term interest rated 27.25 32.75 59.50 54.70 (to Dec.) Note: Dashes in the column for 1983 indicate that data were not available. Dashes in the last column indicate full range (January to December). a. Financial service only. b. Annual relative change at the end of the period indicated. c. From a speech by IMF Western Hemisphere director, reproduced in Banco Central de Chile, Boletin mensual, November 1983. d. End of period rate on three-month certificates of deposit. Source: International Monetary Fund, International Financial Statistics, and Banco de Mexico, Indicadores econ6micos, various issues. important effect on the trade balance, which turned positive in the second quarter of 1982-mainly at the expense of reduced imports, which in turn derived from the imminent recession. Many private firms had difficulty in servicing their foreign debt, and the scarcity of foreign exchange con strained the purchase of imported intermediate goods. This situation resulted in a decline of industrial production beginning in the third quarter of 1982. In August 1982, a two-tier exchange system was imposed with the intention that the capital account (or the speculative movements) would take place at a freely determined rate and the service account and part of the trade account (imports) at the lower rate of about 48 pesos per dollar. As it was recognized that this measure might accelerate capital flight, it was also decided to make dollar-denominated assets in Mexican banks 176 DEBT IN LATIN AMERICA payable only in pesos (at a rate of 69.5 pesos while the free market was fluctuating between 100 and 120 pesos per dollar). In August 1982, Mexico also declared itself temporarily unable to service its foreign debt and obtained a postponement of principal repay ments. At the same time, negotiations for a restructuring of foreign debt maturities began, and Mexico engaged in talks with the IMF. On Septem ber 1, 1982, the nationalization of commercial banks was decreed and controls on the foreign exchange market were imposed, with adverse effects on the already low level of private production and investment; property rights had become unclear. In addition, the private banks owned a large portion of the Mexican industrial sector shares, and what was going to be done with them has never been explained. In 1982, the public sector deficit rose to 18.6 percent of GDP, and deficits in the service and "errors and omissions" accounts amounted to more than $10 billion and $6 billion, respectively. The trade account surplus reached nearly $7 billion, but growth of real GDP was negative ( - 0.2 percent), and the inflation rate (average CPI) was 60 percent. In Decem ber, just after the new administration took office, foreign exchange operations were liberalized, with a controlled rate of 150 pesos per dollar and a preferential rate of about 96 pesos (which would eventually catch up to the higher rate) for debt service payments and some imports. The main objectives of the adjustment program undertaken by the new administration, in agreement with the IMF, were a reduction of the public sector deficit in 1983 to 8.5 percent of GDP and the reduction of inflation and of the current account deficit. The latter target implies lower imports, higher oil exports, and a debt rescheduling. The assumption was that GDP growth, in real terms, would be zero. Available data suggest that the main objectives of the program are being met (see Buira 1983; International Monetary Fund 1984). There was a trade balance surplus of $10.8 billion until October of 1983 (mainly a result of lower imports, whose value declined by 40 percent with respect to 1982). An agreement concerning public sector debt amortization was reached in order to postpone $20 billion of repayments due from August 1982 to December 1984. There was also new financing from U.S. syndi cated sources in the amount of $5 billion and from official sources of $2 billion to $2.5 billion (Brau, Williams, and others 1983). The growth rate of tbe nominal monetary base and of the money stock have been declining (to about 46 percent and 29 percent, respectively, by November 1983), suggesting in view of the close relationship between monetary and fiscal policy that exists in Mexico, that the growth in the public sector deficit has been slowing down. Inflation has also declined, although it was still about 100 percent on average in 1983. There is also LARRY A. SJAASTAD AND OTHERS 177 some evidence of a very important decline in production, indicating that real GDP probably fell significantly in 1983. Indeed, industrial production was about 10 percent lower in the first semester of 1983 than it had been a year earlier. Increasing unemployment is also reflected in the coverage of workers by the Social Security Institute. The adjustment policy has a number of drawbacks, including the serious recession and the postponement of debt repayments. The restruc turing of debt mentioned above implies that obligations for the $20 billion will have to be met starting in 1987 and for the $5 billion starting in 1986. Very high debt service payments will be due in those years unless other arrangements are made. Other dangers are present, such as private sector distrust of economic policy and the possibility that the government will be unable to reach agreements with the labor unions. In short, the adjust ment program requires an important reduction in consumption that may or may not be feasible, in view of the rates of growth experienced in recent decades. Venezuela since 1982 The adjustment facing the Venezuelan economy is rather different from that in other debtor countries. Venezuela did not experience sig nificant external imbalances until 1982, and inflation and monetary ex pansion have not been as high as in other cases in Latin America. In fact, the trade account registered a surplus (albeit declining) until 1982 (see table 6-25), and it does not seem likely to have turned into a deficit in 1983. In the first quarter of 1983, foreign assets of the central bank stood at $8.6 billion, whereas total external debt was on the order of $34--35 billion. Nevertheless, in February of that year, it was decided to abandon the convertibility of the bolivar (which had been fixed at the rate of 4.3 Bs/dollar for years) in order to impose a three-tier exchange rate system. This system consisted of a low rate of 4.3 Bs/dollar for public and some private debt service payments and some "essential" imports, another rate of 6 Bs/dollar for "necessary" imports, and a free rate for all other foreign exchange transactions. Shortly thereafter, the finance minister began negotiations to convert short-term debt into medium-term debt. At the same time, the government announced its economic adjustment plan, which was at variance with many of the recommendations made by the IMF. The plan's main features were: · Indefinite maintenance of the three-tiered exchange rate system and stabilization of foreign exchange outflows to sustain central bank reserves at $8-9 billion 178 DEBT IN LATIN AMERICA Table 6-25. Venezuela: Selected Economic Indicators, 1980-83 Item 1980 1981 1982 1983 Range Trade account (US$ billion) 8.714 7.840 3.199 Imports, Lo.b. (Bs billion) 45.375 50.682 50.056 8.887 (to Mar.) Exports (Bs billion) 82.507 86.388 70.583 51.941 (to Sept.) Percentage change In imports 8.7 11.5 8.6 In exports 34.6 4.8 18.0 Current account (US$ billion) 4.728 4.000 - 3.455 Capital account (US$ billion) 0.164 1.882 -2.182 Short-term capital (US$ billion) -1.896 -2.692 -4.567 Errors and omissions (US$ billion) -1.129 -2.139 -2.526 Change, intI. reserves (US$ billion) 3.823 -0.012 8.215 Percentage change' In monetary base 6.6 17.1 17.5 66.3 (to Sept.) In money, Ml 18.3 9.5 5.6 26.5 (to Sept.) In central bank claims on govt. -21.3 83.4 79.6 15.2 (to Sept.) In claims on govt.lmon. base 8.0 12.4 19.0 22.0 (to Sept.) In GDP -1.7 0.4 0.6 Unit value of oil exports 100 116.1 116.1 98.1 Inflation (ePl) 21.5 16.0 9.6 5.7 (to Sept.) Exchange rate 4.3 4.3 4.3 4.3 Preferential 6.0 Free 8.0-9.0 (to Sept.) Flow of credit to govt. from central bank (Bs billion) -0.65 1.736 3.039 Jan.-Sept. 3.558 1.641 Note: Dashes for 1983 indicate that data were not available. Blank cells in the last column indicate full ranges (January to December). a. Annual relative change at the end of the period indicated. Source: International Monetary Fund, International Financial Statistics, various issues. · Ultimate unification of the exchange rates at a level of less than 7 Bs/dollar · Budget reductions (not specified) not including reductions in wages in state enterprises or in the public sector payrolls · Price controls to remain after the (decreed) general price freeze of sixty days, which began with the devaluation · Reduction of 1983 imports by $4 billion (25 percent), mainly by prohibiting the entry of many luxury products · Inflation at about 15 percent in 1983 · Negotiations to refinance about $10 billion in foreign debt due in 1983 in order to convert it into five- to seven-year loans. LARRY A. SJAASTAD AND OTHERS 179 In March 1983 Venezuela declared a deferral on amortization of public external debt (which was subsequently extended to October 1983); the short-term debt involved was about $11 billion. Nevertheless, interest payments were not going to be suspended. Total public sector debt at that time stood at $27 billion, whereas private sector debt was on the order of $7-8 billion. Private sector debt payments would be eligible for the preferential exchange rate (4.3 Bs/dollar) only if principal payments were to take place over a three-year period starting in January 1984. A key difference between the Venezuelan crisis and others is that the central bank was holding a significant amount of reserves at the time the adjustment program started and negotiations for debt rescheduling be gan. Indeed, in March 1983, reserves at the central bank were more than 25 percent of total external debt and more than 30 percent of public sector foreign debt. Although Venezuela had to meet obligations of about $10 billion during 1983, private capital flight may well have been the factor that finally pushed the government to impose exchange restrictions and to announce the adjustment plan. In 1982, the short-term capital and errors and omissions accounts showed outflows of $4.6 and $2.5 billion, respec tively, whereas the overall balance of payments resulted in a fall of $8.2 billion of international reserves. In addition, there is evidence of a further fall of more than $1 billion in January and February of 1983. The Vene zuelan authorities thus seem to have acted in February 1983 in anticipa tion of a liquidity crisis such as had occurred in Brazil and Mexico by implementing an economic policy package before they were forced to do so under IMF conditions. As of April 1983, capital flight had ceased and international reserves were above their February levels. There is evidence of an increase in the trade surplus, mainly at the expense of lower imports, which was used to finance public sector debt service. In the foreign exchange market, the preferential rates remain at 4.3 and 6 Bs/dollar, and the free rate fluctu ated at about 8-9 Bs/dollar. In July 1983, some government officials were reportedly projecting GDP reductions of about 2.2 percent and 2 percent during 1983 and 1984, respectively, together with significant declines in domestic gross invest ment. The balance-of-payments deficit is believed to have reached $4 billion in 1983, despite the contraction of imports. In addition, inflation was being projected at 20 percent and 30 percent for 1983 and 1984 and international reserves fell to $7.5 billion at the end of 1983. By September 1983, it was evident that Venezuela would not resched ule its foreign debt until the new administration took office the following January. In the same month, creditor banks reportedly stated that they would not consider rescheduling Venezuela's debt until all public sector 180 DEBT IN LATIN AMERICA interest payments had become current.} Despite the intentions embodied in the plan, the private sector had not, as of September 1983, been able to obtain dollars at preferential rates to meet its debt service obligations. If this situation continues, the total amount of Venezuelan debt in need of rescheduling in 1983-84 could be on the order of $23 billion. Trading the Way Out Under the world's normal capital market conditions, the capacity of the countries discussed above to service their foreign debts is seldom questioned. Except for Chile, where the foreign debt-GDP ratio is about 85 percent, all of them present rather "normal" levels of foreign in debtedness by historical standards. If real interest rates and maturity terms were also in accordance with historical precedents, we would certainly not speak of a "debt problem" today. The problem, of course, is that, under the present capital market conditions, these countries are being asked to pay nominal interest rates in excess of 10 percent (real rates being even higher), and amortization rates in excess of 20 percent. For the worst case-that of Chile-this requirement implies debt service payments close to 25 percent of GDP, and for the group in general it is a burden very difficult or impossible to sustain without serious domestic political distress. In the recent past, all of the countries in question have demonstrated a rather impressive ability to generate substantial trade surpluses, but these surpluses are still below the levels required by the present world capital market. The adjustment realized so far has been facilitated by the huge U.S. trade deficit but has also been harmed to some extent by protection ist policies, particularly those of the European Economic Community. Unfortunately, many domestic policies implemented by these coun tries do not contribute to facilitating the always difficult process of adjustment. There is no doubt that adjustment in this context means increased savings. A very popular economic policy instrument used to generate trade surpluses, however, has been to restrict imports. This measure has had the adverse effect of precipitating a reduction in indus trial activity and has also increased unemployment rates; on the other hand, exports have not increased and have even fallen below previous levels in some cases. In sum, the adjustment has mainly consisted of sharp reductions in imports. Perhaps the most important policy mistake has been the closing of domestic capital markets, thereby insulating residents from the strong incentives to save that were afforded by the world capital market's efforts to finance the U.S. budget deficit. Clearly there is a way LARRY A. SJAASTAD AND OTHERS 181 out of the debt problem: the banks can roll over the present debt in order to bring the maturity profile into a feasible configuration. On the trade side of the problem, developed countries, in particular those of the European Economic Community, will have to ease their protectionist policies. Finally, the countries that are in trouble will have to avoid insulating their residents from the incentives to save. The nor malization of the world capital market is also imperative, not only to make feasible the service of present debts, but also to make available in the future the resources that will be needed to finance the development of the region. Notes 1. Information about these governments does not, however, give an adequate picture of the financial status of the whole public sector, which includes various parastatals not covered by the available statistics. 2. "Billion" means "thousand million." 3. By mid-September the government was expected to request that the standstill on interest payments be extended for 120-180 days; it was due to expire at the end of the month. Arrears of the public sector stood at about $85 million, but private sector arrears exceeded $400 million. References Brau, E., R. C. Williams, and others. 1983. Recent multilateral debt restructurings with official and bank creditors. Occasional paper 25. Washington, D.C.: International Monetary Fund (IMF). Buira, Ariel. 1983. The exchange crisis and the adjustment program in Mexico. In J. Williamson, ed. Prospects for adjustment in Argentina, Brazil, and Mexico. Washington, D.C.: Institute for International Economics. Funda<;ao Getulia Vargas. 1984. Conjuntura economica (January). Gil-Diaz, Jose. 1983. DEI ajuste a la defiaci6n: La politica economica entre 1977 y 1981 (Chile). Washington, D.C.: International Monetary Fund; processed. International Monetary Fund (IMF). 1984. International Financial Statistics (January). Langoni, Carlos G. 1981. The strategy offoreign sector adjustment. Seminar on the outlook of the world economy. Brasilia: Banco do Brazil. World Bank. 1980. World Tables. 2d ed. Washington, D.C. Chapter 7 Government Deficits, the Real Interest Rate, and Developing Country Debt: On Global Crowding Out Deepak Lal and Sweder van Wijnbergen THE TWO MAJOR PROBLEMS in the global economy clouding the future prospects of developing countries are those of rising protectionism in industrial countries and the debt crisis. Many observers hope that the current recovery will be sustained and will resolve these problems. Against this cyclical view of current problems, the present chapter ex plores other explanations suggesting that these problems will not dis appear with the cyclical upturn and will require more fundamental changes in policies in both developed and developing countries. Development economics has traditionally emphasized the trade link age between developed and developing countries, namely that growth in developed countries is a major determinant of exports and therefore of the growth of developing countries. This so-called engine-of-growth view of the role of trade in development (as propounded for instance by Lewis 1980) has been empirically questioned by Kravis (1970) for the nineteenth century and by Riedel (1984) for the post-World War II development experience. The far-reaching changes in the composition of developing country exports (toward manufactures and away from pri mary products) in the post war decades have meant that domestic supply rather than foreign demand factors have been the major determinant of developing country export performance and thus of their growth (see Lal 1983, chap. 2). This statement does not mean that the rising protection ism of developed countries (triggered in part by their poor economic performance in the 1970s) may not pose a serious future threat to the growth of developing countries' exports and income. These risks and their consequences, however, have been discussed relatively frequently. In contrast, another important link between developed and developing countries is emerging and is the focus of this chapter. Although the trade link may have become weaker, the integration of world capital markets and the explosion in commercial bank lending in the 1970s, stimulated by the recycling of financial surpluses in the Orga 182 DEEPAK LAL AND SWEDER VAN WIJNBERGEN 183 nization of Petroleum Exporting Countries (OPEC) that arose from the two oil price shocks of the past decade, have provided another important and growing link between economic prosperity and growth in developed and developing countries. This chapter is concerned with charting and analyzing the implications of certain trends in global savings and invest ment balances and with determining how these are affected by public policies, in particular fiscal policies, in both developed and developing countries. The two problems of protectionism and debt that currently plague the global environment and thence the prospects for developing countries, are in turn linked to certain longer-term trends and "structural" weak nesses of developed and developing countries that have been exposed by the supply shocks of the 1970s. These are the possibility of an emerging global shortage of savings; a fiscal problem of rising structural public sector deficits in both developed and developing countries associated with certain structural features: the aging of the population in developed countries and its "greening" in developing countries; the real wage resistance of workers in many industrial countries; and the explosive growth in social expenditures in most countries of the Organisation for Economic Co-operation and Development (OECO).1 In the first part of the chapter we chart these trends and structural weaknesses. In the second through the fourth parts we outline their interrelationships. In the fifth part we present a simple three-region model of global saving-investment balance, which is calibrated with data for the 1970s, and in the final part we summarize our postulated global interactions. Trends in OECD Real Wages, Social Expenditures, and Fiscal Deficits The deteriorating economic performance of the global economy and the industrialized countries in particular is well documented (see OECO 1983). Two structural aspects of industrialized economies were exposed by the supply shocks of the 1970s. The first was the real wage growth that workers in industrial countries had come to expect as a right and that they were willing to enforce at the cost of a declining share of profits in the late 1960s. The second was the commitments that most OECO governments had increasingly made to various groups to serve their notions of social justice. Both features influenced public policy in the postwar decades. Accord ing to a common belief, except for minor recessions that could be 184 DEFICITS, INTEREST, AND DEBT smoothed by suitable demand management policies, the postwar boom would be unending. Its continuation would allow both increased real wages and social expenditures to be financed without the need for dif ficult choices concerning tradeoffs between wages and profits or between consumption and investment (see Crosland 1950 for the classic statement of this view). Increasingly since the 1960s, microeconomic interventions (industrial subsidies, regional subsidies, and in the 1970s various forms of protectionism) to maintain workers in particular occupations and loca tions at income levels above the value of their marginal product were justified as a legitimate tradeoff between "economic security" and eco nomic growth. (For a rough quantification of these microeconomic distortions for Europe, see Curzon-Price 1982.) The supply shocks of the 1970s and the universal slowing down of productivity growth in the OECD countries exposed the unreality of these assumptions and the unviability of the policies they had engendered. Most obvious and best analyzed were the discrepancies between the real wage that workers expected and the lowering of the "full employment wage" caused by both the terms-of-trade losses suffered by the OECD countries in the 1970s and by their worsened growth prospects stemming from the productivity slowdown. Sachs (1979, 1983) has shown (see tables 7-1 and 7-2) how slowly the real wage in different industrial economies and particularly in Europe adjusted downward toward the new "equilibrium" real wage. The United States was an exception, as it succeeded in creating 20 million extra jobs in the 1970s, whereas in Western Europe employment changed by only 2 million. With the maturation of the baby boom generation, the labor supply increased, producing levels of unemployment unprecedented since the 1930s. The distortion in relative factor prices resulting from the relative rigidity of real wages provided producers with an incentive to substitute capital for labor (see Scott and Laslett 1978) while the accompanying squeeze on profits attenuated the means to finance investment to create future em ployment and growth. The rise in commodity prices and the associated structural changes in the economy, moreover, led to a reduction in the "effective" capital stock as the expected return on capital employed in oil-intensive and other raw material-intensive industries declined (see Baily 1981 and Bruno 1982, 1984).2 Actual profits in manufacturing also declined (see table 7-3). In these circumstances, conventional demand management policies cannot avoid the transitional unemployment that would in any case be caused by a higher real wage adjusting only slowly toward that "war ranted" by the changed circumstances. The conventional method of reducing such transitional unemployment is premised on the assumption DEEPAK LAL AND SWEDER VAN WIJNBERGEN 185 Table 7-1. Rates of Growth in the Product Wage and in Labor Productivity for the Manufacturing Sector and the Aggregate Economy, by Country, Selected Periods, 1962-78 (annual average in percent) Sector, measure, Ger- United United and period Canada France many Italy Japan Kingdom States Manufacturing Product wage 1962--69 5.0 4.8 5.6 7.4 10.8 4.6 3.4 1969-73 3.6 7.0 7.5 9.7 12.6 5.6 3.6 1973-75 0.1 6.2 6.8 4.4 0.6 4.6 0.1 1975-78 n.a. 4.8 504' 2.0 8.9 1.4 3.0 Labor productivity 1962-69 4.5 6.3 5.9 6.8 11.2 4.5 3.1 1969-73 404 5.4 4.8 6.9 8.7 4.1 3.2 1973-75 -004 2.8 5.2 004 -1.8 -1.3 -0.3 1975-78 4.5 6.1 5.0 4.1 7.3 1.2 3.0 Aggregate economy Product wage 1962--69 3.6 5.1 5.0 7.8 8.5 3.2 3.1 1969-73 2.0 5.5 6.3 7.9 12.2 3.7 2.6 1973-75 1.5 5.1 4.8 6.0 8.6 4.9 0.2 1975-78 1.8 5.2 2.7 1.2 2.7 1.5 2.3 Labor productivity 1962--69 3.3 5.2 5.3 7.4 9.9 3.1 2.7 1969-73 3.2 5.7 5.2 6.6 9.1 3.9 2.6 1973-75 0.7 2.6 4.0 3.0 3.9 0.7 0.3 1975-78 2.0 5.0 4.5 1.3 4.1 2.0 2.1 Source: Sachs (1979). Table 7-2. Annual Percentage Changes in Real Hourly Compensation for Manufacturing, Selected Periods, 1960-82 Country 1960-73 1973-79 1979-81 1980 1981 1982 Canada 2.8 2.5 -0.6 1.0 -1.1 0.8 France 5.3 4.3 1.8 2.8 0.8 3.0 Germany 6.4 5.3 2.2 2.9 1.5 0.3 Japan 8.2 2.3 0.5 -1.5 2.5 2.1 United Kingdom 3.7 3.7 3.8 4.9 2.7 204 Unites States 1.8 0.9 -0.9 -1.6 0.1 0.2 Source: Sachs (1983). 186 DEFICITS, INTEREST, AND DEBT Table 7-3. Real Rates of Return on Corporate Capital, by Country, 1962-76 (percent) Year or United United period Canada France Germany Italy Japan Kingdom States 1969 9.8 11.5 20.7 12.2 30.6 10.1 10.2 1970 6.6 11.5 18.2 11.5 28.1 8.7 8.1 1971 8.4 10.6 15.6 9.4 23.5 8.7 8.4 1972 9.4 12.3 13.4 9.9 20.2 8.6 9.2 1973 11.4 11.8 12.9 n.a. 15.7 7.2 8.6 1974 11.4 10.8 12.7 n.a. 13.1 4.0 6.4 1975 8.2 5.2 10.0 n.a. 13.9 3.4 6.9 1976 8.1 n.a. 11.4 n.a. n.a. 3.6 7.9 Average 1962-64 7.9 9.7 19.3 10.4 28.2 11.9 12.0 1965-69 9.6 10.0 19.5 11.4 27.9 10.6 12.2 1970-73 9.0 11.6 15.0 10.3 21.9 8.3 8.6 1974-76 9.2 8.0 11.4 n.a. 13.5 3.7 7.1 n.a. = not available. Source: Sachs (1979). that workers suffer from some money illusion and that a lower real wage and unemployment rate can therefore be obtained by boosting demand and the inflation rate. With the disappearance of money illusion, how ever, there is no long-run tradeoff and possibly not even a short-run tradeoff between inflation and unemployment. The increased labor mili tancy in Europe in the late 1960s was illustrated by the explosion in money wages in the 1964-70 period, particularly in West European countries (see figure 7-1 and Phelps-Brown 1983; Sachs 1979; and Soskice 1978) at a time when productivity was slowing down. This militancy can be taken as a sign that, with rising inflation, workers' money illusion had progressively disappeared, and since the 1960s workers were willing to strike to maintain their real wage. The resulting unemployment can then be ascribed to the rigidity of real wages rather than to a lack of effective demand. 3 The social commitments of governments posed equally serious prob lems. Regardless of whether or not the level and coverage of social commitments were justified in view of the growth expectations of the post-World War II Golden Age, the worsened prospects in the 1970s required some downward adjustment in the levels and coverage of these social benefits even with unchanged social preferences concerning equity. 4 Instead, until the late 1970s, social expenditures grew rapidly in most OEeD countries, with most of the growth being due to improved levels of DEEPAK LAL AND SWEDER VAN WlJNBERGEN 187 Figure 7-1. The Origins of Trade Union Power Italy 1,183 1,016 France 692 ,. Belgium 652 ," United Kingdom 649 United States 295 1960 1965 1970 1975 1979 Note: Rise of hourly rates or earnings, mostly in manufacturing, in nine countries of the OEeD, 1960-79, showing change in rate ofrise about 1969-70. (Ratio scale: number at end of each curve gives average hourly rates or earnings in the first three quarters of 1979, expressed as an index in which 1960 100.) Source: PhelpS-Brown (1983:156). benefits (see figure Al-2, table AI-3, and appendix 1 in this book for a summary of the evidence), In light of the recent project evaluation literature (LaI1974, 1980; Little and Mirrlees 1974), a "critical consump tion level" of income can be defined, at which a marginal increase in publicly funded income transfers is considered socially as valuable as a 188 DEFICITS, INTEREST, AND DEBT marginal change in public revenue. The worsened growth prospects of the 1970s required a downward adjustment in this critical consumption level. The net effect of both a real wage and a critical consumption level higher than warranted by the new circumstances is a socially suboptimal consumption-savings balance for the economy. There would now be a premium on domestic savings such that a rational government should attempt to raise the domestic savings level through fiscal policy. Instead, the emergence of a social premium on savings also saw the growth of large fiscal deficits in many OECD countries (see figure A1-4). The decomposition of these deficits into their cyclical and structural components is controversial, although we cannot assume-as people usually do-that deviations of actual GDP in the 1970s from the trend levels based on performance in the 1960s represent a cyclical shortfall that can be corrected by expansionary fiscal policy. The cyclical adjust ments should be made with reference to the peaks of the cycles in the 1970s and not to the hypothetical trend that would have occurred if overall economic performance had matched that in the Golden Age of the 1950s and 1960s. Putting the matter differently, in making the cyclical adjust ments to the deficits, we cannot assume that the natural rate of unemploy ment has remained at its 1960s levels in most OECD countries. If it has drifted upward, the cyclical element in the budget deficits will be lower than has been estimated by many commentators. s Even in terms of conventional accounting in the 1980s, however, the fiscal deficits in some OECD countries, particularly in the United States, have an increasing structural component (see figure 7-2). It is to the causes and implications of these structural deficits that we now turn. The Financing of Social Expenditures, Budget Deficits, and Crowding Out The implications of the rising budget deficits in OECD countries are best seen by examining their relationship to gross domestic savings in the area over time (see figure 7-3). In 1970, OECD budget deficits absorbed less than 1 percent of gross savings; by 1975 this figure had risen to 44 percent and by 1983 to nearly 52 percent. The relationship with respect to the more nebulous notion of net savings was even worse (see figure A1-9). The resulting crowding out of private expenditures, particularly invest ment, that this public draft on available domestic savings represents is the most serious long-term trend in these countries. In subsequent parts of this chapter, we explore its implications for the global savings-investment balance and for world interest rates and thereby outline its effects on developing countries. DEEPAK LAL AND SWEDER VAN WIJNBERGEN 189 Figure 7-2. Budget Deficit in the United States, 1980-89 220 200 180 160 ~ .:;l 140 "0 "0 cI'i 120 :::J 100 '0