A FOV IVA. '1[11 rg U, 'L., 11114 131 5 in, m 9 ii;jll 31, oi 11 W4 R r.1'. T4:I 'il 1014 Ild.: .Ji 4A, AAH11101,110 Will IlMill"A ii V"W1 1E tw I TO l ,l I Ij -g' 4. A;6, "pr to k`vv Of% Policymaking in the Open Economy EDI Series in Economic Development EDI Series in Economic Development Maxwell L. Brown, Farm Budgets: From Farm IncomeAnalysis to Agricultural Project Analysis.Johns Hopkins University Press, 1979. James E. Austin, Agroindustrial Project AnaLysis. 2nd ed.Johns Hopkins University Press, 1992. William Diamond and V. S. Raghavan, editors, Aspects of Development Bank Managemnent.Johns Hopkins Univmrsity Press, 1982. J. Price Gittinger, Economic Analysis ofAgricultural Projects. 2nd ed. Johns Hopkins University Press, 1982. Gerald M. Meier, editor, Pricing Policyfor Development ManagemenL Johns Hopkins University Press, 1983. J. D. Von Pischke, Dale W. Adams, and Gordon Donald, editors, Rural Financial Markets in Developing Countries. Johns Hopkins University. Press, 1 983. J. Price Gittinger, Compounding and Discounting Tablesfor Project Analysis. 2nd ed.Johns Hopkins University Press, 1984. K. C. Sivaramakrishnan and Leslie Green, Metropolitan Management: TheAsianExp erience. Oxford University Press, 1986. Hans A. Adler, EconomicAppraisal of Transport Projects:A Manual with Case Studies. Revised and expanded edition. Johns Hopkins University Press, 1987. Philip H.- Coombs and Jacques Hallak, Cost Analysis in Education:A Tbol for Policy and Planning. Johns Hopkins University Press, 19t°7. J. Price Gittinger,Joanne Leslie, and Caroline Hoisington, editors, Food Policy: Integrating Supply, Distribution, and ConsumptionLJohns Hopkins University Press, 1987. GabrielJ. Roth, The Private P0ovision of Public Services. Oxford University Press, 1987. Rudiger Dornbusch and F. Leslie C. H. Helmers, editors, The Open Economy: Toolsfor Poticynakers in Developing Countries. Oxford University Press, 1988. Policymaking in the Open Economy Concepts and Case Studies in Economic Perfornance Edited by Rudiger Dornbusch Published for the World Bank Oxford University Press Oxford University Press OXFORD NEW YORK TORONTO DEULI BOMBAY CALCUITA MADRAS KARACHI KUALA LUMI'UR SINGAPORE HONG KONG TOKYn NAIROBI DAR ES SALAAM CA'E T-OWN MFLBOURNE AUCKIAND and associated companies in BERLIN IBADAN © 1993 The International Bank for Reconstruction and Development I THE WORLD BANK 1818 H Street, N.W. Washington, DlC. 20433 USA. Published by Oxford University Press, Inc. 200 Madison Avenue, New York, N.Y. 10016 Oxford is a registered trademark of Oxford University Press All rights rescrved. 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 1993 The findings, interpretations, and conclusions expressed in this study are entirely those of the authors and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Library of Congress Cataloging-in-Publication Data Policymaking in the open economy: concepts and case studies in economic performance I edited by RudigerfDornbusch. p. cm.-(EDI series in economic development) Includes bibliographical references and index. ISBN 0-19-520884-6 ISBN 0-19-520985-0 (pbk.) 1. Economic policy. 2. Monetary policy. 3. Commercial policy. 4. Fiscal policy. 5. Foreign exchange. 6. Debts, Public. I. Dornbusch, Rudiger. 1I. Series. HD87.P63 1993 338.9-dc2O 92-33405 CIP Contents Preface i- 1. Introduction Rudigr Dornbusch The Objectives of Economic Policy 2 The Economic Policy Environment 3 The Importance of Economic Policy for Performance 8 About this Book 10 Note 16 Selected Bibliography 16 2. Monetary Policy Rules for a Small, Open Economy 17 Michael Bruno The Basic Monetary Framework 18 Monetary Policy under Fiscal Excess and Fiscal Reform 24 Monetary Targets and Tools under Normal Conditions 28 Monetary Policy in Prctice: Some Recent Israeli Examples 31 Notes 37 Selected Bibliography 38 3. Financial Reform and Liberalization 39 Joaquin Cottani and Domingo Cavallo Financial Repression: The Traditional Approach 39 Problems with Domestic Financial Liberalization 48 Problems with External Liberalization and Dollarization 55 Conclusion 60 Notes 62 Selected Bibliography 62 4. Financial Factors in Economic Development 64 Rudiger Dornbusch and Aljandro Reynoso The Financial Repression Paradigm 66 Budget Deficits and Inflation 73 v vi ontents The Effects of Dericits and of Hiigh and Unstable Inflation 80 Conclusion 84 Appendix. The Mexican Financial System: From Financial Repression to 2 Money Market 85 Note 89 Selected Bibliography 89 5. Exchange Rate Policy: Options and Issues 91 Rudiger Dornhusch and Luis Tellez KuenzLer The Role and Influence of Exchange Rates in the Economy 91 Alternative Exchange Rate Regimes 95 Multiple Exchange Rates for Commercial Transactions 97 Black Markets and Trade Misinvoicing 107 Dual Exchange Rates and Unified, Flexible Rates 119 Conclusion 124 Notes 125 Selected Bibliography 125 6. Protection in Developing Countries 127 Paul Krugman The Costs of Protection 127 Arguments for Protection 132 Forms of Protection - 136 The Extent of Protection 139 Cost-Benefit Analysis of Protection 143 The Political Economy of Protection 145 Conclusion 147 Selected Bibliography 148 7. The Other Side of Tlx Reform 149 Arnold C Harberger The Moral Flavor of Mx Reform 149 Designing a lAx System for Efficiency and Fairness 151 The Value Added Tax 152 The Uniform lTriff 158 The Taxation of Income from Capital 163 Conclusion 167 Notes 168 Selected Bibliography 171 8. Adjusting to a Terms of Trade Shock: Nigeria, 1972-88 172 Michael Gavin Nonmonetary Aspects of a Terns of Trade Disturbance 174 Economic Policy: Money, Exchange Rates, and the Budget 192 Case Studyr Nigeria 198 Conclusion 215 Notes 217 Selected Bibliography 218 Contents vii 9. Bolivian Trade and Dcvelopment, 1952-87 220 Juan-Antonio Morales Economic History, 1952-87 220 The Public Sector 230 Indebled InclustrializaiLion and Other Policies 232 The Policy Reforms of August 1985 245 Conclusion 248 Notes 249 Selected Bibliogrphiy 250 Index 253 Preface THE ECONoMIC DEVELOPMENT INSTITUTE (EDi) was established by the World Bank in 1955 to help mobilize the Bank's knowledge and expe- rience for the purpose of strengthening development decisionmaking in its member countries. The EDI meets this objective by organizing courses and seminars for officials who make or influence develop- ment policy and management; by carrying out institution-building activities to help training institutions in member countries strengthen their capacities to make and manage policy, and by publishing train- ing materials on investment and development policy. In addition, the EDI makes these materials available for independent study. In princi- ple, wi materials are written in nontechnical language so that they will be understandable to interested persons outside the economics profession. This volume examines several areas of policy that critically influ- ence economic performance: monetary policy, exchange rate policy, financial reform, tax reform, and foreign capital. The conceptual dis- cussions of these topics are complemented by several country case studies that illustrate the effects of various policy options. The book does not offer specific policy prescriptions for every conceivable situ- ation a country might face, but the writers give valuable advice based on their experience. Although intended for use in courses and semi- nars at the EDT and its partner insktitutes, the study should also be of interest to government officials and policymakers in the developing world who want to have an overview of the various policy issues. Pro- fessional economists may also find it useful. This collection of essays on economic policy in the open economy follows on and complements The Open Economy: Tools for Polijymakers in Deueloping Countries, a volume edited in collaboration with the late Leslie Helmers for the EDI. In 1986 Leslie Helmers took the initiative to organize a group of policy studies that would address economic policy issues faced in developing countries. His untimely death after ix x Preface the first volume was published made iL impossible for him to compicte the task to which he hiad brought so much enthiusiasm. It is appropri- atc, therefore, to dedicate this volume to his mcmory. Special thanks are also due to Beth Annc Wilson for her editing of the volumc. Introduction m Rudiger Dombusch MILTON FRIEDMAN is fond of observing that there is no such thing as "no policy"; whatever governments do, they are doing something. If they do not print money, that is a policy, and if they print money, that is aiso a policy. The same goes for all other areas in which government has any involvement: protecting domestic industry by trade restric- tions or running an open economy, borrowing abroad or not borrow- ing, accepting foreign direct investment or not allowing any, and so on. What governments do affects economic performance. Good policies enhance growth, and poor policies slow it down; very poor policies can set development back by decades. Of course, a country's resources and the international environment in which it trades will have a criti- cal effect on its standard of living. But the extent to which these resources can raise a country's standard of living depends on eco- nomic policy. We cannot look at economic performance over time-as measured by, say, growth in real per capita gross domestic product (GDP)-and infer whether a country has pursued good or bad policies. Consider table 1-1, which shows the growth experience of four countries. Did Zambia do poorly in economic policy and Korea do well? The data can tell us only that Korea experienced a nearly sevenfold increase in per capita income, whereas Zambia is today 20 percent below where it was in 1960. But more often than not, we find that countries which did very poorly not only experienced adverse shocks but also responded poorly in terms of economic policy, whereas countries which did exceptionally well not only escaped unfavorable shocks but also exploited the favorable environment by creating a policy setting con- ducive to stability and growth. To some extent, case studies can sup- port these claims, but ultimately there is also the need for faith in the belief that good policies pay off, slowly and steadily, and bad policies do not pay and can be dramatically expensive. - 2 Rudiger Dornmbuch lkble 11. Real Per Capita Income, Selected Countries and Yean, 1960-90 (111(10 - 100) C-ountry 1960 1970 1980 1990 Indln 1(10 116 127 177" Kore, Rep of 1(0 172 320 G83 I'eru 100 130 145 102 ZainlIa 100 123 106 801" ;1. 111110. b. 1987. Source: Im (vairious Insucs) We will argue in this book that policy accounts for much of the difference in economic performance and that this difference is at the discretion of policymakers. But we go further to make the following argument: the poorer a country and the more adverse the conditions it faces in terms of resource availability, export prices, or world market access, the more important it is to implement good policies precisely to minimize the severe costs of an unfavorable environment. Rich countries can afford to waste resources; poor countries cannot. Rich countries may be able to neglect saving and investment (at least for a 'while); poor countries cannot. Thus, good economic policy is critical for developing countries. What remains is to define good economic policy. There is no manual of conduct that offers advice for all occasions. But a substantial body of principles is on hand, and case studies can help policymakers analyze what the problem is, what options are available, and which option best meets the need. This book seeks to aid in this process. The Objectives of Economic Pblicy There are three basic objectives of economic policy: prosperity, equity, and the twin objectives of stability and continuity. The pursuit of prosperity, for better or worse, is the deepest motiva- tion of individual economic behavior. Individuals not only want to survive; they also want to improve their standard of living. Whatever philosophers say about the merit of a society of consumers, individ- uals do want to increase their material welfare. Governments can advance the standard of living by creating the environment that is most hospitable to developing a country's resources. Equity is no less important an objective than prosperity, although it is frequently neglected. In some countries, notably in Latin America, income distribution is highly unequal (see table 1-2); the injustice and social discontent that inequality fosters are clearly visible. Economic InMToduction 3 Itble 1-2. Income Distribution, Selected Countries, 1990 (percent) Share of top 20 perena Share of bottom 20 pereet Country ~ of houeshokls of househldsk Brazil 62.6 2.4 India 41.4 8.1 Morocco 39.4 9.8 Source:.World Bank (1990). prosperity must reach all people, and any economic policy must address both equity and prosperity. Attempts to equalize income distribution often lead to policies that seek to dampen the consequences of shocks for poorer groups. Iron- ically, these policies may.have the opposite effect. An overvalued exchange rate, meant to keep import prices low to-help consumers, may ultimately lead to a currency crisis (see, for example, Dornbusch and Edwards 1990). Equalizing the income distribution of a country is tricky because if it is pursued too vigorously or with poor means, the effort may be mostly frustrated; prosperity might decline without much improve- ment in equity. But it is also true that a startling lack of equity, quite aside from the moral offense, can invite political instability, which in turn hurts economic performance. The status of stability ar:d continuity is somewhat subsidiary. For many economists these twin objectives are more a means for achieving prosperity than desirable goals in themselves. Fc; example, low.infla- tion and stability of property rights make for a good business environ- ment and hence for prosperity. But we must take seriously the possi- bility that people psychologically value stability and continuity per se. Whichever of the two reasons applies, stability and continuity rank among the objectives cf policy. The Economic Policy Environment The institutional setting in which economic life takes place provides the environment in which policies operate. Its most important charac- teristics include the presence or absence of property rights, the extent to which these rights are accepted or challenged, and the resources available. The institutional environment also includes provisions that determine the interaction of individuals and the state. At one extreme, the state has extensive discretionary powers to intervene in the econ- omy and dictate- in detail people's economic lives. Socialist countries are examples. At the other extreme, the scope for governmental inter- vention is-severely restricted by the country's constitutional limita- 4 RudigerDonmbusch tions, and individual economic freedom is maximized. The United States and Switzerland might be examples of such a system. Superficially it mighL appear attractive to have a situation in which the state is entitled to intervene in economic life, dictating who gets what and who does what. Regardless of how powerful government is, however, people will find a way to circumvent the restrictions. If their savings are at hazard, they will keep them abroad; if prices are con- trolled, goods may not be produced. The best economic system is one in which individual self-interest is the driving force but market out- comis in terms of individual income and wealth are mitigated by policies that enhance equality. The more a government is bound by institutions, the less need there is for individuals to assume a defen- sive posture against possible governmental action. The following three rules describe the basic institutional setting for what might be described as a "social market economy!" * The right to own property is acknowledged by the culture and protected by a well-functioning legal system. (Property rights in a Lroad sense include all forms of contracts entered into by eco- nomic agents.) * Competition-both within and across borders-is accepted as the basic rule in markets. Competition is enforced by allowing entry into all and any markets, by vigorously opposing monopolization, and by opening the economy to world trade. Competitioni makes the most of resources. - Equity is a distinct social objective and need not emerge from the operation of competitive markets. mne government promotes equity through a tax system that strikes a balance between equity and efficiency and, above all, by investing in education and health, which are key determinants of economic advance. With these principles in mind, microeconomic and macroeconomic policies focus on creating a favorable climnate for economic activity. Microeconomic Policy The maintained hypothesis for microeconomic policy is that free mar- kets best maximize the real income a country can derive from its resources. There are clearly specific instances in which markets fail. The restriction of competition by oligopoly. or monopoly is an instance. Market failure also occurs when spillover effects are not priced in the market. In this case, activities that have favorable externalities are not carried forward sufficiently, and activities with negative externalities are maintained at excessive levels. Introduction 5 From the perspective of the open economy, well-conceived micro- economic policies can be especially effective in two areas: trade and foreign investment. Trade policy determines whether a country is closed-with firms producing -or the domestic market, which can be very small, and often at scales that imply very unfavorable cost conditions-or open, in which case more of the country's resources are devoted to producing exports. Trade policies determine not only the level and dispersion of tariffs but also the existence of quotas and licenses and the extent of trade in technology, design, accounting, advertising, transport, and legal services. There is abundant evidence that countries that choose to trade openly perform better. Foreign direct investment used to be controversial. Today a more informed view is held. In addition to allowing the savings of one country to promote the development of another, foreign direct invest- ment often brings with it technology that cannot be easily bought or rented in the market, superior management that may be in short supply in a developing country, and access to markets in industrial countries. The smaller the developing country is, the more critical it is to take advantage of scale economies available when doing business with mul- tinationals. Policies that encourage foreign direct investment should not limit the fields in which the rest of the world can invest (why restrict, say, manufacturing. to domestic capitalists?). Such policies should provide prompt and efficient procedures for handling new investment, an effective tax system that does not punish foreign investment with double taxation, and a. clear set of rules that protect the withdrawal of profits and principal. These are just two examples of microeconomic policies that affect the use of resources in an open economy. They are discussed here to suggest that countries do make choices about their business environ- ment and that these choices are critical to the effective utilization of resources. New domestic concerns such as environmental issues and labor standards also have implications for trade policy. International differ- ences in the level of regulation or standards create incentives for "social dumping." Poor countries may become attractive locations for "dirty jobs" and "poor jobs" A country must ask whether it wants to become the world's dumping ground for toxic waste and hazardous production and whether its lack of child labor laws should become the basis for an industrialization strategy. On the surface, the answer is clearly no. But regulation and standards must strike a balance between the severely negative effects of environmental degradatioin or poor labor standards, on the one hand, and the need to make a living in a 6 RudigerDornbuscla very competitive world, on the other. Two more remarks should be added. First, denying that tradeoffs exist is poor economics; a country that sets very high standards must be aware that these regulations have some costs. Environmental degradation that later needs to be cleaned up at grcat expense may be a very bad idea. Similarly, "cheap" labor may mean high accident rates and poor productivity. Second, we must recognize that lack of regulation is also expensive. Therefore, a serious cost-benefit analysis, both in labor standards and in environmental regulation, involves investment choices. Macroeconomic Policy In contrast to the focus on market competition and regulation, the four main- goals of macroeconomic policy are controlling inflation, maintaining a stable and competitive real exchange rate, exercising fiscal prudence, and operating efficient capital markets. Some com- ments on each of these follow. INFLATION. An environment of high and unstable inflation deters productive economic activity. Capital market horizons shorten, and economic decisionmaking is distorted. Spurious gins and losses related to the vagaries of inflation rather than to effort and produc- tivity become the rule. Inflation at high and variable levels represents a basic malfunctioning of the economy. A primary rule of macro- economic policy, therefore, is to create a monetary and fiscal environ- ment conducive to price stability. There may be considerations of public finance that favor a moderale level of inflation, but given the risk that such a strategy might lead to high and unstable inflation, more than caution is required. EXCHANGE RATE. A key link between a country and the rest of the world, both in goods and assets markets, is the exchange rate. The profitability of production and investnent in a country is based on their costs in dollars in relation to world prices. Poor exchange rate policy risks misrepresenting true opportunities and thus misallocat- ing resources. For example, an exchange rate that is kept competitive to gain access to cheap imports (financed by borrowing abroad) may be reasonable in a period of high and productive investment, but it is certainly not a reasonable strategy when used to support a consump- tion boom. In fact, as the experience of Latin America in the 1970s shows, once the financing stops, a policy of overborrowing invariably ends in.a foreign exchange crisis and a massive devaluation. The crisis, in turn, risks igniting inflationary pressures, and massive devaluation represents a cut in the standard of living that brings difficult political Introduction 7 consequences The lesson is that exchange rate policy is an area in which mistakes must be avoided and that a competitive, stable real exchange rate is the goal. FISCAL PRUDENCE. In choosing fiscal policy the government must decide on the size of the budget deficit and how it should be financed-through domestic debt, foreign debt, money creation, or a combination of methods. Here I will make a remark on the microeconomic implications behind deficit policies that affect the structure of the tax system itself, with its incentive effects and equity considerations. A good fiscal sys- tem applies cost-benefit analysis to determine the public sector effects of taxes, spending, and production. Today it is widely recognized that a broadly based system with moderate taxation is desirable and that subsidies have become too pervasive. One source of large budget defi- cits has been government involvement in industry. The current move toward privatization represents a healthy swing of the pendulum. There is a special need to consider privatization when the public sector faces constraints on borrowing and experiences a high margin- al cost of fiscal resources. Turning industries such as telecommunica- tions and airlines over to the private sector solves the investment problem and thus avoids bottlenecks. Next, a key macroeconomic decision concerns the size of the budget deficit and how to finance it. The budget need not be balanced every year, and small deficits clearly do not destroy macroeconomic stability. But the question must be asked: how will the deficit be financed? The existence of capital markets offers a temptation to borrow to the full- cst, but ultimately loans must be serviced and even repaid. Moreover, excessive borrowing may end in complete credit rationing, with even emergency loans no longer available. The alternative to borrowing is printing money. In a growing econ- omy, some money printing is acceptable, bit the scope for it is sharply limited. Beyond a certain threshold, money creation becomes infla- tionary, and not much farther down the road it becomes a possible source of violent inflationary instability. Budget deficits, therefore, are' a problem. The trouble is that often policymakers do not recognize the problem until they have used up all available credit and are reduced to printing money. CAPITAL MARKETS. Because the growth of an economy is often linked to the size of its capital stock, the effect of governmental deci- sions on investment is key. The only way a country can add to its productive capacity is by investing. Investment may take the form of an expansion in the amount of human capital or of physical capital. 8 Rucliger Dornbusxeh One way or another, it requires resources, and these must come from savers, either domnestic or foreign. A well-functioning capital market serves to mobilize and allocaie savings among competing uses. Moderate .and stable inflation is a prerequisite for a well-functioning capital market and so is, of course, a vigorous system of property rights. Without well-functioning contracts there can be no financial intermediation. The Importance of Economic Policy for Performance The importance of economic policy in advancing or setting back a country's economic performance is brought out by A comparison between Argentina and Spain (table 1-3). During the postwar period Argentina's wealth of human capital and natural resources gave it the head start that enabled it to thrive for decades. Until the 1970s its real product per capita was above Spain's. By the mid-1970s, however, Argentina had fallen back from the stellar position it held at the turn of the century (see Maddison 1989). During the mid-1970s, Spain mod- ernized, opened, kept down inflation, cleaned up public finance, and prepared to join the European Community. Argentina, instead, indulged in cycles of populism, experimented with policies that over- valued the currency, and imprudently liberalized financial markets. As a result, in the 1980s Spain moved sharply ahead, supported by d.omestic adjustment and a favorable European environment, while Argentina disintegrated to unimaginable extremes-a process that is still under way. Whereas Spain's per capita income grew at an average annual rate of 2.3 percent in the 1980s, Argentina's shrank by 2.9 percent a year (see figure 1-1). Argentina today represents a situation of political and economic distress-possibly pauperization-which is not unlike that of Germany in 1923. The extent to which differences in performance can be attributed to economic policy remains an open question. More striking than the ¶lble 1-3. Real Product Per Capita, Selected Countries, 1950-90 (]975 international dollars) Year Argentina Spain 1950 1,877 1,163 1960 2,124 1,737 1970 3,231 2,751 1980 3,209 4,264 - 1990 ' 2,484 5,561 a. Estimate based on real per capita GDP growth rates. Source: Summers anrd Heston (I984). Iniroductwion 9 Figure 1-1. Real Per Capita Income, Argentina and Spain, 1973-89 (1980 =100) 125i_ 120 -/ 115 110 105 Spain 100 95 90 80 75 1973 1975 1977 1979 1981 1983 1985 1987 1989 Someu: IMF (various issues differences in growth between Spain and Argentina are the very dif- ferent growth experiences of Asia and Latin America. In some way the economic success of Asia must be linked to the fact- that countries there saved, invested, innovated, and sought out world markets while maintaining economic stability. And the problems of Latin America reflect the extent to which countries there borrowed, spent, destroyed economic institutions, and neglected the crucial role of trade in rais- ing the standard of living. lIble 1-4. Per Capita CDP Growth, Selected Countries, 1950487 (annual percentage change) Region and country 1950-73 19 7.-,7 Asia 3.1 2.6 India 1.6 1.8 Korea, RepL of 4.1 6.2 Latin America 2.4 0.0 Argentina 2.0 -0.9 Brazil 3.8 2.2 Source: Maddison (1989). 10 Rudiger Dornmbusct Economic policy is part of the explanation for superior economic performance. History, resources (including human capital), location, the external environment, and luck all have their share. Experience shows, however, that mismanagement can overcome the most favor- able conditions and bring a country down and that superior manage- ment can greatly improve a country's chances of getting ahead, even when the preconditions are not especially favorable. About This Book This volume reviews how policy choices critically influence economic performance. Given the large scope of policies, the work focuses on several key areas. Monetaty policy. In a small, open economy there is not much scope for independent monetary policy. The latitude for noninflation- ary money creation is limited by growth in the real demand for money. If the central bank overestimates its degree of indepen- dence and selects too rapid a path of domestic credit growth, an exchange rate crisis and inflation are inevitable. If money cre- ation and inflation get out of hand, the next question is how best to stabilize (see also Bruno and others 1988). * Exchange rate policy. Policymakers must choose from the full range of possibilities-from floating rates to fixed rates. A fixed rate is too inflexible, given shocks to the real economy or disturbances in the price level. A floating rate may be too unstable except in the rare economies in which fundamentals are never in question. A crawling peg with occasional adjustment of the real exchange rate to accommodate changes in fundamentals may come closest to ensuring external competitiveness without creating oppor- tunities for counterproductive speculation. * Financial reform. The financial sector should function so as to mobilize domestic resources for growth. Effective financial inter- mediation ensures that household savings are channeled to domestic productive investment rather than to inventories or dollars. What are the requirements for effective financial inter- mediation, and what challenges arise in the process of financial reform? The debate has focused on the need for positive real interest rates, but that emphasis may well have been overdone. The lesson is that real interest rates should not be consistently negative (see also Polak 1989). Once more, Argentina, as shown in figure 1-2, serves as an example of what is clearly wrong. * Protection. In the 1950s the Economic Commission for Latin America preached import substitution as a way for countries to industrialize and avoid the growth-inhibiting effects of an exter- Introduction 1I Figure 1-2. Real Value of an Investment, Argentina, 1983-90 (lanuary 1983 = 100) 170 160IG 150 _ 140- 130' = 120 -Lv 110 100 ,~90 ~80 asN ~70 60 so 40- 30- 20- 10 1983 1984 1985 1986 1987 1988 1989 1990 Seuwn IMF (various isucs). nal constraint. Today, in large measure as a result of the poor experience with import substitution and the closing of econ- omies, the pendulum has swung strongly in the opposite direc- tion. Protection is being questioned radically, and opening the economy is increasingly seen as a key step toward using resources more effectively and attracting foreign capital. Tax reform. A productive and efficient tax system is the cor- nerstone for macroeconomic stability. Sound tax policies elimi- nate the need for inflationary money creation and offer maxi- mum incentives for resources to be allocated without waste. A government does need revenue to supply public goods and ser- vices that the market does not supply according to political and social priorities. The principal problem is to design a tax stric- ture that minimizes the resource cost of taxation. Given this crite- rion, should the tax system be selective, or broadly based? Should rates be differentiated, or unifonnrm?axation that is broadly based improves efficiency by lowering marginal rates without sacrific- ing revenue. Moreover, it enhances the enforcement of tax collec- tion and thus reduces evasion. Theory says that differentiated £2 RudigerDomnbuch systems are optimal, but the practitioner urges a broad base and a uniform rate. Fovreign capital. Should a country rely on external resources to supplement domestic saving? And if it does, what form is best- debt or foreign direct investment? In the aftermath of the debt crisis the case for borrowing seems weakened by the apparently negative experience of the 1970s. But that is a misreading of the experience. Two crucial questions must be asked. What form should the borrowing take-debt or equity? And what are the resources used for-to finance budget deficits, consumption, and capital flight, or to finance productive investment? Since the nineteenth century, external capital has been an important instrument for economic development in Africa, Asia, and Latin America. The occasional debt crisis must not detract from the basically positive experience with foreign borrowing. This book addresses both the conceptual issues involved in these topics and how problems arise in practice. As in the earlier volume (Dornbusch and Helmers 1988), the conceptual approach is comple- mented by case studies that review country experiences. The main emphasis is on money, capital markets, and exchange rates and on trade and taxes in relation to resource allocation. Two case studies- on Bolivia and Nigeria-round out the discussion. In the lead essay (chapter 2), Bruno discusses rules for monetary policy in an open economy. He emphasizes that money has two kinds of effects: it influences nominal variables and, specifically, the evolu- tion of prices over time, but it also affects real variables, at least in the short run. This double influence poses a problem because control of inflation, given the real effects of money, becomes costly. Hence, choices must be made. Ultimately, a central bank cannot give a coun- try prosperity simply by printing money; that is an illusion. The best! contribution the central bank can make to a country's economic per- formance is to establish immediately rules for the behavior of money and credit and thus to create a stable economic environment. It is.not enough to announce that rules are desirable. We need to learn about what kinds of rules are useful in an open economy. We also need to look at some experiences, as Bruno does, to see how monetary policy works out in practice. In the 1980s many countries undertook financial reform and liberal- ization. The inspiration for these actions came in part from experiences with highly distorted capital markets that increasingly malfunctioned. Reform was also attributable to the constant efforts of scholars who doc- umented that economic performance is negatively related to financial repression (see McKinnon 1988; Polak 1989; World Bank 1987). Intmduction 13 In chapter 3 Cottani and Cavallo discuss financial reform and liber- alization from the perspective of Argentina's experience. They describe the link between financial repression and public finance. Governments finance deficits by creating money. To do so without much inflation, they try to repress financial markets. But, as the expe- rience ofr Uruguay shows, more often than not the attempt is unsuc- cessful. In fact, negative real interest rates lead to misallocation of credit and to capital flight. The authors clearly recognize that finan- cial repression is a bad idea; interestingly, they give financial liberal- ization only two cheers, not three. Chapter 4, by Dornbusch and Reynoso, attempts to qualify what we know about the effects of financial repression. The qualification runs in the following direction: very negative real interest rates and extremely overvalued currencies have a disastrous effect on economic performance because they are an invitation to refrain from saving or to move one's savings offshore. Such a situation must surely be avoided. But the essay also argues that the posited relation between real interest rates and growth does not hold outside the circumstances described above. Empirically, it is simply not the case that countries with higher real interest rates perform better. Moreover, even if there were evidence in support of such a relation, one would still have to ask .whether policy should try to raise real interest rates. It may be, of course, that real interest rates are high because the profitability of capital is high, as has been argued by Gelb (1989). A complement to the discussion of capital markets and real interest rates is offered by Dornbusch and Tellez Kuenzler, who in chapter 5 look at the foreign exchange market. 'Under conditions of capital mobility, there is a tight link between interest rates within a country and abroad and the expected rate of depreciation. If interest rate differentials fall short of expected depreciation, capital flight is cer- tain. Policies to depress real interest rates will quickly run into trouble in the exchange market. The chapter evaluates various exchange mar- ket arrangements toward which countries gravitate as they try to seg- ment the exchange market. These arrangements include multiple exchange rates for commercial transactions and dual rates that sepa- rate commercial and financial transactions. The chapter comes out strongly against multiple rates, not only because they distort trade but also because they involve serious budgetary problems. Dual rates are found to be useful tools provided that the premium of the free rate for capital transactions does not significantly exceed the commercial rate. But by no means are dual rates a panacea that relieves domestic mismanagement of its external consequences. The discussion of multiple rates has already drawn attention to the costs of resource allocation when prices are distorted from their world 14 Rudiger Dornbusch levels. The essay by Krugman on trade policy (chapter 6) addrcsses this issue squarely. The question "Can broad trade restrictions help a country push its economic development further or faster?" is given a sympathetic hearing and then answered witlh a strong, unwavering Nol Neither in theory nor, certainly, in practice is there a case for a small country to adopt protection as a strategy for growth. The case for trade liberalization-turning away from the accumu- lated protection of the 1950s, 1960s, and 1970s-has been made by Chile, the Republic of Korea, Mexico, and Turkey. Today these coun- tries are examples of highly successful liberalization. And the success of these countries is encouraging others to turn away from their increasingly negative experiences with protection and to open up. (For further discussion, see Dornbusch 1992.) Harberger's essay on tax reform (chapter 7) is a masterpiece in political economy. Its central message is that being overly ambitious is counterproductive. Whereas discussion of tax policy typically has con- centrated on the presentation of elaborate optimal tax structures, Harberger focuses on how to get results. He concludes, "Tax reform in today's world is ... more appropriately focused on the giving of robust signals that will improve the private sector's allocation of resources than on the niceties associated with formal, technical opti- mization" But the statement should not be misconstrued as selling economics short. On the contrary, Harberger, here and in other writ- ings, sets out a strong case for uniform taxation as a way of striking a balance between the administration of taxation and the efficiency of the tax structure. Two case studies complete this volume. Chapter 8, by Gavin, reviews Nigeria's adjustment to a terms of trade shock. The great merit of the paper is that it links the terms of trade to a wide range of macro- ecouomic variables and issues. When a country's export prices fall, everything must adjust. And when the government owns the resources-the prices of which have declined-the adjustments have to reach all the way to monetary and fiscal policies. With a favorable shock, there are basically two options: to spend' the resources while they last. or to set aside a portion and build up a reserve against less fortunate days. Needless to say, Nigeria failed thoroughly. The resources were. dissipated to such an extent that after oil pnrces declined, adjustment became especially difficult. Worse, handling the transition poorly-that is, late and incompletely-ultimately added to the burden of adjustment. The case of Bolivia, described by Morales in chapter 9, brings to the fore the link between budget deficits and money creation. When a country runs a deficit, it can finance the excess of spending over taxes by borrowing at home, borrowing abroad, or printing money. If the Introduction 15 former two sources of finance have dried up, the third comes on with all its risks. In Bolivia in the mid-1980s that meant hyperinflation. Morales shows how the country gradually drifted in that direction until, in the end, control was lost and money creation translated into explosive inflation. The stabilization is equally interesting. It was uncompromising, harsh, and ultimately successful. Price stability was reestablished, even if growth of per capita income was slow to return. The essays in this book seek in their diversity to convey a central message: economic policy cannot be intelligently conducted without a healthy dose of economics. There are choices to be made. Economics helps us decide on the objectives and to assess and compare the avail- able instruments or mechanisms. Above all, economic theory and case studies teach us what measures will not succeed-money-financed def- icits, protection, and financial repression, to name just a few. In concluding, we present here Harberger's thirteen rules for good economic policy managemenL.' 1. Avoid poor technicians in policymaking. 2. Keep budgets under adequate control. 3. Keep inflationary pressures under reasonab!e control. 4. Take advantage of international trade. 5. Keep in mind that some types and patterns of trade restrictions are far worse than others. 6. When import restrictions become excessive and reducing them is politically impossible, attack them indirectly by increasing export incentives. -7. Make tax systems simple to administer and (as far as possible) neutral and nondistorting. 8. Avoid excessive income tax rates. 9. Avoid excessive use of tax incentives to achieve particular objectives. 10. Use price and wage controls sparingly, if at all. 11. Remember that only rarely can a cogent rationale be found for - quotas, licenses, and similar quantitative restrictions. 12. Take a technical rather than ideological view of the problems associated with public sector enterprises. 13. Make the borderlines between public and private sector activity clear and well defined. Fischer (1987) has argued for adding 14. Don't overvalue the exchange rate. The reader will want to bear these rules of good management in mind as the issues are addressed in the essays in this book. 16 ludiger Dornusc/h Note 1. Some oF the precepts are paraphrased to shorten the cxposition. See I-larberger (1984) for the detailed development. Selected Bibliography Bruno, Michael, Guido di-Tella, Rudiger Dornbusch, and Stanley Fischer, eds. 1988. Inflation Stabilization: The Experience of Israel, Argenitia, Brazi4, Bolivia, and Mexico. Cambridge, Mass.: MIT Press. Dornbusch, Rudiger. 1992. "l`ade Liberalization in Developing Countries: Journl ofEconomic Perspectives 6(1):69-85. Dornbusch, Rudiger, and Sebastian Edwards. 1990. "Macrocconomic Popu- I ism 'Journal of Devlopment Economics 32(3):247-77. Dornbusch, Rudiger, and F. Leslie C. H. Helmers. 1988. The Open Economy: 7bTos for Policymaers in Developing Countries. EDI Series Economic Dcvelopment. New York: Oxford University Press. Fischer, Stanley. 1987. "Economic Growth and Economic Policy." In Vittorio Corbo, Morris Goldstein, and Mohsin Khan, eds., Growth-Oriented Adjustment P-rograms. Washington, D.C.: International Monetary Fund and World Bank. Gelb, Alan H. 1989. "Financial Policies, Growth, and Efficiency." Policy, Research, and External Affairs Working Paper 202. World Bank, Financial Policy and Systemls Division, Country Economics Department, Washington, AC. Harberger, Arnold C. 1984. "Economic Growth and Economic Policy." In Arnold C Harberger, ed., World Economic Growth. San Francisco: Institute for Contemporary Studies. IMF (International Monetary Fund). Various issues. Internaional Financial Statis- tics. Washington, D.C. McKinnon, Ronald. 1988. "Financial Liberalization in Retrospect: Interest Rate Policies in Lucs' In Gustav Ranis and T. Paul Schultz, eds., The State of Development Economics: Progress and Perspectives. Oxford, UK.: Basil Blackwell. Maddison, Angus. 1989. 7he World Economy in the 20th Century. Paris: Organ iza- tion for Economic Co-operation and Development. Polak,Jacques J. 1989. Financial Policies and Development. Pads: Organization for Economic Co-operation and Development. Summers, Robert, and Alan Heston. 1984. "Improved International Compari- sons of Real Product and Its Composition." Review of Income and Wealth 30(2):207-62. World Bank. 1987. World Development Report 1987. New York: Oxford Univer- sity Press. . 1990. World Development Report 1990. New York: Oxford University Press. Monetary Policy Rulesfor a Smagl, Open Economy M Michael Brno MONrrARY POLICY CONCERNS the determination and allocation of money and credit in terms of quantity and price, that is, interest rates. The first important distinction in this connection is the difference between the form in which money is issued and mneasured, namely, in units of nominal domestic currency (pesos, liras, and so on), and the way in which the quantity of goods that the money buys is measured, namely, in real terxns (tons, square feet, or GDP in real constant prices). The real economy deals with the production and sale of goods, the quantity of exports and imports, and the employment of labor, where relative prices matter (the real wage, the real exchange rate, and so on). The nominal economy deals with items such as the nominal wage, the nominal exchange rate, and money and credit, whose value will usu- ally be inflated along with the general level of domestic-currency- denominated prices. Money, and therefore monetary policy, may stabilize or inflate the price level, depending on whether it is contractionary or expansion- ary. However, because the government may exercise direct and indirect control not only on the nominal aggregate but also on the ex'change rate and the wage rate, the link between monetary policy and the price level is complex. The close link between money and the exchange rate suggests that the best way to view exchange rate policy is as part of the larger complex of monetary policy in the open economy.' In addition to affecting nominal magnitudes in the economy,zmone- tary policy may also have real effects, at least from a short-run per- spective. The most obvious example in a closed economy is the expan- sionary role that an increase in money and credit and a reduction of interest rates may have on real output when the economy is in a slump. In an open economy, monetary policy may have added real effects by promoting short-run capital flows in or out of the country. Moreover, especially in less developed financial systems, many instruments of monetary policy exist that may have direct real allocative effects, for 17 18 Michael Bruno example, instruments of credit rationing and credit allocation by spe- cific destinations (such as subsidized export credit). Although effective in reallocating resources, such instruments may undermine the effec- tiveness of monetary policy in its more conventional stabilizing roles. More generally, monetary policy can have real effects because central banks may be expected to target their varied monetary tools at other objectives, such as employment levels, foreign exchange reserves, or specific industrial objectives. The degree of independence of monetary policy and of the central bank differs widely from country to country and may change dras- tically over time in a particular country. The most important element is the degree to which budget requirements and fiscal policy objectives dominate monetary objectives. Monetary institutions and operating rules evolve over time, depending on ti-e development of financial and capital markets. Monetary policy is usually depicted in macroeconomic textbooks as the way in which a vwell-advanced, industrial economy controls the money supply. This assumes the existence of separate, and at least minimally independent, fiscal and monetary authorities; the existence of well-developed markets for short-term debt instruments; and a clear definition of the monetary objectives. As economists are increas- ingly learning, however, even the typical industrial country's financial structure no longer conforms to the standard model. Nevertheless, it is instructive to understand the workings of this model in order to trace the gradual evolution of monetary institutions in a developing econ- omy. We shall also discuss situations in which the underlying institu- tional structure is much less developed. In this respect, the experi- ences of some economies that have been or still are in a state of transition are instructive. Our main illustrations will be taken from the Israeli case. The Basic Monetary Framework In a typical small, open economy the public will hold an array of financial assets-currency in circulation, demand deposits (in banks), interest-earning time deposits, foreign currency (or foreign- currency-linked deposits), short-and long-term bonds, various forms of long-term savings funds, and so on. People will hold only some of these assets for transaction purposes, and the rest are held as a store of value or as longer-term savings. Monetary policy is usually designed to affect the quantity and price only of financial assets used mainly for transactions.. Expansion of the aggregate money supply over and above the level desired (at a given cost) may increase the demand for goods and thus cause inflationary pressure (assuming Monelary Policy Rules for a Small,. Open Economy 19 the economy's output cannot be increased in the short run) or a loss of foreign exchange reserves or both. The choice of a suitable aggre- gate for the targeting of monetary policy is problematic, because the dividing line between one type of money and another is somewhat arbitrary. For example, short-term government debt may be almost as liquid as bank deposits. Likewise, foreign-exchange-linked deposits may, in some economies, play a very liquid role. For the moment we shall confine ourselves to the narrowest definition of the means of payments (M), consisting of currency in circulation (CU) and demand deposits in banks (DD).2 Before discussing how money is supplied in the economy, consider a simplified version of the banking system in terms of the balance sheet of the central bank and the consolidated balance sheet of the commercial banks (figure 2-1). The central bank's assets include the authorities' foreign exchange reserves or its net foreign assets (NFA), as well as all the domestic securities that the central bank holds against credit that it hands out, which we shall call central credit (CC), to distinguish it from credit that commercial banks give out, which we shall call bank credit (BC). For the moment, let us assume that all central credit goes to the government and that all commercial bank credit goes to the private sector. (Later on we shall change these assumptions, in particular to allow for central bank credit to both the private sector and the banks.) On the liability side, the central bank issues all the currency in circulation (CU) and holds the reserves owned by the commercial banks (R) (which for that reason also appear on the asset side of the balance sheet for the commercial banks). The central bank's two liability items make up the monetary base, or high-powered money (B). Figure 2-1. The Banldng System: A Simplified Version Cemunia bat Omiu bank Liabiies Asse Liabiias Assets *DCmapId (DD) Bank reseves (R) Net forein Short-term bank credit (BC) Currency (CU) Cedit (CC) Monetary base, or high-powered money (H) 20 Mihael Bruno The commercial banks' liabilities are deposits held by the public, which for the moment we shall assume consist only of demand deposits (DD). Against these, the commercial banks hold their reserves (R) in the central bank and issue short-tenn bank credit (BC) to the private sector (ignoring other types of credit for the moment). Simple balance-sheet-equality rules tell us that for the commercial banks we have: (2-1) Deposits (DD) = Bank reserves (R) + Bank credit (BC) whereas for the central bank we get: (2-2) Monetary base (H) = Bank reserves (R) + Currency (CU) = Net foreign assets (NFA) + Central credit (CC) Remembering our definition of money (Al) as the sum of currency in circulation (CE) and deposits in banks (DD) (M = CU + DD), and consolidating the balance sheet of the central bank with that of the commercial banks, using equations 2-1 and 2-2, we get the following simple, consolidated, monetary identity of the open economy. (2-3) Money or means of payments (= Total domestic credit (De) + Net foreign assets (NFA) where total domestic credit (DC) is defined as the total credit issued by the commercial banks (BC) plus that issued by the centml bank (CC), namely:5 (2-4) M =BC + CC + NFA =DC + NFA Equation 2-4 is an extremely important financial balance equation for consideration of monetary policy in the open economy, and it can be used in a number of ways. What it says first and foremost is that money can be created in several ways: by banks giving credit to the public (providing they. are "allowed" to do so), by the central bank giving credit to the government, and by the sale of foreign exchange to the central bank. The last two forms of increase in the quantity of money are ways of expanding the monetary base (H) (what is often called the "printing" of money), whereas the first component, credit issued by the commercial banks, is one over which the monetary authorities may have an indirect influence through the exercise of monetary policy on the banks. For example, credit may be controlled by stipulating r, which is the ratio of bank reserves to deposits (r = RIDD), from which it follows that BC [1 - r]DD). All of the above relationships apply to monetary expansion and obviously apply sym- metrically for the contraction of money and credit. Suppose foreign exchange reserves are to be kept constant. In that case the basic monetary identity, equation 2-4, says that any change (or Monetay Policy Ruzesfor a Small, Open Economy 21 stabilization) of money depends entirely on an equivalent change (or stabilization) of total domestic credit (to the private and public sec- tors).4 Likewise, if foreign exchange reserves change (because of, say, a deficit or surplus in the current account) the only way to keep the quantity of money constant is by changing total credit in the opposite direction (and in an equivalent amount) to that of the foreign exchange inflow or outflow. If a change is denoted by delta (A), we have. AM = 0, which leads to ADC =-/±NFA. This is called full sterilization of foreign exchange flows. Now we can immediately see the dual role that monetary policy can, in principle, play in the open economy. It may affect or stabilize domestic prices, or it may be used to stabilize the balance of payments by mechanisms that will be discussed later. Let us go one step back and consider how the domestic sources of increase in the money supply operate under the assumption of no change., n the country's net foreign asset position (that is, keeping NF constant so that DC-= M), as would be the case in a closed economy. The simplest form of monetary expansion is one in which the gov- ernment borrows from the central bank to finance fiscal expenditure. In that case the central bank increases its credit to the government (CC on the asset side of the central bank balance sheet) and "prints" money, that is, it increases the monetary base (H). The way this could take place in practice is for the government to pay for its additional expenses with cash (an increase in CC) or by writing checks that the public deposits in banks (thus increasing DD). The latter also automat- ically increases commercial bank reserves in the central bank (R). At this point both high-powered money (H) and money (AM) increase by the same amount. If, however, the monetary authorities operate under a rule that requires a fixed reserve ratio and the government payment is not in pure cash, there is scope for additional monetary expansion because banks can now issue additional bank credit (BC) up to the amount that will return r, the ratio of reserves to bank deposits, to its. initial level.5 So far, we have shown that an increase in fiscal expenditure leads to a monetary expansion that is either of the same magnitude (if done in pure cash that the public is willing to keep holding) or bigger (if causing an increase in r and DD), providing the banks now avail them- selves of the opportunity to lend more to the public. In either case, we can say that the monetary authorities are fully accommodating to the fiscal expansion by not exercising any independent monetary policy. How can the monetary authorities neutralize the effect of the fiscal expansion on the money supply? They can do so by increasing the required reserve ratio (r) so that M (and BC) are constant. Thus the authorities can exercise monetary policy by changing the reserve ratio 22 Michlael Bruno either up (for monetary contraction) or down (for monetary expan- sion). Other tools will be discussed below. So far, we have ignored the effect of the initial fiscal expansion on the economy. If the economy is in a state of underemployment, the increase in demand (by the government) may also activate unem- ployed factors of production and bring about an increase in real income and product. In that case, the authorities would not need to exercise monetary restraint. If, however, real output cannot grow and the monetary authorities want to avoid the inflationary consequences of an increase in demand that cannot be matched by higher real output, money and credit must be reduced appropriately. Alternatively, the government could achieve the fiscal expansion not by borrowing from the central bank, but by borrowing from the pub- lic, that is, by selling bonds. Selling more bonds entails reducing their price (raising the interest rate on the new bonds), which will also raise interest rates elsewhere in the economy. When interest rates rise in the process of inducing people to buy more bonds, people will want to hold less money because the opportunity costs of holding it will have increased. In this case, the economy will exhibit less inflationary pres- sure, and less of a monetary contraction will be required to avoid the price increase. The higher interest rates will in any case show in less investment in inventories and other physical assets. Fiscal expansion will have crowded out investment. Suppose, however, the government insists on financing the fiscal expansion by borrowing from the central bank. The central bank itself could carry out a neutralizing operation through open market operations-that is, by selling government bonds that it holds in its portfolio, thus directly reducing bank reserves, money, and credit and raising interest rates in the economy. A similar effect is obtained when commercial banks borrow from the central bank (so as to increase their reserves). The central bank gives access to funds at a discount window. The central bank effectively reduces the money supply when the discount rate offered at the window is raised.6 The banks then reduce their borrowings from the central bank (and thus t}i; money base falls) and raise interest rates to their own borrowers. So far, we have conducted the analysis as if we were in a completely closed economy. What is the meaning of our initial working assump- tion that net foreign assets (NFA in the centml bank balance sheet) are kept constant? It means that the central bank avoids extra buying or selling of foreign exchange on a net basis, whereas some of the opera- tions that we have described above could otherwise cause such net changes. For example, fiscal expansion may cause increased demand for imports, or higher interest rates may induce capital inflows. The first effect might cause downward pressure on exchange reserves (and Monetary Policy Rules for a Small, Open Economy 23 a depreciating effect on the exchange rate), whereas the reverse would be true for capital inflows, when they are feasible. The assumed constancy in NFA can be maintained only if the exchange rate is allowed to float. But a floating exchange rate policy is highly problematic for a small, open economy, especially if its wage and price system is not fully flexible. Pressure on the exchange rate (namely, forces making for devaluation), especially when resources in the export and import substitute industries are fully employed, will cause a cost-push effect on prices and wages. But reverse forces for an appreciation will harm exports (and import substitutes) and will not have a symmetrical downward pressure on prices (and wages) because of their downward rigidity. For this reason, we shall confine ourselves to the case of pegged or crawling exchange rate regimes. In this case we can no longer assume the automatic constancy of NFA. Suppose now that we are in an open economy with a fixed exchange rate. The effect of monetary expansion or contraction is no longer confined to the link with prices (or output) that has already been describecd. An upward pressure on domestic prices implies pressure for real depreciation (at a given nominal rate), a rise in the current account deficit, and a reserve outflow (NFA falls), which in itself has a contractionary effect on the money supply (consider the central bank balance sheet). A rise in domestic interest rates, however, may induce more borrowing from abroad, if allowred, causing a capital inflow that will have an expansionary effect on the domestic money supply. Expansion will be avoided only when foreign capital flows are under strict controls. To sum up, in an open economy operating under a fixed exchange rate, monetary policy affects not only domestic prices but also foreign exchange reserves. When there is a tendency for a reserve fall (because of a deficit in the current account or capital flight), a monetary con- traction, although having a depressing effect on output (and a possibly disinflating effect on prices), may at. the same time protect foreign exchange reserves from such a fall. Likewise, changes in floreign exchange reserves coming from the balance of payments will have a direct effect on the domestic money supply and may thus require countervailing pressure from the exercise of monetary policy. Finally, the effectiveness of monetary policy in an open economy under a fixed-exchange-rate regime depends on the degree of substitution between domestic and foreign. assets and liabilities. When there are no foreign exchange controls at all, any domestic monetary intervention may be completely neutralized by an opposite effect coming from a capital inflow or outflow. The domestic interest rate will be equal to the foreign interest rate (with allowance for expectations of a devalua- tion and some risk factor), and monetary policy will be ineffective. At 24 Michacl Brumn the other extreme, when there are no capital inflows at all, there will be complete sterilization. Most real-life cases lie somewhere in between. Even with fairly strict foreign exchange controls, there is room for capital inflows or out- flows through the current account (holding back repatriation -of export proceeds, incorrect invoicing of import costs, and so on), thus causing partial substitution between domestic and foreign assets. As the economy develops and gradually opens up its capital account, the degree of substitution rises, and the independence of monetary policy becomes more questionable unless the exchange rate is not kept strictly fixed (that is, if some small-scale floating within a band is allowed). So far, we have conducted our analysis within an economy that has reasonably well-developed financial institutions, where government debt is fairly widely held, and where financial transactions (including lending and borrowing operations) are done within fairly competitive markets to which there is more or less general access. In most coun- tries that are considered here, this is not the case. When financial markets are segmented or thin, the exercise of conventional. monetary tools is fraught with difficulties. Monetary contraction may cause excessive increases in interest rates in some markets rather than others, and if credit is rationed, the contractionary effects on eco- nomic activity may far outweigh the possible deflationary effects on prices. Likewise, the speed of adjustment to monetary policy, whether contractionary or expansionary, may be very slow. It is thus important to look at the operation of a variety of monetary tools under imperfect financial markets, inflationary budget policies, cartelized banking sys- tems, and so on, before coming back to the operation of monetary policy under more reformed systems. Monetary Policy under Fiscal Excess and Fiscal Reform The kinds of problems that imperfect markets and institutions may cause for the operation of monetary policy, especially under inflation- ary conditions, are best discussed using actual country examples. We shall use examples and episodes from the Israeli experience? In Israel, as in many other cases, the central bank's traditional role as the main monetary policy authority in charge of price stability was often dominated by other tasks, such as maintaining full employment or promoting investment and growth (especially in the export sector).; Until 1985 a large and persistent government budget deficit had severely limited the Bank of Isrel's ability to exercise independent monetary control. In addition, the bank had to cope with a highly concentrated and oligopolistic banking system. Moletary Policy Rulesf a SmalL Open Economy 25 A major limitation imposed on monetary policy in a period of high government deflcits is the absence of restrictions on central bank credit to the government. In Israel that absence made for a highly accommodative monetary policy during the period of high inflation from the early 1970s to 1985. In 1985 the government introduced new banking legislation, and by 1987 such credit was limited to small bridging loans within the fiscal year. In the face of large and persistent budget deficits the authorities tend to use monetary tools to prevent the crowding out of preferred sectors of production, such as agriculture or exporting industries. This is done by providing such preferred users with cheap, restricted credit. While alleviating demands from these sectors, the introduction of directed credit narrows the monetary and free credit base on which conventional monetary policy can operate. It also causes wide spreads among vuxious rates of interest. A major monetary tool used in Israel until quite recently is required reserve ratios. The required reserve ratio against various deposits in local currency was close to 50 percent during 1977-87, which is quite high when compared with ratios in industrial economies. Banks financing directed credit were granted exemptions from such require- ments, thus reducing the net reserve requirement by 10 percent or more. The actual reserves held for most of the period were below these. requirements despite the high fines that had to be paid and the inter- est rates paid on the discount window (by which some of these defi- ciencies were bridged). This made for a highly volatile money multi- plier. In addition, formal changes in reserve requirements in Israel had to get government approval, a cumbersome and slow process. An example of a financial innovation that increased monetary accommodation during the high-inflation period in Israel was the introduction in 1977 of foreign-exchange-linked domestic accounts (Patam accounts). The required reserve ratio on such accounts varied between 80 and 100 percent When exchange rate policy follows a crawling peg for balance of payments reasons, such accounts can divert some of the demand for pure foreign currency assets, thus leaving the central bank with greater foreign exchange reserves. Such deposits, however, tend to be a highly liquid asset whose value rises automatically with inflation, thus making monetary control much less effective. Dur- ing 1984, at the height of inflation, these Patam accounts formed 65 percent of all short-term assets held by the nonfinancial private sector. With the July 1985 reform, their liquidity was reduced (allowing new entry only for a minimum deposit period of one year), and their share in total short-term assets fell to 25 percent by 1987. Obviously the exist- ence of such indexed deposits seriously hampered the macroeconomic policy response during the oil price and exchange rate shocks. 26 Mic/tael Bruno During much of the period, the Bank of Israel used bank credit as its intermediate monetary target rather than some measure of the quan- tity of money, This is understandable, given the need to control spec- ulative waves during times of expected devaluation, when the public tends to reduce domestic currency deposits and moves to foreign exchange or foreign-exchange-linked assets. Stabilization of domestic credit rather than a monetary aggregate helps to reduce the demand for foreign exchange assets by raising domestic interest rates. More- over, banks have greater control over credit than over their deposits. Finally, the difficulty of controlling credit expansion by any other monetary tool (such as reserve requirements) has often led to the application of credit ceilings to individual banks. Whereas this may be an effective tool for a short-term monetary contraction, with time it may distort the allocation and price of credit throughout the economy. Our final example-of an action to be avoided-is the use of central bank intervention to support the price of government-indexed bonds in an attempt to reduce the cost of government borrowing. Such inter- vention weakens the central bank's control over the monetary base and the quantity of money. Suppose the government wants to sell a certain quantity of bonds over and above actual demand, which has fallen. To prevent a fall in the price of bonds (that is, a rise in their yield), the central bank buys up the excess (increases the monetary base) independent of whether economic conditionsjustify such mone- tary expansion. Israel abandoned this policy of interest rate pegging after the financial crisis of 1983. The m.ore general question of whether the nominal interest rate rather than the quantity of money or credit should be a target of monetary policy is discussed later. Until the 1985 reform, monetary policy was on the, whole highly accommodative to inflationary shocks. The government tried to affect interest rates, either directly for specific kinds of preferred credits or indirectly for free credit, by setting the interest rate on the loans the central bank extended to the commercial banks. During some periods the government tried to control quantities directly by means of credit ceilings. Such a highly segmented system inevitably leads to wide dif- ferences between short- and long-term interest rates as well as among the short rates charged to different users of credit. These differences in turn cause considerable distortions in the allocation of money and capital throughout the economy. In addition, direct control of a large segment of credit implies that the conventional role of monetary pol- icy in stabilizing prices and the balance of payments is confined to a relatively narrow part of the credit market. Limiting the stabilizing power of monetary policy allows large fluctuations in both quantities and interest rates in that market. Under these adverse conditions the stabilizing role of monetary policy is hampered further, because the Monetary Policy Rulesfor a Small, Open Econmy 27 excessively high marginal interest rates that result invoke public criti- cism and exert pressure on the central bank to reduce rates and thus become even more accommodative. The July 1985 stabilization program enabled the government to initiate reforms in the operation of monetary policy, starting with reforms that substantially reduced the budget deficit. (The reduction in itself loosened the pressure on monetary policy to be accommoda- tive.) An outward manifestation of reform is new legislation that no longer allows direct government borrowing from the central bank except for a temporary loan that does not exceed 1.6 percent of the annual budget and that must be repaid by the end of the fiscal year." The only loophole that remains is the government's ability to borrow abroad and sell the proceeds to the central bank. The stabilization of bond prices has also been terminated. Although a substantial portion of directed credit to exporters remains, it is now financed by bank borrowing on international markets, with interest rates charged close to international rates. Note that this method of financing still subsi- dizes borrowers in the export sector as long as domestic interest rates remain high. Another element in the reform of monetary policy has been a shift from rate fixing and ceilings to the control of monetary aggregates to unify interest rates for short-term credit across the economy. The greater flexibility in the money market results from a new, simplified structure of central bank loans to the commercial banks; more of ehese loans are now sold by open auction rather than at predetermined prices. Likewise, there has been a return to the widespread use of short-term government bills (those that mature in one to twelve months) sold by auction and which the central bank can at times repurchase before the date of maturity. Other advances have been a gradual reduction of reserve requirements on demand deposits -to about 20 percent and unification of reserve requirements across dif- ferent types of fixed-term deposits. The balancing of the government budget also implies that the gov- ernment no longer expands its long-term, net, indexed debt and that an increasing portion of its existing debt, which is recycled, has become tradable on the secondary market, thus facilitating open mar- ket operations by the central bank. All these changes? which would not have been possible had large budget deficits continued, permitted gradual loosening of the rather rigid monetary system and unification of interest rates. Quite a few problems remain, however. As long as domestic inflation is higher than inflation in the countries of Israel's main trading partners-by 1988 Israel's annual inflation rate was still 16 percent, whereas it was some 3 to 5 percent in the trading partners' countries-ceilings on 28 Michael Bruno foreign credit cannot be reduced. An attempt to introduce a flexible capital import tax instead has failed because it could only be applied to a very narrow tax base. As a result, differences remain among dif- ferent types of external and domestic credit. Another problem is the wide spread that persists between the inter- est rate that banks pay on time deposits and the rates that they charge on their credit to the public. This spread is caused by the highly concentrated nature of the Israeli banking system, which is conducive to oligopolistic practices. Another contributing factor is that the gov- ernment still controls the quantity and price of a large portion of the banks' borrowing and lending operations. Bank profits from regu- lated lending are very low, thus motivating the banks to compensate themselves on the "free" portion of the market, where they charge excessive margins. The authorities hope to alleviate this problem through further reform of the financial system, which would include the gradual abolition of directed credit, further reduction of the gov- ernment's role in the capital market, and an increase in the degree of competition in the banking system. Monetary Mhrgets and hols under Normal Conditions The previous section covered some of the structural problems that imit the effectiveness and independence of monetary policy when government deficits are high, or there is a need to achieve multiple objectives, or both of these conditions exist. Let us now return to some of the issues that have to be considered under more normal circum- stances, assuming the existence of a relatively well-functioning money market and a reasonably balanced government budget. MonetaLry policy is only one of a set of instruments (such as fiscal policy and incomes policy) that authorities use to achieve multiple objectives for employment and growth, current account balance,. and price stability. We shall now assume that other instruments take care of employment, growth, and current account objectives, so that the main task of monetary policy (including exchange rate policy) is to achieve stability in prices and the balance of payments. In principle, the tools at the disposal of the monetary authorities are transactions in domes- tic assets that they hold or create, transactions in foreign assets (buying or selling NFA), and administrative measures in either of these fields (for example, fixing domestic credit ceilings or handling foreign exchange controls). One important choice is selection of a key anchor for the price system. Is it primarily the exchange rate or is it some monetary aggre- gate? In a small, open economy the relationship between the general price level and the exchange rate is considerably tighter than between Moneta7y Policy Rulesfor a Small, Open Economy 29 money and prices, primarily because import prices play a large role in the input-output system. Any increase in the cost of imports (say, through a devaluation) quickly feeds into the general price level, which is further enhanced when there is also some formal linkage between another important cost element, wages, and the price level. For this reason there is an advantage in anchoring the price level primarily to the exchange rate and using incomes policy to keep a wage level that is consistent, given the nominal exchange rate, with the required real exchange rates determining international competitive- ness and the current account. What this means, however, is that an important potential tool of monetary policy-transactions in foreign assets-is subjugated to the pegging (or stabilization) of the exchange rate. In other words, the central bank buys and sells foreign exchange at a predetermined price and thus loses control over NFA in its bal- ance sheet. But in that case, transactions in domestic assets. which affect domestic interest rates, must be conducted in a manner that supports the exchange rate and sustains a reasonable level of NFA. If, for example, the monetary authority is tempted to maintain too low an interest rate when the public expects a devaluation, capital will flow out of the country, the exchange rate will have to be adjusted, and its credibility as an anchor will be lost. As the balance of payments fluctuates more under a pegged- exchange-rate regime, the government must hold larger foreign exchange reserves than it would otherwise need. Another conse- quence of the choice of a pegged-exchange-rate regime is the stronger bias in favor of keeping foreign exchange controls. Such controls may be required anyway as long as the internal reform processes have not been completed (see Bruno 1988), but the controls will introduce other distortions in the economy. On balance, however, the choice of the exchange rate as a key anchor for the price system and the use of monetary policy to support it have considerable advantages.9 The next issue that comes up in the practical planning of monetary policy is the choice of the monetary aggregate that is most suitable as an intermediate target for policy. The choice of that target may differ from one country to another and within one country over time. An important first step is to choose the monetary aggregate that repre- sents transactions demand for money and that bears a relatively stable demand relationship with respect to nominal income or product. The aggregate used in Israel, for example, is an Ms figure, comprising ordinary means of payment (MI) and time deposits, in both domestic currency and foreign-exchange-denominated terns (Patam accounts). Assuming a relatively stable relationship with respect to nominal income, the planned growth rate of the chosen nominal aggregate over any period of time should then approximately equal the sum of 30 MiciarLt BnJino the growth rates of the targeted price level and of rcal income (or producL). Considering the basic monetary identity for the opcn econ- omy (M = DC + NMA), the authorities could then use a domestic credit (DC) aggregate as their intermediatc target, given a specified level of expected reserves (NPA). Using the Israeli examplc again, having cho- sen a definition of Ms and a planned level of reserves based on the expected figures for the current account and long-run capital flow, the diffcrence-the net domestic credit (NDC)-is then targetecl.14 What would be the implication of holding onto. a DC or an NDC target? Suppose the current account deficit increases unexpectedly, leading to a fall in NFA in the hands of the central bank. If the authori- ties attempt to keep to a predetermined NDC target, the money supply will fall (remember that M = (N)DC + NFA, thus AM = ANFA in this case). The public, which continues to demand a given quantity of money for transaction purposes, will then have to sell bonds, thereby driving down their price, that is, raising the interest rate. This in itself may activate a capital inflow and will attenuate the fall in reserves. In addition the rise in the interest rate will slow down domestic demand, cause a real depreciation (if prices are flexible downward), reduce the current account deficit, and thus bring about an upward correction in foreign exchange reserves (NFA). A similar analysis applies to the opposite case of an unexpected increase in the current account sur- plus leading to a rise in NFA. The example above shows that using an M aggregate as a target instead of (N)DC is an inferior procedure if the objective is to stabilize foreign exchange reserves. The relative disadvantage of keeping M stable is particularly clear when a devaluation is expected, leading to a reduction in M and a fall in reserves. Expanding M in this case (that is, following an accommodative policy) will keep interest rates low, caus- ing reserves to fall further and threatening the stability of the exchange rate.- The strict prevention of change in NFA, however, will cause greater fluctuations in the interest rate than will the stabilization of NDC. Should policymakers be indifferent to changes-that may take place in the interest rate? Too strict a maintenance of credit limits may at times cause such steep increases in the interest rate that the fall in real production may outweigh the possible gain in short-term stability. Obviously, this is a problem of weighing one bad outcome against another. In practice, even if monetary policy follows certain quantita- tive targets, policymakers should consider its implication for interest rates. By contrast, an inadvisable move would be to go to the other extreme of adhering to strict interest rate targets (which was a policy followed in Israel, for example, during- much of the high-inflation period). Adhering to strict interest rate targets has the great disadvan- Monetary Policy Rules fora Small, Open Economy 31 tage of releasing control over the quantity of money and credit and thus of giving up the stability of monetary targets. An intermediate course to follow is to set a narrow band for the quantitative targets and also keep some maximum and minimum interest rate constraints in mind (these should not be published, however). Strict adherence to quantitative targets may not work simply because the transaction demand for mnoney tends to be unstable. Financial innovation and other unforeseen disturbances keep the velocity of money (the ratio of nominal income to the stock of money) from being constant. This outcome has been* borne out by the recent '..perience of several advanced industrialized countries. The planning of monetary policy can be conducted on an annual basis and on a quarterly and monthly basis, right down to a weekly correction. Annual targeting should be consistent with the annual planned government budget. Quarterly adjustments have to be made in view of changes that take place either in actual fiscal balances or in the output, price level, and balance of payments data. Monthly and weekly financial programming would usually concentrate on the analysis of the major components of the monetary base, or high- powered money (H). These components include two factors that can be taken as exogenous to monetary policy in the very short run, namely, government net injections of liquidity and net sales of foreign exchange. A third component of the monetary base is that which the central bank itself plans to inject into the banking system, either through a monetary loan (or a "discount window,' for which the inter- est rate and fine structure can be varied at will) or through the sale and purchase of government short- and long-term securities. By mak- ing assumptions about various monetary multipliers (for example, required reserve ratios) one can then try to compute the implied changes both in monetary aggregates and in interest rates that the banking system is likely to charge. The shorter the planning period, however, the more unstable are the multipliers and the more attention is given only to the movements in the monetary base. Monetary Policy in Practice: Some Recent Israeli Examples The functioning and role of monetary policy is illustrated by examin- ing the Israeli economy, starting with the July 1985 stabilization program. The stabilization program ended a twelve-year period of deep crisis: two- and (later) three-digit inflation, recurrent balance of payments problems, and very low growth. At the center of the program was a cut in the government budget deficit from 12 to 15 percent of GoP to a virtual budget balance and a freeze on the exchange rate (after an 32 Michaet Bruno initial devaluation of 20 percent), on credit, and on prices, on :ondi- tion that the trade unions agree to a temporary wage freeze. EIventu- ally the program also included an agreement on wage policy. The program imposed a series of supplementary measures on the financial and capital markets, including abolition of the liquid foreign- exchange-linked accounts (Patam accounts). As figure 2-2 shows, the monthly inflation rate dropped dramatically from 15 to 25 percent to some 1 to 1.5 percent and was still at that level at the end of 1988. We will now consider the monctary component of the policy package as well as the program's immediate aftermath (for additional details see Bruno and Piterman 1988, on which some of the following discussion is based). Apart from the removal of the initial source of inflation-the gov- ernment budget deficit-an important part of such a stabilization plan is the determination of one or more nominal anchors and secur- Figure 2-2. Interest Rates and Inflation, Isael, 1984-88 28 24 4 Spla 1s 1* Intcmt ratr '4 . 20 1984 1985 1986 ;-1987 1988 Note. Thc intcret r2ais ic fficstivccost ofrovcdrafRacfidciis inlocal currcncyL a. Consulmncr pricc indcx. -.Sonurce Bank of Isracl dam 16~~ ~ ~ ~ ~ : - .- - . .. Monetary Policy Rulesfor a Sall4 Open Economy 33 ing them through social agreements and appropriate policy measures. The main anchors chosen were a combination of the exchange rate and the wage rate (plus a temporary price freeze), with monetary policy as an auxiliary anchor. Under sharp disinflation, the demand for both ordinary means of payment (MI) and also for the more gen- eral aggregate of domestic-currency-denominated unindexed assets (M2) is bound to increase quite rapidly, and using any of these as nominal anchors would be a mistake (this was apparently done at the early stages of the Argentine Austral Plan). Figure 2-3 shows that both Ml and Ml rose from mid-1985 to almost the end of 1988, whereas Patam deposits, represented by the difference between the Al3 and M2 curves, declined. For the stabilization program, the authorities pre- ferred to rely on the overall volume of bank credit. Control-over the level of credit, however, was indirect (in the preceding years, the Bank of Israel had controlled the price of credit rather than its quantity). A proposal to set credit ceilings for each individual bank was not Figure 2-3. Nonindcexd Financial Assets and Resident Foreign Currncy Deposits (Pate Accounts), Israel, 1984-88 4 3.5 _ se, Total assets 2.5 / Nonindexd assets ~s 2 1.5 0.5 Means of payment 0 1984 1985 1986 1987 1988 Note: At Dcember 1984 prices; montly averagc for quarters. Residcnt foreign cuny deposits (Pata accounts) arc reprcsentcd by dth difierence between the total and the nonindcxed assets a. A billion is 1,000 illion. Soul= Bank of Israel data 34 Michael Bruno adopted at the time; instead, the Bank of Israel used the intercst rate charged on the monetary loans to the commercial banks as an indirect means of trying to achieve an overall credit restraint target on the commercial banks. This target was defined as the level of credit that accommodated the initial price shock minus a cut of 10 percent. Fiscal and incomes policies were aimed chiefly at an initial reduc- tion of privaLe disposable income. The public tends to offset such policy by borrowing, especially if it perceives the policy as temporary. The steep decline in the expected return to foreign currency (after the exchange-rate freeze) was anticipated to lead to higher capital inflows. Also, a sharp improvement of bank liquidity was expected as indexed bonds and the Patam deposits (on which the reserve requirement was 100 percent) were converted into domestic currency deposits, which had lower reserve requirements. To prevent credit expansion under these circumstances, the Bank of Israel raised reserve requirements on domestic currency deposits and maintained an extraordinarily high interest rate on its discount win- dow loans. The quantitative target on credit was indeed achieved, but at a cost of steep real interest rates, which, in annual terms, reached an average of 168 percent on overdraft facilities during the first six months after the program's initiation. That nominal interest rates were reduced only slowly in the first few months partly reflected the monetary authorities' initial lack of trust in the speed of success of the sharp disinflation strategy (see figure 2-2: the upper curve represents the nominal interest rate on overdraft facilities and the vertical height between the two curves gives some measure of the ex post monthly real rate). These high real interest rates caused undue hardship to the business sector. Whereas high real interest rates are necessary in the course of sharp disinflation, they probably need not be that high. In any case, the nominal (and real) interest rate was gradually lowered during the second half of 1985 and until March 1986, even though credit increased in excess of the target. By that time a boom in demand and a deterioration of the private sector's current account, together with a. rapid increase in credit, caused the Bank of Israel to raise the prime rate by 3 percent a year. In retrospect the choice of commercial bank credit as the target of monetary policy at the time of rapid disinflation may have been inap- propriate. As argued above, with a fixed-exchange-rate regime, the authorities must pay more attention to the total domestic sources of monetary expansion: credit to the government and to the public from both the central bank and the commercial banks. Thus a concept like net domestic credit (NDC), which was introduced in 1987, could have been employed. Figure 2-4 shows the curve for M3 and for the NDC Monrtary Policy Rulesfor a Smnall, Open Economy 35 that is derivable from it after the subtraction of foreign exchange reserves (this is measured by the height difference of the two upper curves). The bottom curve gives the part of the NDC that goes to the private sector, and the difference in height between the two NDC curves is the part that goes to the government. As the figure shows, the latter contracted dramatically (along with the cut in the budget deficit). In January 1987 a 10 percent corrective devaluation was carried out as well as a new agreement between workers and employers to waive part of the cost of living adjustment. Apprehensive about impending price increases after the devaluation (forecasts of a renewed outbreak of inflation were widespread), the Bank of Israel raised nominal inter- est rates on monetary loans and the interest rate on commercial bank deposits with the central bank by I percentage point a month, announcing that these raises would be temporary. Afterward, it turned out that price increases were much lower than feared and that real Figure 2-4. Money Supply, Net Domestic Credit, and Net Domestic Credit to the ?rivate Sector, Excluding lkasury Bills, Isrnel, 1984-88 Netdoinesticcredit (NVDC) e. 2 . 1 NDC to te private sector 0 1984 1985 1986 1987 1988 Note: At Dcccmber 1984 prices; end-of-month balances;. M, = Cash +Demand +Tunc depoits+Rcsident foreign currcncy dcposits (Patam accounits). A=: = M- Forcign cxchangc rcscrvs. Souae: Bank of Israel data. 36 Michael Bruno interest rates rose again substantially (see figure 2-2 for the first half of 1987). The devaluation was successful in increasing export profitability, but it upset the equilibrium in the asset market. During the months following the devaluation there was a capital inflow of about $1 bil- lion' I (see the increasing gap between the two top curves in figure 2-4 after the beginning of 1987). This capital inflow began primarily as a result of the sharp decline in the cost of foreign currency credit after the devaluation, when the public realized that another devaluation was not imminent. This inflow tnight have caused monetary expan- sion, additional demand pull, pressure on domestic prices, and a rapid real, appreciation. On the one hand, the policy of absorbing liquidity prevented a rapid choking off of capital imports (because of the higher domestic interest rates), but on the other hand, it helped to restrain domestic demand and moderate the price increases following the devaluation. This moderation was highly valuable in reinforcing the public's expectations that price stability would be maintained. Interest rates were reduced again after a few months, and domestic credit was allowed to increase. In the second half of 1987, when domestic demand slackened off, the time was ripe for a sustained reduction in nominal (and real) interest rates, a process that was continued through most of 1988 as the economy went into a deepening recession. The tools used were gradual increases in the quantity of monetary loans and reductions in their cost and, in August 1988, a direct intervention by the Bank of Israel with the managers of the top commercial banks: the Bank of Israel called upon them to contract their high interest rate spreads and cut their lending rates by an additional 5 annual percentage points (on top of the 6 percent cut in the preceding six months).'2 The operation would have been successful had the hoped-for accompanying correction in the exchange rate been made. But because of the impending election and the subsequent long delibera- tions over the formation of a new government, neither this correction nor a negotiated agreement between the trade unions and the employers' association could take place. Expectations of an impend- ing devaluation intensified, and from July 1988 until the end of November 1988, some $2 billion in private capital flowed out. Mone- tary policy largely accommodated this outflow and allowed only a small increase in domestic interest rates for fear that too high an interest rate hike would deepen the recession and would be difficult to reverse quickly (given the oligopolistic structure of the banking sys- tem), once the expected exchange rate correction was carried out suc- cessfully. After another correction of the exchange rate in December 1988 and January 1989, the same amount of foreign exchange flowed Monelary Policy Rules for a Small, Open Econmy 3 7 back in, and the process of sharp reductions in interest rates was resumed. We have thus considered a few examples of monetary policy in action. Obviously, well-specified rules are one thing and practice is often another, but the main underlying principles are closely linked to our earlier general discussion. Conditions in different countries and in one country over time vary; therefore, monetary policy rules cannot be applied equally well under all circumstances. In the end, there is nothing better than one's own mistakes to learn from, unless one can also learn from others' mistakes. Notes 1. The wage rate is, under some circumstances, at least as important a nominal anchor for the price level as is money or the exchange rate. For policy, however, the wage rate is usually considered under a separate heading: wage or incomes policy. 2. This is usually termed MI, whereas M2 includes short-ternm government debt. M*. may also include such items as foreign-exchange-linked indexed deposits. 3. From the definition of M and from equations 2-1 and 2-2, we get M DD +CU= (R +BC) + (H-R) =BC+H=BC+-(CC+NFA). 4. In that case, the situation is the same as it would be in a closed economy; the term NFA does not appear, and we simply get M = BC + CC = DC. 5. If all the initial injection of money (AM) is done through r and not CC, the increase in credit will be AH(1 - r)Ir. In this way, there will be monetary expansion by the amount AM = AHIr, which is the monetary multipler effect. The formula will be a bit more complicated if individuals now also wish to readjust their cash holdings in relation to their bank deposits, but this rela- tively minor point can be ignored here. 6. Note that we have now extended the role of CC in the balance sheet of the central bank to include credit to the banks. 7. A recent, detailed study of monetary policy and- institutions in Israel (on which some of the discussion is based) is given in Cukierman and Sokoler (1988). For a good account of similar developments and problems in Italy, see Padoa-Schioppa (1987). 8. Italy introduced a similar reform in 1981. This was the discontinuation of the commitment of the Bank of Italy to act as residual buyer at treasury bill auctions. It was termed the "divorce" (see Padoa-Schioppa 1987). 9. With time, as reforms of internal financial, capital, and labor markets are sufficiently advanced, there may be a case for more equal sharing of the anchoringjob between the exchange rate and the monetary aggregates. This can be achieved by making the exchange rate more flexible, for example, making it move within a specified band. Such flexibility might reduce the fluctuations of NFA and give greater autonomy to monetary policy. *10. The NDC is a net number because one has to subtract from private and public bank credit that pan of the banking sector's liabilities that consist of bank deposits not included in the definition ofM3 as well as the net foreign liabilities of the commercial banks. These items are not included in our sim- plistic balance sheet representation, but can in practice be quite sizable. 38 Miclael Bruno 11. Throughout this book, a billion is 1,000 million. All dollar amounts are US. dollars and are currcnt, unless otherwise specified, 12. This action illustrates one important tool of monctary policy that has so far not been mentioned, namely, "moral suasion:' Direct talks with bankers or the governor of the central bank "raising his or her eyebrows" in public can at times be quite effectivc tools provided they arc used rarcly. Selected Bibliography Bruno, Michael. 1988. "Opening Up: Liberalization and StabilizationP In Rudiger Dornbusch and F. Leslie C. H. Helmers, eds., 7 te Open Economy: Tools for Policymakers in Developing Countries. Ncw York: Oxford University Press. Bruno, Michael, and Sylvia Piterman. 1988. "Israel's Stabilization: A Two-Year Review." In Michael Bruno, Guido di-Tella, Rudiger Dornbusch, and Stanley Fischer, eds., Inflation Stabilization: The Experience of Israel, Argentina, Brazi4 Bolivia, and Mexico. Cambridge, Mass.: MIT Press. Cukierman, A., and M. Sokoler. 1988. Monelary Poli 0. This relationship assumes that there is no collateral for bank loans, but the results do not change qualitatively if loans are partially collateralized. 8. Thus, for low values of r,,, R's lies below R'2, which in turn lies below R,l. Selected Bibliography Barandiaran, Edgardo. 1987. "Financial Liberalization in LDcs: A Review of Experiences:' World Bank, Economics and Finance Division, Washington. D.C. Cho, Yoon Je. 1986. "Inefficiencies from Financial Liberalization in the Absence of Well-Functioning Equity Markets:' journal of Monq Credit, and Banking 18(2):191-99. de Juan, Aristbula 1986. "From Good Bankers to Bad Bankers!" World Bank, Country Economics Department, Washington, D.C. de Melo,Jaime, andJames Tybout. 1986. 'The Effects of Financial Liberaliza- tion on Savings and Investment in UruguayY"Economic Deveklpment and Cal- lurda Change 34(3):607-40. Dfaz-Alejandro, Carlos. 1985. "Good-bye Financial Repression, Hello Finan- cial Crash!"Journal ofDvelopment.Economics 19(1-2):1-24. Dornbusch, Rudiger 1985. Special Exchange Rates for Capital Account Transac- tions. Working Paper 1659. Cambridge, Mass.: National Bureau of Economic Research. Edwards, Sebastian. 1984. The Order of Liberalization of the External Sector in Developing Countries. Essays in International Finance 156. Princeton, N.J.: Princeton University Press. Ferndndez, Roque. 1983. "La crisis financiera argentina:' Desarrolo Econ,miro 89(June):79-97. FIEL (Fundaci6n de Investigaciones Econ6micas Latinoamericanas). Various issues. Indicadores de coyutatura. Buenos Aires. Financial Reform and Liberc iuation 63 Fry, Maxwell. 1982. "Models of Financially Repressed Developing Economies" World Development 10(9):731-50. - ~. 1985. "Financial Structure, Monetary Policy, and Economic Growth in 1long Kong, Singaporc, Taiwan, and South Korea, 1960-1983" In Vittorio Corbo, Anne 0. Krueger, and Fernando Ossa, eds., Expori-Oriented Develop- ment Stmtegies. Boulder. Colo.: Wcstview Press. Galbis, Vicentce. 1979. "Inflation and Interest Rate Policies in Latin America, 1967-1976:' imFStaff Paper 256(2June):334-66. Ciovannini, Alberto. 1985. "Saving and the Real Interest Rate in LDCS./'Journal . of Developmenl Economics 18(2-S August): 197-217. Gurley,John. and Edward Shaw. 1960. Money in a Theory of Finance. Washington. D.C.: Brookings Institution. Hanna, Donald. 1986. "The Chilean Financial System, 1974-1982: The Regu- latory Rolc in Financial Collapse" Harvard University, Department of Eco- nomics, Cambridge, Mass. Hlarberger, Arnold. 1976. "On Country Risk and the Social Cost of Foreign Borrowing by Developing Countries:" University of Chicago, Department of Economics. IMF (Intcernational Monetary Fund). Various issues. International Financial Statis- tics. Washington, Dc. Jaffee, Dwight M., and Joseph E. Stiglitz. 1990. "Credit Rationing:" In Ben- jamin M. Friedman and Frank H. Hahn, eds., Handbook of Monetary Economics. Vol. 2. Amsterdam: North-Holland. Leff, Nathaniel, and Kazuo Sato. 1980. "Macroeconomic Adjustment in Devel- oping Countries, Instability, Short-Run Growth, and External Dependency:" Review of EconomticsandStatistics 62(2 May):170-79. Lipsey, Richard, and Kelvin Lancaster. 1956-57. "The General Theory of Sec- ond BesC' Redew ofEconozic Studies 24(63):11-32. McKinnon, Ronald. 1973, Monre and Capital in Economic Development. Tshing- ton, D.C.: Brookings Institution. . 1982. "The Order of Economic Liberalization: Lessons from Chile and Argentina:" In Karl Brunner and Allan H. Meltzer, eds., Economic Policy in a World of Change. Amsterdam: North-Holland. 1986. "Financial Liberalization in Retrospect: Interest Rate Policies in wzcsY Working Paper. Stanford University, Department of Economics, Stan- ford, Calif. McKinnon, Ronald, and Donald Mathieson. 1981. How to Manage a Repressed Economy Essays in International Finance 145. Princeton, N.J.: Princeton Uni- versity Press. Rosende, Francisco. 1986. "Institucionalidad financiera y estabilidad econ6m- ica' Cuadenos de &onoma 23(68):77-99. Shaw, Edward. 1973. Financial Deepening in Economic Development New York: Oxford University Press. Simons, Henry. 1936. "Rules versus Authority in Monetary Policy"Joumnal.of Political Economy 44(February-December):1 -30. Stiglitz,Joseph, and Andrew Weiss. 1981. "Credit Rationing in Markets with Imperfect Information "American Economic Review 71 (3June):393-4 10. Tobin, James. 1963. "Commercial Banks as Creators of Money:' In Deane Carson, ed., Banking and Monetary Sudies Homewood, Ill.: Richard D. Irwin. 4 Financial Factors in Economic Development m Rudiger Dornbusch and Alejandro Reynoso FINANCIAL FACTORS have been assigned strategic importance in eco- nomic development. But different factors have been isolated for dif- ferent regions. In Asia the role of an unrepressed financial market in mobilizing savings and allocating investments is emphasized. In Latin America the emphasis is on the role of inflationary finance-the scope for using deficits to enhance growth and, increasingly, the con- nection between high and unstable inflation and poor economic per- formance. This chapter reviews and contrasts the two approaches. Our analysis concludes that the evidence does not support vigorous claims for the benefits of financial liberalization. Equally, however, we note that the scope for inflationary finance is small and the risks are larger than commonly accepted. Growth in per capita income derives from the accumulation of physical capital and more efficient use of resources. Efficient use of resources is supported not only by the application of superior tech- niques but also by policies and institutions. Financial factors influence economic development by affecting the extent to which savings become available and the intermediation of these savings to invest- ment opportunities that bring the highest return. Economic history is rife with allusions to joint-stock companies as a decisive innovation in the implementation of capitalist production and distribution. Alex Gerschenkron and subsequently other authors have stressed the importance of finance. The Stanford School of John Gurley, Edward Shaw, and Ronald McKinnon has emphasized finance as a determinant of successful economic development, but although their views have become dogma, little evidence exists to support their pervasive claim. The Republic of Korea in the 1963-82 period experienced an average growth rate of output per capita of 4.8 percent-1.6 percent from capital accumulation and 3.2 percent from more efficient resource use. No growth-accounting exercise is available that would show us how much of this growth can be attrib- 64 Financial Ractors in Economic Developmnmt 65 uted to a favorable financial environment. The role of financial factors thus remains largely speculative. We argue here that financial factors are important, but probably only when financial instability becomes a dominant force in the economy. In this respect financial Factors, like the foreign trade regime, probably do not make much difference to die level of per capita GDP unless there is considerable distortion. This view is sup- ported by Denison's "guesstimate" (1985) that the cost of all trade restrictions in the United States in 1957 was no more than 1.5 per- cent of the level of gross national product (rNP). By implication, the impact on the growth rate was almost negligible. Of course, an extra 1.5 percent of GNP is well worth having, but in most cases it would be misplaced emphasis to put the trade regime or finance on a par with capital accumulation, technology, scale economies, or education. Although we believe that there is no significant difference in eco- nomic performance when a stable real interest rate is -1 percent instead of +2 percent, the financial regime can become a dominant determinant of performance when it deteriorates significantly. Argentina, for example, is sliding back as inflation and finance are increasingly dominating the economy, and the same is true of Peru. When hyperinflation takes over and foreign exchange crises dis- rupt the price system and shorten the economic horizon to a week or a month, normal economic development is suspended. Moreover, capital flight, which is difficult to reverse, puts saving outside the home economy. Attention should focus on these extreme cases and explore in greater depth the thresholds at which financial factors become significant, or even dominant, and the particular channels through which this occurs. This argument leads to a discussion of the limits of deficit finance, the risks of overexposure to external debt service, and variations in the ability of different couuntries to adjust rapidly and smoothly to changes in financial resources. -Superior growth performance, in this perspective, may reflect adaptability rather than financial deepening. This point is best illustrated by comparing Asia and Latin America in the period 1960-80 and in the 1980s (table 4-1). Finance does matter for the mobilization of resources, but this ordinarily accounts for little in changing growth rates. The more important fact is macroeconomic: poor finance leads to inflation and external bottlenecks, which in turn bring about restrictive macro- economic policies that slow down growth and investment. A pro- tracted period of poor macroeconomic policy in turn casts a shadow over the future, because it slows down or diverts abroad the supply of capital and the incentives to invest and innovate in the home economy. 66 Ruidiger Dorn&useh and Alejandro Rleynoso 'Able 4-1. Economic Performance in Asia and Latin America, 1960-80 and 1980-87 (average annual percent) Per eapita Financial Period and regon fnJlaion g,rot/i Investmendarn' deepeningb 1960-80 Asia 8.2 2.6 20.4 70.4 Latin America 27.6 3.3 21.5 25.8 1980-87 Asia 6.0 3.0 26.5 44.9 Latin America 102.3 -0.9 20.2 -8.3 a. Financial deepening is measured by the cumulative percentage changc in the ratio of M2 to cur. In the lO9s the change rcfers to 1980-86. Source: IMF (various issues). The Financial Repression Paradigm Financial repression is a major impediment to economic develop- ment. If growth is to come from investment, then three conditions must be met; namely, firms (or the government) must be willing to invest, savings must be available, and these savings must be chan- neled to those who plan to invest and who face the most attractive investment opportunities. The financial struct are and institutions can support or disrupt this process. A repressed system, especially in conjunction with high and unstable inflation, is said to interfere in a number of ways with development (see especially McKinnon 1973, Lanyi and Saracoglu 1983, and Fry 1988 for a discussioii and references). Savings vehicles are underdeveloped or the return on savings is negative and unstable, or both. There are two immediate consequences. First, the low and possibly negative real return on savings depresses the savings rate. Second, any money that is saved tends to go into self-finance, relatively unproductive assets (primarily inflation hedges), or foreign exchange. * Financial intermediaries that collect savings do not allocate these savings efficiently among competing uses. Because interest rates on loans are regulated, rationing occurs, which reduces the productivity of investment. * Firms are discouraged fom investing, because poor financial policies reduce the returns or make them excessively unstable. In particular, unstable inflation, price controls, and overvaluation of foreign exchange add to the risks of doing business and, as a result, depress investment in productive assets. Beyond depressing investment, an unstable financial business environment and the Firancial Factors in Economic Develof5ment 67 rationing implicit in a repressed system also induce the socially wasteful use of resources for rent seeking (see Krueger 1974). This is the case because financial repression creates a ready environment in which firms can secure large transfers from the public sector. A good morality tale is a story of sin and redemption. Korea, as we shall now see, shifted from repressed financial markets to financial reform, a shift that coincided with and was perhaps instrunmental in bringing about a dramatic change in economic development. The Korean Example Korea experienced low growth and increasing financial instability after the Korean War. In 1963-64 its performance deteriorated fur- ther. Sharply higher inflation, in conjunction with a ceiling on inter- est rates, reduced real asset returns. The ratio of M2 to GDP declined by almost 5 percentage points. John Gurley, Hugh Patrick, and Edward Shaw studied the situation and noted: Adequate mobilization of capital in Korea will require a major overhaul of the financial system. While financial reform is crucial to achieve the Korean objective of stable growth, our judgment is that tax reform will have to shoulder an even larger burden than financial reform to raise the ratio of domestic saving to national income within the coming few years. The financial system will need recuperation from past repression and abuse. This is no excuse for delay in financial reform. Indeed it only makes more necessary the need for financial reform now. (Reproduced in Cole and Park 1983, pp. 298-303.) Under the heading "Prerequisites for Financial Reform and Develop- ment," they proposed the following: * Persuade savers that they will not be taxed by inflation. * Maintain the equilibrium value of the foreign exchange rate of the won; do not allow it to become overvalued again. * Release domestic interest rates on deposits so that savers are induced to save, and in a financial form, and so that funds can be allocated to investment on a more rational basis. On the basis of their recommendations, the government of Korea introduced extensive fiscal, financial, and external-balance reforms. As shown in table 4-2, saving, investment, financial deepening, and growth all improved dramatically. The shift toward positive real deposit rates probably deserves much of the credit. 68 Rudiger DornbuscJs and Aljandro Reyvnoso Xble 4-2. Financial Indicators, Korea, 1960-74 (average annual percent) Category 1960-64 1965-69 1970-74 Real curb loan rte 31.1 44.4 28.2 Real deposit rate -0.7 14.3 3.6 M2.1GDP 12.3 21.2 35.0 National savings rate 4.9 12.9 17.4 Gross fixed investmentiG31 12.2 21.4 22.6 TaXCSIGDIP 9.3 12.0 13.8 Growth 5.5 10.0 9.2 Source: Cole and Park (1983); Bank of Korea data. Are There Lessons to Be Learned? That Korean economic performance improved sharply after 1965 is beyond question. The discussion remains open, however, as to whether financial reform was the chief or essential agent of change (see Cole and Park 1983). Skepticism focuses on the fact that high real deposit rates, to some extent at least, only moved resources from the curb market to the banking system. That resource allocation was improved as a result, or that saving increased in response to the higher yield on bank deposits, has not been shown. People are still debating the efficiency of the investment selection made by Korea's banking system in the 1970s, so whether the shift toward organized financial markets represented a clear improvement rather than only a redirection of savings flows is not clear. The large- scale investment in heavy and chemical industries in the 1970s was certainly facilitated by the mobilization of resources in the formal financial system. These investments were supported by credit sub- sidies and are now widely recognized as mistaken because of their low productivity. If this view is correct, financial deepening, which mobilized the resources for this mistake, must have had negative aspects. In addition, Korea's increased saving is a reflection of the governmenfs fiscal correction, real depreciation that promoted export growth, and guarantee programs on foreign borrowing that encouraged capitl inflows. The immediate question is what lessons to draw from the financial repression paradigm and the specific example of Korea. Should financial policy focus on generating significantly positive real returns on deposits, thus seeking to generate high rates of growth in the real size of the banking system? Are growth, financial deepening, positive real interest rates, and the productivity of investment Lightly correlated in historical and cross-sectional experience? The answers Financial Factors in Economic Development 69 are clearly no. Paying positive real interest rates on deposits is not a universal panacea for growth, as some of the literature on financial repression might lead one to believe. Only when financial instability becomes large and persistent do the connections between financial reform and growth performance become tight. Somze Key Relations Reconsidered The financial repression paradigm in some ways seems like supply- side economics-a kernel of truth and a vast exaggeration. Let us examine briefly the theoretical propositions and empirical evidence developed in support of the financial repression paradigm (see Fry 1988 for a review of the literture). Proposition 1: Positive real dposit rates raise the savings rate. It is wbll known from the theory of saving that the offsetting income and substitution effects of increased interest rates imply that their net impact on saving must be ambiguous. Within a framework of target saving, increases in real interest rates reduce the effort to accumulate the necessary savings. It is therefore surprising to find so strong a belief that higher interest rates mobilize increased saving. In the U.S. case, with the best data and innumerable attempts to document signs of the effect, confirmation of this belief has eluded virtually all studies. Evidence from other industrial countries points in the same direction: no discernible net effect. In the case of developing countries, the lack of data and their very poor quality make establishing the facts much harder. Fry (1988) reports a cross-sectional time-series regression of fourteen Asian countries in which the real deposit rate is a significant, although quantitatively unimportant, determinant of saving. Depending on the estimate, an increase of 10 to 25 percentage points in the real deposit rate is needed to raise the national savings rate by 1 percentage point] Giovannini (1985), by contrast, does not find a significant relationship between saving and real interest rates for Asian coun- tries. Reynoso (1988) finds evidence for a Laffer curve, with no significant effects when changes in the real interest rate are around zero. In some case studies, major stabilization programs do, however, appear to affect the savings rate. Some ready explanations exist. First, during a financial crisis saving is channeled into foreign assets through misinvoicing of trade. Accordingly, in these cases national account data easily underestimate true saving. Second, stabilization is associated with fiscal reform, which directly raises the national sav- ings rate. Third, durable purchases are recorded as consumption. Therefore, in a period of financial instability, a shift into durables- 70 Rudiger Dornmnach and Alejandro Reynoso and following stabilization, a sharp reduction in durable purchases- has the appearance of a dramatic increase in saving. In fact, however, true consumption (measured by nondurables and the services of durables) need not have changed much. Proposition 2: Financial deepening and growth are positively related. We already saw in table 4-1 that the correlation between growth and financial deepening as measured by the change in the M2/GDP ratios is not tight. Figure 4-1 shows a cross-section of countries; it is appar- ent that by judicious choice of sample any partial correlation can be generated. A first and important point is that financial deepening need not correspond to the M21GDP ratio. Deposits in nonbank institutions are an important outlet for financial saving and so is the money market. Between 1970 and 1987, the M2IGDP ratio in Korea was practically stagnant (41.3 percent compared with 39.1 percent), but the M3IGDP ratio doubled from 46.0 to 94.4 percent. The focus on M2IGDP ratios Figure 4-1. Financial Deepening and Economic Growth, CrossSection of Countries, 1965-85 7 6 0. t';~ 6 * ,-.@ .* S -2* -3 -5 -3 -1 1 3 5 7 Annual g7owth of de" -mg NWo: DeSpening is the raio of quai monecy to GNP. SburYc: IMF (various years).- 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~; Fitnacial Falors in Economic DeveLoJment 71 misrepresents the picture. Brazil has a market for financial assets that is, of course, not part of the narrow money supply. The shift to high inflation is reflected in a decline of M2 in relation to GDI' (from 19.4 to 12.9 percent during 1975-87), whereas a comprehensive measure of money, including the financial market, shows growth from 37.5 to 47.5 percent. The point of these examples is that differences in financial structure create an obstacle to any simple analysis of the relationship between financial development and economic growth. Propositian 3: Increases in real deposit rates raise investment. The theory here is difficult to pin down. The only immediate link to economic growth is the potential one discussed above: higher real interest rates for deposits raise saving and hence the equilibrium rate of invest- ment. An additional link, suggested by McKinnon (1973), involves the complementarity of money and capital: because investment proj- ects are lumpy, investors must accumulate their investment balances in the form of deposits until the required principal is reached. The more attractive the return on deposits, the more willing investors are to engage in the accumulation process. .We have seen that there is virtually no evidence that higher interest rates mobilize increased saving (and hence investment). As for the additional possible link, it is difficult to see how this view is very different from one that looks in a straightforward way at the effect of real interest rates on saving. After all, the economic choice is between consumption and saving. Econometric models that introduce the savings or investment rate into the equation for real money demand to test this theory are peculiar at best. Proposition 4: Increases in real deposit rates promote growth. Once again, the immediate link to growth is that higher real interest rates raise domestic saving and hence increase the available supply of resources for investment, a proposition we have already questioned. Two additional links can be considered. The first deals with exter- nal resources: eliminating ceilings on active and passive interest rates can bring about an inflow (or reflow) of external savings. Whereas large firms always have the possibility of borrowing abroad, this is not the case for smaller economic units. The removal of ceilings allows the domestic financial system to draw in resources that would not otherwise be available. We distinguish here between the rechan- neling of domestic savings from informal to formal financial markets and the net availability of external savings. The latter have a more difficult task finding their way to finance domestic investment through informal markets. Accordingly, financial reform can poten- tially raise external finance. The second link comes through the quality of investments. Experts often argue that a repressed financial system allocates savings ineffi- 72 RJudiger Dorntnssc/ and Alejandro Reynoso ciently, because rationing leads to the financing of investments that are below average in quality. This argument is suspect because eco- nomic agents have powerful incentives to merge with banks to seize the underpriced savings: they would not have an incentive to invest inefficiently. Indeed, much of the growth of informal markeLs reflects the laundering, for improved efficiency, of credits obtained from the repressed financial system. A Test for Linage Economists have tested the effect of increased real interest rates on the efficiency of investments by relating the incremental capital- output ratio to real deposit rates. Even though these relationships have been established frequently (see Fry 1988), what they reflect is not clear. Consider the ncoclassical growth model. We can write the growth rate of per capita income, y, as: (4-1) y = oe($t - n) where a is the distributive share of capital, ,l is the share of invest- ment in income, 0 is the capital-output ratio, and n is the growth rate. of the labor force. Which factor is influenced by financial repression, and how long does it take for financial repression to affect the parameter? If the capital-output ratio is raised, the avenge investment has been less efficient and the impact on growth could be significant. Let a = 0.7, 1 = 0.2, and n = 0.025. If q5 = 2, the growth rate is 5.3 percent; if 4, = 3, the growth rate is 2.9 percent. Thus, the productivity of the capital stock does make a large difference. This is not the right guide, however, to the benefits of financial liberalization. A more efficient allocation of investments has only an extremely gradual effect on the average capital-output ratio, taking decades rather than a year or two Moreover, the cumulative change may fall very much short of our example. To discern a growth effect, it is preferable to focus directly on financial stability. For a cross-section of forty-one countries, we used averages for the period 1965-85 to compare growth of per capita income with the investment rate and the rate of inflation. Specifi- cally, we were interested in the effect of high inflation on growth. The equation we used was y -1.67 + 0.0005Y + 0.15K - 0.016p R2 = 0.30 (-1.41) (1.35) (2.33) (-1.80) where y is the per capita growth rate, Y is the level of per capita income in 1965, K is the cumulative change in the capital-labor ratio, rFinanchia Factors in Ecenomic Deuelomnt 73 and p is an inflation dummy applied to inflation rates in excess of 20 percent. The regression supported the view that high inflation inter- [eres with growth. So far we have asked whether a liberalized financial system pro- vides more opportunity than a repressed one to mobilize resources for growth or to allocate them more efficiently. We have concluded that the cmpirical support for liberalization is episodic. Let us now consider the alternative perspective on financial factors in economic development, namely, the role of inflation and deficit finance. Budget Deficits and Inflation The discussion of inflation and its link to development finance in developing countries raises several questions. Why is inflation in Asia moderate, whereas in Latin America it is, at best, chronic and often acute? What are the disturbances and practices that set off an infla- tion process? What factors make inflation beyond a certain threshold an accelerating process? The experience of Latin America is one of inflation rates accelerating to 1,000 and more percent, even though the government deficits that are being financed are not much larger than those in Asia. Thus identifying the source of inflation differen- tials is important to obtaining a better understanding of the limits of inflationary finance and of the disturbances and institutions that make these limits especially tight. Inflation represents the interaction of four factors: * Deficit finance (which governs growth of the money supply) * Financial institutions (which determine the demand for money) * Shocks to the budget * The ability to react to these shocks by corrective fiscal measures The combination of these four factors may imply moderate and stable inflation, or it may imply near hyperinflation. Wh.ch of the two-low or extreme inflation-is, of course, critical for economic development. As we shall argue, high and unstable inflation leads to a drying up of resources available for development. This is because asset holders are unwilling to accumulate domestic claims, and firms are not prepared to accumulate productive assets in the inflating country. Inflation Policy Developing countries in Asia differ from those in Latin America in two major respects: their fiscal and inflation performnance and their distribution of income. In Latin America deficits and inflation are 74 Rudiger Dornbusch and Aljandro Reynosa chronic, whereas in Asia deficits tend to be limited to the govern- mcnts' ability to finance them in a noninflationary manner. The relatively equal income distribution in Asia contrasts sharply with the extreme inequality in Latin America. The difference in income distribution in the two continents influences the ability to achieve rapid adjustment of fiscal and real exchange rate positions when these are needed to avoid bottlenecks. This may not be the only reason for the differential ability to adjust, but it is ccrtainly an important one. It must be emphasized that differences in performance are not merely a reflection of dissimilarities in fiscal discipline. Korea, for example, ran largc fiscal deficits and experienced external shocks and debt service problems as recently as 1981. In this respect, its problems were no different from, say, Brazil's. The difference lies primarily in the adjustment to the shock. In one case the adjustment was startlingly rapid; in the other, the consequences of hyperinflation are still being acted out. Consider a simple model of the adjustment problem. We want to sketch a model of the extent to which a government offsets or dampens an inflationary shock. Let the government minimize a loss function, L: (4-2) L = - *)2 + A22 where r denotes the actual rate of inflation, r* denotes the historical rate, X is the marginal political cost of adjustment, and A is the adjustment effort. The actual inflation is the historical rate plus the shock (A) less the impact of the adjustment effort on inflation, or: (4-3) r = ir* + A - AI2, where # is a constant. Then the inflation rate under the optimal adjustment effort will be (4-4) 7r =7r* + ctA;a =XI(X + 2/2). We are interested in the coefficient a, which would differ across countries. The higher the marginal political cost of adjustment (X) and the less effective adjustment is in dampening the inflationary impact of shocks (that is, the smaller the ), the less adjustment effort will be supplied and therefore the higher the rate of inflation will be. This will tend to raise the inflation rate over time (r* will increase), and thus different countries' inflation performance will drift apart over time. The main task now is to identify the shocks and the channels through which they exert inflationary consequences. rinancial Factors in Erwnomic Devlo/nnent 73 Cliaracterisics of thte Infltion Process A high-inflation process has two characteristics. The first is that money creation finances a significant part of the budget deficit. The second is that indexation arrangements link current inflation to past inflation. MONEY CREATION. In the tradition of Mundell (1971), the budget def- icit is a fraction g of real income, and the demand for high-powered money is a linear and increasing function of inflation. A fraction ,i of the deficit is financed by creating moncy. This gives us a relationship between the growth rate of high-powered money (g) and the budget deficit: (4-5) =j3g(p + '7r) where p and vq are parameters of the velocity equation. In steady state, with a growth rate of outputy and a unitary income elasticity, we obtain an inflation rate equal to: (4-6) X= (fpg - y)I(l - Ag). The model makes three basic points. First, the link between inflation and the budget deficit financed by money creation is highly nonlinear. A minor increase in the deficit, when the deficit is high, raises in a major way the inflation rate required to finance the budget. Second, the financial structure affects the inflationary impact of money-financed deficits. The more sophisticated the financial struc- ture, the higher the coefficients p and v7 are and, accordingly, the higher the inflation associated with a given deficit is. Put another way, inflationary finance thrives on a repressed financial system. We return to this point in the context of financial liberalization below. Third, growth dampens the inflationary impact of deficit finance. A percentage point decline in the growth of income raises inflation by a multiple that is higher when the deficit is higher and velocity is more responsive to inflation. A major downward shift in real income growth can therefore be an important contributing factor to increased inflation. - Thbles 4-3, 4-4, and 4-5 show the revenue from money creation obtained in Asia and Latin America. INDEXATION ARRANGEMENTS. With indextion the adjustment of rel- ative prices becomes difficult. When wage indexation is adjusted at given intervals, the easiest means of cutting real wages is to allow infla- tion to accelerate. In this manner, over the indexation period, the real wage is eroded and hence its real value declines. But indexation 76 Rudiger Dornbusch and Atejandro Reynoso 'Tble 4-3. Seigniorage, Growth, and Inflation in Asia and Latin America, 1960-86 (average annual percent) Seigniorangew Growth Inflation Latin Latin Latin Period Asia America Asia Amfrica Asia Americ 1960-78 1.4 3.2 5.9 6.1 7.9 28.4 1979-86 1.5 4.5 4.9 2.5 9.3 116.6 a. The change in ligh-powercd moncy as a percentage of Gcr. Note: Figures in each category are income-weighted avcrages for six countries (listed by continent in table 4-7). Source: Fischer (I 982), updated by the authors. arrangements accelerate inflation when the periodicity of adjustments ;hortens. When an inflationary shock-say, a devaluation or the removal of subsidies-reduces real wages beyond a threshold, the response is often to shorten the indexation interval. For a given aver- age real wage, a cut in the interval to half doubles the rate of inflation. This shortening of adjustment intervals is an important driving force of accelerating inflation. Adjustment periods decline from annual to half-yearly to quarterly to monthly, and then the entire economy converges on the dollar. As every lagging agent in the econ- 'lAble 4-4. Economic Performance, Selected Countries in Asia and Latin America, 1960-87 (average annual percent) Period and country Inflation Growth Irnvestnunt/cDP NIICA"JDP 1960-80 India 7.0 1.2 22.4 -1.3 Korea, Rep. of 14.3 6.5 23.4 -8.0 Philippines 9.7 2.8 22.9 -2.2 Argentina 78.9 1.7 17.6 -0.1 Brazil 40.3 5.4 22.7 -1.8 Mexico 9.5 3.3 21.3 --1.0 1980-87 India 9A 3.1 24.4 -3.5 Korea, Rep, of 8.9 7.1 29.4 -0.9 Philippines 16.1 -2.1 22.1 -0.7 Argentina 279.3 -2.5 14.2 3.5 Brazil 153.3 0.4 18.7 3.8 Mexico 69.5 -1.4 20.2 5.5 a. The noninterest current account measured by net exports (excluding factor pay- ments) in the national accounts. Source: IMF (various issues). Financial Factors in cononoic Development 77 TIble 4-5. Seigniorage, Growth, and Inflation in Selected Countries in Asia and Latin America, 1960-86 (average annual percent) 1960-78 J979-86 aunlty Growth Ilfalion Sdpgiorr"g Growlh Inflatian Seigniorage" India 3.8 6.8 1.0 4.1 9.1 1.9 Korca, Rep. of 9.3 13.8 2.2 6.8 10.8 0.5 Malaysia 7.8 3.2 1.2 5.5 4.3 1.2 Pakistan 3.7 7.6 1.4 6.7 7.5 2.0 Philippines 5.7 8.4 0.8 1.2 17.9 1.1 Singapore 9.2 3.8 2.5 6.5 3.4 J.7 Argentina 3.3 57.2 6.2 -0.5 282.3 11.1 Brazil 8.3 36.5 3.2 4.2 131.3 2.4 Mexico 6.6 8.0 1.6 2.7 55.3 54 Peru 4.6 15.2 2.6 0.8 91.0 8.4 Uruguay 1.9 51.7 4.8 0.5 54.4 4.8 Venezuela 5.6 3.3 1.1 -0.2 12.6 1.2 a. The change in high-powered money as a percentage of GDP. Sourcer Fischer (1982). updated by the authors. omy shortens its lags to try to catch up with the average inflation rate, the average explodes. The nonlinearity of the inflation to the budget (reflecting the endo- geneity of the financial structure) and the shortening of indexation periodicity are channels through which inflation can accelerate once it reaches high levels. Intensifying the budget effect on inflation is the endogeneity of real tax revenue. Because the tax structure is less than fully indexed, high inflation erodes the real value of govemment reve- nues. This, in turn, increases the need for money creation to finance the deficit. Attempts to index taxation and speed up collection can help to dampen this process but have virtually 29 chance of offsetting the impact of a 200 percent inflation. The endogeneity of tax revenue is altogether inconsequential at rates of inflation of 20 or even 30 percent, but each becomes decisive at 100 or 200 percent. This helps to explain why so many countries in Latin America have recently moved to extreme rates of inflation. It remains to identify what disturbances initiate the process and why high inflation tends to become so unstable and explosive. Factors and Practices That Cause High InJlation Apart from an obvious lack of fiscal discipline, three factors have caused major inflation. Their importance is enhanced by their ten- dency to occur together. 78 Rudiger Dornbusch and Alejandro Reynoso DEBT SERVICE SH-IOCK. In the 1970s many developing countries bor- rowed heavily and as a result accumulated debt service burdens. In the early 1980s the debt shock halted the flow of lending. This interrupted the policy of paying interest on old loans by borrowing new money and automatically rolling over the principal. Debtor countries had to start making transfers abroad. In the budget the automatic financing of debt service by foreign loans was replaced by the need to finance at least part of the debt service domestically. Changes in taxes and cur- rent spending were unpopular; thus most of the adjustment took the form of either cutting public sector investment or financing the defi- cit domestically. To the extent that the deficits were financed by money creation (to avoid crowding out or bankruptcies associated with high interest rates), high inflation was the outcome. In many coutntries where some inflationary finance had been the rule, financing the deficit by creating extra money proved an express lane to extreme inflation. Bolivia is a case in point, as is Argentina. Over and above the problem of making external transfers from the budget, a debt service shock has a secondary burden. The real depreci- ation that is required to generate a trade surplus will raise the real value of external debt service in terms of the tax base. Thus for a country that has a debt service of 6 percent of GDP, a 20 percent real depreciation increases the debt service burden by 1.2 percent. The point is simply that a dollar of interest payment now costs more tax dollars. Thus depreciation occurs (except in cases where the govern- ment is a net earner of foreign exchange). FINANCIAL LIBERALIZATION. An immediate reaction to accelerating inflation is agitation in the financial sector for liberalization. Finan- cial repression worsens the social costs of inflation. Allowing banks to offer interest-bearing liabilities permits the financial system to per- form its intermediation task and thus minimize the costs of living with inflation. But liberalizing the financial system may imply reducing the government's revenue from money creation, which in turn means increasing the rate of inflation even more. Not liberalizing may entail costs, specifically, the possibility of capital flight, but the alternative of liberalization may also be perilous. The link between financial liberalization and inflation has been modeled in equation 4-1. Financial liberalization offers asset holders interest-bearing returns; nonbank financial institutions (financieras in Latin America) are allowed to offer checkable interest-bearing lia- bilities (the "overnight") and use the proceeds to hold short-tern commercial paper or government debt. Accordingly, financial disin- termediation occurs as deposit resources shift from traditional inter- Financial Factors in Economic Development 79 mediaries (the banking system) to the money market. The demand for the monetary base is reduced because banks lose deposits. A reduction in the demand for the real monetary base mises the rate of inflation consistent with financing a given budget deficit. Thus financial liber- alization raises the inflation rate unless there is an accompanying reduction in the budget deficit. Hierro (1988) has documented the quantitative importance of this effect. Of course, if liberalization also involves bankruptcies of financial institutions, as is often the case, the financing requirements also increase. Thus there is a tradeoff between financial liberalization and sei- gniorage; a period of fiscal crisis may not be the right time to intro- duce financial liberalization. Of course, that choice may not really exist: if a country fails to liberalize the financial market by offering interest-bearing domestic assets, the result is capital flight or dollariz- ation, with the same or worse consequences for inflation and intermediation. EXCHANGE LOSSES AND QUASI-FISCAL DEFICITS. A third major source of accelerated inflation is the widening in fiscal deficits result- ing from exchange losses on exchange rate guarantees or exchange rate operations, usually of the central bank, but also from the interac- tion of inflation and financial subsidies. Aittmpts to take advantage of secondary-market discounts on the country's external debt, by buybacks or debt-equity swaps add another important source of increased financing requirements (see Blejer and Chu 1988). In some cases in Latin America the increased financing requirements resulting from these operations amounted to several percent of GDP. In 1982-87 Argentina's quasi-fiscal -deficit averaged 1.7 percent of GDP. Much of this deficit stemmed from exchange rate guarantees given by the government in 1982. Demands for repayments of these debts- were hastened by the expectation of further depreciation. Exchange rate guarantees, rather than high interest rates, seemed the cheaper way at the time; in retrospect, they were the source of a mas- sive increase in inflation. Peru's quasi-fiscal deficit in 1985-87 averaged 2.1 percent of GDP (in addition to the regular deficit in the budget). Part of the quasi-fiscal deficit arose from large credit subsidies implicit in credits conceded at low interest rates. But the major part arose from multiple exchange rates with massive discrepancies between buying and selling rates for foreign exchange. Quasi-fiscal deficits, because of their sheer size and because they are financed by printing money that nobody wants to hold, are extremely in.flationary. In addition, they are an obvious mis- allocation of resources. 80 Rudiger Darn busel and Aljandro Rcynoso The Effects of Deficits and of High and Unstable Inflation Large budget deficits and their financing by high and unstable inflation have major effects on economic development. First, and most obvious, the appropriation of resources by the government reduces resources available to the private sector. If the public sector uses resources to finance investment, and if crowding out because of the inflation tax displaces primarily private consumption, this appropriation might well be conducive to development. Mundell (1971) has already questioned the connection between the reallocation of resources and inflation in the discussion of inflation taxation for growth. The Latin American experience of the 1980s highlights the narrow limits to inflationary finance and the dramatic costs when inflationary finance goes wrong. These remaining costs arise primarily from capi- tal flight and the misallocation of resources as a result of uncertainty. Capital flight and Dollarization Financial repression and high inflation together create an atmosphere in which asset holders seek protection by holding dollar-denominated assets, if possible, or shifting their assets abroad. The timing of a wave of capital flight may well be linked to an obvious overvaluation of the exchange rate as, for example, in Argentina in 1979-80 or in Mexico at the end of the six-year presidential terms, as shown in figure 4-2. Even without such a trigger, a history of large negative returns on assets produces capital fligb., Figure 4-3 shows the cumulative perfor- mance of a deposit in Argentina (translated at the official exchange. rate) in relation to a US. deposit and makes the same comparison for Mexico. Clearly, Argentina does not offer a favorable long-run finan- cial return, and the same has been true for Mexico during the past decade. Steady capital flight is the inevitable result. A recent estimate shows the extraordinary size of capital flight from Latin America: from 1975-85 the cumulative capital flight from Africa is estimated at $28.5 billion, from Asia $18.3 billion, and from the western hemisphere $106.6 billion (Deppler and Williamson 1987). To judge the size of capital flight, note that total Latin American debt in 1987 equaled $300 billion. Allowing the banking system to offer dollar-denominated (or dollar- indexed) deposits provides an alternative to the actual shift of assets abroad. Available data on the dollarization process in Mexico and Peru give us insight into the dynamics. The shift into dollar deposits is not a once-and-for-all process triggered by a dramatic event. On the contrary, the shift can be approximated by a traditional portfolio choice model based on relative rates of return and dynamics that are repre- Fxgure 4-2. Real Exchange Rate, Mexico, 1970-88 (1980-82 = 100) 120 110 100 g0 80 '70 60 50 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 Soume: Morgan Guaranty darn. Figure 4-3. Cumulative Return on a Bank Deposit in Mexico a-d .Argentina in Relation to a U.S. Deposit, 1960-87 (deposit= 1.00) 1.9 1.8 1.7- 1.6 1.4- 1.3 1.2 1.1 ~0.9 ~0.8 0.7 0.6 Agnta 0.5 0.2- 0.1 0 1960 1970 1980 198'7 Smrnz. imP f-~rious issucs). 8) 82 Rudger DombafrlL and Aleandro Reynoso sented by the logistic process. This point is illustrated in figure 4-4, which shows the evolution of dollar deposiLs as a share of total deposits in Peru (for a further discussion see Dornbusch and Reynoso 1988). Dollarization is not an instant reaction to the slightest policy mistake. On the contrary, there is substantial inertia in asset holdings. But once people learn to shift their assets into safer, dollar-denominated deposits, bringing about a reversal is difficult. A return to moderate rates of inflation is not rapidly rewarded by a complete reflow into local currency assets. When dollar deposits are not available at:d when economic and polit- ical instability is pervasive, the response is a shift into foreign currency assets in the form of currenc r or to real and financial assets located abroad. For example, in Peru, where access to new dollar deposits was eliminated, the large divergence between inflation and depreciation on the one hand and the return to deposits on the other made capital flight irresistible; the depreciation rate averaged 57 percent percent per quarter in 1987, whereas the deposit rate averaged only 5.4 percentl No reliable estimates of foreign holdings of US. currency exist, but Figure 4-4 The Process of Doliarization in Per, 1976-5 (percentage of total quasi moncy) 80 70 60 II so ~40 30 20 I0 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 *Senmce: Central Bank of Peru data. Financial rbctors in Jcotomic Development 83 some indication of foreign deposits in the banks of industrial countries is available. Table 4-6 shows the large size of dollar holdings in countries that have experienced financial instability. Table 4-4 illustrates an interesting difference between Argentina and Brazil. Brazilian capital flight, until recently, was relatively moderate because the domestic financial market was allowed to adapt. The exist- ence of a relatively indexed short-term money market prevented the massive flight of capital that occurred in other Latin American coun- tries. But even in Brazil a form of capital flight was apparent in the shortening of maturities in financial markets to the point where today the entire public debt has a one-day maturity. The next step, increas- ingly apparent, is flight from the overnight market to the dollar. Governments that face the risk of capital flight must make a strategic decision: Should they contain the flight by high interest rates on domes- tic assets? Should they create dollar-linked domestic assets (for exam- ple, Maex-dollars)? Or should they continue financial repression and attempt, even with little success, to contain capital flight by controls? A number of considerations bear on the choice of policy, most obviously, whether controls could actually stop capital flight-an issue that is viewed with almost uniform -skepticism. Of course, if capital flight can be prevented, the country does not need to acquire external capital, which would require a trade surplus, and thus more resources remain available for domestic absorption. In this sense, domestic dollarization is preferable, and even high interest rates might be, but they carry their own risks. Both create an easily accessible domestic substitute for assets that yield seigniorge, and in this wiy they increase the inflationary effect of a given deficit. Moreover, a rise in interest rates would enlarge the domestic deficit and thus aggravate financing requirements. The strategic question, then, ti whether the reduction in seigniorage is lower with capital flight ALble 4-6. Cross-Border Bank Deposits, Selected Countries, 1987 (doltars) Averageper Country capila hotdinis Argentina 277 Brazil 85 Egypt 65 Korea, Rep of 14 Mexico 225 Peru 89 Philippines 24 Venezuela 745 Source IMF (various issues). 84 Rudiger Donbusc/l and Alejandro Reynoso or with domestic dollarization. The answer presumably is that capital flight involves large transaction costs which could prevent some capital loss; therefore, a country may be better off accepting capital flight than instituting dollarization. Furthermore, donmestic dollar deposits also create the risk that if a major depreciation is required at some point, the banking system is likely to suffer, which may lead to a tendency to over- value the exchange rate. Mexico's experience with domestic dollar deposits illustrates these considerations. Brazil's experience illustrates another point: high interest rates are. not a good substitute for fiscal correction. High interest rates on a debt in local currency (or indexing of the domestic debt to goods or foreign exchange) can postpone the consequences of a continuing, large budget deficit but cannot make them disappear. The steady accu- mulation of domestic debt, and the shortening of maturities to what is virtually a spot market, creates a situation in which eventually the entire public debt is matched by interest-bearing, checkable deposits. Seigniorage has all but vanished, and the precariousness of the public debt is an invitation to a funding crisis, which arises when the govern- ment cannot roll over the debt. The Misallocation of Resources High and unstable inflation also has a major cost in terms of resource allocation. The inflation-induced distortion of the economy is not limited to the printing of new menus every day or month. The uncer- tainty about inflation and the policy reactions to accelerating inflation are major sources of distortions. When there is high and unstable inflation, productive factors are devoted to exploiting financial (and hence, by definition, zero-sum) opportunities rather than to innovating in production and trade. Firms' planning horizons shrink, and the possible introduction of controls to slow down the accelerating inflation forces economic agents into a defensive posture, in which investment in productive assets in the corporate sector becomes overly risky. Firms increasingly hold paper assets, and individuals overaccumulate foreign assets or consumer durables. The uncertainty that causes firms to hold paper assets rather than to invest in real resources translates into trade sur- pluses that finance capital flight or premature debt reduction. Conclusion Our discussion of the financing of economic development has empha- sized three sources: Financial rnctors in Economic Development 85 * External resources that can be tapped by a favorable investment climate or direct borrowing in the world capital market * Liberalization of financially repressed systems to enhance private saving * Development financed by public sector dcficits We have argued that evidence on the beneficial effects of removing financial repression remains open to challenge: the evidence is epi- sodic except when asset returns are significantly negative. The scope for deficit finance as an engine of economic development is extremely limited and extraordinarily hazardous. When overdone, inflationary finance acquires its own dynamics, which cin set back the develop- ment effort by a decade or more. Latin America today provides a striking example of the risks of budget deficits and of earlier excessive reliance on external finance. But to conclude from this experience that financial liberalization would have promoted high growth is a mistake. On the contrary, finan- cial liberalization (including the promotion of capital flight at the official exchange rate) in the face of a poor fiscal position continues to be a major factor in accelerating inflation and ihstability. Argentina is an example of a country altogether destroyed by excessive inflation, which has put an end to net investment and has led households to shift their assets abroad. Brazil is on that same path, and only Mexico has narrowly avoided it. Following a decade of financial instability, mobilizing resources for growth in Latin America is, of course, extremely difficult. The path that will return the region to rapid, long-run growth is awesomely orthodox: realistic exchange rates, balanced budgets, and a favor )le investment climate. Economists in the heterodox mode (and their progressive friends) might easily reject this advice, arguing that the, working poor cannot be made to bear the burden of a decade of mistakes, but the evidence suggests that without an early return to orthodoxy they will bear an even larger burden because capital is mobile, whereas labor is not. ' Appendix. The Mexican Financial System: From Financial Repression to a Money Market In Mexico in the late 1950s and the 1960s, a very regulated and spe- cialized financial system supported an impressive record of growth and price stability. By contrast, Mexico in the 1970s and 1980s faced the serious consequences of inflationary finance, large deficits, and a policy of inflexible nominal interest rates. Rapid disintermediation, mostly reflected in capital flight, reduced the financial system's ability 86 Rudiger Dorn bmsch anl Atejdtiro Rleynwso to promote macroeconomic stability, productive investment, and growth. The foundations of the Mexican financial system date back to 1924. In that year the Banco de M6xico was created as the institution in charge of regulating the circulation of currency, foreign cxchange transactions, and the rate of interest, and it was the only bank with the privilege of issuing paper money. During the 1920s and 1930s, com- mercial banks played a very small role. Instead, a number of spe- cialized national institutions were established in an effort to facilitate the movement of loanable funds to sectors for which the market did not provide credit. The General Law of Credit Institutions and Auxil- iary Organizations of 1941 set the new rules that would prevail, with only minor changes, for almost forty years. This law defined the finan- cial entities that would be allowed to operate and their sphere of action. Commercial (deposit and savings) banks were permitted to issue sight and short-term deposits denominated in pesos and in dol- lars and to make only short-term loans. Financieras were to capture long-term savings and to offer intermediate long-term credits. Some other institutions, such as insurance and mortgage companies, were subject to regulations similar to those governingfinancieras. The Banco de M6xico set deposit rates, which would remain fixed for long inter- vals. Under this scheme, there was no money market, and the stock market remained insignificant. Open market operations were not pos- sible, and the central government received its domestic financing f-rom the reserve requirements imposed on banks and Jmancieras by the Banco de Mexico. Between 1940 and 1960 the combination of inflation and low, fixed, nominal interest rates was reflected in slow financial deepening, dol- larization (see table 4-7), and capital flight. Inflation avenged 11.9 percent in the 1940s and 7.8 percent in the 1950s, while checking accounts did not pay any interest, and the rate on time deposits stayed bhble 4-7. Mexican Economic Performance and Financial Deepening, 1940-55 through 1980-87 (perccnt) Period Cvi' Injiation"t Deepm*h JIYawc 1940-55 5.7 11.6 24.3 12.0 1955-70 6.5 4.6 45.8 17.5 1970-80 .6.5 17.9 45.6 22.7 198047 -0.2 81.1 41.9 20.1 a. Annual average for period. b. Liabilities of the financial sectorlGDP (end of the period). c. IlYis the economy's investment ratio- Souvrce Banco de Mdxico (various years). Financial 'ewtors in Economic Detelopment 87 below 3 percenL during the cwo decades. Low resource capturing and high reserve requirements limited the availability- of financial savings for private use. In the 1940s, deposit banks were by far the most important credit institutions, whilefinancieras took more than a decade to consolidatc (see table 4-8). In the meantime, active movement of funds within the system was used to finance long-term projects. For example, funds were channeled from commercial banks to privatefinancieras, often by way of fiduciary institutiorns, and from there either to ultimate bor- rowers or to national credit institutions. From 1960 to 1971 financial deepening accelerated noticeably. Total resources captured by commercial banks andflnancerras increased at an average annual rate of 17.1 percent, whereas nominal income grew at only 10 percent. The financial system's total liabilities reached 45 percent of GDP by the end of the period. The co-mbination of mod- erately positive real interest rates and inflation and exchange rate stability contributed to the consolidation of the system. The average rate of inflation was 4.5 percent per year, the exchange rate remained fixed at 12.5 pesos per dollar, and the interest rate on time deposits' paid a constant 4.5 percent in nominal terms. : A central characteristic of this period was fiscal discipline. The bud- - get deficit was kept low, and an increasing proportion of public spend- ing went to capital formation. The combination of fast deepening and small deficits considerably reduced the need for inflation finance and external borrowing. The expansion of the. Echeverr(a and Lopez Portillo-governments (1971-82) put an end to fifteen years of price and exchange rate stabil- ity (see Cardoso and Levy 1988; Dornbusch 1988). Inflation came back without any adjustment in the financial 'tructure. It averaged 15.4 percent a year between 1972 and 1976 and 23.4 percent a year from 1977 to 1981, while the interest rates on'time deposits remained at 4.5 percent and the return onfinanciera bonds never increased above 13 percent. As a result, dollarization and capital flight reappeared in full, force'and financial deepening stopped. The banking sector was the liable 4-8.. Structure of the.Mexican Financial System, 1940-80 (percentage af total liabilities to the nonfinancial sector) l2st it2ulion 1940 1950 1960 1970 1980 Deposit banks 82.9 62.9 46.9 37.5 35.8 *- Savings banks 0.4 8.0 11.8. 16.3 19.6 - - Financieras 0.0 14:1 30.9 39.4 43.1 Mortgages' 9.9 7.4 7.2 5.6 1A Others - 6.8 7.6 3.2 1.2 0.9 Sourcc Banco de Mdxico (various years). 88 Rudiger Danmbuscds and Alejandro Rvynoso most scriously affected. Banking liabilities, which had come to repre- sent 34 percent of GDP in 1971, fell to little morc than 25 percent of GDI' in 1981. The combination of rapid growth, disintermediation, and incom- plete financial markets biased banks, private enterprises, and the gov- ernmnent toward borrowing abroad. Between 1970 and 1982, external liabilities expressed as a fraction of total indebtedness of the private and public sector jumped from 12 to 21 percent and from 56 to 75 percent, respectively. After the devaluation of 1976 the government realized that reform of the financial system was necessary. Some changes were introduced in the last months of 1977, but the experience of 1981 and 1982 proved them inadequate. The first innovation came in October 1977 with the creation of the petrobono, a government bond whose value was indexed to the dollar value of a certain quantity of oil. In January 1978 the government issued the first Certificates of the Federal.Trea- SUluy (CETs), which gave the government direct access- to'the money Market and, for the first timne, provided the central bank with an instrument that enabled it to perform open market operations. At first ctcrEs were offered to the.public at a fixed discount rate because of the. authorities' reluctance to allow the market to determine interest ratcs.: The first move toward an instrument with freely determined returns did not occur until October 1982, when the auctions of CErEs began.; By the end of 1982, cErr and petrobonos represented less than 3.5 percent of Mr'' * After the 1982 crisis, Mexico faced: the burden of a large external, * debt, a lack of foreign credit resources, and insufficient domestic sav- ings. A drastic.change:in the way the markets. operated became imper- ative. President de la.Madrid (1982-88) promoted major reforms to : the Central Bank Law, the General Law of Insurance Companies, and the Stock Market Law. As a result of these reforms the operations and instruments of the banking system were diversified and the Banco' de M6xico ceased to act as an intermediary between t,e federal govern- ment and financial institutions. Most important, the reforms gave rise: to a more complete and active money market,4as most regulations.on the issue and trading of'nonbanking financial instruments were relaxed. So far, the flexibility of the, new scheme has proved effective in promoting financial intermediation, slowing down capital flight, and accommodating the borrowing needs of the public sector. The Mexi- can .experience of the 1980s demonstrates that if inflation control isc' not perfect, it is essential to create a domestic money market, at least at' the wholesale level. Without the provision of a money market, capital flight becomes a major policy concern. Financial Pactors in Economic Development 89 Note I. Some observers note that the labor-capital distinction describes the choices for fiscal adjustment in an overly narrow fashion. They note that taxation of immobile land has as yet not bcen used on any scale to avoid accumulating heavy tax burdens on labor. Selected Bibliography Banco de MExico. Various issues. Indicadores econdmicos. Mexico City. Blejer, Mario, and K. Y. Chu, eds. 1988. Measurement of Fiscal Impact: Methodologi- cal Issues. Occasional Paper 59. Washington, D.C.: International Monetary Fund. Cardoso, Eliana A., and Santiago Levy. 1988. "Mexico:' In Rudiger Dornbusch and F. Leslie C. H. Helmers, eds., Tue Open Economny: Tootsfor oblicymakers in DevelopingCountries. New York: Oxford University Press. Cole, David C., and Yung-Chul Park. 1983. FinancWiat Developmnent in Korea. Cam- bridge, Mass.: Harvard University Press. -Denison, Edward F. 1985. Trends in American Economic Growth, 1929-82. Wash- ington, D.C.: Brookings Institution. Deppicr, Michael, andJohn Williamson. 1987. "Capital Flight: Ooncepts, Mea- surement, and Issues" In StaffSudkisfor the WorldEconomic Outlook Washing- ton, D.C.: International Monetary Fund. Dornbusch, Rudiger. 1988. "Mexico: Stabilization, Debt, and Growth:' Eco- nomic Policy 3(2):231-73. Dornbusch, Rudiger, and Aleindro'Reynoso, 1988. "The Dynamics of Dollar- ization in Mexico and Peru:' Massachusetts Institute of Technology, Depart- ment of Economics, Cambridge, Mass. . Fischer, Stanley. 1982. "Seigniorage and the Case for a National Money"Jour- nal of Political Economy 90(2):295-S13. Reprinted in Stanley Fischer, 1nd i- - Inflation, and Economic Polcy. Cambridge, Mass.: Mrr Press, 1986. Fry, Maxwell. 1988. Money, Interest, and Banking in Economic Development. Balti- more, Md.:Johns Hopkins University Press. Giovannini, Alberto. 1985. "Saving and the Real Interest Rate in LDCSCJou4nal -ofDevelopmentEconomics 18(2-3):197-217. Hierro, Jorge. 1988. "Financial Liberalization and Inflation" Massachusetts Institute of Technology, Department of Economics, Cambridge, Mass. IMF (International Monetary Fund). Various issues. International Financial Sat4is- tics. Washington, D.C. Krueger, Anne 0. 1974. "The Political Economy of the Rent-Seeking Society. American Economic Review 64(3):291-303. Lanyi, Anthony M., and R. Saracoglu. 1983. Interest Rate Policies in Developing -Countries. Occasional Paper 22. Washington, D.C.: Intemational Monetary Fund. McKinnon, Ronald. 1973. Mony and Capial- in Economic Development. Washing- ton, D.C.: Brookings Institution. Mundell, R. A. 1971. Monetary Theoy. Pacific Palisades, Calif.: Goodyear. Reynoso, Alejandro. 1988. "Saving and Real Interest Rates: The Laffer CurveC' Massachusetts Institute of Technology; Department of Economics, Cam- bridge, Mass. 5 -- Exchange Rate Policy: 0Options and Issues m Rudiger Dornbusch and Luis Tellez Kuenzler THIS CHAPTER discusses alternative exchange rate arrangements. We present the options, some country illustrations, and a firm set of rec- ommendations. Our chief conclusion is that economies with a prolif- eration of multiple rates or black markets waste resources. Whatever policymakers believe about their ingenuity and creativity in managing multiple rates, the attempt almost always falters on political grounds. The end result is a drain on the budget and a massive misallocation of resources. A streamlined exchange rate system must strike a balance between two considerations: microeconomic efficiency and macroeconomic stability. We argue that a dual-rate system (with a market-determined rate for capital account transactions,at a moderate premium over the official rate) best serves that objective. We start with a general discussion of the role of the exchange"rate in the economy. The next section discusses multiple rates. From there, we move to black markets and the associated problem of misinvoicing of trade flows. A comparison of unified flexible rates and dual exchange rates follows. The Role and Influence of Exchange Rates in the Economy Exchange rates (the price of foreign currency in terms of home cur- rency, say, pesos per dollar) are one of the key linkages between a small, open economy and the rest of the world. In a world of- barter, because there are no monies, there are also no exchange rates. In reality, every country in the world uses money, and most countries use their own monies.' Hence there are also exchange rates. The exchange rate links a country's macroeconomy to the rest of the world through the asset market and the goods market.- It also connects individual countries through the microeconomic channels of.the goods market 91 92 Rudiger Dornbusch and Luis Tellek Kuenzer and the asset market. Throughout the discussion below we refer to the home currency as the peso (pesosldollar). Macroeconomic Linkages Linkages between the rest of the world and the macroeconomy arise in two ways. In the goods market, the exchange rate establishes linkages between prices in the home economy and the given prices in the world' market. The higher the exchange rate, other things being equal, the higher the price of foreign goods in home currency. For a given level of domestic costs and prices, a higher exchange rate makes foreign goods less competitive in the home economy and makes the home economy's goods more competitive in the rest-of the world. The link- age to the goods market can be formalized in terms of the real exchange rate, R. The real exchange rate-is the ratio of domestic prices in pesos to foreign prices in pesos, or, in symbols, (5-1) R PYeP* where P is the domestic price level in pesos, e is the exchange rate (pesosldollar), and P* is the foreign price level in dollars. The real exchange rate measures competitiveness by showing the number of units of foreign goods required to buy one unit of domestic goods. An increase in the real exchange rate, or- a. real appreciation, means that it takes more units of foreign goods to buy one unit of domestic goods. Thus real appreciation is tantamount to a loss of competitiveness. Conversely, if the rtal exchange rate declines, we speak of a real depreciation. The home economy becomes more com- petitive because it now takes fewer units of foreign goods to buy one unit of domestic goods. Much of the discussion in this chapter deals with how the choice of an exchange rate regime can influence a country's competitiveness. For example, if domestic prices are rising because the government is pursuing an inflation policy, but the exchange rate is fixed, then equa- tion 5-1 will show real appreciation. Inflation under fixed exchange rates is tantamount to declining .competitiveness; sooner or later (mostly sooner) that leads to problems of financing a growing trade deficit. This raises the question of how a country should conduct its exchange rate policy. The first answer is to avoid overvaluation or a loss of competitiveness. 'But imagine that noneconomic considem- tions make a move in the exchange rate undesirable despite domestic inflation. Often a system of multiple exchange rates for-various sepa- : rate classes of goods seems to' bean answer; however, we will see that multiple rates have many shortcomings and are rarely desirable.- Excha7tge Rate Policy: Options and Issues 93 There is a separate macroeconomic concern with exchange rates. A. depreciation of the exchange rate tends to raise the domestic level of prices-that is, it is inflationary. This happens directly because the prices of imported goods in home currency rise when the currency is depreciated. But indirect channels also exist: domestic firms can afford to raise their prices when competitors' prices rise, and, in the labor market, workers may ask for wage increases when higher import prices (especially for food) raise their cost of living. Thus exchange rate movements tie in very strongly with inflation policy. Governments, which are naturally aware of this connection, would prefer to stop exchange depreciation so as to avoid the inflationary impact, but if they do so in the face of domestic inflation, they risk a loss in competitiveness. Once again, multiple rates often appear to be an answer. They appear to be an answer because they allow a competi- tive real exchange rate for some goods (say, the export sector) without risking the inflationary impact of import price increases associated with depreciation. As we will see, however, a policy that allows real exchange rates for different groups of goods to diverge widely inevita- bly leads to large costs, because it involves a misallocation of scarce resources., Often there is also an important impact on the budget. Specifically, if the government buys foreign exchange at a high price from exporters and sells it at a low price to importers, a budget deficit is inevitable. The budget deficit, in turn, needs to be financed by higher. taxes or by inflationary money creation. Thust exchange rate regimes are also linked to inflation. The asset market also has an exchange rate linkage. Domestic wealth holders must choose how to hold their wealth. They can hold real assets (houses and land) or claims on real assets (stocks of corpora- tions). They can hold domestic financial assets, such as deposits at banks or government bonds; or they can hold a range of foreign assets, from deposits in banks in Miami or Geneva to U.S. Treasury bills. The choice among assets will depend on the tradeoff between risk and return. One simple rule is clear: when foreign assets yield a higher return (in home currency) than domestic assets and are no more risky,. then many domestic residents will shift sooner or later (probably sooner) into foreign assets. That much is clear, but'how do we compare the return on a dollar deposit in Miami with the return on a peso deposit in Mexico? To make a comparison we have to translate the returns on the two assets into the same currency, which introduces the effect of the expected changes in the exchange rateover the invest- ment horizon. For example, if the interest rate in Miami is 25 percent a year in dollars and the pesoldollar rate is expected to depreciate at the rate of 150 percent a year, the peso return on a Miami dollar 94 Rudiger Dornbwsch and Luis TlUez Kuenzler deposit is 175 percent. if the return on peso deposits is only 160 percent, wealth holders would prefer dollar deposits. Domestic resi- dents would shift their wealth, at least partially, abroad. Conversely, if Mexican interest rates exceed the peso return on an investment in Miami, there would be no tendency for capital to flow out. In general, we can express the asset market linkage in terms of a critical condition for domestic interest rates. For investment in peso- denominated assets to be competitive, the peso interest rate, i, must exceed the dollar interest rate, i*, plus the expected percentage rate of peso depreciation, u. In symbols, (5-2) i>i*+-: If this condition is not met, but capital can move across borders in an unrestricted fashion, then a country would soon find that it loses reserves selling dollars to domestic residents who prefer them as an investment over the low-yield, domestic currency assets. The issue for exchange rate policy becomes whether and how domestic and foreign asset markets might be segmented, or how the exchange rate policy that is chosen for trade can be separated from the exchange. rate relevant for calculations of asset market profitability. Capital controls or dual-exchange-rate systems, as we will see, are a potential means of achieving this end. Mieroeconomic Likges- The microeconomic linkages of exchange rates involve resource' allocation. The issues here closely minor the macroeconomic issues already discussed. When the real exchange rate makes the economy highly competitive, resources are drawn into the tradable goods sector (the notion of what a tradable good is expands). Many goods become exportable, and fewer goods are imported. Goods that used to be imported are domestically produced' or even, become exportable. In the factor markets this is mirrored by a new allocation of resources. When: the real exchange rate is highly competitive, the economy is highly trade-oriented; employment of capital and labor in the export sector and in import-competing industries is high. Conversely, when the real exchange rate is highly uncompetitive, resources are-primarily employed in the home market, imports are high, and export produc-- tion is unprofitable. The distribution of income between different groups -or sectors: often -depends on the level of competitiveness. For example, if the country has a traditional export sector (say, agriculture or mining), then a very competitive exchange rate (meaning that the exchange rate is high in relation to wages) will make traditional exports extremely Ex&dtang Rate Pol: Options and Issues 95 profitable. Owners of the traditional industries earn rents, and workers perceive that their purchasing power is very low. Conversely, when the exchange rate is highly uncompetitive or overvalued, workers' wages have a very high purchasing power (if workers can find ajobl), whereas finns in the traded goods sector are unprofitable and hence cut back on investment. Because of these sectoral and income distribution issues, govern- ments often look for ways to avoid what they perceive to be adverse effects of a competitive exchange rate. They may introduce multiple exchange rates to tax traditional export industries or to raise arti- ficially the purchasing power of wages in terms of imported food. Policies of multiple rates, if pursued for a long time, can ultimately distort the use of resources and lead to low investment and investment in the wrong sectors. There is also a microeconomic counterpart to the asset market link- age. In economies in which domestic returns invariably fall, short of returns abroad, capital flight will take place. This may take time, but in the end people do get their money out-if necessary, by faking trade invoices. That means fewer resources are available for investment in the home economy. Moreover, those who' get their money out often do so at the expense of those who cannot (we will see examples of this below). Hence distribnation issues are important here, as well. Alternative Exchange Rate Regimes We can classify alternative exchange -rate regimes in terms, of two dimensions. The first dimension has to do with whether the crchange market is unified or segmented. In a unified market all transactions occur at the same rate; in a segmented market different transactions take place at different exchange rates. The second dimension con- cerns the supply of foreign exchange. In one: classification, it is: rationed to those who have specified eligible transactions, and recip- ients of foreign exchange are required to surrender it at a specified' price. The alternative is a system in which foreign exchange is freely - available at a specified price from the centl bank, and owners or recipients are not subject to a surrender requirement (see table 5-1). ]hble 5-1. Alternative Exchange Rate Regimes- ;bregn ,,xhinge mariet Foreign xhane supply Un#d Segmnd.- Rationedlrequisitioned I II: Unrestricted Iva a. In practice, IV does not exist as a system. 96 Rudiger Donbusch and Luis Rilez Kunzer Among developed countries, case III is the most common: exporters have no need to surrender foreign exchange, and importers can buy. it freely ill the market. There will typically be no difference (except in Italy or France, on occasion) in the treatment of trade transactions and financial transactions. We refer to such a system as unrestricted con- vertibility. In this case, the central bank may be actively involved in the market by fixing the exchange rate, may be totally absent with a flex- ible exchange rate clearing demand and supply, or may intervene, buying and selling to influence.the level of the equilibrium exchange rate. Most developing countries do not have unrestricted convertibility. Even highly efficient surplus countries such as the Republic of Korea maintain restrictions on payments as, indeed, did Japan in the 1970s. For most developing countries, exchange rate regimes tend to fall into the other boxes. The most straightforward is case L. Here the govern- ment draws up a list of specified transactions (an import list, an authorized transfer list, and an authorized capital outflow list) for which foreign exchange is supplied at a specifiedr te. Exporters have to surrender all foreign exchange they receive at the same specified rate. For unauthorized transactions, no foreign exchange is available (and hence a black market springs up). This exchange rate regime is a reflection of an overvalued currency: at the specified price, foreign exchange is in short supply; therefore, rationing and requisitioning are required. Although using a disequilibrium exchange rate may help a government equalize income distribution, the system misallocates resources: goods that do not make the import list cannot be obtained, iiand, if the list is poorly done, it may exclude, for example, essential intermediate inputs, notjust ""dispensable luxury goods. When the black market for foreign exchange (and the associated smuggling and faking of trade invoices) becomes a major issue, as happens in a system like case I, governments often move to a case mI regime. Or, they might adopt multiple exchange rates for commercial transactions and maintain the rationing and requisitioning of foreign exchange, or they might have a uniform rate for specified commercial transactions and another for capital account transactions and all those commercial transactions not taking place at the commercial rate. This is the system that prevailed in Mexico, for example, after 1982. The difference between these modalities' primarily concerns the question of whether the authorities administer each of the rates or whether they administer some and allow the exchange rate to find the equilibrium level in the remaining markets, perhaps with some intervention. Thus the seqmentation case will tend to be a mixed system, with an adminis- tered commercial rate, for example, and a free rate for all other. trans- actions. The case of unrestricted, segmented markets (case IV),: with a &ce/aange Rate Pholicy: Options and Issues 97 flexible rate for each separate market, is in principle possible but is not a reality. That case does noL exist, because governmcnts get into the practice of segmenting markets for the purpose of actively main- taining a disequilibrium rate for some particular class of transactions. There is another way of classifying exchange rate systems, based on whetlher the rates are fixed or flexible in nominal or real terms (table 5-2). Fixed nominal exchange rates (case I) arise when a country has a strong economic relationship with a particular industrial country. For example;- in most African countries that were former colonies of France, exchange rates are fixed in relation to the French franc. Sim- ilarly, Panama and several Caribbean islands have fixed dollar parities. Other countries maintain an exchange rate fixed in terms of some basket of foreign currencies rather than a single currency. The distinc- tion is important, because in a world where the dollar-franc rate moves wildly, there are clearly important exchange rate issues for a country that trades with both France and the United States. If that country fixes the rate to one currency, it has wild swings in compet- itiveness in relation to the other. Fixing an exchange rate in terms of a trade-weighted basket of currencies helps avoid some of these vagaries for countries that do not have an overriding relationship with one single industrial country. It does not, however, totally eliminate the exchange rate ':fficulties of developing countries that arise from the instability of exchange rates among major industrial countries. A fixed real exchange rate (case UI) is an exchange rate regime in which the nominal exchange rate is automatically depreciated at a rate that offsets inflation differences between the country and its trading partners.2 Fixed exchange rates are administered rates, whether fixed in nomi- nal or real terms. Flexible rates, shown as case II, are determined by- the market. A government may or may not influenice the rate through intervention or through its domestic monetary policy, but the basic point is that the rate is market-determined. Few developing countries have a fully flexible exchange rate. We now turn to specific policy situations and cases to examine, in greater detail, the costs and benefits of alternative exchange rate systems. Multiple Exchange Rates for Commercial Transactions In this section, we look at an exchange'rate system in which the gov- ernment adminiL ers specific exchange rates for different commercial transactions. We leave issues of capital flows to a separate section. Table 5-3 shows the Peruvian system of the mid-1980s in all its complexity. The system had two basic rates and many variations. One 98 Rudr Domnbsbc and Luis 7l.z ACisender lIble 5-2. Fuied and Flexible Exchange Rates Exchang rate mlied I'Ile Nominal I Real I was the "official rate" that applied to all "essential" imports and at least partially to other transaccions. The other was the "financial rate:' MosL transactions were undertaken at a mix of these two rates. Thus in 1987, for example, petroleum exporters had to surrender 55 percent of their foreign exchange at the official rate and 45 percent at the financial rate. In addition to the two basic rates, there was a super rate for financial transactions. The difference between the lowest and highest exchange rate was 60 percent in May 1987. lhble 5-3. Exchange Rate Regime, Peru, 1984-87 AppliCable exchWe tle n17saction 1984 1985 1986 May 1987 imports Essential 0 0 0 0 Other 0 0- F F F.ports Traditional From mines Small * 0 9515 45/55 35165 Large 0 9515 65135 55J45 -Petroleum 0 0 0 55145 Nontraditional 0 80120 P P Invisibles 0 0 0 0 Authorized transfers .0 0 0 0, F Capital New loans 0 0 0 F Bank accounts 0 0+3% O, * O,F,SS Repatriation n.a. n.a. n.a. S - Rates in effect (intisldollar) Official 5.20 13.95 13.95 15A2 Free 5.38 17.38 17.45 19.28 Premium 5.92 19.12 19.20 -21.21 : Special n.a. n.a. n.a. 24.70 Super special n.a. n.a. n.a. 28A0 n.a. Not applicable. - at 0 = Official; F = Free; P = Premium; S = Special; SS = Super special. Where transactions are undertaken at a mix of official and free rates, the ratio.of the two weightings-O (Official) to F (Free)-is given. Source :IMF files. Exchange Raie Policy; Options and Issues 99 Iwo points should be immediately obvious from the table. The first is that the average rate at which the central bank bought intis from importers is far below the average rate at which it sold intis to exporters. Thus, the exchange rate system had advcrse fiscal conse- quences for the central bank. The second is the large variability in the applicable rates over time. Figure 5-1 shows the result of thcse exchange rate policies (and the accompanying tariff and licensing policies): the real pricc of imports and exports in terms of domestic goods. These prices are calculated' (by the World Bank) as the averages, using all applicable exchange rates and tariffs for each of the separate commodity groups. Note that from 1985 to mid-1987,. the real prices of both exports and imports declined by more than 35 percent. In response to these price changes, the demand for imports increased sharply and export profitability fell off. As a result, a growing foreign exchange problem developed. In late 1987 the government started responding to the problem by increasing the real prices of both exports and imports, but the correction did not last long. Soon the policy of cheap imports resumed. The real prices for exports moved sharply up as export exchange rates increased ahead of domestic prices. Import exchange rates,.on the other hand, lagged far behind the domestic rate of inilation. Why Countris Use Multiple Rates The complexity of the exchange rate arrangements shown in table 5-3 immediately raises the question of why governments opt for multiple rates. Do multiple exchange rates achieve economic efficiency, or are there other reasons why governments use multiple rates? The answer is that far from bringing the economy closer to efficiency, multiple rates are likely to create waste, yet they are a way of affecting resource allocation and income distribution.. By interfering with free market prices a government can reallocate resources, direct resources into alternative uses, and affect the real incomes of different sectors and the factors of production. The basic instruments for redirecting resources are taxes, subsidies, tariffs, and qaotas, on the one hand, and special exchange rates, on the other. Anything that can be done with tariffs can be replicated with a special exchange rate. Thus, if an industry is to be protected against foreign competition (thereby giving the owners of resources in the industry. higher real incomes), a. tariff can do the joh but so can a high exchange rate for this particular type of good. Likewise, if the objec- tive is to raise workers' real incomes in terms of-a particular traded good (say, food), this can be done by a consumption subsidy. But it can 100 Rudiger Dornbadsh and uits Thlsz Ruenzler Figurm 5.1. RcIal ric of Exports and Imports, Pcru, 198-88 (1986 - 100) 1 10 100- 80 _ \/\< 0g~~~ ' f-' i tU . .~~~~U aiso be on b anesecaly lwxcang fRcg import pric 80 Real export pri - 70 60 I I I 1 t I II I I I i L I 11 I I I I 1 1985 1986 1987. 1988 Soume: World Bank data. also be done by an especially low exchange rate for imports of this particular good. In the absence of other instruments, governments find multiple rates to be a ready way. of achieving noneconomic objectives. Govern- ments often fool themselves, however, because they underestimate the. cumulative economic costs of a distortionary exchange rate system. We will.discuss the costs in a later section. Mechanisns for Special Exchange Rates There are a number of mechanisms for special exchange rates., EXCHANGE RATES FOR IMPORTS. Special exchange rates for particu- lar groups of imports involve two separate issues: the purchasing power of wages and industrialization. -dx4arvW Rate* Tie; Oplonsand 16u J01 The typical situation is one in which a country has a trade Imbal- ance. That means that the level of spending is too high in relation to incomc (perhaps because of a budget deficit) and that traded goods arc underpriced (imports arc too chlacp. and exports arc priced so low that exporting is unprofitable). An adjustment is needed, and the clenn way is to have a real deprtciaLion that redluces the purchasing power of income anid increases the competitiveness of the tradable goods sector. Governments arc rightly conccrned, however, thlt an across*the- board incriease in import prices will hsave advcrse political conse- quences: increased prices of imports will reduce Lhc purchasing power of wages and create political problems. Governments therefore look for a more selective way to ration foreign exchangc-one that attempts to skirt the political problem. The immediate step is to dis- tinguish between "essential" and "nonessential" imports. For essential imports, the exchange rate is held unchanged, whereas for nonessen- tial imports, it is increased sharply so as to reduce foreign exchange outlays and thus solve the trade deficit problem. Does such a policy work? Often one finds that the purchasing power of wages in terms of essential goods stays unchanged and that it declines sharply in terms of all other goods. The poorest groups in society might be protected, but the vocal middle class is- being squeezed, and that poses new problems. Of course, we would want to check whether multiple rates at least succeed in attaining the govern- ment's objectives. In many instances they have not and may, in fact, have hurt the very group they were designed to benefit, as well as making everybody else worse off., The other important motivation for special exchange rates on the import side is industrialization. A government may feel that domestic industry should be protected. Rather than pursue, a policy of trade-oricnted growth (often declared to be a low-wage strategy), the government may want to promote more rapid economic growth.:' Industries are designated as plausible development targets and are favored with a special exchange rate. Here the special exchange rate plays exactly the same role as a tariff, except that (in a context of import licensing) using special exchange rates may be easier administratively. As a result of the special exchange rate, domestic firms in the pro- tected industry can charge higher prices and can thus afford to pro- duce at a cost that is higher than the world market's. In time they may go through a learning process, reduce costs, and ultimately become world class. More likely, they will keep producing at high cost, enjoying protection at the expense of consumers and at the cost of a wastb of resources. 102 Rudiger Donmbusch -and Luis Tellez Kuenzler EXCHIANGE RATES FOR EXPORTS. Special exchange rates are used for exports primarily for sevcral reasons. One reason is to encourage indugtrialization. Particular export industries are given an implicit subsidy that allows them to sell at world prices. This typically applies to manufacturing or nontradi- tional export industries. Another reason concerns traditional export industries (oil in Ven- ezuela ancl Peru and agriculture in Argentina). A government may believe that these industries should not enjoy a windfall gain from a general exchange depreciation because their output is not very price- responsive. Thus, the government may use a special exchange rate to tax any windfall profits away from these industries. For example, if the currency is depreciated to eliminate a trade deficit, agricultural exporters will face higher prices in peaos. If their harvests have already been planted (or even harvested), the. entire price increase might rep- resent windfall profits, with no -favorable effect on dollar earnings. Moreover, the increase in pesos of agricultural exporL prices would raise the prices at which the agricultural sector would sell in the home economy, because farmers clearly would not sell below what they could earn abroad. But that would imply a -reduction in the standard of - living for consumers of agricultural goods-hence the politically **- attractive argument for a "compensated devaluation;' namely, a deval- uation for all sectors except agricultural exports. The argument becomes even more attractive when it appliesI to foreign oil companies. Why create windfall profits for them (which - they might take out of the country) rather tbmn administer a low, special exchange rate at which they are required to surrender foreign * exchange revenues? The Costs of Multiple Rates We have seen the political reasons for multiple rates. We nec- look at the costs-the misallocation of resources, rent seeking, and the budget problem. THE MISALLOCATION OF RESOURCES. Multiple exchange rates have important implications for resource allocation. The (competitive) market solution is to allocate resources so that, on the export side, the marginal cost (in relation to the domestic resources required) to pro- duce an extra dollar of foreign exchange is equalized across all export activities. In this way, expanding one industryjust a little and contract- ing another slightly leave the total resource costs of obtaining a given dollar of revenue unchanged. The resource cost of earning foreign exchange is effectively minimized. Exchange lhate Pol icy: Options and lssues 103 In an economy that allocates resources efficiently, the marginal cost of an extra unit of output equals the price a firm receives, and that price equals the value a buyer places on an extra unit of output. This is the crux of microeconomics. When this equality of price and marginal cost prevails, there is no way of reshuffling productive resources- producing more of one good and less of another-in a way that makes one person better off without making someone worse off. This, in a nutshell, is the economic case against multiple rates. Multiple exchange rates interfere with the optimal allocation of resources because for the same good-a dollar-different sectors or groups in society pay or receive different amounts of pesos. As a result, people or activities that get cheap dollars will spend too much on dollar goods. People or industries that receive high peso prices for the dollars they surrender will produce too much of the goods that are earmarked for especially high exchange rates. Resources could be reshuffled, and they would be if the economy went to unified rates, so a. to reduce the resource cost of earning dollars and to spend the dollars earned in a better way. The government resists this "better way" because, although it makes society better off, it may make particular groups (low-wage labor) worse off. Export activities with relatively favorable exchange rates overex- pand in relation to those with less favorable exchange rates. Thus it takes a larger amount of domestic resources to earn a given amount of foreign exchange. This is because in the Dindustries with a favorable exchange rate the marginal cost of earning a dollar can be much higher than in the industries with the unfavorable rate. Equalizing rates would contract high-cost production and expand low-cost pro- duction, thus lowering the total cost of earning foreign exchange. The same reasoning applies to imports. Imports of goods favored by: relatively cheap foreign exchange expand, and those with a high price of foreign exchange contract. Domestic resources are drawn into the areas in which the exchange rate is unfavorable; they cannot compete- where the exchange rate is-particularly favorable. Once again, equaliz- ing the rates would tend to expand -the relatively low-cost import. substitution industries and contract the industries in which import substitution is subsidized (implicitly) by the highly protective exchange rate. Resource misallocation between exportable and import-competing industries also occurs. With a large spread between (average) import and export rates, the export sector will overexpand and import com- petition will be underdeveloped. Exchange rate unification would move resources from earning foreign exchange through -exports to saving foreign exchange through the production of import-competing goods. The marginal cost of producing import-competing goods (and 104 Rudiger Dombusek and Luis Teltez Kuenzler thus the marginal cost of saving a dollar by import substitution) is thus much lower than the marginal cost of earning an extra dollar by expanding exports. Exchange rate unification would lead to the equal- ization of the marginal cost of dollars on the saving and earning sides. Equalization of marginal costs and benefits would not be an issue if exports and imports did not respond to prices. If they were fully unresponsive, the exchange rate system would simply be a way of distributing from one group of households or firms to another. No waste would be involved, just politics. But when activities do respond to prices, the economy is worse off with a system of multiple exchange rates. A given amount of resources produces fewer goods and services than could be obtained if marginal costs and benefits were equalized across all activities. Why then do governments use multiple rates? The reason is a com- bination of two factors. First, governments are interesteid in redis- tributing income across groups, and the exchange rate system offers that possibility. Second, governments often believe (erroneously) that the resource allocation effects of a distortionary exchange rate regime are -small. The reason they cite is that resources do. not easily move across industries (technically, elasticities are low). and consumers are relatively unresponsive to price changes. In fact, very little evidence for low responsiveness of households or firms- exists, certainly when we allow for some time to adjust to a new set of prices. The temptation to use multiple rates is particularly strong when a country's economy has a division between traditional and nontradi- tional exports, and even more so when traditional is tantamount to large, rich, foreign finns or domestic landowners. In the Peruvian. example cited earlier, nontraditional exports (compared with,. say, exports from srnall mines) are heavily subsidized by the exchange rate system. Moreover, the exchange rate for nontraditional exports far exceeds the rate for imports. Thus import competition in nontradi- tional industries is discouraged by a cheap import policy, -but at the same time, the implicit export:subsidy encourages these industries to work for the world market on the export side. The system thus draws resources out of import cor.petition (thereby increasing import spending) while expanding exports. There is a net domestic resource cost, and there may even be a loss of net dollar revenues.. Finding examples is easy. Suppose intermediate inputs, for exam- ple, are allowed a favorable exchange rate, but final luxury goods. are implicitly taxed by-a high exchange rate. In that case, we should not be surprised to- find that the resource cost: of producing luxuries may actually involve a larger. foreign' exchange cost than if they were allowed outright Resources might even be spent to break up finished goods abroad, import them as intermediate goods! at a favorable ExcIange Rate Policy: Options and Issues 105 exchange rate, and reassemble them into finished goods at a consider- able domestic resource cost. Privately, this may be entirely profitable, but from a social perspective, it wastes resources. A particularly extreme example of the misallocation that can result from multiple rates involves negative value added. This is a situation in which an industry is privately profitable but, from the social point of view, should be closed. Consider a specific example. A firm produces chemicals for export. A gallon of chemicals can be sold in the world market for $20. The export exchange rate at which the revenue is converted is 20 pesos per dollar, so that for each gallon exported, the revenue is 400 pesos. Next consider the cost side. A gallon of chemicals requires one person's day of labor at 60 pesos a day and two barrels of oil at $15 a barrel. For imports of intermediate goods the government allows a favorable exchange rate of only 10 pesos per dollar. Hence a barrel of oil-costs the chemical company 150 pesos. The two barrels of oil required to produce one gallon of chemicals will cost the company 300; pesos,, which, together with the labor cost of 60 pesos, brings the total cost to 360 pesos. Subtracting the cost from the revenue of 400 pesos leaves a profit of 40 pesos. Thus, with the particular set of exchange rates, producing chemicals for export is profitable (privatelyl). But now consider the social costs and benefits. To earn the $20 of export revenue per gallon of chemicals there is an expenditure of $30 to buy the oil, which is an indispensable input.f Therefore, of every dollar of chemicals exported, the country loses $10, not even counting the labor effort. Privately, exporting chemicals is profitable because the multiple exchange rate makes oil (in pesos) cheap in relation to the export price (in pesos). Socially, the undertaking is a big waste. If the operation were closed, the economy would be ahead. This example seems farfetched. It is noL In many countries, the exchange rate structure produces this extreme kind of distortion of negative value added. With a unified exchange rate (and in the absence of subsidies), such a possibility cannot arise. RENT SEEKING, RED TAPE, AND CORRUPTION. The costs of multiple exchange rates go beyond those caused by the direct waste of distorted production and consumption patterns. Additional costs result from what Krueger (1974) has called rent-seeking activities. Because multi- ple exchange rates create economic rents for those factors of produc- tion that are specific to activities receiving special protection, eco- nomic agents will devote real resources to obtaining exchange rate protection. In plain language, they will spend resources on lobbying. Lobbying may appear to be a relatively innocuous activity, but it does absorb scarce resources. Not only do lobbyists spend time and Examnge Rate Policy Options and 1sntes 107 Suppose now that trade in dollars is balanced so that X = M. In this case, equation 5-4 for net outlay simplifies to X(e, - e.). Thus, if the average export exchange rate exceeds the average import exchange rate, the central bank will be making losses on its foreign exchange transactions. The larger the discrepancy between the rates, the greater the losses. To finance these losses, the central bank will print money. This money creation, broughL about by the foreign exchange regime, in. turn proves a source of inflation. This is almost certainly the case when the central bank's losses are very large. The World Bank estimates that central bank losses from exchange transactions in Peru were, %s a percentage of CDF, 1.5 in 1985,0.4 in 1986, and 2.0 in 1987. A loss of 2 percent of GDP, financed by money creation, is barely, compatible with moderate inflation. But when the deficit that -requires financing exceeds 10 percent, as it did in Peru in 1987, high and rising inflation is inevitable. Just -how costly the budget financing turned outt to be is apparent from- Peru's experience in 1988. Inflation in August- 1988 reached 2,500 percentl In September, the government unified the exchange rate in a last attempt to stem hyperinflation. Black Markets and Tade Misinvoicing Black markets exist because of a combination of factors. First, the rationing of foreign exchange by quotas, licenses, absolute restric- tions, or special exchange rates gives rise to a demand for foreign exchange to finance illegal imports and outlays abroad. Second, there is a portfolio demand for foreign exchange that needs to be satisfied in a black market if the government does not sell foreign exichange for that purpose, but economic agents want to hold foreign assets. Third, if the black market rate exceeds the official exchange rate (or rates), exporters have an incentive to underreport their true exports and surrender their foreign exchange to the black market rather than the central bank; likewise, firms find a way to overstate the costs of autho- rized imports, diverting part of the foreign exchange obtained at the . official rate toward the black market.5 The Black Aarketfor Dollarts To understand the operaiion of a black market, it helps to separate the portfolio aspects and the current flows in and out of the black market. On the right-hand side of figure 5-2 we show as a hyperbola the stock; demand for foreign exchange, that is, the amount of pesos people want to hold in the form of dollar-denominated assets. The vertical axis shows the black market exchange rate (peisosldollar), and the hori- &xclhange Rate Policy: Options and Issues 107 Suppose now that trade in dollars is balanced so that X = M. In this case, equation 5-4 for net outlay simplifies to X(e. - e). Thus, if the average export exchange rate exceeds the average import exchange rate, the central bank will be making losses on its foreign exchange transactions. The larger the discrepancy between the rates, the greater the losses. To finance these losses, the central bank will print money. This money creation, brought about by the foreign exchange regime, in. turn proves a source of inflation. This is almost certainly the case when the central bank's losses are very large. The World Bank estimates that central bank losses from exchange transactions in Peru were, 1s a percentage of CDP, 1.5 in 1985,0.4 in 1986, and 2.0 in 1987. A loss of 2 percent of GDP, financed by money creation, is baTely compatible with moderate inflation. But when the deficit that requires financing exceeds 10 percent, as it did in Peru in 1987, high and rising inflation is inevitable. Just how costly the budget financing turned ouit to be is apparent from Peru's experience in 1988. Inflation in August 1988 reached 2,500 percent!. In; September, the government unified the exchange rate in a last attempt to stem hyperinflation. Black Markets and Trade Misinvoicing Black markets exist because of a combination 'of factors. First, the rationing of foreign exchange by quotas, licenses, absolute restric- tions, or special exchange rates gives rise to a demand for foreign exchange to finance illegal imports and outlays abroad. Second, there is a portfolio demand for foreign exchange that needs to be satisfied in a black market if the government does not sell foreign exchange for that purpose, but economic agents want to hold foreign assets. Third, if the black market rate exceeds the official exchange rate (or rates), exporters have an incentive to underreport their true exports and surrender their foreign exchange to the black market rather than- the central bank; likewise, firms find a way to overstate the costs of autho- rized imports, diverting part of the foreign exchange obtained at the -official rate toward the black market.5 The Black Marketfor Dollars: To understand the operation of a black market, it helps to separate the portfolio aspects and the current flows in and out of the black markeLt On the right-hand side of figure 5-2 we show as a hyperbola the stock: demand for foreign exchange, that is, the amount of pesos people want to hold in the form of dollar-denominated assets. The vertical axis shows the black market exchange rate (pesosldollar), and the hori- 108 RudigerDornbusciz and Luis Telkz Xumezkr zontal axis shows the number of dollars. The product eB, where e is the exchange rate and B the stock of dollars, thus represents the value in pesos of a given dollar portfolio. The schedule VV shows a given level of the stock demand. This stock demand is determined by the expected profitability of holding dollars rather than pesos and thus depends on interest rates at home and on the expected rate of exchange depreciation. The higher the foreign interest rates and expected depreciation are in relation to the home interest rate, the farther out and to the right is the value of black dollars (in terms of pesos) that residents want to hold. The vertical schedule BB denotes the existing stock at a point. in time. Short-run equilibrium in the black market occurs at an exchange rate eo. The left-hand side of the diagram shows current flows in and out of the black markeL. The schedule SS.shows the current inflows of dollars from export underinvoicing and import overinvoicing. The higher the black market rate is in relation to the official rate, the larger is the flow supply. The flow demand for dollars for tourism and for import smug- gling, DD, is a declining function of the black market rate. The lower the black market price of dollars, the cheaper it is to buy in restricted goods or buy foreign exchange for tourism. The stock and flow markets are connected by the fact that the price at which the existing stock of dollars is held, eo, determines the current amount of inflows and outflows, SO and Doe The way we have drawn the schedule, the difference between the quantity supplied and demanded (SO -Do or C - A) is positive. This means dtat inflows exceed outflows, Figure 5-2. A Model of the Black Market ss A C :' - I DD vv - n . S0 Fa B . - . - Stock of of dolan3 . 0 doUas Exchange Rate Policy: Options and Issues 109 and hence the stock held in portfolios increases. This will show over time as a rightward shift of the stock supply, BB; a declining exchange rate; and declining net inflows until the market achieves the steady state. These bare bones of a black market model help to identify the main determinants of the black market exchange rate (or the premium over the official rate): The higher the interest rates abroad and the expectations of depreciation are, the higher is the stock demand and hence the price of black market dollars. Figure 5-3 shows the black market premium in Brazil. Note the premium of nearly 100 percent at the end of 1986. This premium stemmed from a widespread expectation of a large depreciation against which asset holders sought to protect themselves by a flight into dollars. As the supply of dollars was fixed, the flight into dollars pushed up. the pre- mium. Thus monetary instability is the prime reason for a large black market premium. Political instability works in the. same direction. * An increase in trade restrictions increases the flow demand for dollars for purposes of financing smuggling. Over time, the higher flow demand leads to a rising premium, because the -net inflow into the -market falls off. The black market for foreign. exchange typically finances all imports that are subject to high duties. * Greater ease in supplying the black market by misinvoicing shifts the flow supply curve out and to the right and thus reduces the long-run premium. Misinvoicing There are basically two reasons for misrepresenting trade. data. One is capital flight; the other is the desire to evade import or export duties or to capture export subsidies. These reasons, summarized in table 5-4, are discussed in more detail below.. CAPITAL FLIGHT. Residents of a country where exchange control prohibits the free export of capital for portfolio purposes might attempt to acquire foreign assets for several reasons. The three most important are fear of political instability, anticipated financial insta- bility, and the laundering of assets. Shifting into foreign assets, in terms of currency denomination and location orjurisdiction, shelters the portfolio holder against the vagaries of domestic political and economic disturbances or against inspection by the tax authorities. 110 Rudiger Dornbusch and Luis Tilez Cuenzler Figure 5-3. Black Market Premium, Brazil, 1962-89 110 100 - 90 80 - 70- 60- .50 30 20 10 0 -10 1962 1965 1968- 1971 1974 1977 1980 1983 1986 1989 Source: Bank of Brail data. Interest in sending asseLs abroad in anticipation of political changes is. apparent when one thinks of the extreme cases of Marcos or Duvalier, but such interest is also manifested, in a less extreme fash- ion, among asset holders who fear that their wealth or their previous dealings might become exposed in a regime change. For example, a significant political capital fight from Brazil in 1986 was motivated by the business community's fear. that t.he left-wing parties might gain a large share of the seats in the Constitutional Assembly. Financial instability is perhaps the most important reason for capi- tal flight. When asset returns in a country are negative because interest rates are outpaced by inflation, asset holders are keen to protect.their 'Able 5-4. Principal Motives for Ovezinvoicing aid Underinvoicing Caegy Oernwoiking Underfrvoicing Exports Capturing export subsidies Capital flight Avoiding export taxes Iinports Capital flight Evading import duties Exchange Rate Policy: Options and Issues I1I wealth by moving into dollar-denominated assets. This is the capital flight typical of Mexico in 1980-82 or of Argentina in 1979-83. Figure 5-4 shows the high Argentine black market premium during the latter part of this period of instability. This kind of capital flight is particu- larly vigorous when a large overvaluation of the currency makes it almost certain that holdings of dollars will carry a large return because of the anticipated depreciation of the home currency. Capital flight to launder assets is common whenever a country starts enforcing its tax laws and institutes an improved system for monitor- ing tax compliance. When that happens, offenders have to take their undeclared assets out of the local jurisdiction to avoid detection and prosecution. Imagine what would happen in Argentina, for example, if bank secrecy were abolished and tax laws were, for the first time, seriously enforced. An immediate wave of capital flight would result as the business community sought to shelter its previous fraud. (Laun- dering of assets also applies, of course, to public officials who take abroad the revenue from corruption.) Capital flight for' these various motives needs a vehicle. When exchange control means that foreign exchange cannot be bought. freely and anonymously, then the misinvoicing of trade transactions becomes the means for generating the foreign exchange that is sold in the black market to those individuals seeking to move money abroad. Whenever it is possible to buy foreign exchange freely at the official rate, and with not too many traces left behind, there is no advantage to using trade flows as a means of acquiring foreign exchange. Exporters can generate foreign exchange to finance capital flight by- underinvoicing their exports. Rather than surrender the full value of export revenues to the official foreign exchange authorities, as they are required to do, they understate the export revenue on their invoices and instruct their trading partners to deposit the balance. in a foreign account for their benefit. This balance, in turn, might be held. by the actual exporter or sold in the black mar'ket to someone else who wishes to acquire foreign exchange.. The incentive for an exporter to engage in underinvoicing is quite' obvious. Take the case of Brazil in the last quarter of 1986. The pre- mium in the black market was 100 percent. Underinvoicing an export by 25 percent would mean making an extra 25 percent profit over and above the normal commercial profit. With so !arge a profit margin for an activity that is widely considered to be a sport rather than a crimi- nal activity, underinvoicing of exports becomes pervasive. This is made easier when exports are diversified manufactures rather than homogeneous primary commodities with commonly known world prices. Although it is apparent that export underinvoicing will take place when its rewards are as extreme as in Brazil, one wonders what 112 Rudiger Darnbuscld and Luis lWtez Kuenzler Figure 5-4. Bladc Market Premium, Argentina, 1960-88 280 260- 240- 220 - 200 a 180 160 - 140 120 100 80 60 40 20 0I 1960 1964 1968 1972 1976 1980 1984 1988 &ume: CentrAl Bank of Argentina data. the threshold is in Argentina. Experts maintain that export underin- voicing starts when the black market premium goes beyond 25 percent. Overinvoicing of imports, just like underinvoicing of exports, can serve as a vehicle for capital flight An importer, when applying -for foreign exchange to the authorities, will overstate the value, of the import shipment. Depending on the country's procedure, the authori- ties wril directly instruct payment to be made or they will put at the importer's disposal foreign. exchange in an account abroad. In one case, the importer directly acquires the foreign exchange in excess of actual import costs. [n the other case, cooperation with the foreign shipper is orequired. Whichever is the case, defrauding the foreign, exchange authorities is a-routine activity. Once-again, the proceeds might be held by the actual importer or sold through the black market to other residents who wish to acquire foreign exchange but cannot obtain it freely or discreetly from official sources. International' transactions are natural vehicles for tax fraud, because domestic tax authorities do not have access to trading part- ners' business reco'rds as they do for domestic ones. In the domestic E&clhange Rate Policy: Options and Issues 1)3 market, one person's receipts are another person's outlays, a fact that sharply reduces the opportunity for fraud. IMPORT DUTIEs. High import duties are a common feature in many developing countries. One reason is that trade taxes remain an impor- tant part of public revenues. The other reason is a long-standing pol- icy of import protection. For all of Latin America, taxes on interna- tional trade are 12 percent of total government revenue. In some countries, the share is more than 25 percent, as in Bolivia or Costa Rica. When import duties are high, importers have an incentive to cheat on import duties by underinvoicing. By simply understating the value of imports, the importer reduces' the amount that must be paid in duties. As a result, the average burden of the duty, expressed as a fraction of the true value of imports, is reduced. For example, if the duty is 70. percent, and underinvoicing amounts to 25 percent, an importer pays only 52.5 percent effective duty rather than the 70 that is owed. *So far we have assumed that the importer can freely obtain foreign exchange to finance the part of the import bill that is underinvoiced, but, of course, that rnay not be the case. When there is exchange control, the importer has to buy the extra foreign exchange, in the black market at the cost of a premium over the official rate. To know whether underinvoicing remains profitable, we must compare the belack market.premium with the tariff rate. If the tariff rate exceeds the black market premium, then underinvoicing saves more in.import duties than it costs in the black market to make up the difference in the foreign exchange required. This points to goods with very high tariffs-of 100 percent or more-as the items most likely to attract. fraud on import duties. By contrast, goods with no import duties serve as vehicles for underinvoicing to generate foreign exchange for sale in the black market. We have discussed'how import duties are avoided by understating the value of shipments, but. an important, parallel alternative occa- sionally applies. Often tariff rates differ significantly among different goods that might pass for being the same. Whenever that is the case, the incentive goes strongly in the direction of misstating the exact nature of the imported goods. By representing the commodity as belonging to a low tariff category rather than to the appropriate high tariff category that actually applies, the importer saves tariff costs. EXPORT DUTIEs. In some countries,-the government levies export duties. For example, in Argentina agricultural exports. are typically taxed and so are, on occasion, some manufactures. In such a situation, 114. Rudiger Dornbuscd and Luis 7TIiez Kmn= ter when export duties are ad valorem, the exporter has. an incentive to understate' the true value of an export shipment to reduce the amount of duties paid. Thus, underinvoicing of exports reduces the effective tax rate that the government collects. EXPORT SUBSIDIES. When a government grants export subsidies, for example, to nontraditional cxports, the incentive runs the other way. To increase the effective subsidy rate, the exporter will overstate the value of a shipment. OTHER TAX FRAUD. Overinvoicing or underinvoicing occurs for other, more specific reasons. Some are worth mentioning if only to show that in a society in which tax fraud is common, overinvoicing or underinvoicing will tend to appear in a very significant association with trade transactions. One example is abuse of investment tax credits. Suppose a country pursues an investment strategy and for that purpose allows a tax credit against equipment purchases that also applies to imported equipment. By overstating the value of imports, the firm obtains a larger amount of tax credit or a higher effective investment tax credit. As another example, consider corporate income tax fiaud. A company that routinely imports materials has an interest in overstating the cost of materials to reduce its before-tax profits and hence the effective tax on profits. This is difficult to do within a country, but overinvoicing is an obvious means. One can imagine that in -the case of multinational corporations, misinvoicing of imports and exports is simply a means of shifting their tax liabilities from one jurisdiction to another. By overstating import costs from their foreign headquarters or by understating export value, they shift corporate profits abroad to take advantage of lower-tax rates that may apply there. . EMPIRICAL EVIDENCE. The significant misrepresentation of trade data in developing countries is well known. Bhagwati and Krueger (1974) documented extensive underinvoicing, and more recent work by Gulati (1987) shows that the practice flourished in the late 1970s and 1980s. There are no direct measures of underinvoicing or overinvoicing in international trade. The only method is to compare one country's trade data with the data of its trading partners. Any discrepancy is a guide to the extent of overinvoicing or underinvoicing, although it may represent, to a small extent, other factors. (See Gulati 1987 and Bhagwati and Krueger 1974 for a more detailed discussion of the statistical issues.) The International Monetary Fund (IMF) reports trade data by country and region. Using these data, it is possible to calculate Exchange Rate Poliy: Opti onm and lIssms 115 the amounL of overinvoicing or underinvoicing for a particular coun- try, say, Argentina. One simply compares the value of exports to indus- trial countries and imports from industrial countries reported by Argentina (on the basis of invoices presented to the Argentine author- itics) with corresponding data reported by industrial countries. lable 5-5 shows the average level of underinvoicing of exports for a number of countries. At a more aggregate level we can look at Latin America's trade with the industrial countries. We take 1982 to 1985 as years in which cur- rency disturbances were particularly important. Interestingly, both export and import underinvoicing occurred throughout the period (table 5-6). Export underinvoicing points to continuing, steady capital outflows. Import underinvoicing means that tariff evasion dominates capital flight motives. Thus, looking at the trade account, export underinvoicing finances the tariff evasion by supplying , - black mar- ket dollars required when the import value is understated. It is striking and depressing that a whole continent's tax fraud should be so simple to document. The Mexican data show substantial export underinvoicing and hence evidence of steady capital flight. Persistent underinvoicing of imports points to tariff evasion. The huge underinvoicing of 1983 presumably indicates that the system of rationing foreign exchange was particularly vulnerable to fraud. The Macroeconomic Effects of Black Markets The preceding section showed pervasive evidence of capital flight via underinvoicing of exports. In countries where exporters are required lhble 5-5. Underinvoicing of Exports, Selected Countries, 1977-83 (percentage underinvoicing) C-ountry 1977 1978 1979 1980 1921 1982 1983 Averag Argentina- 20.6 21.2 18.9 17.4 19.2. 21.0 19.5 19.6 Brazil 11.2 9.0 17.1 14.5 11.4 15.9 9.8 12.7 Chile 14.0 23.5 14.8 13.8 9.6 .12.0 6.5 12.8 Korea, Rep, of 2.0 -0.9 -0.7 0.5 -3.5 -6.1 -5.1 -2.5 WMico 77.9 26.8 42.4. 27.6 26.9 26.5 33.8 . 33.6 Peru 20.8 15.6 8.9 14.4 18.1 10.0 9.6 12.9 Uruguay 15.0 13.5 13.7 14.3 19.5 64.1 52.6 27.8 Venezuela 7.6 6.1 0.7 4.2 9.5 8.9 9.4 6.9 Note Because exports are recorded fab, and imports ci.f., a difference of approxi- mately 10 percent is to be expected. Thus only magnitudes above 10 percent should be regarded as capital flight. Source: Cuddington (1986). 116 Rudiger Domrbmuh and Luis Telz Kuenzler ibble 5-6. Average Underinvoicing, Mexico and Latin America, 1982-85 (percen tage ofexparts and imports, respectlvely) categor 1982 1983 1984 1985 Exporl undeinwicing Mexico 25.0 32.0 17.0 17.5 Latin America 9.3 8.5 v 7.3 8. Import underinwicing Mexico 34.0 84.0 39.0 19.7 Latin America 18.9 23.8 27.9 15.8 Source: IMr data. to surrender their full export proceeds at the official exchange rate, underinvoicing implies that the government receives less than the full export revenue. What are the. macroeconomic implications? One. immediate implication is. that the government does. not collect all the revenue that is due. The other is that the red-ced foreign exchange receipts imply the need for extra exchange depreciation, which brings with it a reduced standard of living and inflationary pressure. Each of these has adverse effects on the budget and, from: there, on macro- economic stability. LOSS Or REVENUE. hable 5-7 shows data on the share of trade taxes in total revenue and the ratio of trade taxes to ONP. These data show that import duties are an important element of the government reve- fnue base. Accordingly, widespread underinvoicing is a major issue in public finance and macroeconomics. The averages in table 5-7 conceal the fact that some countries are: much more dependent than others on trade taxes for government finance. In Africa, for example, trde taxes provide almost 50 percent of Mauritnians revenue and 30 percent of Senegal's. In Asia, the reve-. -nue share of trade taxes in Pakistan and Sri Lanka exceeds 30 percent. In Latin America, and especially Central America, the smaller coun- tries derive a large share of their revenue from duties. - Ihble 5-7. 'Tade 'Mixes as a Share of Govenmment Revenue and as a Share of cmp, 1982 Trade taxes as a Revenuefromn trad axes : Region percentge of revemu . as a peentage of CIVP Africa 15.5- 2.9 Asia . 19.1 3.2 Latin America . 14.4 . 2.9 SOUeCC IMF (1986). Exchamge Rate Policy: Options and Issues 117 We saw earlier that underinvoicing of imports for Latin America averages 30 to 40 percent. What is the budgetary implication of these losses in duty revenue? A crude estimate, using table 5-7, is revenue losses on the order of 1 percent of GNP! In countries.where budget balancing and revenue collection are major problems, a revenue loss of I percent of GNP is an extraordinary shock to public finance. LOSS OF FOREIGN EXCHANGE. When a firm underinvoices exports, it withholds -foreign exchange receipts from the exchange authorities. This foreign exchange might be used to build up external assets or perhaps sold to importers who are underinvoicing. For the govern- ment, the reduction in foreign exchange receipts implies that, at the * going exchange rate, there is now a shortfall of export revenue in relation to the foreign exchange required to finance the existing level of imports. The foreign exchange deficit forces one of two policy measures: the imposition of extra protection to cut selectively the level of imports or a depreciation of the currency so as to increase export competitiveness and reduce import dependence. The economy must make up the foreign exchange deficiency caused by underinvoicing of exports either by saving on foreign exchange through restrictions. or- by finding extra ways of earning and saving foreign exchange through currency depreciation. It is immediately apparent that overinvoicing imposes an important social cost. Protection is costly not only for consumers at home but also for.foreign exporters who are barred from a market in which capital flight absorbs the foreign exchange that otherwise would have been available to finance extra imports. Depreciation achieves much the same end, because the only way to close the foreign exchange gap is to sell more and buy less. There is an obvious cost to foreign exporters who are forced out of the home market by the gain in competitiveness of extra depreciation. There are also losses to foreign firms whose products now have- to compete with the more competitively priced imports from the devel- oping country. The simple way to think of this process is that the developing country-must earn the extra foreign exchage that is now diverted to capital flight: extra exports or cuts in import spending i must supply the dollars that are absorbed by capital flight through export underinvoicing. In the country suffering the capital flight, the, costs go much- fur- ther. Real depreciation reduces the purchasing power of wages; that is, after all, how depreciation makes a country more competitive in world trade..Thus there is a redistribution of incomie from wage earners and the public at large to the firms that practice capital'.. flight. There is an economywide cut in real. income, the proceeds of 118 Rudiger DonbusIh aund Luis Tellez Kueuzter which are indirectly and illegally appropriated by the firms that practice capital flight. The cut in the standard of living has inevitable budgetary effects. Spending decreases, and much of the reduction in spending falls on nontraded goods. Domestic activity therefore tends to decline. The reduced level of activity reduces budget revenues and hence contrib- utes. to a deficit. In addition, the reduced standard of living often forces the government to adopt selective subsidies for food or public transport. Thus an increased budget deficit almost certainly emerges as one effect of capital flight. INFLATION. The inflationary effect of the depreciation required by increased capital flight is also important. The inflationary pressure stems from two sources. First, depreciation raises the price of traded goods and the cost of living.- Then, assuming the increased cost of living results.in wage demands, a wage-price spiral is triggered. That wage-price spiral is fed by the money creation that stems from enlarged budget deficits, as noted earlier. Thus an increase in inflation-possibly a major increase-inevitably appears. This inflation, in turn, promotes further capital flight. There is, accordingly, a vicious spiral of capital flight and instability-that:can be stopped only by both working directly on the budget and curbi'ng capital flight. To place the burden exclusively on the budget raises the social costs of making illegal trade transactions unprofitable. The bet- ter way is to make at least some effort at directly stopping the faking of export invoices., AGGREGATE COSTS. To obtain an impression of the aggregte effect of capita! flight on an economy, it is helpful to compare capital flight to a loss of export revenue caused by a decline in export prices. Imag- ine a-country that suffers a 20 percent decline in the price of an export product that accounts for 70 percent of total exports. The total reduc- tion in dollars in export revenue accruing to the central bank is 14 percent. This loss can easily result in a foreign exchange crisis with' devaluation, budget deterioration, and inflation. Underinvoicing of exports at the'rate that is evident for Latin Amer- ica in table 5-6 is strictly equivalent to a fall in export prices, and the social consequences are quite the same. But there is a difference: when export prices decline, the country at large, including exporters, loses.' With- export underinvoicing and capital flight, by contrast, the exporters gain and force not only the direct costs of what they misap-.- propriate but also the excess burden of the macroeconomic adjust- ment required by the foreign exchange shortage onto the country-. ExcanLge Rate Policy: Options and Issues J19 We have emphasized here the macroeconomic costs of export underinvoicing. In addition, there are microeconomic or welfare costs. Underinvoicing of exports can bc thought of as a privately col- lected export tax. Interpreted in this way, there is an identifiable dead- weight loss for consumers as well as a transfer of resources from the community at large to exporters. One might be tempted to conclude that because black markets have adverse macroeconomic effects and lead to a misallocation of resources, they should be vigorously repressed. Economists will give a different answer. Repressing black markets is rarely successful; there- fore, it is better to repress the causes rather than the symptoms. More financial stability and less restrictive trade regimes will vastly reduce the demand for foreign exchange.that gives rise to a black market. Dual Exchange Rates and Unified, Flexible Rates In this section we discuss what exchange rate system is suitable when the capital account plays.an important role in the economy anld when capital controls are ineffective either because of a-lack of idministra- tive infrastructure or because geography (as with Mexico) makes.con- trols implautsible. The discussion compares two alternatives: a system of dual exchange rates or one of unified, flexible rates. Dual Excktange Rates A system of dual exchange rates has only two rates: an "official" rate that applies to most, if not all, trade transactions-and a "financial". or 'free" rate at which all other transactions take place. Mexico produces a good example of such a system.6 From 1955 to 1976 and again from -1976 to 1982, Mexico.maintained a fixed exchange rate in relation to the United States. But as figure 55 shows, in the early 1980s, the real exchange rate appreciated massively, and this led to a payments- crisis.- The government-responded in August 1982 with a devaluation and the adoption of a split exchange niarket. Approved trade transactions were to be financed at the official rate and all other. transactions at a "market" rate. Figure 5-6 shows the premium of the market rate (in New York) over the official rate. It is interesting to note that with the exception of isolated crises in 1982, 1985 (when oil prices declined), and late 1987 (when private external debt payments fell due) the premium was quite moderate. This is the main point: a combination of a realistic com- mercial rate and.a realistic interest rate ensures that the gap between the. commercial and financial rate will be moderate and that distor- tionsareunlikelyto-become major.-: Exc-anv Rate Plicy Options and Issws 121 Figum - Real Exchanp Rate, Mexico, 1970-88 (1980-2' 100) 120 110 100 90 80 70 60 1970 1972 1974 1976 1978 1980 1982 1984 1986- .1988 SsiNr. Moipan Guaranty data. Figue !;6. Free MarkEt Preumium, Mexico, 1982-88 160 140 120 60 40 20 0 1982 1983 1984 198: I19 1987 198 .- smw Cnal-r BaD of Mexco (variou isuc). Exchav Rate Policy: Options and Issues 121 Figure 5-5. Real Exchange Rate, Mexico, 1970-88 (1980-82= 100) 120 110 .100 n90 60D 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 Sourne: Morgan Guaranty data. Figure 5-6. Free Market Premium, Mexico, 1982-88 : 160 140 120 100 ri80 60 40 20 0 1982 1983 1984 1985 1986- 1987 1988 Soue: Central Ban of MCxico (variowus iucs). 122 Ruder Domwbuwh and Luis Telkz Kuenlter Figure 527. Exchange Premium, Venezuela, 1983-88 240 220- 200 160 J140 '~120 100. ~80 60 40 20 0 1983 1984 1985 1986 1987 1988 Sorce Centra Bank of'Vcnezuda datL Figure 5-8. Exchange Rates, Mexico, 1982-85 220 a 18, 160 1 V New York,,-., 120 . - . : -Controlled" 100 60 \~~~~ ~~~~~~~ ~ mo 12}C/ - - - 1982 - 1983 1984 1985 S0uc Ccntral Bank ofMc3dco data. Exchange Rate Po icy: Options and Issues 123 viability of the commercial rate, a large, persistent premium develops, and this leads to distortions. For example, if oil prices decline perma- nently, the commercial rate for an oil-exporting country ultimately has to depreciate. Asset holders anticipate this event, and therefore the free rate (pesosldollar) immediately rises, opening up a persistent premium. Unless the authorities adjust the commercial rate, ineffi- ciency and misinvoicing soon become a major headache. This high- lights the critical role of a moderate premium. Considering the need to keep a moderate premium, is a dual-rate system really useful? The answer is yes. Under a dual-rate system tran- sitory disturbances in the capital account do not affect the macro- economy. At the same time, the fiee rate suggests what the commercial rate should be whenever the premium becomes sizable. The Mexican system involves central bank intervention in the free market along with an exchange rate rule for the controlled rate. Nor- mally the spread between the New York rate and Mexico City rate is very small or even negligible, but when major disturbances do open a gap, the rationing of foreign exchange at the free rate in Mexico leads to a double premium and thus to an extra shock absorber. As long as such a situation does not persit for long, the system is quite desirable in that it minimizes, the exceptionally disruptive effects that would come from either a major collapse of the exchange rate or the news of masslive exchange reserve losses. The key point is- the need for an anchor, and a dual-rate system provides precisely that service. U-uefid, Flexible Rates Faced with complicated distortions in many exchange rate regimes, numerous observers conclude that simplification of the exchange rate system down: to a single rate is desirable.' They argue further that governments are excessively tempted to maintain overvalued rates whenever exchange rates are fixed. Hence they advocate a move to a unified, flexible exchange rate. Advocacy of flexible exchange rates for developing countries is a new development. The traditional view has- been that exchange mar- kets in developing countries are too "thin" and that, accordingly, exchang;e rates might fluctuate excessively. The new view challenges that belief, and its proponents might add that in a comparison of the volatility of one system and the wasteful resource allocation of another, why conclude that volatility is the worse evil? (For a review of the arguments and some case studies see Quirk and others 1987, 1988.) The emphasis on thin exchange markets takes the discussion about- fixed, veTsus flexible, rates in the wrong direction. The United States, 124 Rud*ger Dornbusch and Luis Te:kz Kuenzier under flexible rates, experienced extraordinarily large fluctuations in the real exchange rate during 1980-87, and these rate movements are hard to rationalize in terms of fundamentals. Yet nobody doubts that the world market for dollars is large and deep. The main issue is really the underlying volatility of fundamentals and the extent to which government policy measures are endogenous or exogenous. If government policy-the budget and the money supply-is totally independent of what happens in the economy, and if fundamentals such as the terms of trade vary little, then speculation is easy and stabilizing. Speculators. have little room to guess what the equilibrium price will be. But if policy responds to what happens in the economy and speculators start anticipating policymakers' responses, there is no' assurance that the economy will have a stable equilibrium. In this case, the economy has no stable anchor, and any level of wages, prices, and exchange rate is equally plausible. In such a system, flexible exchange rates can be immensely volatile, as can prices and wages. The exchange rate system actually interferes'with economic performance. The case for unified, flexible exchange rates in developing countries is therefore implausible. Thus the choice is between a unified,-fixed rate, and a dual-rate system. The need for some flexibility in the capital account, without macroeconomic feedback, makes the case for dual rates. But efficiency needs to be addressed by arguing.for a limit on the size of the premium. The best way to handle this issuie is by an informal rule that has two features. First, major, events that are rele- vant to the commercial rate should trigger an immediate devaluation or Tevaluation. Guessing'whether the disturbance is transitory or. per-- manent and the appropriate size of the exchange market is difficult. Mistakes will occur, but this is unavoidable. Second, when major shocks in' the absence of the premium diverge from a moderate bound, depreciation of the commercial rate should gradually speed up to reverse the buildup of the premiumn. Conclusion The exchange rate is a central price linking an economy to the rest of the world. Although exchange rates are always important, they become especially important when misaligined. In that case, they create a wasteful allocation of resources and lead to capital'flight. The econ- omy's productive potential deteriorates, and ultimately the realign- ments force major devaluation, which brings about massive inflation and disarray. Managing exchange rates along simple, stable rules is thus critical. These stable rules must assure that cumulative overvalua- tion is ruled out by a policy of devaluation that keeps in line with domestic inflation. ange Rate Policy: Options and issues 125 The rate system should also be unifonn as it applies to commercial transactions. There may be (in isolated circumstances) a case for taxes or subsidies to promote specific economic or noneconomic objectives, but using explicit taxes and subsidies rather than the disguised form of multiple exchange rates is almost always prefemble. Finally, realism requires governments to recognize that portfolio holders have the option of holding external assets and that they will acquire them if domestic assets are unattractive because of low returns or exchange rate overvaluation. To avoid the distortionary effects of black markets, a realistic exchange rate system shouldb build on a dual rate. A dual rate strikes a compromise between the microeconomic case for unified rates and the macroeconomic need for.an anchor. Notes 1. For a discussion of a cost-benefit analysis of having a national money, see Fischer (1982). 2. There are various ways of calculating trading partners' inflation using export weights, import weights, or both, including or not including services. For a discussion see Williamson (1982) and Wickham (1985). 3. The issue ofred tape goes further when we note that there will be impor- tant incentives to misinvoice products or even to devote real resources to modifying products so as to take advantage of special exchange rate regimes. 4. We need not be concerned with the second term, e,(X -M). If this term is positive, the outlays arise from a trade surplus in dollars and the central bank should contemplate exchange appreciation or unification. If the term is negative, then the central bank faces a-trade deficit in dollars and must finance this by running down reserves. The reserve losses reduce the inflation- ary effect of the multiple rates, but they have their own problem. Specifically, what happens when the central bank runs out of reserves? 5. On the economics of black markets for foreign exchange, see Pechman (1984), Dornbusch and others (1983), and Dornbusch and Pcchman (1985) and the references given there. 6. There are many practical issues in the administration of such a: system. They include whether investment income and services such as tourism should be financed at the controlled or the free rate. They also include the important question of which rate applies to the amortization of private debts.- For a discussion- of some of these questions and for references and a formal model, see Dornbusch (1986). Selected Bibliography E Banco de Mdxico. Various issues. Indicadores econdnmcos. Mexico City. Bhagwati, Jadish, and Anne 0. Krueger. 1974. "Capital Flight from wcs. A Statistical Analysis" In Jadish Bhagwati, ed., llega Transaction in Interna- tiXna Trade. Amsterdam: North-Holland. Cuddington, John. 1986. Capital ight &timat4 Issues, and Explanation& : Princeton Studies in International Finance 58. Princeton, N.J.: Princeton - University Press. 126 RudigerDornbuchs and Luis TeUtz Kuenzler Dornbusch, Rudiger. 1986. "Special Exchange Rates for Capital Account Tmansactions:' World Bnk Economic Review 1(1):8-SS. Dornbusch, Rudiger, and Clarice el'chman. 1985. "The lBid-Ask Spread in the' Black Market for Dollars in Brazil'Journal of Mone, Credit, and Banking 17(4, Part 1, November):517-20. Dornbusch, Rudiger, Daniel V. Dantas, Clarice Pechmnan, Roberto de Renzende Rocha, and Demetrio Simoes. 1983. "The Black Market for Dol- lars in Brazil" Quarterlyjourn a of Economics 98(February):25-40. Fischer, Stanley. 1982. "Seigniomge and the Case for a National Moncy'Jour- na oafPolitical Economy 90(2).295-313. Reprinted in Stanley Fischer, Indexing, InJ?ation, and Economic Potiy. Cambridge, Mass.: MrT Press, 1986. Gulati, Sunil K. 1987. "A Note on Trade Misinvoicing:' In Donald R. Lessard and John Williamson, eds., Capital Right and Third World Debt. Washington, D.C.: Institute of International Economics. - iMp (International Monetary Fund). 1986. Government Finance Staistcs Yearbook 1986. Washington, D.C. Krueger, Anne 0. 1974. "The Political Economy of the Rent-Seeking Society" American Economic Review 64(3):291 -303. Pechman, Clarice. 1984. 0 dolarparalelo no Brasd. Rio dejaneiro: Paz e Ter. Quirk, PeterJ., Benedicte V. Christensen, Kyung-Mo Huh, and Toshihiko Sas- aki. 1987. foating Excihage Rates in Developing Countries. ilmf Occasional Paper 53. Washington. D.C.: International Monetary Fund. Quirk, Peter J, Graham Hacche, Victor Schoofs, and Lothar Weniger. 1988. Poticiesfor Developing Forwrd Foreign Exichang MaTkets. IMF Occasional Paper 60. Washington, D.C: International Monetary Fund. Wickham, Peter. 1985. "The Choice of Exchange Rate: Regime in Developing Countries: A Survey of the Literatuvet iwMStaffPapers 33(2June):248-88. - Williamson,John. 1982. 'A Survey of the Literature on thc Optimal Crawling Peg"Joutnad of De'elapment Economics 11(1 August):39-61. 6- Protection in Developing Countries M Paul Krugman ALTHOUGH ECONOMiST have long advocated free trade as an ideal, nearly all market economies protect at least some industries from foreign competition. Developing countries generally provide higher levels of protection, to a wider range of industries, than do advanced economies. When governments instituted protectionist policies, par- ticularly in the two decades following World War II, economists hoped that they could serve as instruments of development, helping to close the. gap between developing nations and the industrial world.- Today these same policies are widely viewed more as obstacles to advance- ment than as desirable parts of development strategy. Appropriate trade policy remains a subject of considerable dispute, however, and in any case trade liberalization presents difficult problems both of eco- nomic management and of political economy. The Costs of Protection In protecting a domestic industry from import competition, a country risks imposing a series of costs on its economy. These costs may usefully be divided into four categories. First, by raising the price of imports and irmport-substituting goods relative to their world prices, protection distorts incentives, leading the economy both to produce and to consume the wrong mix of goods from the point-of view of economic efficiency. Second, protection fiagments markets, so that the economy typically ends up producing too many products at too small a scale, with a resulting loss of production efficiency- Third, because protection lessens the competition from abroad, it leads to an increase in the monopoly power of domestic firms. Fourth, protection- typically creates "rents"-extra returns to those who receive privileges from the protecting government. Resources that could have been used productively may be dissipated in pursuit of these rents. 127 128 Paul Krgman Against these costs, offsetting benefits may sometimes occur; how- ever, we postpone consideration of the benefits until the next section and begin by considering the costs in more detail. 7he Distortiin ofIncentives The most straightforward cost of protection is that it gives false signals to the economy, which lead it both to produce and to consume the wrong mix of goods. Suppose, for example, that a given set of resources could be used to produce either soybeans for export 'or oil to replace imports. Suppose, also, that the soybeans that these resources could produce are worth $1 billion at world prices, whereas the oil they could produce are worth only $500 million. Then the country clearly ought to use the resources to produce soybeans and buy the oil from abroad. Yet a tariff of 100 percent on oi. imports makes it' more profitable to produce oil, leaving the country $500 million poorer. The point of this example is that the tariff, by driving a wedge between the prices that a country faces on world markets and the prices facing individual firms, also'drives a wedge between the actions that maximize national income and those that maximize the income of individuals. As a result of this distortion of incentives, resources may be diverted away from their most productive uses. The example emphasizes the distortion of production incentives that results from protection. Similar 'distortions, however,. will also occur- on the demand side. Suppose that instead of being able to produce oil, the domestic resources could only produce gasohol,' which is an inferior subsiitute for oil, but that a tariff makes oil so expensive that consumers choose to buy the gasohol to replace.$500 million worth of imported oil. Again the country is worse off: con- sumers have shifted to an. inferior substitute, even -though the resources used to produce that substitute could have been used to produce extports that more than paid for the imports eliminated as a result of the tariff. The 'effects of distorted incentives are most dramatic when imports of intermediate' goods are restricted. In highly distorted economies,. millions of dollars of exportable output may be stranded because transportation networks or processing industries are unable to obtain fuel or spare parts costing only a fraction of the value of the exports. The distortion of production and' consumption incentives that results from protection is often quite substantial. As we will see later, effective rates of protection in developing. countries -can lead to an allocation of resources to sectors in which they are less than' half as productive as they would be if used elsewhere. Yet most students of Protrtirm in Developing Countnries 129 protection believe tiat the costs are considerably larger than the dis- tortion of incentives alone. Thle Fragmentation of Markets Protection is often described as an "inward-looking" policy: it induces a country's producers to look toward the domestic market and tends to make a country more self-sufflicient. A completely protected economy would produce everything it consumed. One problem with such an inward orientation has already been noted: an cconomy may bc able to "produce' many goods more efficiently by producing other goods for export and using the earnings to buy imports. Protection distorts incentives, inducing countries to forgo some of the benefits of special- ization in accord with their comparative advantage. Another impor- tant cost is that developing economies are simply too small to combine self-sufficiency with an efficient scale of production. A sense of the numbers may be useful. The world's two largest free trade areas are the United States and the Eluropean Economic Com- munity (EEC), each of which provides a market in excess of $4 trillion. Trade in manufactured goods between these areas and other industrial countries is fairly free (with a few exceptions), so that the advanced- country market in which manufactured goods can be sold may be said to exceed $10 trillion annually. Meanwhile, only four developing countries-Brazil, China, India, and Mexico-offer home markets with even 1 percent of this purchasing power. Thus, an industry that develops behind a protectionist barrier in a developing country must operate on a scale that is tiny by world standards. The result is that many protected industries are unproductive -because of their very small scale. In many manufacturing industries. the minimum efficient scale of plant is believed to be! 5 percent or more of the US. or EEC market, which means. that most develoning countries cannot support even one plant of efficient scale if the plant is oriented wholly toward the domestic market. Much less can they achieve economies of multiplant- operation or support competition among a number of efficient producers. Instead, production takes place at a small fraction of the scale that manufacturers in advanced countries regard as desirable. In the automobile industry, for example, manufacturers in advanced countries typically Tegard output of sev- eral hundred thousand vehicles a year as necessary to achieve full production efficiency; yet in the protected automobile industries of Latin America, manufacturers have often produced less than 20,000 units annually. In addition to the direct costs in terms of efficiency, the fragmenta- tion of markets that comes with protection in developing countries 130 Paul Krugznan typically leadIs to problems in maintaining effective competition. Althouglh this is closely related to the problem of inefficient scale, it deserves treatment as a separate cost of protection. Monopoly Power In a small economy, which even the largest developing nations are when compared with the major advanced countries, industry will inev- itably be relatively highly concentrated. Because of the economies of firm size, a country with a GNP of $20 billion c; aniot support as many firms as one with a GNP 200 times as large. The same is true of small advanced countries: Sweden's industrial structure, for example, has a far more concentrated ownership than that of the United States. When an economy has relatively free trade and is highly open to the world economy, however, a concentrated industrial structure does not create much monopoly power. Swedish manufacturers cannot sharply raise prices to domestic consumers because if they did the consumers would turn to imports instead. In any case, the typical Swedish manu- facturer sells as much or more in foreign as in domestic markets, and in foreign markets. the firm is usually a small player rather than a monopolist. So despite a relatively concentrated industrial structure, Swedish industry is not characterized by any unusual degree of monopoly power. When a small economy follows a protectionist policy, however, it simultaneously, ensures that domestic firms are strongly. oriented toward the domestic market and that they are insulated, to at. least some degree from foreign competition. The result is that monopoly power is a much greater concern in protectionist developing countries than in advanced countries. At, first glance, monopoly might appear to be purely a distributional issue: the monopolist is better off-at the expense of the consumer, but is the country as a whole any poorer? Monopoly, however, imposes:net costs for two reasons. First, the exercise of'monopoly' power distorts prices and hence incentives: by making their goods expensive, monopolists induce consumers to shift to other goods that yield less satisfaction yet are more costly to produce. Second, the possibility of monopoly profits leads to wasteful efforts to obtain or secure these profits. Too many firms may enter an industry in an effort to share in the monopoly profits that a cartel could provide, dissipating these profits in wasteful duplication and an even more inefficient scale of production than necessary. Or established firms may invest in excess capacity to discourage additional entrants. An* important aspect of monopoly power in both:advanced and developing countries is that it -often seems to be, shared by firms with Protealion in Devebldng Countries 131 some of their workers. In highly concentrated industries that are insu- lated from foreign competition-whether in the US. automobile and steel industry or in the manufacturing sectors of most developing countries-organized labor often seems able to achieve the position of a partner in the exercise of monopoly power, receiving wages that are considerably higher than what the same workers could receive else- where in the economy. As we will argue below, the wage differentials thus created may exacerbate the labor market distortions that are sometimes used to justify protection in the first place. Rent Seeking Protection generally generates some benefits to particular groups. These benefits are a function of the policy rather than of the groups' actual contribution to the economy. The most notable example is the value of licenses to import when imports are restricted: anyone who receives a license can purchase imports at world prices and sell them at a hiigher price internally. More generally, industries that receive protection will at least temporarily earn higher rates of profits, be able to pay higher wages, or both. These benefits conferred through gov- ernment policy may be referred to in general as "rents"' - Economists have argued that a major cost of protection lies in the resources used up by groups attempting to ensure that they receive such rents. The classic example is India's allocation of foreign exchange to firms. In the 1960s, the government allocated foreign exchange for imported inputs, spare parts, and so on to firms in proportion to their installed capacity. This provided firms with an incentive to build excess capacity in an effort to increase their share of the valuable foreign exchange. This excess capacity represented a waste of resources over and above the usual costs of distorted pro- duction and consumption, market fragmentation, and so on. The rent seeking induced by protection may thus be considered an extra cost. Although most examples of rent seeking are less clear-cut, they are nonetheless important. Examples include the tendency of firms to locate in or near the capital city in order to have greater access to political influence, even when doing so is uneconomic; the tendency of workers to leave rural areas for cities, even when they have a high probability of being unemployed, to have a chance at high-paying jobs in protected industries; and the expense of retaining lobbyists to affect decisions. There is an obvious similarity between the costs of protection thatf- are a result of rent seeking and the costs of monopoly that arise from efforts to acquire or secure monopoly power. Indeed, the term "rent qure or seur monop 132 Paui Krugman seeking" itself is now widely used in the literature about industrial organization to refer to the costs associated with monopoly, even though these costs may arise without any government action. In devel- oping countries, however, because protection is a major source (prob- ably the major source) of monopoly power, one may regard all these costs as being of like kind. As we have seen, protection is likely to be harmful. Yet protectionist policies remain widespread in the developing world. No doubt much of this represents a failure to understand not only the costs but also the realities of political economy, in which economic efficiency is far from the only criterion of economic policy. Nonetheless, examining the purely economic arguments. for protection is also important, because these arguments have played a large role in influencing or at least rationalizing protectionist policies. Arguments for Protection Nearly all intellectually respectable arguments for protection, whether for advanced or developing countries, have a common conceptual basis. This basis is the view that the domestic economy is subject to some kind of failure of markets, so that the market left -to itself does not allocate resources efficiently. If a government can correctly diag- nose the market's failure, it can use protectionist policies as a partial correction: the distortion' of incentives that is a result of protection can "lean against" the inherent distortion in the market, moving the economy closer to, rather than farther from, the resource allocation it should have had in the first place. This broad view is often referred. to as the domestic distortions argument for protection. Two points should be immediately apparent First, it is by no means a blanket argument for protection:;a randomly chosen protectionist policy is surely. as likely to aggravate a domestic market failure as to cure one. Thus, this is at best an argument for a: selective, targeted trade policy. Second, if thejustification for an active govemment policy is a failure of domestic markets, trade policies.are unlikely to be the best possible answer. Instead, the best answer will be to fix the market failure directly, if that is possible, and if it is -not, to target policies specifically toward the source of the problem. In gen- eral, trade policies will turn out to be a blunt instrument for dealing with market failure. The general. description of the domestic distortions argument is important, because it helps us to see the common themes among various specific arguments. Most discussion of justifications for pro- tection, however, hinges on more specific kinds of market failure: Protection in Developing Counties 133 distortions of factor markets and possible spillovers resulting from learning and technological change. Factor Market Distortions In many developing countries the labor market exhibits two charac- teristics that suggest a significantly distorted market. First is the pres- ence of large differentials between wage rates in the relatively advanced manufacturing and urban sector and those in the tradi- tional and agricultural sector. Second is the presence of substantial persistent unemployment, especially in urban areas. Each of these has been seen as a possible justification for some kind of protection of the manufacturing sector. The reasons for a large urban-rural wage differential are still dis- puted. It seems unlikely that all of the differential is attributable to the skill levels of workers. The prevailing explanations rest in part on firms' need to. pay high wages to discourage turnover and to encour- age high-quality work (the so-called "efficiency wages" argument) and in part on the market power of unions, reinforced by government actions such as the enactment of minimum wage rates-well above the normal wage in the traditional sector. Whatever the reason for the wage differential,. it is an important source of divergence between private and social costs, leading to a suboptimal level of urban employment. A firm considering hiring an additional worker in the urban sector will do so only if that additional employee yields addi- tional revenues that exceed the urban wage rate; but because shifting a worker from rural to urban employment greatly increases the worker's wage rate, it would be worth doing from a social point-of view even if the value of the urban worker's addition to output were somewhat less - than the wage. Ideally, a developing country could eliminate the wage differential X through institutional reform. Failing that, a subsidy on urban employ- ment would be the next best policy. A subsidy to manufacturing pro- duction would not be as good, because it would have the unintended by-product of shiftingother factors of production to manufacturing as well Finally, a tariff protecting manufacturing from foreign compe- tition could have the desirable effect of increasing manufacturing employment; unfortunately, it would&have the undesirable side effects of also shifting other factors and of distorting consumption. Thus we see that the existence of wage differentials provides a possi- ble justification for protection, but only as a crude (strictly speaking, a fourth best) instrument to be used if no others are available. The existence of substantial unemployment in urban areas seems at first sight to reinforce the case for government intervention, including 134 Paul Krugman protection if nothing better, such as wagc or production subsidies, is available: any such policy may now draw workers out of unemploy- ment, a situation in- which they produce nothing, instead of from traditional occupations that still have some positive product. Some economists, however, have challenged Lhis, arguing that urban unem- ployment in developing countries is itself a consequence of the wage differential: workers move to urban areas to have access to higher paying jobs than are available elsewhere. In this case an increase in urban employment, by inducing more rural-urban migration, may. actually increase the number of urban unemployed. An important critique of arguments for protectionist policies based on labor market distortions is that these distortions are themselves exacerbated by protection. In particular, the urban-rural wage differ- ential, which is a problem in itself and may underlie some of the persistent urban unemployment, probably arises in part because of the monopoly power of industrial unions. As we have argued in the previous section, however, protection increases monopoly power and is therefore likely to make the wage differential larger than it would otherwise be. Quite conceivably, a government that protects its manu- facturing sector helps engender a wage differential that in tum leads to rural-urban migration, creating as a result precisely the labor mar- ket distortions that are often used to justify protection. The Infant Indus ty Argument Among the oldest arguments for protection is the idea that new indus- tries need a period of being sheltered from their established competi- tors in other countries in order to become established. The infant industry argument, which was used to justify the tariffs that protected manufacturing on the Federal Republic of Germany, Japan, and the United States in the past, continues to have popular appeal. That the infant industry argument is a domestic distortions argu- ment for protection may not be readily apparent. En the absence of any domestic market failure, however, there would be no reason to give special protection to infant industries. If an industry were worth developing, in an undistorted economy it would be privately profit- able to do so. For the infant industry argument to be valid, there must be some divergence between the private and social benefits of invest- ing in a new sector. One kind of divergence might be a. failure of the capital market. Suppose that a firm entering a -new industry must go through an extended period of losses before eventually becoming profitable and that for some reason capital markets in a developing country will not finance it through this entry period. Then entry of new firms may be Pratection in Developing Countries 135 facilitated by temporary protection that makes firms profitable even during the breaking-in period. The question is whether the capital markets are in fact inadequate in this way and, if so, why. In practice, there are reasons to doubt the inadequacy: many "infant" industries in developing countries are made up of subsidiaries of multinational firms, which presumably have little problem of access to capital, and even domestic firms often have close links to banking groups that could presumably finance them if the prospect of future profitability were fairly likely. In any case, the substantial imperfections of capital markets in developing countries are in large part the result of govern- ment regulation, so that the appropriate policy would be one of insti- tutional reform rather than protection. Another reason why infant industries may need help lies in external economies of learning and technological change. Creating a new industry in a developing country generally requires some investment in adapting the technology to local conditions. Yet a firm that makes this investment may find that it is unable to appropriate the benefits fully because other firms can follow- its lead and benefit from its efforts. Thus private agents may fail to develop an industry even when this is socially.desirable. In this case, a temporary period of protection can help correct the market failure.. Two points need to be emphasized, however. First, the need for such temporary protection is almost inevitably difficult to ascertain. Nobody knows how to put a value on spillovers of knowIedge between firms with any accuracy, especially in a prospective rather than an established industry. Thus the infant industry argument is a highly speculative one, easily abused to justify protection for other reasons. Second, the justification offered by the infant industry argument is not really for protectionist trade policy, but for an policy specifically aimed at encouraging investment in knowledge. Even a less targeted policy of encouraging production would be preferable to limitations on imports. This is particularly true if one remembers that infant industries could be aimed at exports rather than import substitution. "StrategicArguments" Recent work in the theory of international trade has emphasized the importance of imperfect competition and increasing returns to scale in the world economy. Out of this new theory emerges a new class of arguments for government intervention in international trade. These arguments arise from the potential "strategic" role of government action in shifting the terms of competition'to the benefit of domestic' firms. We shall refer to them in short as "'strategic arguments" The most widely known example is the one for export subsidies as a way of 136 Paul Kn gman deterring foreign competition. The proponents, Brander and Spencer (1985), show that under the right circumstances, a preannounced export subsidy, by inducing foreign firms to produce less, can raise the profits of domestic firms by much more than the amount of the sub- sidy. The result is to raise the national income of the subsidizing country at foreign expense. Simnilar arguments can be used to show that a protected domestic market can give domestic firms a strategic advantage that similarly raises national income. For the most part, trade policy arguments based on strategic needs rely on two assumptions that rarely apply to developing countries: first, the country is assumed to be a major supplier in the market for a good being produced. in an imperfectly competitive market, and, sec- ond, the country's own domestic market is assumed to be a significarnt part of the world market. As already emphasized, developing coun- tries are uniformly small in economic terms, so that they are never major markets for traded goods. They are also rarely major producers, except where the good is a raw material of which they happen to have a large share. Thus the particular strategic trade arguments that have attracted the most attention in discussions of the competition among the United States,Japan, and Europe are largely irrelevant to develop- ing countries. Some related arguments may have more relevance. One possibility pointed out in the new literature is that because of the prevalence of oligopoly in world markets even a small country may be able to affect the prices of its imports. In principle, trade restrictions or a govern- ment marketing board could induce foreign suppliers to provide goods more cheaply than they would otherwise. Also, entry into some industries may yield rents to specific factors of production even though monopoly profits are competed away, so that a policy of export promotion based on strategic needs can still raise natibnal income. These possibilities-are discussed in Krugman (1989). .Overall, however, the new strategic arguments seem unlikely to add much force to the argument for protection in developing countries. This argument still rests primarily on considerations of factor market distortions and infant industries. If these considerations seem dubious when compared with the costs described earlier, this simply says that in the current state of knowledge, the case for protection as a tool of development strategy is not a very strong one. Fonns of Protection . The effects of protection often depend crucially on the specific char- acter of the protectionist policy. We therefore turn next to a review of some of the main forms of protection and their effects. Protection in Developing Countries 137 Ta riffs A tariff-a tax levied on imported goods at the border-is the oldest form of protection. The effect of a tariff is usually straightforward: it raises the price of the imported good by an amount equal to the tariff. There are only two exceptions. First is where the importing country is a large share of the world market for the imported good, in which case the tariff will lower world prices somewhat and be less than fully passed on in domestic prices. This case is of little importance for developing countries. The other is where the good is supplied by monopolistic or oligopolistic firms that may choose to absorb part of the tariff rather than raise prices by the. full amount. The importance of this effect in practice is not clear. Tariffs were the main instruments of protection before the 1930s and in particular were the basis of the import-substituting industrial- ization the United States and Germany pursued in the nineteenth century In most of the world, however, tariffs have become less impor- tant than administrative protection that sets quantities rather than simply taxing goods at the border. Import Quotas and Exchange Controls An import quota sets a limit on the. quantity of a good that may be imported during a given period. Control of foreign exchange, under. which the quantity of foreign exchange available for imports within a particular category is restricted, has a similar effect. The most common misunderstanding about import quotas is the view that, unlike tariffs, they do not raise prices. In practice, by restricting the supply of imported goods to the domestic market, an import quota raises the price that domestic residents are willing to pay for the good by at least as much as a tariff that restricts imports by the same amount. Because a quota raises domestic prices, the right to import a restricted good is valuable: the good may be sold on the -domestic market at a price higher than its purchase price. Thus, to administer a quota, the authorities must allocate the rights to import, usually through a system of import licenses. These licenses are sometimes auctioned off; more often they are allocated through some administra- tive procedure. The first difference between a tariff and an import quota is imme- diately apparent. A tariff yields revenue to the government, whereas a quota, unless it is auctioned off, yields rents to the recipients of import licenses. In highly protected economies these quota rents are often a substantial share of national income, and their allocation is -therefore 138 Paul Krmn a key government policy. As already mentioned, rent seeking is a possi- ble source of additional 'costs to protection. Most economists believe that quotas are much more likely to generate rent-seeking activities than are tariffs. Ain additional difference between tariffs and quotas lies in their effects on domestic monopoly power. Consider a firm that has a domestic monopoly position protected by a tariff. Although this firm can raise its price to a certain point, if it raises its price too high customers will shift to imports despite the tariff; If the same firm is protected by an import quota, customers will be prevented from shift- ing to imports; when the firm's price is increased, the premium charged by holders of import licenses will rise again. As a result, a firm protected by a quota will, other things being equal,.find it more prof- itable to exercise its monopoly power and raise domestic prices. It is in this sense that a quota, which does not directly tax imports, will often raise prices more than a tariff will. Finding reasons for preferring an import quota to a tariff is diffi- cult. The loss of revenue, the costs of rent seeking, and the increase of monopoly power all point toward the conclusion that tariffs are at any rate less costly than quotas for any given-degree:of protection. The popularity of quotas and exchange controls thus lends credence to the most cynical interpretations of protectionism in developing countries: Either the countries do not know what they are doing, or their govern- ments are actually eager to create the private rents associated-with import licensing. Cotent RequiTeenwits and Trae Offsets In some cases, the authorities allocate import licenses according to a formula that makes their receipt conditional on other actions by firms. In this case the quota rents are in effect used to subsidize same other desired activity. The most familiar schemes of this type are con- tent restrictions on sales and the more general category of trade offsets. A content restriction requires. that some minimum fraction of a product category sold in the country represent domestic value added. For example, a country might require that 80 percent of the value of an automobile, sold in the country represent domestic content, with only 20 percent being imported. If the required -local share is larger than what unrestricted firms would have chosen to do, the effect of the scheme is like a tariff on imports whose proceeds are used to subsidize 'domestic production. Producers 'would like -the fight to import, because imports are cheaper than their domestic substitutes; they are Protection in Dvewoping Countries 139 therefore willing to sell domestic production at a loss, because this gives them the-right to bring in imports. A trade offset is a generalization of a content restriction that allows exports to be traded off for imports. For example, one might have a scheme that required only that a firm produce 80 percent of domestic value added per unit of domestic sales, without requiring 80 percent content on those sales themselves. Then the firm might choose to have some exports, say, equal to 10 percent of domestic sales, allowing it to have imports equal to 30 percent. The advantage of trade offsets over content requirements is that they allow more flexibility. Firms can specialize in the products for which the country's cost disadvantage is smallest, and they can also produce a narrower product range to achieve greater economies of scale. Both content restrictions and trade offsets differ from straight import quotas in not generating quota rents. In effect, firnns compete for import licenses by engaging in domestic production, so that the quota rents are redistributed in an implicit subsidy. * . The Extent of Protection Considerable effort has gone into measuring the actual extent of.pro- tection in developing countries. One might think that this would be a straightforward matter of simply cataloging the trade policy measures in place. In fact, the issue is fairly difficult. First, the use of a variety of forms of protection (tariffs, quotas, content requirements, and so on) poses the problem of establishing a-common measure of protective impact. Second, the degree of protection provided to any given activ- ity depends not only on the direct protection given to that activity but also on the indirect effects of tariffs and import restrictions on other sectors. Measuing Protectzonv TariffEqutvaLs.- As we have noted, actual systems of protection mix a variety of tools. To come up with meaningful estimates of the extent of protection, the usual method is to calculate what system of tariffs would have pro- vided industries with the same incendves and then to. measure the extent of protection by the percentage of tariff equivalen:t. thus calculated. The basic idea is straightforward. Suppose that an import quota raises the internal price of the restricted import 60 percent above the world level. Then the import quota is considered equivalent to a 60 percent tariff. Three issues complicate this measurement in practice, two of them issues of principle, one an issue of practicality. 140 Paul KrgVman First, the effects of oLher protective measures may not be fully equivalent to those of tariffs. As noted earlier, most economists believe that an import quota creates more monopoly power and raises inter- nal prices by more than a tariff that limits imports to the same amount. As a result, there may be no meaningful equivalent tariff into which to convert a quota. Using the implicit rent on quota licenses, which is how it will be done in practice, will not tell the full story. Second, complex schemes such as content requirements will be dif- ferent from tariffs in their relative effects on consumer and producer prices. A simple tariff on automobiles will yield revenue to the govern- ment. A local content requirement will in effect distribute that income as a subsidy to domestic producers, so that domestic producer prices rise by more than the rise in consumer prices. The normal convention in measuring protection is to focus on the effects on producer incen- tives rather than on consumers. This represents a judgment about relative importance that is defensible, but not entirely satisfactory. Finally, measurement of tariff equivalents is difficult. A regulation that limits imports does not specify the equivalent tariff that results, so the economist must study, not only the government's actions but also -the effect of these actions in the marketplace. Economists frequently discover that the rates of tariff-equivalent protection. resulting from the administrative restriction of imports are substantially larger than the government had realized, so that when. countries replace quotas with tariffs they often lower the rate of protection without meaning to. Despite these problems, the measurement of protection is a useful exercise, giving at least a rough idea of the quantitative importance of the kinds of trade policy discussed earlier. Effective Protection If all: goods were produced entirely from capital and labor inputs, without the use of intermediate inputs that might be imported or exported, computation of tariff equivalents would complete the job of calculating rates-of protection. Because this is not the case-especially in manufacturing in developing countries, where industries may add only a thin slice of value added to a product-the notion of protection must be elaborated on. The basic tool used to do this is the concept of* effective protection, which measures the degree of protection offered to particular activities rather than simply the tariff rate. Probably the best.way to understand effective protection is to look at a simple example. Suppose that a country can buy complete auto- mobiles on the world market for $10,000. Suppose, alternatively, that it can purchase the parts needed to. construct an automobile for $8,000. Finally, suppose that the domestic automobile assembly indus-- Protection in Developing Countries 141. try, which is inefficient by world standards, cannot assemble a vehicle for less than $4,000. To encourage the domestic assembly of auto- mobiles, the government would have to make the price of an. assem- bled import equal to $12,000, the price at which a domestic firm could sell a vehicle assembled from foreign parts. This can be done by imposing a 20 percent tariff or equivalent on imported automobiles. What rate of protection is the domestic automobile assembly indus- try receiving in this case? The tariff rate on automobiles is 20 percent, but this tariff allows the domestic industry to assemble a car despite the fact that its costs are twice the value added it provides at world prices. Therefore, the 20 percent tariff provides the domestic industry with an effective rate of protection of 100 percent. This example shows that the effective rate of protection provided by a tariff may differ substantially from the nominal tariff rate. It should also be apparent from the example that tariffs on the inputs used to produce a good, as well as the'tariff on the good itself, enter into the determination of the effective rate. For example, a 5 percent tariff on imported automobile parts, other things being equal, would reduce the margin on which domestic assemblers can work from $2,000 to $1,600, that is, it would amount to a negative 20 percent rate of effec- tive protection for domestic automobile assemblers. Overall, assessing effective protection thus requires measuring tariff equivalents not only on an industry's output but also on all its inputs. Evidence on ihe Extent of Protection Table 6-1 presents some estimates of effective rates of protection in several developing countries. Although calculating effective rates does not tell us whether the costs of protection exceed its benefits, evidence Thble 6-1. Effective Rates of Protection in Selected-Developing Countries, Selected Years, 1967-81 (percent) Average ratefor Spread across Coutry and year of estin ste manuqacturing industries Brazil, 1980-81 23 -85 to 219 Chile, 1967 217 -23to 1,140 Colombia, 1979 55 25 to 127 Korea, Rep. of, 1978 5 -38 to 135 Nigeria, 1980 82 -62 to ,119 Philippines, 1980 44. Singapore, 1967 0 -7 to 21 -Not available. Surace World Bank (1987). 142 Paul Krtgrnan such as that summarized in table 6-1 has been widely used to make the case that protection in developing countries is harmful. Three points stand out. First, effective rates of protection for some industries are very high. Recall that an industry receiving an effective rate of protec- tion of 100 percent is in effect producing goods that are worth only half as much at world prices as they cost to produce. One must place very great weight on the arguments for protection described earlier to second-guess market prices Lhis much. Yet effective rates of protection exceeding 100 percent are quite common. This observation suggests that justifying the protective structures of these countries on a cost- benefit basis must be hard. Second, the rates of protection are highly variable across industries within each country. Are governments really well enough informed to know that one industry needs no protection whereas another is worth an effective rate of several hundred percent? Again, one cannot avoid a heavy dose of skepticism. Third, the most successful manufacturing nations in the table, Singapore and the Republic of Korea, have, on average, virtually no protection rates for manufacturing This apparent correlation of rela- tively free trade and economic success lends some weight to the view that protectionism is associated with large costs, although economists have offered other explanations for the relative success of outward- looking economies. A Broader Anaysis of Protection Although the concept of effective protection is a useful tool for mea- suring the protection of particular activities, economists have long recognized that it gives only an incomplete picture of the way in which protection affects a country's industrial structure. There are other. channels-more indirect, but ultimately equally powerful-through which protection of one sector affects the output of others.' One such channel is the exchange rate. Protection, other things. being equal, reduces imports and tends to create a trade surplus. Such an effect is always transitory, however. Either the currency appreciates as a result of the trade surplus, or its rate of depreciation fails to keep up with inflation, which has the same effect. The end result is a stron- ger real exchange rate. This in turn makes unprotected sectors less competitive: Both import sectors that receive low rates of effective protection and export sectors:find that their costs rise in terms of foreign currency, and thus they are actually taxed when some sectors receive high rates of effective protection. World Bank estimates find that every percentage point of taxation of imports translates into a tax of 0.4 to 0.8 percentage points on exports. Protection in Developing Countries 143 Protected sectors also compete with other sectors for scarce inputs. A policy that creates capital-intensive industries behind tariffs or quotas will make capital scarcer, and this scarcity will affect the incen- tives for production throughout the economy. Sometimcs such effccts may be indirect and surprising. Consider the following hypothetical example. Argentina protects its manufacturing sector, which draws labor away from agriculture. The shortage of labor that results forces a shift away from labor-intensive cultivation, such as wheat farming, toward livestock raising; thus, in an indirect way, protecting manufac- tures actually encourages beef exports. In principle, fully specified "general equilibrium" models can keep track of all these indirect effects. Such models have been developed for several countries and ultimately represent a better approach. Because the models are difficult to construct and rely on assumptions that are often disputed, however, they have not yet displaced calcula- tions of simple effective protection as a tool of practical analysis. Cost-Benefit Analysis of Protection * A tariff or import quota normally produces a mixture of costs and benefits to different groups-in the-economy. Higher prices hurt con- sumers while benefiting import-competing producers. In addition, either the government gains revenue from a tariff or the holders of. import licenses gain rents. The net costs of a tariff or import quota are measured by the differ- ence between the consumer costs and the benefits. In the simplest. analysis, net costs arise from the distortion of incentives described above. There is a producer distortion loss because producers manu- facture goods that could be imported at a lower cost. And there is a consumer distortion loss because consumers shift away from. imports on the basis of an internal price, that exceeds the true cost of the goods to the economy. Standard techniques exist for calculating these costs, and, although the amount is sometimes a matter of dispute, the costs: X themselves are not controversial in principle. Many of the effects of protection, however, go beyond these distor- tion losses. On the cost side, the gain to producers may be dissipatcd by firms' efforts to compute monopoly profits. As discussed earlier,- the proliferation of firms of inefficient size and the maintenance of excess capacity in order to deter competition may absorb much of the initial gain. Increased wage differentials and the increased un- employment that results may also effectively wipe out some of the gain.' When protection takes the. form of an import quota rather than a tariff, the rents may also be dissipaLcd Ay finns' efforts to secure d uyim fot oscr 144 Paul Krugman import licenses, whether through the emergence of excess capacity or through resources expended on lobbying. On the other side, there may be benefits from protection. Empirical estimates of the costs and benefits of protection start with the con- sumption and distortion losses caused by distorted incentives. They then proceed to make the best imputations they can of the other effects. In general, researchers have to rely heavily on plausible assumptions in doing this; thus, wide uncertainty surrounds any esti- mates of the cost of protection. Most empirical studies give short shrift to the potential benefits. This reflects thc judgment of the researchers that protection has had little success in countering domestic distor- tions and may have actually aggravated them in many cases. Although this is a plausible view, the estimates of the costs of protection are not only based on many assumptions but also somewhat incomplete. Nonetheless, it is worth considering some sample calculations of the costs of protection (none of the empirical estimates for developing countries find net benefits). Table 6-2 shows estimates of the costs of protection in several devel- oping countries at various times, measured as a share of the gross national product. The estimated costs are in general several percent of GDP, but less than 10. That is, these costs are significant but not overwhelming. What conclusion should one draw from the range of costs? Keeping in mind the caveats about the estimates, the general conclusion is clearly negative for advocates of protection as part of a development strategy. In practice, countries have done themselves significant harm rather than good with protection. Anyone. advocating protectionist, development policies must explain why future results should be any different. The costs of protection are not so large, however, that one could describe protectionism as the principal obstacle to development. The largest estimates of costs are no more than two year's growth differen- tial between the successful outward-riented economies and some of Table 6-2. Estimated Costs of Protection in Selected Developing Countiear. Selected Years, 1960-78 Percentageof Conshy and year ofestimate. GDP Brazil, 1966 9.5 Mexico, 1960 2.5 Pakistan, 1963 6.2 Philippines, 1978 -5.2 Souras: For Brazil, Mexico, and the Philippines, Belassa (1971); for Pakistan, Clarete and Whalley (1985).; I. ~ ~ ~ ~ ~ ~ ~ ~ ~ . . Prolection in Developing Countries 145 the countries where import substitution has had the worst failures. This poses a puzzle: if protection has only the moderate importance we see here, why is the apparent difference in performance between outward-oriented and inward-looking countries as large as it is? From 1973 to 1985, countries that the World Bank classified as strongly outward-oriented experienced real. per capita growth at more than 5 percent annually, whereas those classified as strongly inward-oriented experienced a slight decline in per capita output. Why this is so is still a matter of dispute. Some economists argue that there are additional dynamic costs of protection. Others argue that the correlation between protection and growth is accidental and that other policies- or cultural factors-play the key role in rapid growth. The Political Economy of Protection Theory and evidence both suggest that protectionist policies in devel- oping-countries, at least on the scale practiced by many nations, have costs that substantially exceed their benefits. Yet such policies remain in place, and major-trade liberalizations are infrequent. No discussion of protection in development can be complete without. some explana- tion of why protection persists despite its costs. - Broadly speaking, the political economy of persistent protection has three levels. At the highest. level, liberlization imposes transition: costs that may deter a government: The short-run costs- may.look larger and more certain than the eventual, gains. Less charitably, we may note that even if the nation a& a whole benefits from liberalized trade, important groups may suffer losses, at least at first, and a gov- ernment may- not be willing or able to face the political costs- of income redistribution. Finally, protectionist trade polities, especially those involving substantial administrative discretion, provide oppor- tunities for influence and profit to government officials;.These offi- cials will often oppose reforms that deprive them of that role.; Transitiona Costs The transitional costs associated with trade liberalization arise from two main sources-the balance of. payments and. trnnsitional unemployment. Other things being equal, trade liberalization will lead to a worsen- ing:of the. liberalizing country's trade balance. The public will demand more imports, which will place a strain on the country if it is facing constraints.on its international financing. The import,boom that follows a trade liberalization may, perversely, beexaggerated if the liberalization lacks full credibility. When tariffs 146 Paul Krugman. are reduced and quotas lifted, imports become cheaper. If the public does not regard this state as permanent and expects controls to be reimposed, it will rush to buy imports, especially of durable goods, while it can. In some cases this lack of confidence in liberalization has been a self-fulfilling prophecy, in which an inrush of imports leads the government to reimpose controls. The authorities can avoid the balance of payments effects of trade liberalization if the trade reform is accompanied by currency devalua- tion. Indeed, a devaluation-cum-liberalization strategy is.precisely the orthodox one. Countries facing inflationary pressures are usually reluctant to make the steep devaluations that.they may need to accom- pany a trade liberalization, whereas countries facing external financ- ing problems are reluctant to count on a devaluatuon to offset the increase in imports. Thus, as a practical matter, the balance of pay- ments problem cannot be easily dismissed. Aside from the balance of payments issue, there is the problem of transitional unemployment. This may best be understood by consider- ing our earlier discussion of wage differentials. Suppose that a country has built up a protected manufacturing sector' that pays wages well- above what the same workers could receive elsewhere. With trade lib eralization the wages of those workers will no longer be sustainable at the same level, yet their wage rates may be slow to come down. In the meantime, employment in the manufacturing sector-particularly in the protected sectors-may fall sharply, leading to a rise in unemploy- menL Unlike the balance of payments problem, this transitional unemployment may be unavoidable even with appropriate currency and monetary policies, because it ultimately stems from the need to get some workers to accept a decline in real wages (which one hopes is offset by real wage increases for others).. The transitional problems associated with trade liberalization are certainly real. Yet they do not fully account for the unwillingness of countries to dismantle costly protectionist regimes. We need to turn next to less creditable explanations. Income Distribuiion Any change in trade policy creates winners and losers, with the redis- tribution normally large relative to the net gain. So considerations of - income distribution clearly matter. When a country liberalizes trade, firms-and to some extent workers-in protected sectors lose, whereas the rest of the economy gains. The key question is why the losses often seem to outweigh the gains in political calculations, even though in- terms of national -income the rewards from freer trade should be larger than the costs. One answer Proftaction in Developing Countries .147 might be social concern: If protection generally helped the relatively badly off at the expense of the more affluent, one could understand a reluctance to liberalize. This does not, however, seem to be true in general; if anything, the evidence suggests that protectionist economies have a less equal income distribution than those with freer trade. A better explanation may be a general dislike on the part of political actors for large changes, regardless of who is affected. To remove protection, once it has become established,, would shake up existing political and economic compromises, and leaders are reluctant to take the risk of doing so. Finally, the beneficiaries of protection usually have an organiza- tional advantage over their opponents. The benefits of, say, an import quota are concentrated on a few firms and.labor unions, whereas the costs are broadly diffused over the population. The beneficiaries are usually therefore better informed and. better organized than those who are hurt. Indeed; consumers may not even realize the effect of trade policy on their standard of living. influence and Comrption In discussing protection in practice, it is impossible to avoid the blunt fact that some government officials value protection because of the privileges and power it gives them. The preference of countries for administrative measures rather than tariffs to protect industries, despite an overwhelming economic case against this practice, makes sense only if one thinks of the greater role for officials in an adminis- tration system. Most charitably, one may assert that:officials find ways to convince themselves that a system that makes their decisions impor- tant is desirable. Least charitably, one may assert that quota rents have a way of being appropriated by those who allocate the rights to import. To suggest that bad policy results from the desire of public servants to enrich themselves at public expense is embarrassingly crude, but it is not irrelevant The unfortunate fact is that long-standing protectionist regimes. end up being defended. by "iron triangles" of interested groups- firms that depend on the barriers, organized labor that extracts wages above the level in unprotected sectors, and government officials who gain influence and perhaps profit from their role in controlling trade.:. The strength of these triangles is such that major trade reforms usually occur only following severe political or economic crises. Conclusion Despite movements toward liberalization in some developing coun- tries, high rates of effective protection remain in much of the develop- 148 Paul KInan ing world. Although some intellectually respectable arguments favor limited, selective protection, the protectionist policies actually pur- sued in developing countries are almost surely the cause of significant economic costs. Unfortunately, protection is not continued simply because coun- tries are misguided in their economic analyses. The main problem in the political economy of protection is that any'long-established pro- tectionist regime tends to generate a set of economic and political interests that make it self-perpetuating except under crisis conditions. Selected Bibliography Banuri, Tkriq, ed. 1991. Economic Liberalizatiow ho Panacea. New York: Oxford University Press. Belassa, Bela. 1971. Thle Structure of Potection in Developing Caoutries. Baltimore, Md.:Johns Hopkins University Press. Brander,James L., and BarbaraJ. Spencer. 1985. "Export Subsidies and Inter- national Market Share Rivalry"Journal of Inrnatfional Ecownmics 16(1-2):83- 100. Clarete, Ramon, and John Whalley. 1985. Interactions between Trade Policies and Domestic Distortions. Centre for the Study of International Economic Rela- tions Working Paper- 82550. London, Ontario: University of Western Ontario. Dornbusch, Rudiger. 1992. "The Case for Trade Liberalization in Developing Countries"Jouna of Economic Pepectives 6(l):69-85. Krueger, Anne O., ed. 1988. Development with Trade. San Francisco: Institute for Contemporary Studies. Krugmnan, Paul. 1989. "New Trade Theory and-the Less Developed Countries." In Guillermo A. Calvo, Ron Findlay, Pentti Kouri, and Jorge de Macedo, eds.,Debt Stabilization and Development. Oxford, U.K.: Basil Blackwell. Thomas, Vinod, and James M. Nash, eds. 1991. Best Practice in Trade Policy Reform. New York: Oxford University Press. World Bank. 1987. World Development Report 1987. New York: Oxford Univer- sity Press. 7- The Other Side of Tax Reform - Arnold C. Harberger EcoNOMISTS DEALING with public policy, public finance, ecanomic development, and related topics confront the subject of tax reform time and again. Mostly, the confrontation has taken place in an arena defined by economic treatises, textbooks, and journals. Within that aTena, economic efficiency has been the guiding star, the dominant criterion governing the discussion. This essay deviates from the standard practice just described and views tax reform from different vantage points. Simply put, it exam- ines tax reform fromn the standpoints of people who are involved in different layers of the process-from the finance ministers who spon- sor the reforms, to the taxpayers who have to live with them. In between are at least two other important groups: the administrators who must implement the reforms and the interest groups that are principally affected. My comments about tax reform will not be highly structured -and formal. Rather, consider them as the musings and reflections of some- one who has witnessed many efforts at tax reform and participated in more then a few. The Moral Flavor of Tax Reform To me, the key to the success of many tax reform efforts lies in the mora. dimension.-We start from a point at which an old systemiof gathering revenue has broken down. The taxes that are on the books are subject to widespread and quite open evasion. Administration has become lax. Rumors, and often far more than just rumors, abound of corruption on the part of those who administer the taxes. Usually part of the fault lies in the bad design of existing taxes-bad design by technical standards and bad design in terms of how various- groups of taxpayers perceive these taxes. The seecds of evasion are' 149 150 Arnold C Harberger often sown when different groups of taxpayers come to feel that they have been singled out for unfhir treatment. They then try to "correct" the unfairness by simply not paying the full tax that is legally due. On the whole, when the need for reform is greatest, the most impor- tant sources of the problem go far beyond simple issues of bad design. In the process of the breakdown of a system, quite ordinary people- who generally think of themselves as honorable and hioncst-end up participating in gross evasion of taxes. Why? How? Their standard answer is, "I would be a fool to pay the full tax when nobody else does. Once this idea gets around, it is like a virus; anybody can become infected. The judges in the tax courts may themselves be evading, but even if they are not, they know their brothers and sisters and friends are doing it. So the honest judge becomes the one who looks on evasion as something that almost everybody does and who lets evaders off with a small fine or a warning. The dishonest judge becomes the one who extracts a bribe for doing so. No judge actually nakes a serious attempt to enforce the law as it stands.. The problem can get considerably more complicated if a "tax-and- evasion race" between the fiscal authorities and the taxpayer ensues. The authorities, finding that:the taxpayers are paying only half of what they should according to law, respond by doubling tax rates. But now taxpayers react by evading more, triggering yet another increment in rates by the authorities. The end result of such a vicious cycle is a set of tax rates so high that even the most conscientious taxpayers are driven to evasion. As is so often the case with major policy mistakes, no straightforward solution to the problem exists.' I do not like to use the word "corrupt" to characterize socieiis that end up in the situation just described. Too many fundamentally decent, ordinary people are involved to warrant so harsh a term. I prefer to sayt hat such a'society has lost its moral tone-has become demoralized-in tax matters. Perhaps the biggest task that faces the tax reformer in such a case is to rekindle the society's moral tone: to turn it from a society composed mainly of tax evaders to one in which most people comply with the tax law. It would be naive to think that a spiritual conversion is all that is required here. Most successfully administered tax systems anywhere -in the world incorporate penalties for noncompliance. On examining these tax systems a bit more closely, however, we find that the affected populace typically considers them to be reasonably fair in their design and administration. By fairness I mean "absence of capriciousness" or, in a very basic sense, 'justice:' I do not mean fairness in the sense of exactly the right sharing of the .tax -burden among different groups and categories of taxpayers (what we used to call vertical equity). Nor do I mean. fairness The Other Sie of 7hx Reform 15 in the sensc of people in similar-situations being treated equally by the tax law (what we used to call horizontal equity). As an example of vertical equity, a well-administered, flat-rate income tax could easily fit into my definition of fairness, even though the society might in some sensc prefer (and maybe even in the next election vote to institute) some degree of progression. As an example of horizontal equity, I can imagine a tax law being fair, according to my definition, and yet giving substanLial tax breaks, say, to farmers or to people living in remote regions of the country, even when the society at its next election opportunity would eliminate these preferences. "Capriciousness" is present when large numbers of individual tax- payers experience a sense of injustice concerning the way the system treats them. Mainly this occurs through differential evasion by differ- ent groups, somc groups being able to get away with paying little or no tax, and others (usuaUly wage and salary earners), having no easy escape route, being forced to comply.. But it can also occur in an exacerbated fashion if tax collectors go aroundvsoliciting bribes for special favors. To turn around a demoralized tax system is a major task. It requires working at all levels, from the tax collector down to the most incon- spicuous individual taxpayer. The surest (and most humane) method is to go after the- most flagrant violators first. A good starting -place is with the most corrupt officials on the one- hand, and their biggest and most notorious clients on the other. Severe penalties should be imposed on such people, not just -for their own sake but as a warning to other administrators and taxpayers who are not quite such blatant violators. Step-by-step, the -cutting edge of enforcement is moved closer. and closer to the average administrator and the avenge tax-; payer. As-this happens, more and more of these people modify their behavior so as to conform to the -law. In the process, the demoralized society becomes remoralized. Designing a lix System for Efficiency and Fairness The more I have looked at situations in which reform has been the aim and the watchword, and the more I have reflected;on the solutions adopted. or recommended by wise and observant people, the more I have come to feel that these solutions embodied important criteria. above and beyond pure economic efficiency. Let me begin this topic. by setting out what I believe to be the "expert-consensus package" for the 1990s. I am thinking here: particularly of small, open, developing economies. The experts are those from the countries themselves; from the World Bank, the IMF, and other international as well- as bilateral. e.onomic assistance agencies; and from the segment of the academic J52 Arnold C. Harberger world that deals with the economic policy problems of actual countries. Where has the thinking of these experts been heading over the last decade or so? As I perceive it, the emerging consensus includes a broadly based value added tax, a tariff structure that moves over time toward uniformity, and a set of sumptuary and "luxury" taxes to deal with consumption of special items. At the level of enterprises, there is an increasing recognition of the folly of levying progressive taxes on the income of -business firms. The consensus is definitely for a uni- form rate (perhaps with a minimum exempt amount), with the rate itself established with an i ye to'the tax situations of foreign and multi- national firms that have invested or may invest in the currency in question. The experts ar virtually unanimous in deploring the wide- spread use of tax incentives aimed at all -kinds of (real or alleged) special objectives. The Value Added Mlm The value added tax is an interesting case study in itself. In textbooks and classrooms this tax is almost invariably expounded as if it were a fully general tax on all production'of goods and services in the econ- omy. Given that at each level of production, inputs (including capital. goods) purchased from other levels are deductible in the computation of value added, the fully general tax on all productive activities - becomes a fully general tax on consumption. Typically, textbook and classroom presentations will at this point note that even this tax of maximal generality does not reach the, leisure (nonmarket) time of workers. and their families, so that one cannot assert that it is comr pletely neutral. But their argument for a value added tax nonetheless rests on its comingjust about as close as is possible in the real world to an idealized vision of a neutral tax. There is a sharp contrast between the foregoing picture of a value added tax and what one sees when one looks at value added taxes in practice. The striking feature- of real-world value added taxes is how- much their actual revenue yield differs from-what-would be generted by a fully general tax. Take, for simplicity, a case in vhich a uniform' rate of value added tax is in effect. If it were a fully general tax of the consumption type, its theoretical yield would be equal to the tax rate multiplied by total consumption in the economy. When we perfonn this exercise, however, we find that actual tax yields rarely reach half of the theoretical yield. Why do actual yields fall so far short of the hypothetical yield of a fully general tax at a similar rate? Part of the answer lies in evasion, which is important in some countries. But the bulk of the answer, The Other Side of Tax Reform 153 universally among countries (even those with superb tax administra- tion and hence minimal evasion), is that the tax, far from being fully general,covers an amazingly moderate part of the total consumption. Looking simply at the part of consumption that is counted as such in the calculation of national income, we find that such activities as medical services, educational services, charitable institutions, and so on are universally exempt. So, often, are other professional services and those of craftsmen and tradesmen. Domestic service, which is quite important in many developing countries, is almost never reached by the net of a value added tax. Finally, there is the imputed income from owner-occupied housing, that proverbial wraith of the tax world that seems to elude the tax collector's grasp, no matter what, no matter where. Coverage of the. Value Added Tax Apart from the sectors and activities that are more or less routinely left aside in the laws defining the base of a value added tax, other activities exist whose treatment is the result of more conscious deci- sions by policymakers. In this vein, sometimes all agriculture is left out of the network and sometimes farams below a certain economic size. The areas producing financial services-banks,fina icieras, insurance companies, and so on-are almost invariably left out of the jurisdic- tion of the value added tax. These sectors, particularly agriculture and finance, have the curious attribute that one actually does not know whether one gains or loses tax revenue by leaving them out of the value added tax net. The apparent paradox of actually gaining revenue by leaving a pro- * ductive sector uncovered by a tax is easily explained. The key lies-in. the credit method.,.by which most real-world value added taxes are administered. Under the credit method, each enterprise that is cov- ered must pay tax on the value of all its production (sales). It can then claim credit for the value added iax paid by supplier enterprises on the various inputs (including investment goods) that it bought from them during the tax period. Suppose that we have a flour mill that buys wheat from farmers. The mill always pays tax on the flour it produces. If the wheat farmers who supply it are part of the tax network, they pay tax on-the value of their wheat, and the mill gets to deduct this tax from the amount it owes to the government. If, on the contrary, the.farmers are not included in the tax network; the mill is not able to deduct any tax previously. paid. The mill in effect ends up paying the farmers' tax for them. So far, we have no paradox. The farmers pay. tax on the output if they are in the system, and the- mill pays tax on 154 Arnold C. Harberger the same output if the farmers are out of the system. But now consider the tax treatment of the farmers' inputs. If the farmers are included in the value added tax network, they get to deduct the tax paid on their inputs; if they are excluded from the network, they receive no deduction, and neither do the mills that process their wheat into flour. When the farmers are out of the system, they consti- tute a break in the value added tax chain. Because of this break, neither they (the farmers) nor the food processors who.buy their output receive a credit for the tax embodied in purchased farm inputs. Leaving farmers out of the system does not always yield greater tax revenue. Offsetting the greater tax that the government gets on pur- chased farm inputs is the fact that direct sales by farmers to con- sumers (or to other entities out of the tax network) do (or at least should) result in tax liabilities if the farmers are in the network but quite obviously do not generate any tax liability if they are out of the network. Thus the key to whether the government gets more or less tax revenue by leaving farmers out of the tax network lies in whether the tax embodied in purchased farm inputs exceeds or falls short of the potential tax that would be paid on direct sales by farmers to con- sumers or other purchases outside the network. This principle applies to any activity or sector. The Financial Sector under a Value Added Tax The financial sector is perhaps the most interesting-case, mainly because of the complication of defining inputs. Clearly, the Canon or Xerox copiers, the IBM computers, and the tons of paper used by financial institutions are inputs that are as easy to identify as the wheat used by a mill or the flour-used by a baker. But what about demand deposits, time deposits, and savings deposits? What about: CDs? Are they inputs into banking operations in the same sense-as computers and paper? If so, are interest payments by banks to busi- ness firms supposed to' embody a value added tax that should then be creditable against the value added tax that.they owe? And should banks in turn pay value added tax on the interest they charge on their loans, with this tax then being creditable by firms that are.part of the value added network but not so by firms and families that are outside it? These questions are really conundnrms, ideal for occupy- ing insomniac economists during 'their sleepless nights. Insurance companies are not easier to fit into a scheme of admin- istering the value added tax through the credit method than are banks. The problem here is that the premiums charged do not simply The Other Side of Tax Reform 155 cover purchased inputs plus the firm's value added. They also cover reserves of all sorts against future claims, reserves that in turn func- tion also as investments on which interest, capital gains, and possibly other financial income accrue. Compounding the problems connected with incorporating finan- cial institutions into a value added tax network is the fact that corre- sponding to any treatment accorded to items such as interest received and paid by banks, or premiums received, claims paid, or reserves formed by insurance companies, is a counterpart treatment to be applied at the level of the financial entity's client firms. Public finance experts have studied and restudied these problems, and the invariable answer is thatjust leaving the financial sector out of the network is much simpler. Then we do not have to define its value added for tax purposes. For analytical purposes, we can simply call it x. The analyst can then use the results of, say, our farm example and state that whateverx may be, the business firms that are members of the value added network do not get to claim a credit for tax paid on whatever part of x they buy. On this part of the financial sector, the government gets the same revenue regardless of whether the financial sector is in or out of the network. At the same time, for the services rendered directly by the financial sector, consumers go without paying any tax if the finan- cial sector is out of the network. Finally, if the financial sector is out of the network, it does not get credit for its inputs purchased from firms that are members of the network. Some of these inputs, such as paper, copiers, and computers, are easy to define and identify; others are sub- ject to the same questions as enter into the definition ofvalue added for the financial sector. To any definition of value added,i x, is added a corresponding definition of purchased inputs,y. What we know is that while pan of x (that part sold to entities outside the network) goes untaxed when financial services are left out, part of y (that part pro- vided by firms that are in the network) pays tax that would not be paid if the financial sector were itself covered. Retailing under the Value Added Tax The case of retailing is not nearly as complicated as that of the finan- cial sector. Analytically, it is a straightforward application of the results presented earlier. The problem is that retailers sell an overwhelmingly large fraction of their output to final consumers. Because retailers have no corresponding sales (or at least hardly any) to firms that are in the value added network, nobody else within the network pays tax, as it were, on their behalf. Therefore, leaving retailing out of a value added network means, in general, the outright loss of the tax that would accrue on the basis of the retail sector's value added. For- 156 AmnohI C Harberger tunately, leaving retailing out does nothing to impinge upon the tax already collected at earlier stages.2 Sometimes retail enterpriscs below a certain size (usually measured in terms of annual sales) are excluded from the tax net. Excluding these enterpriscs saves administrative costs but also encourages eco- nomic inefficiency. In effect, this differential treatment amounts to taxing-one mechanism of providing retailing services (the value added by a firm of medium or large size), while leaving untaxed an alterna- tive close substitute (the value added by a smaller firm that provides the same or similar services). Moreover-and once again human inge- nuity comes into play-it stimulates the destruction of medium-size and large firms as their owners create new entities, involving many small firms, to do the same work previously done by a single firm. One partial solution to the efficiency problem, and its corollary the artifi- cial fractionalizing of larger enterprises, is to place on smaller firms of a given type a simple tax (such as 1 percent of sales, as estimated by the- tax assessor), with the firms being allowed to comply in full with the record-keeping and other requirements of the value added tax should they find that alternative preferable. D-fferent Levelsfor the Value Adde Tax Another way of dealing with the difficulty of collecting from vast num- bers of tiny retailers is to leave all retailing out of the-value added net- work. In effect, this creates a value added tax at the manufacturer and wholesale level. The problem here is one of enterprises trying to define themselves out of the tax category. In particular, integrated firms will try to use artificially low transfer prices between their wholesale and their retailing divisions so as to minimize value added up through the whole-' sale level. This minimizes their tax payment, because ihey report that value added is located at the retail level. It is, of course, the task of the tax authorities to track down the sort of abuse just mentioned and to stamp it out. On the whole, conscientious enforcement is considerably easier at the wholesale than at the retail level, simply because of the vast differences in the number of taxpaying firrns. When problems of administration loom large, as in populous coun- tries with low levels of per capita income, and with gross shortages of competent administrators, the alternative of a value added tax at the manufacturer level is sometimes adopted. This tax usually covers the manufacturing sector, broadly defined to include modern food pro- cessing and mining, perhaps even modern forestry and fishing opera- tions. As in all other cases, imports of the affected items are subject to the tax, and the tax is rebated on exports of such items.3 Te Oth1er Side of 7TxReform 1)7 The Strategy of Value Added Tax I hope that you can extract from the preceding commentary some sense of the different considerations that enter into the design of a modern value added tax. From the outset, I have emphasized that the tax is far from being fully general. Whole sectors, such as domestic service and the imputed income from owner-occupied housing, are left out of the tax net without a moment's hesitation and virtually without discussion. Underlying such wholesale sectoral preference seems to be some combination of a "sacred cow" image (spelling severe political difficulties for the government that is rash enough to meddle) plus an inchoate, unarticulated sense of vast enforcement difficulties. Once past the sacred cows, one confronts a series of questions about how the base of the value added tax should be defined. Here administrative issues interplay with those of equity and efflciency. All these questions have come up many times in the context of a value added trx,c with a single uniform 'rate. They are not issues of pure equity, nor of pure efficiency, nor of pure. administrative con- venience. In the end, I would risk the judgment that tax experts facing these problems try to establish a workable, robust set of rules that specify in a simple way what should and should not be directly taxed, where this distinction. is workable from an administrative. point .of view, and where it does not lead to any gross, easily avoid- able production; inefficiencies. This view. carries the presumption (sacred cows aside) that' a productive activity or sector should fall within the net. But experts recognize that sometimes it is not very costly in terms of either revenue or efficiency to leave an activity out. Sometimes the inclusion of an activity can itself have large efficiency costs (particularly if close substitute activities must for one reason or another be left outside the. net), and sometimes it is simply a matter. of fairness to put a given activity on one or the' other side of the line that divides the. covered from the uncovered sector. This view.also presumes that as an economy develops, the answers to the many questions concerning the inclusion and exclusion of sectors will themselves change. Taxes at the zmanufacturer level will. probibly evolve to the wholesale level, then to a partial retail cover- age, and finally (perhaps) to full retail: coverage. Agriculture and. . fishing sectors that were.- initially left out may also over time find themselves included in the tax net. All this is compatible with a general sense of robustness and fairness of the system prevailing at each stage. This, in my opinion, is the way contemporary tax reformers look at the problems surrounding the value added tax. 158 Arnold C Harberger The Uniform Tariff The political economy of tariffs is one of the more fascinating puzzles facing economists. In the first place, I believe one can say without fear of contradiction that tariffs are never, at least for a developing country, the ideal or best solution to any economic problem. Usually tariffs are motivated by a desire to protect (that is, subsidize) some domestic activity. But it is easy to show that a tariff is in fact equivalent to a subsidy to the domestic production of a good, together with a lax (at an equal rate) on its domestic consumption. Hence, if the authorities want to stimulate domestic production of a particular good, the idea of a subsidy makes sense, but why the domestic users of that type of good should be singled out as the group that has to pay for the subsidy is a question that has never had a satisfactory answer. On the whole, one can say that tariffs exist, when and where they do, because the political influence of producers is stronger than that of consumers. High tariffs appear where producer interests are at their most powerful and where the countervailing consumer interests are at their most dispersed and least well articulated. Facing the interests that argue and press for protection, economists have accumulated two centuries of experience pleading the case for free trade. Some may say all this effort has come to little, as in the end we still see all the signs of protectionism around us. But that sells short the efforts not only of economists but also of hundreds of wise and dedicated public servants and more than a few true political leaders. For most of the industrial countries, protectionism is today far less than it used to be, and for the developing countries, perhaps the biggest, most dramatic change in- commercial policy in more than a decade has been movement in the direction of freer, not more restricted, trade. I believe the modern view of tariff policy, and of trade restrictions in general, is more subtle and complicated than the simplistic free- tradism of earlier generations. It is not that the old arguments are wrong or that modem policy economists would negate free trade as a worthy goal. Rather, what I would call the modern view accepts the existernce of protectionist pressure and finds it strategically interesting to cede a little territory to the protectionists in order to preclude their command over a much greater area. Central to this vision is the idea of interest groups using all their powers of threat and persuasion to achieve their protectionist objec- tives. This is in line with modern political economy studies, which help to explain why some pressure groups succeed grandly, others manage to win some protection, and still others succeed not at all. But such studies certainly do not justify the end result as being in the The Other Side of rax Reform 159 social interest. To my mind and that of many others, it is the role and responsibility of the economics profession to articulate society's broader interests and to lead the resistance to pressures by special interesL groups. Wherein does the strategy call for ceding a little? In granting some protection to essentially every import-substituting activity in the econ- omy. How does it thereby gain a lot of territory? By limiting that protection to whatever is provided by a moderate, uniform tariff What is the logic behind this? It provides a natural guarantee against the huge efficiency costs (as well as the gross inequities) involved in the exaggerated rates of effective protection that flow from grossly differentiated tariff structures. What is the key political economy tac- tic that the strategy invokes? Putting each individual protectionist interest group on the defensive. How is this accomplished? By asking them to justify why their group should be singled out for special protectionist treatmenit while all other import substitutes are given the effective protection implied by a standard uniform tariff rate. Of course, a strategy to limit protection to whatever is provided by a me-derate, uniform tariff cannot be implemented unless the govern-. ment adopts and supports it. If the government has already sold out to the protectionist interests, the strategy has not a chance. This strategy is not for such governments, but rather for those that actually have had the general interest at heart yet have been unable to overcome the many political obstacles in the path of liberalization. For them, adopt- ing a uniform tariff has the effect of splintering the opposition into separate groups, each favoring special treatment for itself. This, of course, is the way interest groups typically behave, but fighting for differentiated treatment when such treatment is the norm is one thing-, fighting for it against the backdrop of uniform tariff protection for everybody else, with special treatment just for one's own group, is quite another. In a sense, a uniform tariff policy gives the government-the high ground in its struggle against special interest groups. The government adopts a policy of uniform effective protection of import substitutes, then challenges the opposing interest groups to come up with a clear alternative criterion. In general, they wifl not easily discover such a criterion, and if one is found that favors one group, it is likely to disfavor another (relative to the uniform tariff). Unifor7m Effective Protection The idea of a uniform tariff as a practical, real-world policy measure received great -impetus from the development and elaboration of the concept of effective protection. I think of this as the really great 160 Arnold C Harberger advance of international trade theory of the 1960s, at least in its real, as distinct from monetary, dimensions. To me, the big lesson of effec- tive protection analysis is that without something like a uniform tariff, one is not even able to interpret the protective "content" of a given and known tariff structure. Consider a 30 percent tariff on men's shirts (a favorite example). If the country in question produces its own cotton, a 30 percent tariff on shirts represents 30 percent effective protection. But suppose the country does not produce its own wool, silk, or cashmere and that yarns of these various kinds enter duty free. If the wool to make a woolen shirt accounts (at international prices) for half the cost of the shirt, the 30 percent tariff on men's shirts will amount to 60 percent effective protection of local value added in making woolen shirts out of imported yarn. The same 30 percent .tariff will give 120 percent effective protection to the making of silk shirts, if the silk yarn accounts for 75 percent of the international cost of the shirt. And it will provide an astounding 300 percent effective protec- tion to the making of cashmere shirts, if that particular yarn accounts for 90 percent of the international cost of the cashmere shirt. Not only does effective protection thus vary greatly, even within a quite narrow tariff category such as men's shirts; but also the degree of effective protection moves up and down capriciously with movements in world prices. If silk yarn accounts for 75 out of 100 percent of costs for a silk shirt at some initial point in time (yielding 120 percent effective protection), a fall of a third in silk prices will shift that to 50 out of 75 percent (yielding 90 percent effective protection), whereas a rise of silk prices by a third will make it 100 out of 125 percent, bringing to 150 percent the degree of effective protection provided by a 30 percent tariff on the final, product. This means that even if one wants to aim for a particular rate of effective protection of an opera- tion using imported inputs, achieving that rate in practice is impossi- ble as long as the international prices of the relevant inputs and out- puts are subject to even reasonably modest fluctuations.4 Not all the lessons of effective protection analysis are so dishearten- ing, however. We can use the same analysis to show that a uniform tariff will rescue policymakers from the mare's nest of effective protec- tion rates that seems to be the inevitable outcome of traditional tariff setting. It is easy enough to visualize: if all final products receive 30 percent protection and all imported inputs are likewise subject toa 30O percent tariff, the domestic value. added process that converts the imported inputs into a final product will also be given protection of 30 percent.5 A few technical caveats should accompany any statement concern- ing the virtues of a uniform tariff. The first and most important of these is that although the uniform tariff in principle treats all import- The Other Side of Tax Refonm 161 substituting activities alike, it obviously discriminates against export activities. If the exchange rate in a developing country is 10 pesos to the dollar, setting a uniform tariff of 30 percent stimulates import- substituting activities, paying them, in effect, up to 13 pesos for saving a dollar's worth of foreign exchange. At the same time. export activ- ities receive only 10 pesos for each dollar's worth of foreign exchange they generate. Such discrimination in favor of import substitution and against exports is the inevitable result of protectionist policies. Uni- form tariffs make protection more rational and less costly, but they do not eliminate distortions altogethdr.6: At another level, when imposing a uniform tarift, the authorities must ensure that the tariff costs embodied in the prices of export products are reimbursed when the goods are in fact exported. This treatment has been incorporated in the "rules" of the General Agree- ment on Tariffs and Trade (crA) from its very inception. If such treat- ment is not given, then export goods that use imported raw materials, or components will receive negative effective protection.7 A final caveat is in a sense the obverse of the previous one. If export goods are sold inside the country at world market prices while import goods (and their substitutes) sell for world market prices plus 30 per- cent, then an import-substituting activity can end up being favored if it uses one or more export goods as an input. In principle, for a uniform 30 percent tariff always to yield 30 percent effective protec- tion, the tariff should be accompanied by an internal tax (a quasi tariff) on the. use of exportable goods as inputs into import- substituting activities. No one has ever seriously suggested that this be done. I note it here as a source of deviations from full uniformity- of effective protection ist the real world, even in countries that have adopted uniform tariff policies.8 Tariffs on Imported Inputs: Today's BigDebate As countries drift toward more uniform tariff strucrures, there is rela-- tively little resistance to the reduction of the highest rates of tariffs. These usually fall on finished products of a luxury nature. Sometimes the prevalence of contraband-in such highly protected products causes even firms that ate beneficiaries of protection to question the merits of extremely high tariffs. In any-case, the nature of a differenti- ated tariff structure means that only a relatively few. activities receive the highest rates of protection. Hence reducing the highest tariffs has been a .relatively. easy task for governments that are determined to liberalize their trade restrictions.. By contrast, raising the tariffs on imported inputs to levels that are close to the target level for final product tariffs has.been far from easy. 162 Arnold C. Harberger The difficulty stems from a number of quite distinct factors. First, the interest groups that benefited from the tariff structure in the first place have, in general, also benefited from being able to buy imported inputs that enter with low.or zero tariffs. (Ti.is is the historic pattern of protection: highest on luxury goods, next highest on other final prod- ucts, lowest on certain key "necessities" and on all or most imported inputs.) These interest groups can coalesce on the general principle of keeping tariffs on inputs low, even while they may be forced to accept lower tariffs on final products. Second, many imported inputs do not have domestically produced substitutes. It is hard for people to understand the "nreed" for having tariffs on products that are required for some local productive activity' and that are not actually produced, or even foreseeably producible, at home. Typically, only those who are familiar with the mechanisms of effective protection can see the logic behind that "need,' and they, sad to say, are invariably.a small minority, composed mainly of economists and other technocrats. Finally, even economists typically fail to reach consensus ona the * desirability of raising tariffs on imports, because some economists remain swayed by the arguments for fully free trade and see any rais- ing of any tariff as unwanted deviation from that goal. Some of these economists remain unconvinced, even after all technical demonstra- tions, but others change their view once they see the positive merits of raising tariffs on inputs in the presence of higher tariffs on outputs. A tariff on an imported input of a domestically produced good is a simple, straightforward distortion. It does nobody any good except the local producing interests (if any) that it protects. When the tariff is on an input of a good produced for export, GAnr-rule treatment (called border tax adjustments) should guarantee that it will not interfere with the international competitiveness of. such exports (see above). With respect to the exports that. are given GAiT-rule treatment, tariffs on imported inputs are, in effect, nullified. We come finally to tariffs on inputs used to produce import substi- tutes. The proposition here is that raising such tariffs is always welfare- enhancing, at least up to the level of tariffs affecting the.final product in question. This much should be clear from our earlier example of the tariff on men's shirts: raising the input tariff to 30 percent lowers the effective rate of protection to 30 percent. This represents a benefit (that is, a welfare enhancement), regardless of whether the initial rate - of effective protection was 60 percent (woolen shirts), 120 percent (silk shirts), or 300 percent (cashmere shirts). - What many people do not understand is that raising the tariff rate on the input in question continues being -beneficial considerably beyond the point at which the input tariff:equals the output tariff. h77 Othler Side of Tax Reform 163 However, this result should be intuitively obvious. If by raising the tariff on wool from zero to 30 percent, we can reduce the effective protection of woolen shirts from 60 to 30 percent, should we not be able to reduce that cffective protection all the way to zero by pushing the tariff on wool still higher? Indeed, this is the case. Effectivc protec- tion for the woolen shirt industry reaches zero when the tariff on wool is 60 percent; for the silk shirt industry it reaches zero when the tariff on silk yarn is 40 percent; and for the cashmere shirt industry it becomes zero when the tariff on cashmere yarn is about 33 percent. Now it would be absurd to argue for a set of differentiated tariffs on all sorts of imported inputs as a device for reaching lower rates of effective protection on final products. But the fact remains that the optimal levels for tariffs on inputs, given the tariff rates on their outputs, are higher, not lower, than the output tariff rates. Realizing this, many who espouse free trade on general philosophical grounds are swayed to support raising input tariffs up to the standard uniform tariff level and thus give up their initial resistance to the idea.9 The Lxation of Income from Capital Nowhere in the field of taxation can we find, in my opinion, a sharper contrast between appearance and reality than in capital income taxa- tion. Taxing capital income appears to be an excellent way to intro- duce progression into the tax structure-to make the rich bear a heavier relative tax burden than the middle classes and the latter a heavier relative burden than the poor. The reality is quite the oppo- site. In every small country, at least in the Western world, anyone with funds to invest can easily place those funds in New York or London or Zurich. There is really nothing that policymakers can do to prevent such a "portfolio diversification" on the part of the nationals and other residents of developing countries.10 Because the relevant rate that can be earned abroad is something that is totally beyond the control of any developing country's govern- ment, it must be taken as given in any analysis of developing-country policy.' 1 From this it follows that the heavier the taxation of income from capital in the developing country, the greater the incentive to hold assets. abroad. In the simplest analytical case, one can take the required expected net of the tax rate of return in the developing country to be equal to that available abroad plus a fixed risk premium or differential. In this case, each increment of the tax placed on the income from capital in the developing country leads to a capital -out- flow from that country- The outflow continues until the gross of tax return has risen so as to fully reflect the increment of tax. In this way the net of tax return in the developing country is restored to its equi- 164 Arnod C Harberger librium level (equal to the "given" foreign rate plus the required differential). Thus the higher the tax on the income from capital in a developing country, the smaller will be the stock of capital that is occupied there. Because labor is, in general, much less mobile internationally than is capital, the consequence of having less capital occupied in the country is a lower scale of real wages and salaries. Thus it is really labor that bears the burden of taxation on capital income in developing countries.'2 This is the irony of policies such as heavy taxation of capital income. They have the appearance of being "populist" policies yet end up hurting mainly the wage and salary earners of the country. It follows that if one wants truly to help these groups, one should tread very lightly when it comes to capital income taxation. It would appear to follow from the above that one should, in princi- ple, reduce corporate income taxes to zero. This would generate an "optimal" capital stock and would produce a distinctly positive effect on real wages and salaries, in contrast to the present situation in most countries. Two considerations, however, weigh against this option: first, the presence in most developing countries of a significant sector composed of multinational corporations and, second, the existence of a greater problem of monopoly power (and hence monopoly profits) than is common in more advanced countries because of the small economic size of most developing countries. Income of Multinaticona Companies The first problem-that of multinational companies-derives from the fact that they are, in principle, subject to tax in their home (base) countries on the profits they earn via their subsidiaries located in developing countries. Typically, the multinationals are also eligible in their home countries for tax credits that reflect the taxes they pay to developing-country treasuries. These credits generally apply only up to the amount of tax that would be due to the home country govern- ment on the profits in question. The end'result of this tax treatment is that a developing country, by lowering its tax rate below that applying in a company's home base, simply gives up the tax money. The com- pany still pays the tax, but now it goes to the treasury of its home base country instead of to the developing country where the operations in question are located. The obvious way for.a developing country to avoid simply giving up money to the treasuries of the advanced countries where the multina- tionals. are based is (a) to have a corporate income tax of its own and ne Otl/erSice of Trax Reforwa 165 (b) to seL the rte of tax at a level approximating those prevailing in the most important base countries. At first sight, the above solution appears to handle well the problem of loss of revenue to industrial-country treasuries but to leave unsolved the problem (referred to earlier) of reduced holdings of capital assets by nationals and oth& residents of the developing coun- try. Happily, there is an easy way to handle both of these problems. It consists of maintaining the corporate income tax as part of the legal tax structure but of providing also for the integration of the corporate and personal income taxes as far as local resident taxpayers are con- cerned. Under full integration of the two taxes, the corporate income tax would turn into a simple withholding device. Local shareholders in a local corporation would be asked to declare as their own income their proportionate share of that corporation's profits. Then, once their tax liability was determined, the corporate tax paid on the basis of their shares would be creditable toward what they owe in personal: income tax, in just the same way as the amounts withheld. by employers are creditable against employees' tax liabilities. A similar effect would be obtained by simply exempting from per- sonal income tax the dividends received from domestic corporations plus the capital gains on their shares. This treatment works best when the top bracket rate of the personal income tax is equal to the rate of the corporate income tax.'3 Monopoly ProJits The second issue with respect to the taxation of capital income has to do with mono)poly profits. Many people now recognize that the econ- omies of scale that exist in many activities are operative for smaller firms but do not continue to be significant once the firm (or the productive operation, as the case may be) reaches a certain critical- size. The critical size in most industries is such that monopoly prob- lems are not severe in advanced economies. There, demand is suffi- cient to allow several firms above the critical size to coexist in the market and compete with one another. In smaller countries, particu- larly developing countries, the demand from within the country may' not be large enough to warrant several firms of efficient size. Thus one quite commonly finds in such countries a number of activities in which only one or two firms play a significant role. in these circumstances monopoly power becomes a matter of sig- nificant policy concern. The standard economist's answer to this problem is for small. countries to use the world market as their principal instrument for control of monopolies. To the extent that the potential monopoly power occurs in the tradable goods area, a 166 Arnold C Harberger policy of free trade (or of a relatively low uniform tariff) will do a great deal to limit the monopoly potential that would otherwise exist. The obverse of this proposition is that the freehanded application of protectionist measures is typically an open invitation to the exercise of monopoly power by protected local firms. The monopoly problem is thus greatest under protectionist re- gimes and least under free trade or low uniform tariff policies. Nonetheless, even under the latter type of policy, the governments of smaller countries probably have more reason to worry about monop- oly power than do those of larger countries. Now one thing that a well-administered corporate income tax will do, automatically, is to take its bite out of monopoly profits, just as it will out of any other category of profits. This is the most natural way for a small develop- ing country to approach the monopoly problem. Price fixing and rate fixing are approaches that may work (if well administered) in basic public utilities such as telephone and electricity companies. It is very difficult to administer beyond this rangc, particularly with respect to activities whose output consists of a number of different products, with numerous variants of each with respect to quality, design, and so on. In such cases, price regulation has proved to be unworkable and often directly counterproductive (when the mrionop- oly succeeds in flouting the controls, but only by incurring signifi- cant real costs). Under the circumstances, the course of greatest wisdom for a developing country with a perceived problem of monopoly control is to use the corporate income tax as a device by-which the public at large gains command over at 'least a certain share of such profits, wherever and for whatever reason they may occur. This is the only argument I know of for limiting the degree of integration between the corporute income tax and the personal income tax in a developing country. Full integration by the withhold- ing method would only ensure that shareholders in such enterprises pay their full personal tax on their full profits, including monopoly profits. As one pulls back from full integration, the total tax paid by enterprises gets correspondingly higher. Policymakers concerned with monopoly in developing countries thus face an interesting pol- icy question. How much of the, benefits of full integration are they willing to give up in order to obtain a greater bite into the monopoly profits of corporations? I will not attempt to answer this question here. Let me simply point out that a number of policy devices and packages are available, all of which succeed in taking a greater share of monopoly profits, but at. the cost of making the country a less hospitable place for local residents to invest their capital. A brief listing of a few such schemes follows: 77w O}ter Side of Ta Reform 167 * Make the corporate tax rate higher (say, 40 percent) than the highest personal tax rate (say, 30 percent) with the corporate tax being creditable only up to the percentage rate (here, 30 per- cent) of the personal income tax. * Make only a specified fraction of corporate taxes creditable against the personal. income tax liability of domestic nationals and residents. Introduce integration of corporate and personal income taxes only with respect to dividends, with dividends being grossed up to reflect the profits on which they were based. This would mean that with a 40 percent tax rate, there would be a tax credit of 40 for each 60 percent of dividends received. No tax credit would be received -by local residents with respect to profits retained within the corporation. Introduce integration of corporate and personal income taxes only with respect to dividends, without the dividends being grossed up. In this, case, and again with a 40 percent corporate tax rate, there would be a tax credit of 24 (60 x 0.4) for every 60 percent of dividends received. Clearly, these devices provide a wide range of options for the policymaker. At the same time, it is equally clear that any gain made with respect to getting a greater share of monopoly profits comes at the expense of reducing the incentive for local investors to keep their money in the country. Conclusion The foregoing discussion of policy issues will, I hope, give readers something of the flavor of the issues facing tax reformers in devel- oping countries. On the whole, the scope for tax reform is enor- mous, in the sense that actual tax structures are grossly inferior in almost any dimension to those that could be achieved simply by implementing the. reforms suggested in this paper. Yet, and this too is worthy of note, the: devices on which we have concentrated-a' uniform value added tax; a uniform tariff at a moderate rate; and a moderate corporate income tax that .is fully or partially integrated with a personal income tax, both with similar (and moderate) maxi- mum rates-are themselves not "optimal"- in a formal, technical sense. Tax reform in today's world is, I believe, more appropriately focused on the giving of robust signals that will improve the private sector's allocation of resources than on. the niceties associated with formal, technical optimization. . 168 Anold C. Harbeger Notes 1. Experts are unanimous in advising against the granting of tax amncs- tics as a general policy. Such amnesties typically allow taxpayers to bring their taxes up to date with little or.no pcnalty. Sometimes they even collect no intcrcst and makc no inflationary adjustment on back taxes due. Thus the evaders of the past end up paying less real tax at its total present value than did the compliers who made every payment in full and on time. But the tax-and-cvasion race referred to in the text takes away all moral abso- lutes. If past taxes were set at high levels in anticipation of cvasion, the argument for a partial forgiveness of such taxes becomes stroniger, espe- cially when such forgiveness is thought to clean the slatc for a "new future" in which evcrybody pays taxes due in full. But granting partial forgiveness once can easily leave taxpayers with the idea that evasion yields a positive payoff. This can set in motion yet another tax-and-evasion race, with successive increments of evasion triggering successive upward shifts of the entire tax rate structure. The only happy solution to such a problem is one in which taxpayers somehow are brought to believe that the future will be very different from the past, so that they shift their behavior at some point from wholesale evasion to something close to full compliance. It takes a government with great credibility and moral authority to accom- plish such a feat. 2. An individual's ingenuity blooms in full flower when it comes to tax evasion. One of the great virtues of a value added tax administered by the credit method is that if one stage of production understates its sales (and therefore underpays its tax), the error will tend to be compensated at the next stage, which will, as a consequence of the understatement, receive a corre- spondingly smaller tax credit. This self-correcting feature works well up to the last stage. But what about an understatement of sales by a retailer? Obviously, there is no compensating correction here, and va]ue added tax administrators are well aware of this fact. In theory, cunning retailers could not only fail to pay tax on their own value added but even end up collecting money from the treasury by having sales consistently lower than purchases of inputs. For-, tunately, this would show up as negative value added. In general, value added tax legislation does not permit instantaneous cred- iting of taxes paid on inputs when these exceed the tax due on a firm's output. Rather, the excess must be carried over to future tax periods. In this way, the worst thateven the shrewdest retailers can do is to reduce the tax on their own value added to zero; they cannot, under the treatment, eat into the tax already paid by another. Note also that cases of consistently negative or zero value added are highly likely to trigger an audit by the tax authorities, even when they are no more than mediocre in the conscientious execution of their duties. Persistently negative value added is a phenomenon too conspicuous simply to overlook. Hence, even at the last and most vulnerable stage, the value added tax has some sort of built-in insurance against the loss of revenues garnered at previous productive stages. 3. Imposing:a value added tax on imports and rebating-it on exports does not protect import substitutes or subsidize export activities. These so-called "border tax adjustments" are simply ways to permit the tax to show up as part of the domestic prices of the affected products without giving up the artificial stimulus for users to seek substitute goods through importation and without putting artificial trammels on export activities. Thle Oiter Side of Tax Reform 169 4. The price movements dealt with here should be thought of as movements in rcal or relative prices, not as a reflection of international inflation. An easy rule of thumb is to think of the price of wool, the price of silk, and so on, and the intcrnational prices of the corresponding shirts, as being expressed in relation to an index of world market prices of tradable goods.. 5. If ri is the nominal rate of protection on final productj, r4 the nominal rate of protection on input i, and aGj the fraction of the cost afj accounted for (at international prices) by input i, domestic resource costs can extend up to the domestic currency equivalent of (I - rj) - E,a1(1 + r,) per dollar's worth of final product displaced. The net saving of foreign cur- rency obtained in the process is equal to I- This pattern of protection therefore allows for domestic resource costs of up to [(I - E,a0) + (r - 4a,r)(1-aj) per net dollar of foreign exchange saved. This implies a rate of effecdive protec- tion of (ir. - Eta,,r01/ - Thus, this rate of effective protection will be equal to r,, whenever all the relevant r, are aLso equal to rj. This says that the effective protection of a final product will be equal to its nominal protection whenever the relevant import- ed inputs into its production have tariffs equaling (or averaging) the rate that applies to the final produc. Thus if all final products and all imported inputs carry the rate r*, then all domestic value added (in processes of import substi- tution) receives protection at that same rate. 6. The famous "Lernier theorem'! of the 1930s points out that a uniform 30 percent tariff accompanied by a uniform 30 percent subsidy to exports would in principle be nondistorting. Indeed, it would be equivalent to free trade. As far as trade items (imports and exports) are concerned, a country could equally well (a) institute a uniform 30 percent tariff plus a uniform 30 percent subsidy to exports or (b) devalue the currency by 30 percent while in all other respects maintaining free trade. 7. The effective protection level of an export good xj is found by substitut-- ing the export subsidy rate Zj for the tariff rate rj is, the formula for effective protection. The effective protection rate on an export good Xi is then {Zj - Z,a!,r0)1(1 - Zia,j) Normally, zf will be zero for the typical export good. Effective protection can then be made equal to zero by rebating the tariff r; that enters into its cost structure. Otherwise xJ will end up with negative effective protection. 8. 1 do not believe it is a serious flaw in these cases, at least for the develop- ing countries with which I am familiar. On the whole, export production is far more specialized than import substitute protection. Many important exports by developing countries are agricultural products (wheat, beef, fruits, and so on), which do not typically enter as inputs into import substitute production. Other export products, such as copper and tin, do not play a major role in the costs of import substitute production. Still,others (jute, forest products, and so 170 Arnnok C Harberger on) may be important as inputs into export products but not into import substitutes. 9. This is not the place to elaborate greatly on the abovc point, but it is worth mcntioning that raising tariffs on inputs into the production of import- ables is lrmde creating. This is Lhe only big and important class of trade-creating tariffs. The process works very simply. If imported inputs account for 40 out of 100 percent of total cost (at international prices) of an import substitute, then increasing the tariff on the inputs will, with fixed proportions, result in an incrcase of 100 percent of trade in the final product for every 40 percent of trade in the inputs. Abandoning the assumption of fixed proportions mod- ifies, but does not annul, this result. 10. In my own experience, India of the early 1960s was a striking example. When I was there in 1961-62, capital outflows were essentially prohibited, and. all kinds of controls existed. Nonetheless, a parallel or black market in foreign currency not only prevailed, but thrived. The black market premium was high, howevcr about 50 percent over the official rate of Rs4.75 to the dollar. I took every opportunity to ask people what rate of return they expected to get on the money they had invested abroad (mainly in London). The answer was always the same: they expected to get whatever rate of return prevailed in the London financial market. When asked if the high black market premium didn't get in the way of such an outcome, they replied, "No, we take it out over the black market, and we bring it back the same way:' They did not face a significantly lower expected return, despite all the-controls. All they faced was an added source of possible variability (up or down) in the rate of return, stemming from possible changes in the black..market premium over the period for which the fvnds were invested. 1l. For our purpo-.es it does not matter whether this return is taxed or not taxed in, say, the Jnited States. Interest earned by developing-country nationals on acco'nts in the United States is free of US. income tax, but U.S. -c_vorate inrt:e tax is paid on the profits from their investments in corpo- rate sha.cs. Whatever the tax treatment in the financial market center, developing-country investors'will react to the net of tax return they receive. This is what must be taken as "given" when analyzing developing countries' policies. 12. Most of the time, labor bears more than the full burden of, say, a corpo- rate income tax. This is because the corporate sector produces mainly trad- able goods, whose prices are to a large degree determined in world markets.' The tax therefore cannot push up the international prices of corporate output significantly; the effect of the tax must rather be to push down wages by enough so that the tax can be paid while product prices remain substantially the same. But wages in the corporate sector do not move just by themselves. The effect we are talking about is on the equilibrium level of wages and salaries throughout the economy. Thus if the wages paid in the corporate sector fall by enough to "absorb" the tax, and if wages in the noncorporate .sector fall by roughly the same percentage (reflecting continuing equilibrium in the labor market), then labor will end up bearing more than the full burden of the tax. 13. Over the past two decades or so there has been worldwide recognition of the merits of keeping tax rates relatively moderate, both on personal and on corporate income. Whereas at one time personal income tax rates ranging up to 50, 60, 70 percent and even higher were common, now most maximum rates are between -30 and 40 percent. Likewise, although corporate income - 1hc Other Side of 7hx Reform 171 taxes of 50 percent and more werc once common, today they, too, tend to be clustered in the 30 to 40 percent range. Thus, a developing country today can mcet the two conditions specified in the text by setting its corporate tax rate at about, say, 35 percent and by making this also the maximum rate of its per- sonal income tax scheclule. Selected Bibliography Boskin, Michael T., and Charles E1 McLure, Jr. 1991. World Tax Reform. San Francisco: ics Prcss. Gandhi, Ved l 1987. SuppltSide Tax Policy; The Relevance to Developing Countries. Washington, D)C.: International Monetary Fund, Gillcs, Malcolm, ed. 1989. Tax Reform in Developing Countries. Chapel H-ill, N.C.: Duke University Press. Harberger, Arnold A., ed. 1984. World Economic Growth. San Francisco: Institute for Contemporary Studies. 1990. "Reflections on Uniform lhxation:' In Ronald W. Jones and Anne Q Krueger, eds., 77w Potitical Enomy of internalional Thade. Oxford, U.K.: Basil Blackwell. Newbery, David, and Nicholas Stern, eds. 1987. The Theory of Taxation for Devel- aping Countries. New York: Oxford Univeisity Press. Tanzi, Vito, ad. 1990. Fiscal Policy in Open Deeteoping Economies. Washington, D.C.: International Monetary Fund. Urrutia, Miguel, Shinichi Ichimura, and-Setsuko Yukawa, cds. 1989. Tlhe Potiii- cal Economy of Fiscal Pblicy Tokyo: United Nations University. - a- - 8 Adjusting to a Terms of Trade Shock: Nigeria, 1972-88 M Michael Gavin NIGERIA, the most populous nation in Africa, is one of the few coun- tries blessed with substantial petroleum deposits. Production of petro- leum started in 1958 and expanded briskly during the 1960s. By 1972 Nigeria was exporting nearly 2 million barrels a day of high-quality crude oil. Nigeria was therefore ideally situated to capitalize on the large oil price increase that the Organization of Petroleum Exporting Countries (OPEC) engineered in late 1973 and the subsequent, compa- rably large increase that followed the supply disruptions generted by the 1978 revolution in Iran. When the petroleum market entered its deep and protracted slmp in 1981, however, Nigeria found that what commodity markets give they can also take away. Since then, Nigeria has been grappling with an economic crisis of enormous proportions. As a result of increases in the price of petroleum, the dollar value of Nigeria's exports rose sixfold between 1972 and 1977 and doubled again by 1980. Adjusting for inflation, the value of Nigeria's exports more than tripled between 1972 and 1977 and rose another 50 per- cent by 1980.1 Although oil prices fluctuated widely during the period, the decade after 1972 was one in which Nigeria's international pur- chasing power was increased beyond the wildest dreams of the most unreasonably optimistic planner. By 1986 oil prices had fallen considerably from the dizzying heights reached in the early 1980s, and Nigeria's terms of trade in 1987 were substantially less favorable than in the boom years. Still, as table 8-1 indicates, the terms of trade remained better than they had been before the original petroleum price increase. The data presented in table 8-1 also, however, raise the suspicion that Nigeria captured astonishingly small long-term benefits from the twvelve-year boom. By 1987 Nigeria's output had recovered dramatically from the crisis years 1984 and 1985. Nevertheless in 1987, and in an external environment that was not more hostile than that of 1973, Nigerian GNP per capita was 22 percent lower than it had been in 1973, fourteen yeari before. 172 Adjusting to a Terms of Trade Sko&le Ntigeria, 1972-88 173 -hble 8-1. Nigeria before and after the Oil Boom, 1973 and 1987 Indicator 1973 1987 Terms of tradc (1980 = 100) 22.3 49.1 Real GNP per capita (1980 naira) 577 450 Privatc consumption 356 334 Govcrnment consumption 48 47 Investment .100 49 Sources ofper capita real GDP Agriculture (1980 nairas) 208 124 Industry (1980 nairas) 218 153 Services and other (1980 nairas) 165:. 167 Food production per capita (index) 109.7 103.1 External debt per capita (1987 dollars) 39 269 Debt-export ratio (percentj 32 369 Consumerprice inflation (percent)" 9.7 18.8 Life expectancy (years) 44.9 51.1 Infant mortality (per ],000 births) . 143.2. 102.8 Primary school enrollment (percentage of age group) 51 92 a. Inflation rates are avcrages for the five-year periods ending in 1975 and 1988. Source. World Bank data. Declines in output were pervasive, with agricultural output falling 40. percent, industrial output falling 30 percent, and only the service sector, which includes government employment, registering a slight increase. Food production per capita actually fell over the period, and Nigeria, which had been a net exporter of food, became a large. net . importer during the late 1970s. Both private and government con- sumption were smaller in 1987 than they were in 1973, and per capita investment fell in real terms from N 1,001,980 in 1973 to less than N 501,980 in 1987.: And the view from the financial markets was no more f, artening than the view from product markets. In 1987 individuals ii, Nigeria: owed an average of about $269 per person to foreign creditors. Infla- dion during the five years leading up to 1973 was just under 10 percent a year; during the five years leading up to 1988, inflation averaged almost twice that figure. And whereas the naira traded for $1.52 in . ; 1973 (about $1.20 on the black market), it was worth roughly 14 cents' by mid-1989.; Offsetting these setbacks,.it must be said that Nigeria made substan- tial progress during that time in raising health and educational stan- : dards. Nigeria increased the number. of physicians and nurses per capita by substantially more between- 1965 fand 1984 than did low- income developing economies as a whole. And as table 8-1 documents,.- life expectancy rose and infant mortality declined between 1972 and petnc ro .e,- -e-.w-een 174 Micluxa Cavin 1987, although it should be recognized that both indicators had been improving prior to 1972 as well. Most impressive was a tremendous surge in primary and secondary school enrollments, most of which occurred during the middle 1970s, when the Nigerian government embarked on the expensivc Universal Primary Education policy. But despite these very real gains in the areas of health and educa- tion, it cannot be said that Nigeria in 1987 was, on balance, signifi- cantly more prosperous than it was in 1973. And certainly the gains that should have been made possible by the relatively short-lived but nevertheless immense increase in oil earnings had not, apparently, been realized. What happened? Although the oil boom that began in late 1973 may have fallen short of expectations raised during that period, it gener- ated, on balance,-an enormous increase in the resources available for the development of Nigeria. Why was the payoff from these resources, ostentatiously invested in far-reaching national plans, apparently so low? This essay explores the response of the Nigerian officialdom and economy to the petroleum price disturbances in order to draw lessons for other policymakers facing similar circumstances. To discuss. the adequacy of Nigerian policies, we need some norma- tive benchmark against which to compare policies and outcomes. This is provided by the second and third sections of this essay. In the second section we review some important nonmonetary aspects of adjustment to a change in the terms of trade. We define "successful adjustment" with reference to the adjustment that would take place in a friction- less, neoclassical economy. Particular emphasis is placed on the role of foreign borrowing in the adjustment process. Nonmonetary Aspects of a Terms of 1kade Disturbance Tlhe outstanding feature of a change in the terms of trade is that it alters a nation's disposable income. When the terms of trade improve, a country may consume more than it could previously without increas- ing its foreign debt. Alternatively, residents may choose to save some portion of the higher income and consume more in the future. When the terms of trade decline, however, a country must spend less than it could previously. The reduction in spending may be postponed by the expedient of foreign borrowing, but eventually a nation's -spending must fall into line with its new, lower income. This decline in national income is an inevitable consequence of altered terms of trade; it cannot-be changed by the monetary and fiscal policies, pursued by national authorities. Nevertheless, movements in the terms of trade require adjustment by the domestic economy that may be either facilitated or frustrated by the economic policies pur- -Ma e. . . Adjusting to a 'Is of tadeShoct: Nigeria 1972-88 175 sued by monetary, fiscal, and regulatory authorities. This section explores a number of issues surrounding such adjustment to changes in the terms of trade. Its purpose is to lay out the economic responses that constitute successful adjustment and to provide thereby some rough normative guidelines for the analysis of economic policy that. follows. What does successful adjustment involve? We alluded above to req- uisite changes in national expenditure. Such actions are clearly a cen- tral component of adjustment. But adjustment may also require important changes in the structure of production and the allocation of domestic investments, whichi will in turn require potentially large alterations in domestic relative prices. So one important goal of the following discussion is simply to articulate the impact of changes in the terms of trade on demand, production, and relative prices. An issue of paramount importance is whether gains or losses in national income that are the result of changes in the terms of trade should be met with immediate changes in consumption expenditure or with changes in the rate of saving. If saved, current-income gains from an.improvement in the terms of trade can be translated into * higher future consumption, whereas reductions in the rate of national saving can postpone the decline in consumption necessitated by a - deterioration in the termns of trade. The appropriate response of expenditure to altered terms of trade is therefore inherently a prob- lem of intertemporal choice. Thus the second goal of this section is to- lay out the simple principles that govern intelligent savings elecisions. Here we raise the crucial distinction between temporary and perma- nent movements in the terms of trade. We emphasize the usefulness of access to international capital markets for stabilizing the rate of con- sumption and investment when the. terms of trade, and therefore national income, are highly variable. Because adjustment to a change in the terms of trade is costly and takes time to complete, we further argue that temporary recourse to international capital markets may be. an appropriate response to both permanent and temporary changes *in the terms of trade. As mentioned, adjustment is a costly and time-consuming process, - but the speed with which an individual economy adjusts.depends on the -policies the local authorities pursue. This section focuses on- the impact of foreign borrowing on the speed of adjustment to the terms of trade disturbance. We argue that too much foreign borrowing or * lending results in a:suppression of the relative price signals required to effect adjustment. It may make sense, however, to use foreign bor- rowing to attenuate somewhat the relative price "overshooting" that would occur in the absence of international borrowing and thereby * reduce the speed of adjustment. Excessively rapid adjustment is costly;. 176 Miclhael Gavin- it may generate bottlenecks and expensive mistakes that could be avoided with a more moderate pace of adjustment (see the third section). The last set of issues that we raise in this section concerns uncer- tainty. So far we have assumed that decisionmakers know how long the change in the terms of trade will persist. In reality, of course, nobody knows whether a change ;n the terms of trade will be reversed, will persist, or will deteriorate even more. Thus we cannot assess Nigeria's economic policy without an awareness of the uncertainty that sur- rounds future movements in the terms of trade. This uncertainty. leads us to recommend a conservative approach to adjusting to a change in the terms of trade. Conservative in this context means two things. First, it means a higher rate of saving than would be appropriate under certainty; the rationale for this comes from the theory of pre- cautionary saving. Second, it means avoiding irreversible commitments-or commitments that are costly to reverse-when there is substantial uncertainty about future developments. The prac- tical implication is that domestic investment should respond more tentatively than would be appropriate under certainty and that savings * . should instead be channeled, at least temporarily, into. relatively liquid * foreign assets via current account surpluses. The Effects on Income We begin with the most basic considerations, which can be illustrated with -an extremely simple framework. Consider an economy that is endowed with two tradable consumption goods' that, anticipating the discussion.of Nigeria, we label oil and manufactures. As with most developing countries, the ecbnomy is too small. to alter. the relative price of these goods, which is determined in the global commodity' market. The economy is endowed with Qo urfits of oil and QM units of manufactures each period. Now we consider a deterioration in the terms of trade, wihich is here a decline in the relative price of oil. Income rises if measured'in terms of oil and declines if measured in terms of manufactures. But real income, that is, the purchasing power of the economy's production, declines. This change in income is equal to the change in the terms of. trade times the initial level of exports: AY= AP,(Q-, - CO) where A stands for a change, Y stands for real income, P, is the terms of trade, and CO is the units of oil at a point-in time. This measure of the income effect of changes in the terms of trade- is the right one to focus on in the following sense.2 If the change in the Adjusting to a Terns of Rade Shock: Nigeria, 1972-88 177 terms of trade were accompanied by an increase in transfers from foreigners of the magnitude AY, then the initial consumption point, C, would still be (just barely) affordable, so consumers would be as well off as they were before the terms of trade deterioration. Phrasing the same point rather differently, we can say that AY is the amount by which the trade balance would deteriorate and foreign borrowing would have to increase if the initial level of consumption were main- tained in the face of the terms of trade deterioration. This first-order measure of the pain generated by a change in the terms of trade is easily quantified. Between 1972 and 1980 the real value of Nigeria's exports rose from about $6 billion to roughly $30 billion, a fivefold increase. With a population of roughly 80 million in the late 1970s, this increase in income amounts to about $300 per capita at 1988 prices. By 1986 the real value of Nigeria's exports had fallen to about $7.5 billion, almost their 1972 level. Compare these changes in income with the World Bank's estimate of per capita GNP, which was roughly $1,000 in the early 1980s and $370 in 1987. In other words, the change in income attributable to variations in the real value of Nigeria's oil exports was approximately equal to the level of real income Nigerians enjoy today. This estimate is, of course, very rough and subject to. all the quali- * fications surrounding international income comparisons, but it makes two important points. First, large changes in the terms of trade can wipe out years of economic growth. Second, the direct effect on income of changes in the terms of trade is noL enough to explain the entire decline in Nigeria's economy after the.oil market softened, beginning in 1982. We will therefore look for further explanations in policy mistakes. Finance ar Adjustiment? Having identified the loss in income. generated by a change in the terms of trade, the next step is to determnine the appropriate response. BASIC CONSIDERATIONS. Our first- concern is with the level of total expenditure and therefore the current account balance. The key ques- tion is whether an economy that has suffered a deterioration in. its . terms of trade should borrow from abroad to permit- a higher level of expenditure or should, instead, reduce its consumption expenditure. by the full extent of the loss in income, thereby maintaining balance in its international payments. :We emphasized earlier that this question involves intertemporal tradeoffs, and these tradeoffs result because spending in excess of income leads to an accumulation of foreign debt, which must be serviced in the future. The following equations 178 Micrtael Gauvt help to clarify the intertemporal financing problems that Nigeria must face. (8-1) D1+I -iDI = rD, + (PXCX + CM) - (PXQX + Q&) = Rd, +Mt - P,, where Dg stands for the country's net indebtedness at time , and r is the interest rate on this debt. X denotes exports, M refers to imports, and C and Q stand for domestic consumption and production, respectively. The right-hand side of equation 8-1 is the current account deficit, which can be written as the excess of domestic expenditure over domestic income and also as the excess of imports of goods (M,) and capital services (Rd,) over total exports (P1X)). Equation 8-1 emphasizes that current account deficits must be financed with increases in net foreign indebtedness or reductions in an economy's net foreign assets. Equation 8-1 relates the stock of debt tomorrow to the stock of debt today and the imbalance between domestic income and domestic expenditure. We can also solve equation 8-1 forward to obtain an expression for debt two periods in the future'as a function of today's and tomorrow's spending and income. If we continue to solve forward and make the assumption that international debt does not grow too fast, we.obtain a revealing alternative expression for the intertempor-al budget constraint: 18-i') X(I + r)!E,+, = (l + r)-i(Y,j - Di) i-o i-o where E, is total expenditure on consumption, and Y, is.the value of * -a~icome, from production. Equation8-1-i' states that the present value of expected future expenditure on consumption must be.equal to the present value of domestic income minus previously contracted debt. obligations, DI. So if individuals decide, to spend in excess of national income in one period,. they are, simultaneously deciding to spend less than the value of national income in some future period.- : Another way of writing the same budget constraint is E (1 + r)-i[P41X+i0X,4i - = Dr junO Here we see an important way of thinking about a nation's solvency constraint The present value of expected future trade surpluses must be equal to the existing level of international indebtedness. Thus, if an economy is burdened with international debt and wishes not to default, it must plan to run trade surpluses in the future that equal, in. present value,.the value of the nation's internrational debt obligations.; Adjusting to a Terms of hade Swoic Nigeria; 1972-88 179 To determine the appropriate response of consumption to changes in the terms of trade we need to discuss not only budget constraints but also consumers' tastes. For now we focus on the most basic motiva- tion governing consumption decisions: the desire to spread consump- tion evenly among periods of life. That desire leads to a simple rule for consumption, which we know as the permanent-income hypoth- esis: plan to consume in each period of life a constant amount and set that amount equal to the most that can be consumed without violating the lifetime budget constraint summarized in equation 8-1'. This simple rule for consumption expenditure gives an equally sim- ple rule of optimal savings and therefore, in this simple economy without investment, for the optimal current account surplus. When income is transitorily high, save, which implies current account sur- pluses. When income is transitorily low, dissave, which implies current account deficits. When income is equal to "permanent income;' nei- ther borrower nor lender be. This advice translates immediately into a simple prescription for responding to changes in the terms of trade. If the change in the terms of trade is permanent, it leads to a permanently lower level of income, as in the top panel of figure 8-1. In this case, the best response is for consumption to decline immediately to the new, lower level of national income; only such a policy maintains a smooth path for con- s-umption while still respecting the lifetime budget constraint. If, how- ever, the change in the terms of trade is expected to be transitory then it does make sense to borrow from abroad to cushion the short-run effects on domestic expenditure. The second panel of figure 8-1 depicts the optimal response to a temporary decline in the terms of trade, which generates a temporary decline in national income. PRODUCTION RESPONSES. The above discussion on financing versus adjusting to a terms of trade disturbance embodied a very limited perspective on the process of "adjustment:' which was viewed simply as bringing the rate of consumption expenditure into line with income. In reality, adjustment also includes the need to adjust produc- tion in response to the altered opportunities presented by changes in relative prices. A reduction in the price of oil will lead profit-maximizing pro- ducers to switch out of the petroleum sector and into the manufactur- ing sector, moving from one production equilibrium to another. This response of producers to the change in relative prices serves to offseL the reduction in national income that would otherwise occur as a result of the terms of trade disturbance.- If adjustments were costless, then prompt and complete adjustment would be appropriate to any disturbance to the terms of trade. Such 180 Michael Gavin Figure 8-1. Income, Expenditure, and Current Account Responses to a Change in the `Drms of 'Tade Income , Consumption (C), A pcrmancnt dcterioration Y-=c Time. A tranisitry deterioration Income (1X, Consumption (C), Y cTrad surplus Trade deficit * . . - : 4 adjustment would reduce the income effects of disturbances that are adverse and allow the economy to capitalize more fully on those that are favorable. Except for altering the magnitude of the income effects of relative price changes, nothing in the foregoing analysis of financ- ing versus adjustment would be altered. In particular, foreign borrow- ing would still be recommended. to accompany transitory distur- - bances, whereas permanent disturbances would require, full adjustment of production, consumption, and no foreign borrowing. In the Teal world, however, moving from one production equilib- rium to another takes time and is costly. This has two important impli- cations. First, bearing these costs of adjustment will. not be worthwhile if the deterioration in the terms of trade is expected to last for only a Adjusting to a 'Terms of 7ade S/Lock: Nigeria, 1972-88 10) very short time. Thus rapid, comnplete adjustment in the production sector makes sense in response to permanent changes in the terms of trade rather than Lo transitory changes (see, for example, Mussa 1978). So, the adage "adjust to permanent disturbances, but not transitory disturbances" acquires a fuller meaning: not only does it apply to changes in consumption expenditure, but it applies to changes in production patterns as well. Second, because adjustment of production to changes in relative prices takes time, so that the economy moves only gradually between the production equilibriums, a permanent deterioration in the terms of trade will have a larger adverse effect on income in the short run than in the long run, when production has had time to adjust. This has an important implication: when it takes time to adjust the production sector, even permanent changes in the terms of trade should, in part, be "financed" by foreign borrowing. Figure 8-2 illustrates the logic of this conclusion.. In figure 8-2 we consider a permanent deterioration in the terms of trade, which leads to an immediate decline in the value of national income from Y0 to YV. Over time, however, the production sector adjusts, and national income rebounds partially until income eventu- ally reaches its new, long-run level, Y2. The sustainable level of consumption expenditure will be some- where between the temporarily short-run level of income, Y1, and the long-run level of income, Y2. As figure 8-2 makes clear, this means that Figue 8-2. Income, Expenditure, ane: he Thade Balance with a Slow Adjustnent of Production Income , Consumption (C), y - tl Trade surplus. | . . ~~~~~~~~' Inlcomc R -= : ~~~~~~~Consumption Trade decrcit 1 .: : f -: 0 : . - S - f -.~~~~~~T'im 182 Michael Gavin consumption will be greater than income in the immediate aftermath of the decline in the terms of trade, which implics, in turn, a period of foreign borrowing while the economy adjusts LO the change in relative prices. The extent of foreign borrowing depends on the income gains (Y2 - YI) that come from adjusting production and on the speed with whlici adjustment takes place (sec Gavin 1988 for a more in-depth explanation). If the gains from adjustment are small, then a small amount of borrowing is appropriate, and vice versa. If the economy adjusts to the disturbance quickly, then permanent income is close to the long-run- level of income, Y2, and substantially above the short-run level of income, Y1. This implies that there should be more borrowing immediately following the impact of the change in the terms of trade than there would be if the speed of adjustment to the disturbance were quite slow. This points out a key qualification of the policy advice to adjust to permanent disturbances and finance transitory disturbances. In gen- cral, even permanent disturbances will induce changes in the econ- omy that take time to materialize and that may require resources to effect. When that is the situation, a case can be made for financing part of even permanent disturbances to the terms of trade. Relative Prices and AMjustment Our discussion so far has been based on a model in which traded goods are produced by a country too small to alter their relative prices. Adjustment comprised movements of factors between the export sector and an import-competing sector but involved no changes in relative prices. In reality, however, adjustment to a change in tLb, terms of trade may require dramatic changes in relative prices; and it is by counterproductive obstruction of desirable changes in relative prices that governments make some of their most important policy mistakes. To bring out these aspects of adjustment to a. terms of trade disturbance, we now introduce nontraded goods, the prices.of which, unlike those of internationally traded goods, are determined by the interplay of supply and demand in. the simall economy that we discuss. BASICS. For ease of analysis, let us consider an economy in which. the exported good is not consumel by the domestic population and the imported good is not producea in the domestic economy.3 We let the imported good be the numeraire, so that the price of the exported good is the terms of.trade and the price of the nontraded good is likewise expressed in terms of the importable. p~~~~~~~~~~~~~~~~~~ Adjustiig to a Tems of T7ade Slho/c Nigeria, 19 72-88 183 The supply side of the economy is easy to describe. Factors of pro- duction will move between the export sector and the nontradables sector, depending on the price of the exportable in relation to the nontraded good. When the relative price of nontraded goods increases, the nontradables sector bids factors of.production away from the export sector, and vice versa. We can sulmmarize the produc- tion side of the economy with an upward sloping curve, such as the curve labeled SS in figure 8-3, which gives the supply of nontraded goods as a function of the price of nontradables in relation to the price of tradables. We-will be interested once again in discussing the fact that econ- omies adjust gradually to changes in relative prices. The practical importance of this is that the long-run supply response to a given change in relative prices will be greater than the short-run response. Thus, if we view SS as a short-run supply curve, the long-run supply curve will be a flatter curve such as the one labeled LL in figure 8-3. The demand for nontraded goods depends on total expenditure and the price of nontraded goods in relation to the price of the Figure 8-3. The Terms of bade and the Real Exchange Rate SS: SS' LL - / DD N f , ./ / .- .-. '' '' ' 184 Michual Gavin imported consumption good. Total expenditurc, in turn, is equal to national income net of intcrest payments on foreign debt plus the rate of foreign borrowing, which is in turn equal to the current account deficit. Expressed as an equation, we have (8-2). E, = Y, -rD, + CA, where, as before, E is total expenditure on consumption, Y is the value of national production, D is the economy's foreign debt, and CA is the current account deficit. Using equation 8-2, one can derive a demand equation for any given level of the terms of trade, Px, and current account deficit, CA,.4 Given these two determinants, demand for nontraded goods is a downward sloping function of the price of nontraded goods, which is drawn in figure 8-3 as the curve DD. Equilibrium in the market for nontraded goods obtains, at the intersection of the short-run supply curve SS and the demand curve DD, which determines the equilibrium relative price and output of nontraded goods. Figure 8-3 provides some important insights. First, what is the rela- tionship between the current account deficit and the relative price of nontraded goods? An increase in the current account deficit corre- sponds to an increase in domestic expenditure in relation to income, which implies an increase in demand for consumption goods. Thus a current account deficit is reflected in figure 8-3 by an outward shift in the demand curve from DD to DD'. In the new equilibrium, point A, the relative price and production of nontrded goods have both risen. We therefore have- the conventional relationship between real- exchange-rate appreciation and increases in the current accoUrt defi- cit, where the real exchange rate is defined here as the price oi raded goods in relation to the price of nontraded goods. Such an apprecia- tion leads to a movement of productive resources from the tradables sector into the nontradables sector; as the export sector shrinks, a trade deficit emerges. An increase in foreign interest payments, holding constant the cur- rent account balance, has just the, opposite effect. To generate the trade surplus that is implied by the higher foreign interest payments, domestic expenditure must decline in relation to income. Thus, the demand curve DD shifts down and to the left, leading in equilibrium to a reduction in the relative price of nontraded goods, that is, a depreciation of the real exchange rate. Foreign borrowing, therefore, has an important influence on the real exchange rate in both the short and the long run, an influence that we will -want to bear in mind. In the short run, when the foreign borrowing serves to finance consumption expenditure in excess of income, the demand curve DD'is shifted up and to the right; consumption is high, and the real exchange rate Adjusting to a Tenns of 7ade S/ock: Nigeria, 1972-88 185 appreciates. Then, when expenditure is brought back into line with income to eliminate the current account deficit that corresponds to the foreign borrowing, the demand curve shifts down and to the left. It must actually shift further to the left than where it would have been in the absence of the temporary period of foreign borrowing, because the debt that was accumulated during the consumption binge must be serviced. Thus, we see that a transitory episode of foreign borrowing leads in the long run to the real-exchange-rate depreciation arising from the need to service the associated accumulation of foreign debt. We are now ready to discuss the optimal response of foreign bor- rowing to a change in the terms of trade. The analysis is complicated by the fact that changes in the terms of trade have both income and relative price effects. The income effects were emphasized above; when the terms of trade improve, domestic income rises and con- sumers demand more of both the nontraded good and the imported consumption good. This income effect means that the demand curve DD shifts upward when the terms of trade improve and downward when they decline. However, an improvement in the terms of trade also makes imported consumption goods less expensive in relation to nontraded goods at any given relative price of exports and nontrad- ables. This change in relative prices leads to substitution of the imported good for the now more expensive nontraded good. The income effect conflicts with the substitution effect; which effect will dominate depends on the degree of substitutability between the' imported good and the nontraded consumption good. The:most plau- sible supposition is that imported and nontraded goods are relatively poor substitutes in most developing economies, so this discussion will consider the case in which income effects dominate. When income effects dominate, an improvement in the terms of trade leads to increased demand for the nontraded good. Thus, the demand curve DD shifts up and to the right as in:figure 8-3, and a new short-run equilibrium is reached at point B. Thus, we see that if we maintain balance in the current account, an improvement in the terms of trade leads in equilibrium to an appreciation of the real exchange rate, whereas a deterioration of the terms of trade leads to a deprecia- tion of the real exchange rate. The short-run equilibrium point B is above the long-run supply curve, LL. Because the relative price of nontraded goods is higher in this short-run equilibrium than is consistent with long-run equilib- rium, factors of production will gradually move from the tradables sector-into the nontradables sector. This shifts the short-run supply schedule, SS, gradually to the right, and the economy's equilibrium- moves gradually along the demand schedule DD' until a new long-run equilibrium is reached at point C, on the long-run supply schedule. 186 A1ickael Cavin Thus, in the absence of foreign borrowing, an improvement in the terms of trade leads in equilibrium to real-exchangc-rate appreciation, whereas a deterioration in t-he terms of trade leads to real-exchangc- rate depreciation. Because production responds only gradually, the change in relative prices may be dramatic in the short run, but over time, as production gradually responds, the exchange rate apprecia- tion is partially reversed. Thus real-exchange-rate "overshooting" occurs in response to a change in the terms of trade, not because pricc rigidities occur or markets fail, but because production adjusts only gradually to changes in relative prices. IMPLICATIONS FOR OPTIMAL FOREIGN BORROWING. The. foregoing analysis assumed that foreign borrowing would not respond to the change in the terms of trade and the attendant fluctuations in relative prices. But the analysis is interesting because it reveals an additional motivation for foreign borrowing in response to changes in the terms of trade. During the transition from the short-run equilibrium, B, to the long-ru t equilibrium; C, the relative price of nontraded goods is=. g gradually declining. As Dornbusch (1983) points out in a slightly, dif- ferent context, this grdual decline in the relative price-of nornraded. goods corresponds to an increase in the rate of interest, which makes it desirable for consumers to consume less in the short run in return for higher consumption in the future. This increase in savings trans-. lates in this model into a current account surplus. Thus, a permanent improvement in. the terms of trade should be met with a transitory current account surplus, whereas a deterioration of the terms of trade should be met with a transitory period of foreign borrowing.-. An important implication of the optimal foreign lending and- bor- rowing just described is: that they tend to dampen the real-exchange- rate overshooting that would otherwise occur. The reason for this is that borrowing, and therefore increasing spending, is desirable when the relative price of nonttadables is, low; foreign borrowing tends to raise the prices of nontradables. Similarly, it is optimal to save:when. the relative price of nontradables is transitorily high, which tends' to lower the prices of nontradables compared with what they would be in the absence of the foreign lending. We see, therefore; that one result of foreign borrowing, when it is intelligently, conceived and executed, is to attenuate the otherwise wild fluctuations in relative prices:that nmight arise in the aftermath of a major change in the terms of trade. It is also possible to show that optimal foreign borrowing -tends to.' reduce the speed with which the economy adjusts to a:change in the terms of trade. This.happens exactly. because the foreign, lending and borrowing reduce the response of the real exchange rate to a given: change in tlieterms of trade. This reduction in -the speed with which ' Adjusting to a Terms of 7Vade Shock: Nigeria, 1972-88 187 the economy adjusts to a change in the terms of trade should not be considered a drawback of the foreign. lending; instead, it is an impor- tant advantage. This is because adjusting production is costly, and the cost rises the flster the adjustment is undertaken. If an improvement * in the terms of trade is met with a temporary current account surplus, the price of nontradables increases by less than it otherwise would, the speed with which resources move into the nontradables sector is reduced, and the waste and costly mistakes of excessively rapid adjust- ment are avoided. Of course, an excessive amount of foreign. borrowing could eliminate-for as long as the borrowing was sustainable-the real- exchange-rate depreciation that is the equilibrium response to a change in the terms of trade. In doing so it would likewise eliminate the incentives for the economy to adjust to a change in the terms of trade. We will argue that such overborrowing. was one consequence of the policies the Nigerian authorities followed after petroleum prices declined in the early 1980s. investment Up to now we have assumed.that savings that are a result of changes in the terms of trade should be channeled into foreign investments by running-current account surpluses. Changes in the terms of trade may also, however, alter incentives for domestic investment, which will have an effect on the* optimal response of: foreign .borrowing.' Although a plausible assumption is that an improvement in- the terms of trade would lead to an increase in domestic investment, little'theo- reticaijustification exists for this view if one maintains the assumption of perfect international capital markets. In .this case, the impact of a change in the terms of trade on the optimal' rate of domestic invest- * ment depends, among. other.,things, on the technology with which investment goods are produced and installed and on differences in factor intensities in the goods-producing sectors.. If international capital markets are imperfect, however, domestic 'investment in most developing countries will be constrained by 16w domestic saving A presumption then arises that increases in the terms * of trade that generate increases in domestic saving should lead to-an increase in:domestic investment as well.5 Two questions then arise. First, 'of the savings generated by an imnprovement in the terms of trade, how much should be invested abroad and-how much should be invested in the domestic capital stock? Second, in what sectors should theinvestmenttake place? - - Thex answer'to the first question- depends largely on the tme horizon that we consider.'Just as adjusting the structure of produc- J88 Miciael Gavin tion very rapidly is costly, so is making very rapid additions to the capital stock. The more frenetic the rate of investment, the more likely that bottlenecks will result in delay and waste and that costly mistakes will be made, These concerns are all the more important in developing countries such as Nigeria, in which the economic infra- structure is primitive and managerial and entrepreneurial resources are scarce. Thus it may pay first to 'divert a large share of the savings generated by an improvement in the terms of trade into the accu- mnulation of foreign assets. These foreign assets can then be used to finance capital accumulation at a more modeiate, and presumably less wasteful, pace. The sectoral allocation of domestic investments is an equally impor- tint issue. We noted above that an improvement in the terms of trade will, assuming that income effects are dominant,.lead to an increase in the relative price of nontraded goods. We argued that this would lead to a gradual movement of factors of production from the tradables sector into the nontradables sector. Similarly, the real appreciation: generated by the change in the terms of trade is 'a signal of the desir- ability of allocating a substantial share of domestic investment into the nontradables sector. This makes sense: as an economy grows richer, de~ country's population will naturally wish to consume more of both internationally traded goods and domestically produced..nontradable goods and services. The increase in export earnings, caused, by an improvement in the terms of trade makes it possible to import more of the internationally.traded goods, but increased. production of the nontraded goods and services requires an increase in the amount of productive resources that are devoted to the production of those goods. Investment is, of course; a forward-looking process; depending not only on the current economic environment but also on expected future conditions. One implication of this is that the extent to which investment should be biased toward the nontradables sector depends' on how fang the improvement in the terms of trade is expected to-last. If the improvement is predicted to be permanent, one can expect that the increase in the demand for nontradables will also be permnanent. Then placing a substantial portion of domestic investments into the nontradables sector may make sense. If the improvement in the terms . , of trade is expected to last for only a short time, it may be more logical to. place a larger share of domestic investments in the tradables sector. This is because when the improvement in the terms of trade comes to an end, the demand for nontraded goods will -decline toward where it. was before. the' transitory improvement in the terms of trade. There- fore, having in place a large capital stock devoted to production of nontradables will not be desirable. Mjwuing to a Ternu of Pade Shock. Nigeria, 1972-88 189 Uncertainty In theory we can draw sharp distinctions between temporary and permanent changes in the terms of trade and derive the appropriate response for changes of either kind. In the real world, however, we do not know for ccrtain whether a change in the terms of trade will be permanent or temporary. While we can confidently expect some disturbances, for example, weather-related ones, to be transitory, for many changes in the terms of trade it is simply impossible to know whether the terms of trade will revert fairly soon to their original. level, remain at the new level indefinitely, or even move farther in the same direction as the previous change. This uncertainty is particu- larly salient for the changes in petroleum prices -that buffeted the world economy in the 1970s and 1980s. Nobody knew in 1974 if OPEC would be able to continue extracting the monopoly rents it was then' enjoying. Certainly nobody forecast the Iranian revolution or pre- dicted its impact on oil prices. Finally, the collap'se in oil prices after 1981 was a shock to almost every informed observer of the world oil market. The theoretical literature has not emphasized the uncertainty of the terms of trade as'much as it should have. Although this is not the place- for a complete treatment, the implications of uncertainty are important enough to warrant at least some discussion. Here we will consider: the case for a conservative policy response when there is substantial uncertainty about future movements :in the terms of trade. In this context, the first important meaning of "conservative" con- cerns the level of saving and consumption.-The theoretical basis of reconmnending conservative policies .is the substantial literature: on "precautionary" motives.for saving. When there is uncertainty about future income, the appropriate generalization of the- dictum-Jto' smooth consumption is to plan on a consumption path that sets the marginal utility of consumption today equal to the :expected margi- nal utility of consumption tomorrow. Under certainty, and assuming that tastes don't:change over time, this' is the same as setting: con- : sumption today equal to. consumption tomorrow. But when there is uncertainty, it generally pays to forgo some consumption.in the present *in exchange for somewhat higher consumption -.in the future.6 This is the phenomenon'of 'precautionary savings:' The practical importance of precautionary savings is to provide a; justification for policies that aim for higher saving than would be appropriate if individuals or policymakers could be certain that the terms of trade will evolve as expected. In essence, authorities should - set aside income today to ersure against a bad outcome in the future. 190 Michael Gavin There is a second sense in which policies should respond conser- vatively lo changes in the terms of trade. As stressed above, in many cases the appropriate response is sensitive to whether the change is permanent or transitory. If the improvement in the terms of trade is permanent, then it may be appropriate to expand greatly the nontrad-. ables sector. If the improvement is only temporary, investCng in the tradables sector is more appropriate. If one is unsure whether the change in the terms of trade is permanent or transitory, the possibility exists that one will invest in the wrong sector. This-causes no problems if investments can easily be moved between'sectors, but if reversing investment decisions is very costly or impossible, then one may be stuck with the bad decision for a long time.. Recently a number of authors have pointed out that it may make sense- to -wait until some of the uncertainty about, for. example, whether a change in the terms of trade will sooh be reversed has been resolved (see, among others,,lDixit 1989; McDonald and Siegel 1986). By accumulating foreign assets instead- of investing domestically, it may be possible to reduce the probability of making costly mistakes .in the allocation of domestic investments. Then, when the uncertainty is at least piartially resolved, it will be possible to use the accumulated foreign assets to finance domestic investments. In short, uncertainty abo" t: the terms of trade seems to support policies that encourage high savings, protect the-economy against out- comes that are worse than expected, and place domestic savings in liquid form to avoid locking the economy into mistakes that are'very: costly or impossible to undo. In practice, this means larger current account surpluses in'response to terms of trade improvements than - would be desirable under certainty, and smaller deficits in response to a deterioration of the terms' of trade.' Sumvupy Adjustment.to a deterioration in the terms of trade comprises a reduc- tion in aggregate expenditure arid, under plausible assumptions, a depreciation of the real exchange rate. This depreciation will, in gen- eral, lead to shrinkage of the nontraded goods sector and a. concomi-. tant expansion of the tradables sector. An improvement in the terms of trade will generally imply an appreciation in the real exchange rate, which will: lead. to an expansion of -the nontradables sector and a shrinkage of the tradables sector. When-considering.policyiresponses, the first and most important distinction. is. between disturbances expected to, be permanent and those expected to be transitciry. The more persistent the disturbance, the more rapid and, complete the adjustment of both expenditure and production should be. This rule. * - - - . . . - . - C - - -- Adjtisttng to a Terms of 7'rae Shock: Niria, 1972-58 191 is summarized in the adage "finance transitory disturbances and adjust to permanent disturbances:' This advice must be qualified, however, when an' economy responds slowly to changes in the economic environment. In particu- lar, two reasons may make it sensible to engage in foreign borrowing in the immediate aftermath of a permanent deterioration in the terms of trade. First, as the economy gradually adjusts to the terms of trade disturbance, income will rise. It makes sense to borrow against that higher future income to raise consumption today. Thus, foreign -bon owing may legitimately be considered to finance costly'adjust-. ment to permanent, as well as transitory, disturbances. Second, real exchange rates will fluctuate more in the short run than in the long run, because short-run supply elasticities are lower than long-run, elasticities. Borrowing in response to a deterioration in the terms of trade, and running current account surpluses after an improvement, attenuates this overshooting of the real exchange rate. In doing so, it prevents what would otherwise be an excessively rapid-adjustment of the real economy to the terms of trade disturbance.. While pointing out the impact of optimal foreign borrowing, this - discussion incidentally brings out some implications of excessive reliance on foreign :borrowing to finance a deterioration in the t'orms Of tracde. First, a high rate of foreign.borrowing is associated with an overvaluation-of the'real exchange rate, and the overvaluation draws resources from the tadables.sector, where` .they belong, into.the nontradables.. Thus, 1 $rrowing too much in the wake of a deteriora- tion in the terms of ! .-ade will lead to a suboptimal rate of adjustment of the production sector to the terms of trade disturbance. Second, foreign borrowing cannot persist indefinitely, and when the current account is eventually corrected, the trade balance will. have to improve by even more than it would have in the absence of the. excessive foreign borrowing, -because interest on the acc'umulated "'-; ''foreign debt will have to be serviced. iThis will 'require a greater depreciation of the real exchange rate than would have. been:. neces-. sary in the absence of the excessive foreign borrowing. The result will be a lower. standard of living than would otherwise have: been possible. An even more fundamental qualification of the traditional policy. advice derives from the simple fact that one cannot know in advance whether a change in the terms of trade (or any other change. in the.. external environment) is transitory or pernanent, or will perhaps be augm'ented by a subsequent move in the same direction. In the face ; ' ' of such: uncertainty, a case.can be made for Spolicies that sacrifice consumption in the present to guard against highly adverse out-. comes in the future.' 192 Micihael Gavin Economic Policy: Money, Exchange Rates, and the Budget In the previous section we focused on a small barter economy charac- terized by continuous market clearing and populated by residents with rational expectations and full access to a perfect world capital market. The purpose of that focus-was to develop normative guide- lines for the management of a terms of trade disturbance. We empha- sized the appropriate response of foreign borrowing, although we also discussed other aspects of the adjustment process. In a sense, we may think of the previous section as defining the "external balance" toward which the authorities should steer their economies. This earlier analysis is, however, an. important step away from a serious discussion of economic policy. First, real economies are mone- tary, and it is through monetary and. exchange rate policy that govern- ments exert much of.their influence over the macroeconomy. Second, it is misleading to think of a government determining policy for for- eign borrowing and then letting expenditure and relative prices adjust to the policy. Instead, governments pursue monetary, fiscal, and regu- latory policies, and the level of foreign lending or. borrowing is deter-. mined as one part of the economy's equilibrium. If a government wishes to maintain the "external balance" as defined in the previous section, it must set its monetary. and exchange rate policies accord- ingly. The purpose of this section is to clarify the link between the . govemrnment's policy instruments and macroeconomic-outcomes. We. do so in the context of a simple monetary extension of the. previous model. Consider a small economy, very similar to the one discussed in the previous section, in.which three goods are produced; Oil is produced in some exogenously given amount, Q, and sold on world oil markets at an.exogenously given price, P1, in terms of a second tradable good,- which we label; as before, "manufactures!'7 Manufactures are, in equi--. - librium, imported, so the price of oil, P, is the terms of trade.' In addition to manufactLures, the economy produces nontraded goods, which in the case of Nigeria include.services, many agricultural crops,. and those parts of the manufacturing sector that are protected- by quotas or prohibitive tariffs and face no direct foreign competition. As discussed in the previous section, production of the-tradable good and the nontradable good depends on their relative prices: =m QAIAEIPNv) := Q(EIPN) Production of manufactures :() is increasing-in the real exchange:- rate, defined here as- the nominal exchange rate (E) divided by the Adjusting tooa Teins of 73'ade S/wdckNigNra, 1972 -8 & 193 domestic currency price of the nontraded good (ENj), and production of nontradables (Qv) is decreasing in the real exchange rate. As in Nigeria, the government collects all the..revenue from petro- leum production. In addition, it collects tax revenue from the domes- tic-privatc sector of amount T in t'erms of the nontraded go od. The government spends on the nontraded good and manufactures and must pay interest- on the outstanding foreign debt. We define the real government deficit as the amount by which'it must increase its mone-,' tary and debt obligations to the rest of.the world. This is equal to: RD ef = GN + (E(PN) GM + rD) - [EP.QJ Pv + T + irn] The first term in brackets is total, government' expenditure, which .includes spendig on the nontradable, the non-oil tradable, and real interestobiaions on the outstanding government debt,~ rD. TheSc ond term in-brackets is total revenue, which includes revenue from the sale of petroleum, taxes levied on the domestic population, and infla- tion tax revenue, win. Here ir is the rate of inflation for the prices of' nontrad ed goods, and m is the supply of high-powered money deflated by the price of nontraded goods.. This' equation emphasizes. the direct ink, between the. terms of trade and the government. budget; in Nigeria, for example, roughliy 75 percent of government, revenue comes from the oil sector_. It also empha sizsthe fiscal dimension of infatin; nflation effectively cancels th'e governmen' moeayla- bilities to the pivate sector,thereby effecting atransfer of wealth from Pr~~~~~~~~~~~~~~~~~~~~. the private tothepublicsector.Itconstitutesataxof the private sector that is just as real as a profits tax or import tariff, and, as we will see later, recourse to the inflation tax is an important part of Nigeria's recent economic liistory. The government budget deficitomust be financedeither by emitting interest-bearing debtor by issuing money. AthoughntherNigerian gov- ernment does borrow; both- from foreigners an.domestically; it also tends to respond to budgetary imbalancesb isusing money- T , a loose but important budg-etary link exists between the ~terms. of trade and the doxiiesticg money supply, which i exactly: the reverse of the relationship under a "classical" fixed-exchange-rate regime. Under a classical fixed-exchange-rate system, defined here as a system in which domestic authorities allow th ly to adust automatically in. response to changes in foreign reserves,, a deterioration in the terms of trde would lead to reserve outflows, which would, in turn reduce the domestic money supply. In Nigeria, however,titiwould be more realistic .tionp t e is the of trhee fcea- to suppose thatgoods,'d defic t generated by theermsoferddeete- rioration would lead, at least in the'shorti run, to anhinese ine money. supply, because the aau es tue therprintingeto finance tthe fiscal deficits.8 a 194 Michael Gavin Our second departure from. the last section is to abandon the assumption that consumers operate in a world of perfect foresight and frictionless capital markets and to assume that private expenditure is determined entirely by current income and wealth. This assumption of completely.myopic behavior is certainly as much a caricature of the world as was the extremely rational behavior modeled above.9 Remem- bering that the government owns the oil sector, we see that income in the private sector is equal to the.value of domestic production of manufactures and nontradables: - Y-= Qm(E1Pv)+ Qv '* This measure of income is, in tum, a function of the real exchange rate. We can therefore write demand for tradables and nontradables as a function'of the real exchange rate, domestic taxes, and real cash balances: C v = CN [EPv1N Y(EIPN) - T, MIPN - CM = Cm [EIPN, Y(EIPN)- T, MlPNI- We will assume that private demand for both goods is increasing in disposable income and: therefore decreasing in the level of domestic * taxation, T. Demand for the nontradable is assumed to. be increasing in the real exchange rate, EIPN, because an increase in the relative * price of manufactures creates an incentive for consumers to substitute in favor of nontradables. Similarly, we expect the demand for trad- ables to be decreasing'in the real exchange rate. Finally, we assume that demand for both the tradable and the nontradable is an increas- ing function of the stock of real money balances held by the. public.. There are many .explanations for why increases in money lead to increases in domestic demand, and they all boil down to reduced formsuthat look more or less like the equations above. Equilibrium in the goods market requires that production of-non- tradables. be equal to domestic demand for nontradables; this condi- tion is expressed in equation 8-3:. (8-3) Qv,(EIPv) = C(ElPN, T, AIPN) + GN Notice that the goods-market equilibrium depends on.the fiscal vari- ables.(specifically, the level of taxes and the government's purchases of -: nontraded goods), monetary variables, thc'exchange rate; and the money supply. Note also that the equilibrium condition (equation 8-3) depends only on "real" quantities. In-particular, a 10 percent increase in the money supply accompanied by a 10 percent increase in the price of nontraded goods will leave-the equilibrum condition unaffected if there is at the same time a 10 percent devaluation of the : nominal:exchange rate. Adjusting to a Term of 7ade Shock: Nieria, 1972-88 195 Second, we note that the economy's current account can be expressed as total income (public and private) less total expenditure (public and private). Expenditure includes not only expenditure on goods but also interest payments on foreign debt, D. Imposing equilib- rium in the market for nontraded goods, this can be written: (8-4) CA = lP, + QA (EIPN) - CM(EIPN T, MIPN) - - rD In our model, the money supply, the exchange rate, and fiscal variables are given by government policy. Monuteay Policy, Exchange Rates, and External Balance We consider first a change in monetary policy. We begin in a position of internal and external balance, 'When the economy is disturbed by an expansion of the domestic money supply. As noted above, both.equa- tions 8-3 and 8-4 depend only on relative prices and real money balances. Thus, if both the exchange rate and the relative price of nontraded goods increase in proportion to the increase in the money supply, the economy remains.in internal and external balance. The exchange rate is usually not adjusted immediately, however. As long as it is held fixed, the relative price of nontraded goods increases,. but not by qs much as. the increase in the money supply. Therefore, real money balances - increase. When the nominal exchange rate is not,. changed, the. increase.in: the price of nontraded goods implies an appreciation in the real exchange rate. This appreciation, with the increase in real balances, leads to an increase in demand and a reduc- tion in the supply of traded. goods.. The increase of demand over supply in turn implies an external deficit. The counterpart of the current account deficit is a reduction in private sector saving associ- ated with-the expenditure effects of high real balances.- : To defend the nominal exchange rate in the face of the external * deficits generated by the expansionary monetary policy, the central bank has to intervene in foreign exchange.markets. Thus the.external deficits generated by expansionary monetary policy lead to reductions * in the central bank's international reserves. If the central bank runs out of, foreign exchange, the external deficits associated with the expansionary monetary.policy become impossible to finance, and;the exchange rate has to be devalued. When this happens, the economy moves to a new equilibrium, in which the domestic currency prices of. both the traded good and the nontraded good have increased in pro- portion to the rise in the money supply... Two points about dev-aluation warrant some emphasis. First,-we see here the inflationary cost of devaluation; however, it is not the deval- uation itself that bears responsibility for the inflation, but rather the 196 Michiael Gavin - expansionary monetary policy that made the devaluation necessary. Second, failure to devalue aftcr a monetary expansion delays much of the inflation and in doing so makes future devaluation and inflation predictable. This expectation of devaluation induces domestic resi- dents to try to protect their wealth by obtaining foreign assets, thereby causing-a deterioration of the balance of paymeints not only on the current account but also on the capital account. We shall see that, although Nigeria's external deficits were primarily the result of cur- rent account transactions rather than capital flight, the deficits were aggravated by outflows on the current account when people realized that devaluation was in the offing. Fiscal Policy and the External Equilibrium We can quickly summarize the effects of fiscal policy. Consider, for example, an increase in domestic taxes or, equivalently, a reduction in.' government transfers to the private sector. At given relative prices, these changes reduce private sector demand for both the traded and the nontraded good. If the nominal exchange rate is held fixed, the economy moves to a point at which the domestic price level has'fallen, the real exchange rate has depreciated, and the current account has: moved into surplus. Thus governments can use contractionary fiscal policy to eliminate, an external deficit, but such a policy does not substitute for an appre- ciation in the real exchange rate.. Real depreciation is required because, if relative prices do not change, the fiscal contraction leads to. an excess supply of nontraded goods, which leads, in. turn, to unem- ployment. Real depreciation induces firms in the -tradables sector to. hire labor that has been released from the nontradables sector, thus establishing equilibrium for domestically produced goods' and at the same time increasing the size of the current account surplus generated by the fiscal contraction. Changes in government spending have effects similar- to changes in : - . government taxes, except that-one needs to specify-whether spending on the nontradable or the tradable has changed. Reductions in spend- - '- ing on the nontradable reduce domestic.-demand. At the original exchange rate- and price of nontraded goods- an excess supply of non- tradables emerges. Elimination of this excess supply requires a reduc- tion in the price of nontraded goods, that is, a real-exchangerate depreciation that coaxes resources from the nontradables sector into the tradables sector. The'expansion in output of tradables leads to a ' : trade surplus. Thus reduced government spending leads to an increase in.the trade surplius, even if the spending reduction falls on domestic goods. Adjusling io a 2rms of fade Shwck: Nigeria, 19 72-88 197 The effect, however, is an indirect one that operates through an induced decline in the price of nontraded goods. If relative prices take somc Lime to decline, then the immediate effect of the spending reduction may be to generate an excess supply of domestically pro- duced goods and a transitory period of higher unemployment. Even- tually the unemployment goes away, but at the expense of a depreci- ated real exchange rate. The governmenL can avoid these effects by concentrating:spending reductions, to the extent possible, on internationally traded, as opposed to nontraded, goods. The domestic equilibrium is unchanged, but a current account surplus emerges. There are, how- ever, limits to the wisdom and the feasibility of such a policy. For the government as well as the private sector, traded goods are imperfect substitutes for nontraded goods. Consider the following particularly relevant example. In many developing economies, investment is a very import-intensive activity; such economies simply do not produce many of the required capital goods. During an attempt to eliminate an ongoing fiscal and external deficit, governments are-often tempted to .cut investment spending, not only because such reductions have a smaller impact on current standards of living but also because invest- ment spending is import-intensive, which means that a given improvement in the current account'can be achieved with a smaller required adjustment in the market for domestic goods. Although, as discussed in the first section, there may be good reasons'why a govern- ment should meet a deterioration in the terms of trade with a reduc- - tion in investment spending, a desire to postpone'required adjust- . ments in labor and product markets is probably not one of them. The Terms of Trade in a Monetary Economy Finally, let us analyze an exogenous change in the terms of trade. Once.. again, we imagine that'an economy that begins in a position of inter- -nal and external balance is disturbed by an adverse movement in the terms of trade. At the initial equilibrium, prices, the external accounts have now moved. into deficit. Reestablishing external balance requires-' depreciation. On the assumption, however, that government spending ' does not respond immediately, the short-run domestic equilibrium -remainsbut is now accompanied by external deficits. The domestic' counterpart of this external deficit is a domestic fiscal deficit, caused ; by the loss of government export revenue. Irz the first section we discovered several circumstances under which a deterioration in the terms of trade should be met. with a transitory - period of foreign borrowing. If, however, consumers- are as myopic or: constrained in capital markets- as we have assumed in this: section, 198 Mich ael Gavin there are strong grounds for presuming that they will adjust their expenditure insufriciently and therefore that thc current account defi- cit that emerges will be undesirably large. Elimination of the external deficit requires some combination of fiscal contraction, monctary contraction, and nominal excchange rate depreciation. The above dis- cussion of the separate instruments can bc applied directly to. the problem of eliminating the deficit generated by the-terms of trade deterioration. Case Study: Nigeria In this section we apply the principles outlined earlier to a dit.cussion of Nigeria's recent economic history, which divides naturally'into sev- - eral subperiods. We begin with a.very brief discussion of Nigeria's' economic development before the 1973 oil boom. It was dulring the period between Independence in 1960 and the 1973 oil shock that some. of the economy's weaknesses first appeared. V/e then turn to a discussion of the five-year period following the first of the two major- increases in oil prices.that rocked the world economy in the 1970s and 1980s. Shortcomings in the response to that price increase were revealed when oil markets began to soften in 1978 and early 1979. In. 1980 the Nigerian authorities were saved by the second round of oil price increases that followed upon the-supply disruptions associ- ated with the Iranian revolution and the Iran-Iraq war. Though sharpm this increase in oil prices was more short-lived than the earlier' distur- bance, and by 1981 oil markets. had begun to weaken considerably. They found. a plateau during 1982-85, when Nigeria's external posi- ' tion was almost,, though not quite, as favorable'as it had been during the 1973-79 period. The last stage in the oil cycle began in 1986, when- the bottom fell out of the oil market. We break the period since 1979 into two stages, distinguished not.by conditions in the oil market but Tather by Nigeria's policy response to'- those conditions. 'The first period, 1980-83, was a period of almost- lurid fiscal and monetary excess, carried out essentially with no regard for budgetary realities. The legacy of this period continues to haunt. attempts to. stabilize the Nigerian eco.,omy today. The second- period, 1984. to the present, was one in which the government reversed those fiscal and monetary excesses-at first tentatively, then more forcefully during Babangida's regime.. Act 1: The Formation of the Oil Economy During the period leading up to the first oil shock, Nigeria's economic, record was quite favorable. If the IMF data are to be believed, Nigeria's: ., . - - - . :, ....................................................... - ......... -2- AIvwUnsmg to a 7krms. of Tdrad Shoclk Nigeria; 1972-88 199 econom-y grew 9 percent a year duiringte190 and about 5.5 percent a yea.r dur-ing the fir-st twelve year-s after Independence-this de'spite the disruptions caused by the fouir-year civil war of 1967-70. Up until about 1960 agriculturle domiinated the Nigerian economy. Farming of dom-estic crops was the main.occupation, and -domestic agriculture.was supplemented by a vigorous export-oriented agriwi Cultural sector. Niger-ia was for many years a dominant producer of cocoa, palm products, and other tropical agricultural products. Upon Independence in 1960, a new political class, with a new set of priorities, obtained contr-ol of the government (see Bienen 1988 and Kirk-Grecene and Rimnmer 1981 for a discussion of the political- attitudes of the Nigerian authorities). There Wats a.'strong and under- standable desire to see the benefits of de-velopment directed toward Nigerians'and away from the colonial elite, and development was understood above all to entail -industrialization. In practice, these priorities led to the development ofVa highly interventionist state that used its power- to protect local producers by subsidization, protection from foreign competition in commodity 'markcts, .and, in the "Nigerianizatio n" program of the 1970s, protection in.:factor markets. The interventions had two effects that concern us here. First, pro-. tection c5f urban manufacturing led to the development of an ineffi- cient manufacturing; sector that relied highly on' imported internmedi- ate goods. The external resources that fueled this, sector could come only from the oil sector, for during the 1960s and 1970s, policymakers' neglect of the agricultural sector had led to a dramatic~ shrinkage of Nigeria's once robust agricultural exports (see table 8-2). . - By 1972 production of most major export crops had shrunk dramat-' ically. This was partly because of the ravages of the civil war. It was ¶ihble 8-2. ikaditional Exports from Nigeria, 1964-78 (thousands of ton v). crop 1964-66 .1970-72. .1976-78.. Cocoa beans .218 232 193 Grounidnuts . .554 179 1 Grounidnut oil 95 57 Palm kernels 409 21317 Palm oil 145 10 2 Rubber 71 5] 31 Raw cotton _ 21 ..1, 4 Lumbe 594" 20 14"1 Negligible. a. Thousands of cubic meters. Source: Kirk-Greene and Rimnmer (1981). 200 Michael Gavin in part a natural response to the increase in oil wealth during the 1960s, as factors of production were pulled out.of the non-oil export sector and put into domestic crop production. But the decline of export agriculture was also caused by a tendency for government cornmodity boards, ostensibly created to stabilize crop prices, to tax farmers by paying them substantially less than the world price. The effect of this policy was the creation of an economy, that was exces- sively reliant on petroleum, not only to'pay for imported consumption goods but also to keep the manufacturing sector running. The risks inherent in this policy would become apparent in the crisis of 1981-84. The major development in Nigeria during the 1960s was the exploi- tation of its petroleum deposits. Production began in 1958 and by 1966 had reached 418,000 barrels a day. By 1970 production exceeded. 1 million barrels a day and in 1973 slightly more than 2 million barrels a day, roughly equal to the present capacity. Not only was the quantity produced increasing gradually, but so was the price in the period leading up to the first oil shock; and by 1973 petroleum revenues accounted for.almost 85 percent of Nigeria's exporrs, 15 percent of: GDP, and roughly 60 percent of fedenrl government revenue. Until 1972 the Nigerian -authorities followed relatively'restrained -financial policies. There was a bout of wartime deficit finance in the late 1960s, but by 1971 the deficits had vanished. Between'-1960 and' 1972 consumer price inflation averaged about 5 percent a year,'close. enough to that of Nigeria's major trading partners -to make, an exchange rate of $1.40 to the' naira sustainabie from 1960 to 1972.10., Relatively minor current account deficits: were largely financedby foreign investment, and the government's internal and- external indebtedness-was low. Act 2:. The First Oil Boomz, 1972-78 Petroleum prices, like most commodity prices had followed a gener- ally.rising trend during, the late 1960s. Then,. following an Arab embargo of oil exports to countries deemed supportive of Israel, the world price.of oil tripled almost overnight. As table 8-3 and figure 8-4 . indicate, Nigeria's terms of trade increasdd dramatically, in 1974. The first column of table 8-3 gives the World Bank's' measure of the - - terms of trade, which tripled between 1973 and 1974. The second column shows the U.S. dollar value of Nigeria's exports. To facilitate historical comparisons, the third column gives the value of Nigerian, exports in real terms, in which the deflator is the.price of US.- indus- trial goods."' We see that the purchasing power of Nigeria's exports doubled in 1974 to almost 20 billion:1988 dollars. Mdjusting to a Terms of deSck: NigrMa, 1972-88 201 bable 8-3. Terms of flde and International Payments, Nigeria, 1970-88 (millions of 1988 dollars) TM" of Current Internadionat Foreign 7fade Exports Exports account resemes . debt Year (1) (2) (3) (4) (5) (6) 1970 23.4 1,248 3,621 -368 202 567 1971 19.8 1,889 5,303 -406 408 651 1972 18.9 2,184 5,862 -342 355 732 1973 22.3 3,607 8,571 -8 559 1,205 1974 65.3. .9,698 19,396 4,987 5,602 1,274 1975 57.7 8,329 15,245 42 5,586 1,143. 1976 62.7 10,122 17,684 -357 5,180. 906 1977 64.0 12,431 20,475 -1,018 4,232 3,146 1978 57.3 10,508 16,056 -3,785 1,887 5,091 1979 70.7 . 16,774 22,776 1,372 5,548 6,235 1980 100.0 25,741 30,617 5,104 10,235 : 8,888 1981 108.7 17,961 19,587 - 6,220 3,895 12,039 1982 100.9 12,088 12,924 -7,241 1,613 12,908 1983 96.6 10,309 10,887 -4,337 990 18,586 1984 97.2 11,827 12,192 114 1,462 18,664 1985 89.8-. 13,369 13,850 2,623 1,667 19,522 1986 44.4 6,599 7,041 371 1,081 24,470 1987 49.1. 7,702 8,011 362 1,165 28,714 1988 - 7,419 7,419 -555 - 651' _ Not available.. Source wF (various issues); World Bank (1989a, 1989b). To put this disturbance into some sort of perspective, note that the $6 billion increase in the value of Nigeria's exports amounted to about N 4. billion, which was, in turn, more than one-fifth of the 1974 Nigerian GDP. The disturbance was even more significant for the pri- mary recipient of the price increase, the Nigerian government. lTble 8-4 shows the impact of the oil price increase on the Nigerian govern- ment's budget.. With the increase in the value of its' oil taxes, the governmenets fiscal situation was transformed. Petroleum taxes were about 8 percent of GDP in the 1972173 fiscal year, rising to 9t percent in 1973174 and 20 percent in 1974/75. There they remained until 1978. In- 1974 the gov- V ernment was simply unable to adjust spending rapidly enough to keep up with the growth in revenue. Some increase in spending occurred, but not enough to prevent the emergence of a large fiscal surplus., The counterpart of this fiscal surplus was a movement of the econ- omy as a whole into surplus compared with the rest of the world. Table: 8-5 and figure 8-5 show that in -1974, Nigena swung into a surplus on- 00 0 0 the current account; of its international payments. Combined with 202 Miciael Gavin Figure 8-4. Nigeria's Dnrms of 1kade, 1966-87 120 100 - 80 - 60 - 1 I 2 01.I ii rnium;n 1' ";-- 40 - i iii i i -~ ~~~2 ':1 R , _1 = i''"1 il§ -i ; - 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 Source: World Bank data. some net capital inflows, Nigeria ended 1974 with an increase in inter- national' reserves of more than $5 billion. As the end of 1974 approached, the Nigerian government faced all the uncertainties and decisions that were discussed earlier in this essay. The government was unclear whether the change in the terms of trade would be permanent or transitory.-It had to decide whether to accumulate more foreign. reserves or, instead, spend the higher income. If the income should be. spent, was investment or consump- tion the road to take? If invested, which sectors should receive the resources? l ble 8-4. Oil Receipts and the Fiscal Situation, Nigeria, 1970-1978/79 (percentage orGDP) Spending es- Fiscalyear - Oil receips :.. otherreceipts . .. rlscal baanc.e- 1970171 2.20 3.9 -1.6 -1191 . 1971172 5.60 5.2 04 - 36P 1972173 8.30 9.2 -0.9 -83 - 1973174 9.10 7.4 1.7 189- 1974/75 - 20.00 13.3 6.7 1,2482' 1975)76 20.20 27.0 -6.83 - -1,436''' 1976177 19.80 26.7 -6.9 -1,870 - 1977178 19.10 25.8 -6.7 -2,134a 1978179 13.75 - 17.3 - -3.5 -1,851 . a. Millions of nairas. SourwccPintD (1986). MAdjsting to a Terms of Tade Shock: Nigena, 1972-88 203 lible 8-5. International Ttade and Payments, Nigeria, 1972-78 (billions ofdollars) Otiler Curret 7tade current account Net capital Overal Year balance atcount balance movem-nts balance 1972 0.82 -1.16 -0.34 0.48 0.14 1973 1.89 -1.90 -0.01 0,13 0.12 1974 7.22 -2.32. 4.90 0.22 5.12 1975 2.85 -2.80 0.05 0.14 -0.02 1976 2.64 -3.00 -0.36 -0.02 -0.41 1977 2.71 -3.73 -1.02 0.19 -0.95 1978 -1.18 -2.61 -3.79 1.67 -2.34 Source: IMF (various issues).: The answers arrived very quickly. Driven by a political imperative to spend the oil income quickly, the government adjusted expendi- ture to the new, higher level by 1975. As table 8-4 indicates, by that year the federal budget had moved from surplus to a substantial deficit. This deficit persisted at almost 7 percent of GD? for the next three years. The fiscal authorities chose, implicitly or explicitly, to act as though the improvement in the terms of trade was permanent and Ito disregarrd the risk of a fuLure reversal of the current, favorable trends. The composition of government expenditure, especially in the : _early years, was fairly heavily weighted toward. investment. Current government expenditure grew rapidly, reflectir.g increases in public sector wages,12 subsidization of utility prices, and increased spending *: Figure 8-5. Curm nt Account, Nigeria, 1966-88 (billions of dollars) 6 4 -2 -6 S*u..e. IMF (:arious iSSU:). 204 Miclade Gavin on educationi, public health, and other social programs. Although current expenditure between 1973 and 1976 rose by a factor of about 3.5, capital expenditure rose by a factor of more than 9. Figure 8-6 shows that real gross investment more than doubled between 1973 and 1977; most, but not all, of this increase was attributabie to the public sector. The response of investmenL spending is the bright spot of Nigeria's economic management. Saving some fraction of the income gains' from a terms of trade improvement is, we argued earlier, a sensible policy when the terms of trade might return to their original level. It is extremely implausible that Nigeria did not have sufficient viable projects with returns higher than the rate of interest on foreign assets. From that perspective, investing domestic savings in domestic capital made, sense. But,did the Nigerian.authorities give themselves enough time to find the right projects and to implement the invest- ments in a reasonably efficient manner? There are strong grounds to suspect that the answer is no. When the magnitude of the petroleum windfall became clear, the military government urged-the administra- Figure 8-6. Gross Investment, Nigeria, 1967-87 (billions of 1980 naira) :17.5 -=-.- ~::: 12.5- 10 -. ' 1967 1969- 1971 1973 1975 1977 1979 1981 1983 1985 1987 S.ne IMF (vaious issucs) - Adjustiing to a Terms of 7'rad Shiock: Nigwir4 1972-S8 205 tors simply to scale up expenditure on high-priority sectors such as tran.sportation and education. Without the executive capacity to mon- -itor and execute the rapidly growing public investments, the projects were inevitably marred by unnecessary waste and fraud. A notorious example is the cement scandal of 1976 when, as a result of corrupt practices, more than 400 ships canrying imported cement tied up the .harbor at Lagos for many expensive months. Furthermore, the con- centration of petroleum wealth in the hands of th-e government pro- vided a degree of freedom for bureaucratically attractive but uneco- nomical investments, of which the inevitable steel complex under construction at Udoja is a particularly expensive example.' There are also grounds for questioning the sectoral allocation of the government's. investments, which were heavily: oriented toward the provision, of n'ontradable services, notably education, health,. and transportation. It is not so much that these priorities were demonstra- bly wrong as that committing the investments so early was incautious. If the oil boom had turned out to be permanent, such an allocation may have been appropriate. But the risk was always present that the oil boom was going to be transitory,in-which case pla'nners would regret the absence of ,investments in*- the tradables sector, -which could have provided an alternative sour'ce of export revenue. Combined with the fiscal response to the oil boom was a strong monetary response. The huge, rise in international reserves. that occurred in 1974 was not totally sterilized by:-the monetary authorities. Reserve money, almost tripled in 1974, -whereas m roney 'held bythe public roughly doubled. In addition;,h icldeiista emegdi the three years after 1974 resulted in rapid monetary gowth in those' years. as well. By the end of -1977 the money supplyi had risen -sixfold from the 1973 level. inflation responded as one would expect. Between 1973 and 1978 the consumer price index rose by a factor of 2.7, an average inflation rate.. of 22 percent a year. With th oia xhnerate approxi- mrately fixed, the real exchange rate appreciated strongly,during die peri'od, contributing to the external'deficits. By 1978 these excessively' expansionary budgetary-and monetary, policies were -taking their toll on the economy. In 1978 the budget deficit was about 7 percent of GDP for the third year in a. row. With private spending relatively robust and with a slight softening of the oil market in 1978, the-current.account of the bala'nce of paLyments moved into a substantial deficit, and the government was forced for the first time to obtain loans from' foreign banks. ~More borrowing was pro-- jected for 1979. Before turning to the, 1979-83 period, let us briefly summarize and, evaluate the policy response to the first oil boom: 206 Mickael Gavin Essentially none of the increased income was saved in the form of foreign assets. The cumulative current account during the five- year period 1974-78 amounted to a deficit of $131 million. This was an incautious policy. A more conservative policy would have been to save some of the increased income in the form of foreign assets, which could easily have replaced export earnings in the wake of a decline in the oil market. * There was a substantial savings response to the increase. in oil' income, but the savings were all invested domestically. These domestic investments were positive developments and could have. laid, the groundwork for higher income in the 1980s, but the evidence suggests that-the rapidity with which the public sector investment program expanded lent itself to corruption and expensive mistakes. The government adjusted its expenditure to the new, higher level of income with alacrity and within a year of the disturbance was quite overextended. Persistent, large budget deficits emerged and were financed in part by the issuance of money. The inconsistency: of the government's fiscal, monetary, and exchange rate policies was becoming increasingly clear by late 1975. Recognition of this inconsistency spurred a sharp depreciation of the parallel-market exchange rate during .1975. The depreciation continued, with occasional interruptions, until 1978 (see figure 8-7). Act 3: The Second Oil Boom and theGathering.Crisis, 1979-83 In 1978 the Nigerian government faced a difficult economic situation. The external accounts were clearly out of balance: substantial foreign borrowing had been required to finance the current account deficits, and more borrowing was projected fdr 1979. These external deficits were largely the result of two factors: first, the government's sluggish fiscal response to the weakening oil markets and, second, the accumu-- lated effect on the real: exchange rate of rapid monetary growth- and the attendant domestic price inflation, unaccompanied by exchange rate depreciation. *The government responded with some measure of fiscal retrench- . ment. Tlble 8-4 shows that in 1978179 the fiscal deficit fell by roughly half, to 3.5 percent of GDP, despite a significant reduction in the value of revenue from the petroleum sector during that year The excess of. domestic spending over the non-oil revenue fell from the 27 percent of GDP it had reached in the previous three years to only 17 percent of G-GDP in 1978179. Much of that retrenchment came from reductions in 'public sector investment (see figure-8-6, which shows a substantial Adjusiing to a Terms of 7ade Sitock: Nigeria 1972-88 207 Figure 8-7. Official and Black Market Exchange Rates, Nigeria, (dollars pcr naira) 2.0 1 .5 1.0 . ~~~~~Black rarketrae.I 0.5 ..........-.: 0 1970 1972 1974 1976 1978 1980 1982 1984 1986. 1988 Sruir: IMF(vamious issucs); WoTid Cu47W.- rbook, 19868?. decline in the rate of real investment from the veTy high levels reached in. the mid-1970s). The government did nothing significant, however, about the real exchange-rate appreciation that-had occurred during., the previous years, and the desirable but somewhat. half-hearted fiscal measures of 1978/79 are unlikely to have succeeded if oil markets had remained as weak in 1979 as they were in 1978. In any event,, the Nigerian government was saved from painful adjustment by the (Iranian) bell. In the beginning of the year, the shah.. of Iran was deposed in an uprising of fundamentalist Muslims. The attendant disruptions in. petroleum markets led to a second massive increase in world oil prices, which raised Nigeria's terms of trade by- almost 25 percent in 1979 and another 40 percent in 1980. Table 8-3- ; shows that the increase in the real value of Nigeria's exports between 1978 and 1980 was, in absolute terms, about as large as, the increase. that occurred between 1972 and 1974. As.happened in the-earlier period, the current account moved out of the 1978 deficit into a surplus in 1979 and an even larger surplus in 1980. By the end of 1980 Nigerian foreign reserves had:reached $10 billion-more than $120 for every man, woman, and child in the country and about 14 percent, of the World Bank's esti'mate of per capita GNP during that year.- -Once again, the Nigerian authorities faced the question of how to disposeofthis massiveinfusionofoil wealth.This time,onemighthave 208 MicaweL Gavin expected a more conservative policy, as the authorities now knew from painful experience that oil prices could fall as well as rise and as they had felt some of the painful consequences of adjusting to such price declines. However, in the years following the second oil shock, the authorities responded with policies that were significantly more incau- tious and extravagant than were those following the first oil shock. lTble 8-6 provides data on the fiscal accounts during this period. As in 1972-74, the oil price increases led to a vast. expansion of govern- ment revenue from the petroleum sector, revenue that rose from 14 peircent of GDP in 1978179 (see table 8-4) to 22 percent in 1979 and 23 percent in 1980. The fiscal discipline that was instituted in 1979, how- ever, disappeared during 1980, and although the deficit declined dur- ing these two years of the second oil boom, the government remained in deficit during both years. The expansion of government spending during 1980 and 1981 is impressive: in 1980 total expenditure rose by .65 percent, encompassing increases in domestic subsidies, new capital expenditures, and a resumption or acceleration of construction proj- ects that were suspended or slowed during 1978.. If the failure to maintain some measure of fiscal discipline during the two boom years was unwise, the Nigerian authorities' response to the softening of oil markets, which began in the middle of 1981, was disastrous. The decisions taken-or rather not taken-'during 1981, 1982, and 1983 are chiefly responsible for the crisis in Nigeria. We noted that the fiscal restraint instituted during 1978 was relaxed in 1980 and 1981, a relaxation that was accompanied by an increase in oil revenue and an overall dedine in the fiscal deficit. In 1981 the oil market softened considerably. Tble 8-3 shows that the real value of Nigeria's exports, measured in. 1988 U.S. -dollars, fell by one-third, from more than $30 billion in 1980 to less than $20 billion in- 1981. Table 8-6 relates thefiscal consequences of this decline in oil pnces. Revenue from the petroleum sector fell by almost 5 percent of GDP in 1981 and continued to fall during 1982 and 1983. bkble 8-6. Oil Receipts and the Fiscal Situation, Nigeria, 197983: (percentage of cmr) Spendinglass- FisCal yar Oil receipts other receipts Fiscal balance 1979/80 22.3 24.2 -1.9 _757a 1980 23.1 23.4 -0.3 . -143: . 1981 18.5 27.4 -8.9 _4,734a - 1982 16.7 24.7 --8.0 ; -4,524a; 1983 . 12.0 - 23.0- -11.0 -6,650: a. Millions of nairas. Soruce. Pinto (1986). Adjusting Co a Terms of Trade S/ck: Nigeria, 1972-88- 209 The fiscal response to this decline in government resources was negligible. Indeed, during 1981 the non-oil budget expanded signifi- cantly, with the excess of domestic spending over non-oil revenue -increasing to a historical high of more than 27 percent of Gor. The result was a fiscal deficit in 1981 of almost 9 percent of GDP. This deficit remained about the same in 1982 and increased to I1 percent of GDP in 1983, as small fiscal contractions were overwhelmed by large reductions in the government's petroleum revenue. The fiscal deficits were accompanied by a generally accommodative monetary policy and, most important, no significant change in the nominal exchange rate. Domestic prices rose 60 percent during 1980- 83, responding to lax monetary policy, the inflationary effects of a halfhearted depreciation of the exchange rate, and the lagged effects of the extraordinarily expansive monetary policy pursued in the years just preceding 1980. The nominal exchange rate was altered insignificantly, compared. with the high rate of domestic price inflation, and the real exchange rate therefore appreciated significantly. By 1984 :the real effective exchange rate, as measured by the IMF, had appreciated almost -125 percent compared with 1976, when the measures first became avail- able. If anything, this figure probably understates the:degree of over- valuation; between 1976 and 1984 the Nigerian economy had accumu- lated a crushing debt burden, and, as we discussed earlier, the -servicing, of such,a debt burden implies a depreciation of the real exchange rate.13 The simple macroeconomic model laid out in the.previous section makes a prediction that expansionary fiscal policy, inflationary' mone- try policy, and. deteriorating terms. of trade will lead to- current account deficits, and table 8-7 shows that current account deficits -materialized in-this period on a massive, scale. During 1981 the cur- rent account moved from the $5.billion surplus recorded in 1980 to a deficit of $6 billion. In 1982 the deficit widened to more than $7 * billion, and although the deficit shrunk in 1983, it was still clearly excessive. The fact that the deficits recorded during 1981-83 were excessive is .apparent from foreigners' unwillingness to finance them voluntarily. During 1981 the current account, and. baance of payments deficits were financed entirely from previously. accumulated foreign reserves. ; These reserves therefore fell from more than-$10 billion at the end of 1980 to $6 billion at the end of 1981. The lower level.of reserves was insufficient to finance the $7 billion current account deficit that emerged during 1982, and foreignersmwere'apparently unwilling to increase their lending to Nigeria at the pace required to, finance what the. country could nrot finance by drawing down reserves. The result f ~~~~~~~~~~- E-. E t 210 Michuel Gavin TAble 8-7. International 'ade and Payments, Nigeria, 1979-83 (billions of dollars) 01/an Cturrent fade current accoutd Net capita Overall Year balance accounn balance movements balance 1979 4.07 -2.70 1.37 1.83 3.66 1980 -11.11 -6.00 5.10 -0.76 4.69 1981 -0.91 -5.31 -6.22 0.72 -5.50 1982 2.71 0.53 -7.24 1.65 -5.59 1983 -1.08 -3.26 -4.34 1.37 -2.97 Soure: IMF (various issucs). was a rapid accumulation of arrears: involuntary extensions of credit by Nigeria's trading partners, extensions that continued during 1982 and 1983. Estimates of the amount of arrears vary, with the Nigerian government understandably more reluctant than claimants to grant the validity of claims, but the accumulation of arrears was by any calculation large. Foreign exporters claimed obligations amounting to about $10.5 billion at the end of 1985, whereas importers' records showed only $7.7 billion (Nigeria 1986). TIble 8-8 gives the IMFtS estimates. Table 8-8 shows that in 1982 and 1983, the height of Nigeria's pay- ments imbalances, arrears were being accumulated at the rate of $2 billion-$3 billion a year, so that a substantial fraction of the payments imbalances in those years was "financed" by accumulating arrears. During the entire 1981-87 period, the payments arrears totaled a staggering $9 billion, roughly one-third of Nigeria's international debt in -1988. The authorities responded to the payments imbalances during this period by imposing increasingly restrictive quantitative controls on bble 8-8. Arrears in Payments, Nigeria, 1981-87 (millions of dollars) Year Unoffidl Offtcial Total 1981 58 .. 58 1982 708 2,843 3,551 -1983 883 1,219 2,102 1984 624 424 1,048 1985 221 246 467 1986 - . 454 454 1987 . .. 1,395 1,395 Total 2,494 6,581 9,075 Zero or negligible. Sourec IMF (1988). Converted to dollars using dollarIsDn (special drawing rights) in IMF (various issues). 2)2 Mictael Gavin Figure 8-& Black Markc Premium for Nigerian Nair, 1970-86 500 ,400 30.0 -200 100 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 So.w Ccntral Bhn} of Niga data;, Uwl COrnqw Iarbow, 2986-87. The new military government decreed that it would not tolerate "cor- ruption. squandermania, misuse, and abuse of public office for self or group aggrandizement" (In*ernationa Cuirency Analysis 1989, p. 143). The military government banned political parties, instituted a curfew, and began to grapple with the task of reconstituting Nigeria's crippled economy. Act 4:Adjustment 1984 to the Present In its first two years the military government made substantial prog-- ress toward stabilizing the budget and the economy. The government began negotiations aimed at settling trade arrears accumulated duing- the previous three years, it slowed somewhat the rate of monetary- growth, and, most important, it reestablished control over the budget and sharply reduced the fiscal deficit (see table 8-9).- The government sharply reduced expenditure, especially invest- ment expenditure, from 1983 levels. During 1981-83, capital expendi- ture averaged slightly over N 5. billion a year; in 1984 this was: slashed to N 4 billion, and in 1985 investment spending was further reduced to less than N. 3.5 billion (Nigeria 1986). Despite a continually weaken- ing oil market, which drastically reduced oil sector revenue, the Nigerian government increased overall taxation from 10.5 percent of GDP in 1984 to roughly 16 percent -of coP in 1986, thus maintaining a fiscal deficit of roughly 2.5 to 4.0 percent of GDP during 1984-86. This 212 Michael Gavin Figure 8-8. Black Market Premium for Nigerian Naira, 1970-86 500 400 Soo J200 100 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 Sowre. Ccntral Bank of Nigeria data; KrM Cufnenq earbooh, 1986-87. The new military government decreed that it would not tolerate "cor- ruption, squandermania, misuse, and abuse of public office for self or group aggrandizement" (Iternational Currenc Analysis 1989, p. 143). The military government banned political parties, instituted a curfew, and began to grapple with the task of reconstituting Nigeria's crippled economy. Act 4:Adjustment, 1984 to the Present In its first two years the military government made substantial prog- ress toward stabilizing the budget and the economy. The government began negotiations aimed at settling trade arrears accumulated during the previous three years, it slowed somewhat the rate of monetary growth, and, most important, it reestablished control over the budget and sharply reduced the fiscal deficit (see table 8-9). The government sharply reduced expenditure, especially invest- ment expenditure, from 1983 levels. During 1981-83, capital expendi- ture averaged slightly over N 5 billion a year, in 1984 this was slashed to N 4 billion, and in 1985 investment spending was further reduced to less than N- 3.5 billion (Nigeria 1986). Despite a continually weaken- ing oil market, which drastically reduced oil sector revenue, the Nigerian government increased overall taxation from 10.5 percent of GDP in 1984 to roughly 16 percent of GOP in 1986, thus maintaining a fiscal deficit of roughly 2.5 to 4.0 percent of GDP during 1984-86. This Adjwuving to a h'rmsf 7Vado Shock: Nigeria, 1972-88 213 Ibblc 8-9. Federal Budget, Nigeria, 1984-87 (l)tCIEIKC oflalc1t;m'l) Ym.r lrrunnee IMInditurne" lialuicre 1984 10.5 14.7 -4.2 1985 11.4 13.9 -2,5 1986 15.6 19.1 -Y.5 1987 16.0 25.0 -8.9 a. I ncudesle Icllng miii uN rclpiy:cnts. Source; INII (varlou9s Issues). major shift toward budgetary balancc, combined with some inten- sification of foreign exchange controls, led to a dramatic improve- ment in the country's external balance. Moving from deficit into near balance in 1984, the current account in 1985 registered a surplus of more than $2.6 billi (see table 8-10). This occurred despite external interest obligations (much of the "other current account" category in table 8-10) that had by 1985 reached roughly $1.5 billion, roughly 10 percent of exports. This. improvement in external finances was achieved through a drastic compression of imports, which in turn reflected sharp reductions in public spending, and through reductions in private spending, as tax increases and reductions in domestic subsidies reduced private sector incomes. By 1986 the real value of imports in Nigeria was less than one-third the 1981 level, and in 1987 real imports amounted to less than 25 percent of the 1981 level. Indeed, real imports in those years averaged less than in 1973-74, before income from the first oil boom was-spent. Thus the Buhari regime must be credited with bringing budgetary and external accounts toward balance. As a later government was to argue, however, the Buhari regime's policy was one of "austerity without structural adjustment" (Nigeria 1986, p. 13). The forceful measures required to wean the private sector from its reliance on 'ILble 8-10. International Trade and Payments, Nigeria, 1984-88 (billions of dollars) OtJher Current Trade current aceount Net capital Overall )Yar balance accounti balance moveenits balance 1984 2.98 -2.87 0.11 -1.08 -0.97 1985 5.74 -3.12 2.62 -3.87 -1.25 1986 2.54 -2.17 0.37 -1.52 -1.15 1987 3.52 -3.59 -0.07 -2.69 -2.76 1988 2.42 -3.52 -1.10 -3.95 -5.05 a. Incudes external intcrest obligations. Source Iiw (various issues). 214 Michael Gavin the government requircd to wean the private sector from its reliance on the government and the economy as a whole from its reliance on the petroleum sector were not utken. In particular, the government refused to consider exchange rate depreciation, and the real exchange .rate continued to appreciate through 1984 and into 1985 (see figure 8-9). Although the real exchange rate remained at this inflated level, there could be little hope of developing a viable non-oil export sector. The austerity measures that were responsible for the improvement in the current account were also associated with immense reductions in domestic investmnent, the real value of which after 1983 was lower than in any year since 1971 (see figure 8-6). In short, although the govern- ment managed to bring some semblance of balance into the economy's external fina-nces, it did so at the clear expense of prospects for eco- nomic growth. In August 1985 another military coup took place, and Major General Babangida rose to power. Under Babangida economic policy under- went a drastic change of direction. His goverk..nent sought to diversify: the economy from its excessive reliance on the petroleum sector, to achieve fiscal and balance of payments equilibrium, to lay the basis for noninflationary economic growth, and to "lessen the dominance of unproductive investments in the public sector, improve the sector's efficiency, and intensify the growth potential of the private sector" (Nigeria 1986, p. 1). The Babangida regime's identification of market solutions to Nigeria's economnic woes marked a major departure from the approach Figue 8-9. Exchange Rate Depreciation, Nigena, 1971-88 Oflicial rate 6 Parallel rate 2 9 . -- I -I I I --- - I I I I I; I 1 I 1972 1974 1976 1978 1980 1982 1984 1986 1988 - S-W,Ccutr3 Bank of Nigeria dam; lUrd Gunny Iarboek, 1986-87. Adjustingiora Terms of 7'rdeShwockNigeria, 1972-88 215 .of previous governments, military and civilian alike. The earliest, most dramatic, and most fundamental change in policy was the acceptance that the market should determine the exchange rate. In late October 1986 the government abolished foreign exchange licensing for many activities and began to aucion foreign exchange, allowing traders to set the price. The first result was a.very substantial depreciation. Figure 8-7 shows that the official rate of exchange rapidly fell fromn $1.00 to roughly $0.25, approximately the black market rate during the previous two years. It stabilized.there until late 1988, when it began to decline even further. InJuly 1989 the naira traded at about $0.14. This devaluation of the exchange rate had a massive effect on the real exchange rate and therefore on production incentives. We saw that in 1984 the real exchange rate was grossly overvalued, at about double its 1976 value. By the end of 1988 the real effective exchange rate was only .30 percent of its 1976 value and was clearly lower in real terms than it had been at any time in Nigeria's recorded history.xCombined with this radical change in production incentives was a substantial liberalization of domestic markets and of the international trading regime. The com- modity boards that had so perniciously. insulated domestic producers from world prices were abolished as of the end of 1986, interest rates ' were liberalized and rose in real terms from their persistently negative levels into a range that would reward savers for their thrift, subsidiza- tion of petroleum products was ended, and the cumbersome bureau- cracy involved in enforcing the import quota machinery.was relaxed. It is too early to tell whether this dramatic change in direction will be successful or even whether it will survive. A hopeful sign exists: the dollar value of nonpetroleum exports rose significantly in 1987. The amounts of nonpetroleum exports remain tiny, however, both com- pared with petroleum exports and compared with their pre-1980 levels. Fonner policy mistakes have so eviscerated the non-oil.export sector, and the dominance of the state.in the economy.has become so entrenched, that diversification from oil and reorientation of the econ- omy toward private sector activity will require years of sustained com- mitmentto present policies. So far, the Nigerian authorities have shown no'sign of backing away from their. determination to prevent overvaluation by letting market forces determine the exchange rate. The budget deficit widened signifi- cantly, however, in 1987; only time will tell whether this reflects tempo- rary factors or a resurgence of past policies. Condusion The theoTetical section of this essay lays out a number of arguments why, in the aftermath of an improvement in the terms of trade, 216 Michael Gavin authorities ought to take steps to ensure a substantial rate of savings and, at least in the short to medium run, to invest a significant fraction of those savings in the form of foreign assets. One argument for such a policy is that the substantial uncertainty surrounding com- modity prices means that changes in the terms of trade may not last as long as expected. The accumulation of foreign assets or viable invest- ments in the export sector provide insurance against unexpected, but possibly adverse, movements in the terms of trade. A second argument for such a policy is that extremely rapid rates of: adjustment and extremely high rates of investment lead to costly mistakes and waste and may foster a confused environment in which fraud is less easily detected and therefore more prevalent. The accu- mulation of foreign assets can slow the rate of adjustment to a more manageable pace and reduce the incidence of costly miscalcuilations, bottlenecks, and, perhaps, corruption. The Nigerian authorities erred, first, in failing, to adopt such con- servative policies. Foreign assets were accumulated only in the imme- diate aftermath of the oil price disturbances; spending the revenue as rapidly as-it was accruing was physically impossible. Within a year-of both disturbances the Nigerian authorities had- brought expenditure into line with the higher level of income, and very soon thereafter the public sector and the economy as a whole became seriously overex- tended. In the first boom especially, much of this expenditure was in the form of investment. But the investments seem.to have had a very low rate of return; despite the high rates of investment recorded in the 1970s, real GDP in 1987- remained at about its 1974. level, and substantially below that level in per capita- terms.Some reasons'for this apparent low rate of return on investment may be found-in-the frenietic environment in which the investment was undertaken. Some may be found in the tendency for governments to invest in projects with high prestige or political payoff, but with little economic rationale. . . - - The second and far more serious mistake was the authorities' fail- ure to respond to the terms of trade declines that followed the-end of each boom. The crisis in 1978 was short-lived, because oil prices increased again in 1979 and 1980. But'the authorities were not res- cued from their mistakes in 1981-83. Adjustment to-a deterioration in' -the terms of trade involves reductions in domestic expenditure. In Nigeria, however, the authorities promoted excessive expenditure both by spending too much themselves and by subsidizing the private sector. Adjustment to a weakening in the terms of trade demands real- exchange-rate depreciation, but the authorities promoted real- exchange-rate appreciation by implementing highly inflationary mon- etary policies and then failing to adjust the nominal exchange rate. Adjustingto a Term ofadSlocNigeri 1972-88 217 The corrections to policy have managed to bring the Nigerian economy closer to equilibrium in its external payments, but coming as they did years after they were due, they did so only after an enormous foreign debt was accumulated, a debt that bedevils Nigeria's attempts to grow in the 1990s. This legacy of the past decade's policy mistakes is a vivid argument for adopting a cautious approach to disturbances in the terms of trade. Notes 1. Here we make no attempt to disentangle the separate effects of changes in price and quantity. If commodity markets are competitive, so that price equals marginal cost, then changes in the 'quantity of exports have (to a first order of approximation) no welfare effects, because the price at which the marginal units are sold exactly equals their marginal cost. Even at the rela- tively low prices in petroleum markets today, the price of oil is substantially above marginal cost for producers such as Nigeria. Because of these scarcity rents, changes in the quantity of oil exported have large welfare implications. 2. To be more precise, this distance is the first-order approximation of the change in real income caused by'a change in the terns of trade. Two second- order considertions make this measure only approximate: the fact that the c consumption basket may change in response to the change in relative prices and the fact that the composition of production . may change. We discuss changes in the composition of production later. 3. This is a natural assumption for a country like Nigeria, which is, so highly specialized in the export-of a single, primary commodity. There is, of course, some consumption of petroleum products within Nigeria, but it is a very small fraction of total production. 4. A detailed derivation is not necessary for the purposes of this paper. Interested readers arc referred to Gavin (1990), where the model underlying this discussion is laid out in substantially more detail. 5. A fairly large body of literature is available on the impact- of changes in the terms of trade on optimal rates of investment and the current account. For recent papers see Brock (1988), Murphy (1989), and Sen and Turnovsky (1989). 6. An early reference is Ldand (1968). Zeldes (1989) and Skinner (1988) give evidence on the empirical significance of precautionary savings. Not all utility functions generate a demand for precautionary savings. The precau- tionary savings motive is intuitively so compelling, however, that we implic- itly confine ourselves in this discussion to utility functions that are con- sistent with precautionary savings. 7. This should be thought of as a composite of all the non-oil traded ' goods that Nigeria purchases and sells. In particular, it includes exportable agricultural commodities that are produced in Nigeria and that 'were once such an important part of its foreign trade. 8. This remark is designed only to highlight the fiscal dimension of mone- tary policy. There is no implication that Nigernan monetary policy is deter- mined only, or even mostly, by the terms of trade. As we will see later, the- - Nigerian monetary authorities have shown themselves perfectly able to imple- 'ment.very expansionary monetary-policy when the terms of trade are highly favorable. 218 Miclael Gavin 9. There is some evidence that farmers in developing countries do distin- guish between permanent and transitory changes in income in the way sug- gested by the theory (see, for example, Paxson 1989 and Bevan, Collier, and Gunning 1989). At the samc timc it is well established that, cven in industrial econrbmies with well-developed capital markets, consumption is more sensi- tive to current income than is consistent with simple versions of the perma- nent income hypothcsis. 10. The naira was introduced injanuary 1973 and was set to equal the value of half a Nigerian pound, the previous currency unit. Before 1973 what I refer to as the naira exchange ratc is the naira equivalent of the Nigerian pound exchange rate. 11. This is used as a proxy for the dollar price of Nigerian imports. It is an imperfect proxy because Nigeria also imports from Europe. But movements in the relative price of US. and European exports are likely to be minor compared with the oil price changes that dominate the Nigerian terms of trade, so a more perfect measure would not materially affect the analysis. 12. In 1975, the government awarded public sector employees a very large wage increase. This wage increase had major budgetary effects,'amounting to almot a quarter of the preceding year's total spending, and contributed sub- stantially to the high inflation Nigeria experienced in the latc 1970s (see Bienen 1983). 13. In other respects, 1976 provides a good benchmark against which to evaluate the 1984 real exchange rate. The current account was in balance in 1976, and oil revenues in 1976 and 1984 were, in real terms, roughly compara- ble (see table 8-3). 14. There are no official sources of information on this market. The data reported in this paper are reported in World Currency Yearbook, a standard, if not official, source. Selected Bibliography Bevan, David, Paul Collier, andJan Gunning 1989."'The Kenyan Coffee Bo0om of 1976-79:' Oxford.University, Department of Economics, Oxford, UK. Bienen, Heny 1983. "Oil Revenues and Policy Choice in Nigeria' World Bank Staff Working Paper 592. Washington, D.C.' Brock, Philip L 1988. "Investment, the Current Account, and the Relative Price of Nontraded Goods in a Small, Open Economy." Journal of Interna- tionalEconomics 24(3):235-53. Dixit, Avinash K. 1989. "Intersectoral Capital Reallocation under Price Uncer- tainty"Joumal of lnternatonal Economics 26(3)309-25. Dornbusch, Rudiger. 1983. "Real Interest Rates, Home Goods, and Optimal External Borrowing: Journal of litical Enomy 91(i):141-53. Gavin, Michael. 1988. "Income Effects of Adjustment to a Terms of Trade Disturbance: Asynmetries in the Harberger-Laursen-Metzler Relation?" Columbia University, Department of Economics, New York. - . 1990. "Structural Adjustment to a Terms of Trade Disturbance: The Role of Relative Prices2Journal oflntemational Economic 28(314):217-43. iMF (International Monetary Fund). 1988. Balance afPayments Statistics; Yearbook, 1988. Washington, DLC. - - . Various issues. International Finantcial Statistics. Washington, D.C.- Adjusting to a Terms of Trade Shock- Nigeria, 1972-88 219 International Currency Analysis, Inc. 1989, World Currenc Yearbook, 1986-87. New York. Kirk-Grcenc, Anthony, and Douglas Rimmer. 1981. Nigeria since 1970: A Pbliti- cal and Economic Outline. New York: Africana Publishing Company. Leland, Hayne E. 1968. "Saving and Uncertainty: The Precautionary Demand for Saving." QuarterlyJdurnal of Economias 82(3):465-75. McDonald, Robert L, and Daniel I Siegel. 1986, "The Value of Waiting to Invest" Quarterlyjoumnal ofEconomics 101(4):707-27. Murphy, Robert. 1989. "The Terms of Trade, Investment, and the Current Account:" Discussion paper. Boston College, Department of Economics, Boston, Mass. Mussa, Michael L. 1978. "Dynamic Adjustment in the Heckscher-Ohlin- Samuelson Model rJournal of Politial Economy 86(5):775-91. Nigeria, Federal Republic of. 1986.. 'Information Memorandum' Central Bank of Niigeria and Ministry of Finance, Lagos, November. Paxson, Christine. 1989. "Transitory Trde Shocks and Saving: The Case of Thailand:' Discussion paper. Princeton University, Department of Eco- nomics, Princeton, N.J. Pinto, Brian. 1986. "Nigeria during and after the Oil Boom: A Policy Compar- ison with Indonesia: World Bank, Country Policy Department, Trade and Adjustment Policy Division, Washington, D.C. Sen, Partha, and Stephen F. Turnovsky. 1989. "Deterioration of-the Terms of Trade and Capital Accumulation: A Reexamination of the laursen-Metzler Effect"CJouma ofInternational Economics 26(3):227-70. Skinner, Jonathan. 1988. "Risky Income, Life Cycle Consumption, and Pre- cautionary Savings:'Journal of Monetay Economics 22(2):237-55. World Bank. 1989a. World Debt Tabler External Debt of Delping Cut. 1988-89 ed. 2 vols. Washington, D.C. 1989b. World Tables. Baltimore, Md.:Johns Hopkins University Press. Zeldes, Stephen P. 1989. "Optimal -Consumption with Stochastic Income:: Deviations from Certainty Equivalence" Quarterly Journal of Economics 104(2):275-98. 9:: Bolivian Trade and Devekopmen4 1952-87 m Juan-Antonio Morales BoLivIA is the most extreme example;of the Latin American economic crisis of the 1980s. During 1982-86 the growth rates of GDP were negative every year, and in 1984-85 Bolivia suffered from hyperinfla- ion. Inflation has now been successfully stabilized, but a host of inter- nal and external factors are hindering economic recovery, which is slower than expected. This essay provides-an overview of Bolivia's economic policies.. Bolivia is a small, open economy dependent on a few export prod- ucts and on a small number of large trade partners. The perfonnance of exports and normal access to foreign financing are crucial to Boliv- ia's welfare. Another characteristic of Bolivia's economy is the size of its public sector. Bolivia's governiments have been so overburdened with economic functions that they have on the whole performed poorly. The cost of the public sector and. its regulations has uniduly affected investment rates and. the patterns of foreign trade and indebtedness. The international debt crisis of the early 1980s put an. especially pronounced strain on the Bolivian economy. Exacerbating the effiec of increased interest payments on international debt, export earn- ings collapsed in 1985-86. EIternal factors were also involved. To many Bolivian observers the hyperinflation appeared to be more than a transitory monetary disarray. To them. it constituted an indictment of the development model Bolivia had followed for the preceding thirty years and the culmination of accumulated eco- nomic policy mistakes. Economic Historm 1952-87 Table 9-1 and figure 9-1 depict the main characteristics of the 1957- 87 period with regard to growth and overall economic performance. 220 T3ble 9-1. Key Indicators of Development, Bolivia, 1953-87 Indicator 1953-61 1962-71 1972-78 1979-81 198285 1986-87 Average annualgrowth rates GDP -0.4 5.9 5.4 0.9 -3.0 -0.8 Population 2.2 2.4 2.6 2.8 2.8 2.8 Per capita GDP, -2.6 3.5 2.7 -1.8 -5.6 -3.5 Consumer prices 56.6 5.1 17.4 32.6 982.7 107.1 Exports .-9.5 11.6 19.5 13.2 -9.1 -14.5 Import capacityb -5.9 5.5' 8.2 -4.3 -14.0 4.5 : Average ratios t%3 > 0 - |'Investment/GDP 15.8 19.0 16.5 14.0 9.1 7.2 Gross domestic savingslcDr . 6.1 11.8 18.4 13.3 10.7 ExportsluDP 13.3. 19.7 21.9 18.8 15.3 12.9 Exports of tin and natural gasl merchandise exports . 63.5 68.7 56.5 65.9 81.7 73.6 Current account deficitlcw -- . 3.1 2.3 7.8 4.8 10.7 - Public external debtIGDP . 36.1 45.2 63.0 80.0 93.2 Public external debtlexports 199.3 194.4 229.5 361.5 516.1c -Not available.. a. Preliminary. 1. Defined as the sum of export receipts and long-term foreign capital inflows deflated by Import price index. . c. Value for 1986.. Source: IMF (1987); UDAPE (1987); Ac>ha and Huarachi (1988); World Bank (1988); Central Bank of Bolivia (various issues). 222 Juan-Antonio Morales Figure 9-1. Annual Growth Rttes oGDP, Bolivia, 1953-87 8 6 4 2 A -6. -B.. -1 I '''i I -I II I iI, I I ' I -I I I 1 t I I I I Iii i i I '1- I i- i :- i - 1955 1960 1965. 1970 1975 1980 1985 Sosc CentraI Bank of Bolivia (various issues). The subperiods in table 9-1 are somewhat arbitrary but reveal some dominant patterns. The subperiod 1953-61 is marked, by low eco- nomic growth rates and high inflation. In 1962 the Bolivian economy. initiated a sustained takeoff: high rates of growth of CDP and low inflation rates characterize both the 1962-71 and 1972-78 subperiods. The latter subperiod was also chairacterized, however, by a rapid accu- mulation of external debt The subperiod 1978-81 was defined. by extreme political instability and by the beginning of difficulties in servicing the foreign debt. The economic crisis reached its height during 1982-86. During the final subperiod, 1986-87, Bolivia made a successful attempt at stabilization and fought hard to achieve eco- nomic recovery. What made possible the sustained growth during the relatively long period 1962-78? Econometric estimates by.Morales (1988) support the . view that investment rates, growth in the capacity to import, and polit- ical and macroeconomic stability were the main factors behind the GDP growth rates. The capacity to import is given by the sum of export proceeds and foreign capitaI inflows, computed in constant prices. Ram (1987) finds that export performance in itself was a significative Bolivian Trade and Dewlopmen4 195247 223 explanatory factor of Bolivian economic growth; however, his results are difficult to replicate. Table 9-2 shows the main changes in the structure of production between 1952 and 1987. The most salient aspect of the table is the increase in the share of services II in GDP. Services II consists princi- pally of the production of nontradables in foreign trade. Notice that production in the primary sector (agriculture, mining, and hydrocar- bons) was reduced as a share of GDP until around 1977. During the crisis years, this share increased somewhat. The manufacturing sector was greatly affected by the crisis of the 1980s, as can be inferred from the strong fall in its share of GDP. Geography is an important determinant of the pattern of Bolivia's foreign trade. A large share of the population is concentrated in the highlands. The country has no direct access to the sea and only a few waterways free from natural obstacles. Roads and railways are limited because.the topography is such that building costs are steep. Thus transportation is very expensive' This implies that exports have to be of high value added .per unit of weight. - Bolivia's export base has been heavily.dependent on a handful of, commodities: tin and other minerals, petroleum, and, more recently, .natural gas. In the mid-1970s illegal exports of cocaine and its deriva- tives made a strong appearance; they have important distorting eff.xcts on -the legal economy. Table 9-1 shows that exports of tin.and, since* 1973, of natural gas accounted for more than 55 percent of total (legal) merchandise exports during most of the period. With the onset of the 'Mible 9-2. Sectomal Composition of Gflp, Bolivia, Selected Years, 1952-87 ( (percent) Sector 1952 1962 1972 - 1977 7982 1985" 1987" Primary 46.5 39.9 29.1 25.8 352 32.2 . 33.3 Manufacturing 14.2 14.1 14.6 15.9 12.0 9.8 10.8 Services lb 4.6 5.4 5.4 5.9 4.0 4.2 3.7 Services IIc 34.7 40.6 .51.0 .52.5 462 50.7 -50.5 Indirect taxesd - - 2.6 3.1 1.6 GDrt 100.0 100.0 100.0 100.0. 100.0 100.0 100.0 Not available. Note: Percentages are computed from basic data in real terms. a. Preliminary. b. Electricity, gas and water, construction. c. Commerce, transportation and storage, communications, finance, general govern- ment, other sevices.e d. Included in each sector-for 1952,1962, 1972,1977. e. Basic data on GDP are at market prices. Source Bolivia (1970); Central Bank of Bolivia (various issues). 224 Juan-Antonio Morales crisis in the late.1970s, tin and natural gas made up an even greater percentage of exports. Geographical concentration also increased, as can be observed in table 9-3. Note that in 1985, as much as 56 percent of Bolivia's exports went to Argentina. With this limited export base, Bolivia's vulnerability to international market conditions is not surprising. Despite the crucial role of exports in the Bolivian economy, their long-run performance has been poor. Betwecn 1952 and 1985 the volume of exports grew at the meager average annual rate of 1.5 per- cent. Attempts at import substitution were short4lived and had little impact. The Bolivian strategy appears to have been geared more toward terms of trade improvements than to a long-run expansion of - the production of tradables. Growth depended to a large extent on improvements in the.terms of trade. In 1986, however, the terms of trade took a nosedive (see table 9-4). ..During 1952-87, exports and imports made up a relatively high. percentage of the national economy. Foreign trade-GDP ratios can be observed in table 9-5. The ratios show a generally stable trend, except .over short periods associated -with severe disruptions in the economy. An important feature of Bolivia's development from 1952 onward has been its continued heavy dependence on foreign capital and- transfers. During the 1960s the. flow of foreign funds took essentially two forms: bilateral or multilateral foreign aid through the official development banks to finance social overhead projects and subsidize small loans to the private sector; and direct foreign investment, heavily concentrated in the hydrocarbons sector. In the 1970s Bolivia gained access to credits from the international commercial banks, as ¶kble 9-3. Exports from Bolivia, by Destination, Selected Years, 1960-85 (percent) Destination 1960 1970 1980 1985 Argentina 5.7 4.7 '23.7 55.9 Brazil 5.9 04 3.5. 0.7 Chile 0.4 1.1 4.5 1.6 Peru 0.2 2.7 3.1 1.9 Other LAFrAILAIAa countries 0.0 0.0 1.9 O- 06 Total LAFrALA.AM countries 12.2 8.9 36.7 60.6 Rest of the world 87.8 91.1 63.3 39.4 Total 100.0 100.0 100.0 100.0 a. Latin American Free Trade AssociationLacin American Integration Association. The former became the latter in the late 1970s. Sourme: Central Bank of Bolivia (various issues). Bolivian Rade and Dwezpw/ent, 7952-87 225 Ibble 94. Trnns of MBade Indexes, Bolivia, 1952-87 (1980 - 100) Year &Eport unit valsues mport unit values -h7ns of trade 1952 22.9 28.1 81.5 1953 21.5 22.7 94.7 1954 20.3 22.5 90.4 1955 19.0 22.5 84.6 1956 19.5 24.0 81.1 1957 16.9 20.9 80.7 1958 16.4 19.5 83.9. 1959 16.7 21.1 79.2 1960 . 19.9 17.8 111.6 1961 20.5 19.9 103.1 1962 21.1 19.7 107.4. 1963 21.1 20.7 102.1 1964 21.8 27.7 78.5 1965 21.6. 32.9 65.5 1966 20.7 30.0 69.0 1967 21.3 29.0 73.6 '1968 22.0 28.6 76.9 1969 22.4 30.8 72.6 1970 23.2 32.3 .72.0 1971 22.5 32.9 . 68.4 1972 25.2 34.9 72.1 1973 25.5 40.7 62.8 1974 50.7 49.9 101.5 1975 45.5 . 56.8 80.0 1976 48.3 59.1 . 81.8 1977 56.7 - 65.9 86.1 1978 63.1 73.0 86.5 1979 75.0 88.3 84.9 1980 100.0 100.0 100.0 1931 96.7 101.9 94.9 1982 93.3 99.3 94.0 1983 96.2 , 95.8 100.3 1984 99.1 - 98.5 100.6 19855 98.8 93.5 - 105.7 19861 76.3 88.9 85.8 1987a- 79.3 92.0 86.2 a. Preliminary. Sourcc pAL (1987, various years).. 226 Juan-Antonio Morales lible 9L5. 'lade as a Percentage of GDP, Bolivia, 1952-87 Annual Annual Annual aven7ge averag 7btal ap Year Exports overperiod imports over period trae ovperiod 1952 17.2 18.6 35.8 1953 5.9 .7.9 15.8 1954 4A . 5.3 9.6 1955. 4.8 8.1 5.8 9.4 10.6 17.5 1956 13.4 16.0 29.3 1957 22.5 30.8 53.3 1958 16.0 25.4 41.4 1959 20A 27.1 47.5 1960 .16.2 17.7 23.9 24.6 40.1 42.3 1961 16A 23.3 39.7 1962 16A 24.9 41.4 1963 17.4 26.5 43.9 1964 20.7 . 23.5 44.2 1965 21.5 18.5 26.6 25.0 -48.1 43.4. 1966 21.9 26.3 48.2 1967 22A 26.0 48.4 1968 19.7 24.0 45.7 1969 19.3 . 23.6 42.9 1970 20.2 20.7 20.3 24.0 40.5 44.7 1971 17.3 20.0 37.3 1972 17.2 19.9 -37.1 1973 22.6 23.5 46.0 1974 28.7. 22.2 50.9 1975 213 21.4 26.9 22.5 48.2 43.9 1976 22.5. 24.7 ;47.2 1977 225 ; 25:1 47.3 1978 18.6 25.3 43.9 1979 19.3 25.2 44.6- 1980 25.6 -21.7 20.2 24.1 45.8 45.8 1981 22A : 21.7. 44.1- 1982a 2153 14.8 36.1 .1983S 20.1 15.7 35.8 1984a 19.0. 12.9 31.9 1985" 16.5 19.9 14.6 15.9 31.1 35.8 1986P. 14.5 18.1 33.3 1987A 113 12.9 19.1 19.0 30.4 . 31.9 Note: Percentages arc derived from national accounts data in current prices for the years .1952-81; percentages for 1982-87 are estimated from balance of payments data. a. Preliminary estimates. Source: For 1952-81, Central Bank of Bolivia (various issues); for 1982-87, UDAPE (1987), Afcha and Huarachi (1988). Bolivian 71ak and Dawloenpmt, 19J2-87 227 did oLher countries in the region. The burden of the heavy indebted- ncss Bolivia incurred in the 1970s continues to be a major obstacle to economic recovery. The Revolution of 1952 and Its Aftermath The revolution of 1952, led by Dr. Vfctor Paz-Estenssoro, is the main landmark in Bolivia's recent history. In its aftermath the government nationalized the threc largest mining firms and undertook extensive agrarian reforms. Simultaneously, it implemented a policy of import substitution and production diversification. The need to lock in the wealth and income changes of the revolution gave birth to state cap- italism, which would prevail until 1985. The political and economic changes of the early 1950s were accompanied by the start of a popula- tion movement from the highlands to the eastern lowlands, particu- larly the Santa Cruz region. The move to the east brought about important shifts in economic and political power. The period 1952-56 was one of high social mobilization, huge dis- ruptions in production, and inflation that reached 178 percent by 1956. As inflation worsened and foreign exchange and staple foods * became scarce, ihe government gave top priority to stabilization. The stabilization plan, announced by the government in December 1956, indeed ended the high inflation but also led to significant modifica- tions in the development policies of the revolution. After inflation had stabilized, the government emphasized export- led.growth, and the country's dependence on foreign capital inflows increased. Central policies from 1956 to 1964 were the recovery of the : tin mines and new investments in the petroleum sector. Policies of import substitution were abandoned, as were all remnants of popular mobilization. There was an inflow of foreign aid, which largely financed physical infrastructure projects. The foreign aid, especially the resources of the Alliance for Frogress program (launched by Presi- dent Kennedy in 1962), had a significant impact on investment rates. These in turn affected GDP growth rates, and beginning in 1962 the economy started to recover. After twelve years of civilian rule that had started with the revolution of 1952, General Rent Barrientos led a military coup that deposed. President Paz-Estenssoro. Barrientos generally upheld the economic policies implemented after the stabilization program of December 1956. But there were some changes. First, a resurgence of the private sector was made possible through increased inflows of foreign direct investment in the petroleum sector and the return of many Bolivian entrepreneurs who had earlier fled-the revolution. Second, substantial' credits and grants were made available to the private sector. 228 Juan-Antonio Morakes 77e Populist Interlude Barrientos died in a helicopter crash in 1969 and was succeeded by his civ;;ian vice president, Luis-Adolfo Siles-Salinas. Shortly afterward, General Alfredo Ovando overthrew Siles-Salinas in a military coup. With Ovando and later with General Juan Jos6 Torrez, who succeeded Ovando after som'e confusing coups and countercoups, Bolivia entered an era of open populism. Ovando unwisely nationalized the Bolivian Gulf Oil Corporation and started to build the controversial state-owned tin smelters. Torrez followed an even more populist course, trying to co-opt the labor movement into government. Real wages reached an all-time high. The government also made a deter- mined effort to collect more income taxes, but with relatively little success. Banzer's Bonanza Years Torrez lasted less than a year and was deposed in a coup led by the conservative General Hugo Banzer in August 1971. Banzer ruled for. seven years, a long time by Bolivian standards. He presided over an era of high economic growth, with an average annual GDP growth rate of 5A percent. Two main factors led to the high rates of growth and .investment. First,'a significant improvement in the terms of trade from 1973 onward caused export prices to double in 1974. Second,. Bolivia's access to foreign capital;markets increased. In addition, the prolonged period of political stability was undoubtedly a major force behind the economy's good performance. The Banzer boom was tem- porary however. It was based on a short upsurge of export prices and.: high foreign indebtedness. Policymakers did not appreciate the transi-. tory nature of the high prices and favorable interest rates. ''The growth rate became smaller after: 1978 and turned negative in 1981. 'his was the result of political uncertainty and reduced effi- ciency in the use of internal and external resources, which started in the last years of tile Banzer regime. 7he Years ofPolitical Chaos. Banzer failed in his efforts to build a large political;base. The pres- sures of the Carter administration to restore democracy in Bolivia and the perception of impending economic troubles, whose first symp- toms had already appeared by the end of 1977, made Banzer abandon -his stated purpose to stay in power until 1980. He called general elec- tions in 1978; however, a clear winner did -not emerge, and the elections were annulled. A period of political chaos -ensued, with Bolivian 7rade and Development, 1952-87 229 inconclusive elections, interim civilian presidents, and military coups and countercoups. The political turmoil undermined the govern- ment's capacity to recognize the economic crisis that was steadily building up and to take appropriate action.' Bolivia's fortunes reached a low ebb with the military government of Luis Garcia-Meza in 1980-81, when Bolivia was cut off from international recognition and financial markets. The Hyperinflation The military government of the second half of.1982, unable to cope with the deteriorated economy and the growing pressures for a return to democracy, decided to reconvene the congress elected in 1980. Acting as an electoral college, the congress elected as president Her- nmn Siles-Zuazo, who had received the most votes in the 1980 elec- tions. Siles-Zuazo inherited an economy that was already severely strained. Indeed, the economy had experienced a negative rate of growth in 1982, and.the avenage annual inflation rate in the year preceding Siles's inauguration was already 170 percent. The start of Siles's presidency coincided with the onset of the international debt crisis. One effect of this crisis was that net resource transfers abroad stood at 2.4 percent Of GDP in 1982. Siles-Zuazo was completely unable'. to arbitrate between the demands of rebuilding a crumbling economy and the demands of his electorate, who wanted higher real wages and higher employment and investment in the public sector ..In addition, he faced opposition to his policies in a congress that he did not con- trol. His policies and his timid (and sometimes belated)'attempts at stabilization, if anything, worsened an already difficult situation. Dur- ing his presidency, inflation jumped from the hundreds to the thou- sands: from April 1984 to August 1985, prices increased by a factor of 623, the average increase being 50 percent a month. Along with the hyperinflation, real per capita GDP declined by'some 23 percent between 1982 and 1985. Shortages of food and other essen- tial goods and services were everyday occurrences,, strikes and work stoppages were frequent, and several threats to democracy surfaced.; The hyperinflation and the economic crisis were as much the result of difficult external conditions, coupled with poor domestic economic design, as of political weakness. Siles-Zuazo was forced to call early elections., Elections took place in July 1985,.and former president Victor Paz-- Estenssoto was. elected. Three weeks after his .inauguration, he unveiled a package of stabilization measures and policy reforms. The stabilization package halted inflation almost immediately,-and the pol- icy reforms set the stage. for significant changes in Bolivia's produc-: 230 Juan-Antonio Morales tion and foreign trade. For more detailed information about this period of hyperinflation and stabilization, see Sachs (1986, 1987a, 1987b); Morales (1987a, 1987b, 1987c). In the last quarter of 1985 international tin prices dropped dramati- cally, and by the end of 1985 the price of tin was half what it had been four months earlier. To compound matters, in the first quarter of 1986 oil prices dropped, which affected the price of Bolivia's other main export, natural gas. In 1986 export prices fell 23 percent, and the decrease in terms of trade was 19 percent. The difficult external con- ditions continued throughout 1987, hampering the prospects for a strong economic recovery. The Public Sector Bolivia's development between 1952 and 1987 cannot be fully under- stood withotut examining its public sector. Even after the reforms of *1985, the share of the public sector in total investment is probably the largest of any market economy of the Westem Hemisphere (see table 9-6). Political scientists such as Malloy (1970) 'view-the' heavy public sector investment as a manifestation of state capitalism. State capitalism has a number of features. First, and most important, the public sector performs the bulk of capital accumulation, both in social infrstruc-' ture and through state enterprises. Second, the governmentumanipu- lates the system of incentives and penalties to the private sector. Third, there is a large bureaucracy; which frequently pursues its :own ends. -lhble 9-6. Main Characteristics of the Public Sector, Bolivi 1958-845 (percent) Categor . - 1958-61 1961-71 1972-78 1979-81 1982-65 Share of public. investmentt in total . investment 42.8 54.7 51.0 58.2 49.9 Share of foreign trade taxes in central government revenues 51.4". 61.2. 60.9; .55A. Consolidated public sector deficit (share of our) - _ 7.9 8.4 17.5 - Not available. Notc Figures are averages over each period. a. Over period 1964-71. b. Over period 1974-78. Soxrce Morales (1982); World Bank (1983); xcF.(1986); Morales and Sachs (1987). Bolivian 7Tide and Developmenl4 1952-87 231 Fourth, the state often develops a symbiotic relationship with favored segments of the private sector to carry out a particular distributional agenda. The development plans prepared between 1952 and 1985 contain the philosophy underlying the Bolivian state capitalism model. The most influential development plan was the Socioeconomic Strategy for National Development (ssND), unveiled during General Ovando's regime in 1970. In the SSND the guiding principles of the state capital- ism model were clearly spelled out. Investment in sectors defined as strategic was reserved for-state enterprises. The SSND included such a restrictive system of incentives and penalties that it hampered the normal development of the private sector. In addition, the SSND emphasized natural-resource-based industrialization, with heavy pub- lic investment. The SSND took the position that only that type of indus- trialization could generate the big surpluses (or profit margins) to give the economy the big push it needed. Although President Banzer aban-, doned the SSND in 1972, the SSND had a powerful influence on Bolivian policymakers that went well beyond 1972. Its nationalistic approach appealed both to the Bolivian left and to the right-wing officers in the military. State Enterprises An examination of the government's role in the economy is incom- plete without a discussion, of the state enterprises.-Since 1952 the export activities of the mirning company, Corporaci6n Minera de Bo- livia (CoCMIBOL), and the petroleum company, Yacimientos Petrolfferos ' Fiscales Bolivianos (YPrF), have provided most of Bolivia's foreign exchange. This has given the two companies a privileged status that is frequently abused by both management and workers. To be fair, how-_, ever, these crucial export companies have also suffered from the mis- management of macroeconomic policy, especially from grossly over- valued exchange rates. From 1962 to 1982 the largest state enterprises generated small surpluses on their current accounts, but these savings were too small to finance their capital expenditures. During the high inflation years.. of 1982-85, the current account deficits of.the consolidated public sector became important for two reasons: the interest servicing of the large foreign debt and. the growing employment (and the resultant wage bill) in the public sector. If the public enterprises' investment record was dismal, their* errployment policies were litte better until recently. Employment in the public enterprises grew at an annual rate of 4.6 percent between 1970 and 1982, a period of.strong military presence in government.. 232 Juan-Antonio Morales .The military typically tried to gain popular support through. patronage. The return to democracy in 1982 did not reduce the rate of job growth in the most important public enterprises. Overstaffing and completely misaligned prices for the public sector were the main causes of the big deficits of virtually all state enterprises during the crisis years, 1982-85. Govenmnwnt Financing One striking fact about government financing in Bolivia is the nar- rowness of the tax base. Most taxes are related to the performance of the foreign sector: they are either export taxes, import tariffs, or "roy- alties" on petroleum and mineral products,' which are largely exported. Thus, the volume of taxes collected is closely related to what happens in the vulnerable export sector (table 9-6 shows how depen- dent tax revenues are on foreign trade).'Such a narrow tax base can result in punitive tax rates for the few taxpaying activities. Punitive taxation of the mining sector in the past and of the petroleum sector today has long-run costs. The narrowness of the tax base and the extreme difficulties, politi- cal and other, associated with the need to finance growing claims for a more favorable income distribution partly explain the growth of the public sector. During 1952-85 the search for tax revenues became tantamount to a search for investment opportunities for the public sector. Successive governments hoped that they could use the profits generated by public sector investments to pay for more social pro- grams and development projects. The easy access to foreign loans in the- 1970s permitted the govemment to postpone the necessary reforms to'broaden the tax base. Had tax reform been carried out in that period, Boliv;a could have better withstood the 1982 shock of sudden severance from foreign credits. -ndebted Industrialization and Other Policies -Frieden (1981) defines indebted industrialization as the strategy of rapid growth takeoff based on foreign financing of large-scale public investment projects. (The strategy is similar,to the state capitalism model.) For most of the period 1952 to 1985, Bolivian governments- followed this strategy. The overall deficits of the consolidated public sector were financed with foreign savings, either aid or loans, and to a lesser extent by the forced savings produced by inflation. During the periods of great disruption, the latter form of financing predominated over the former.' Bolivian 7Tade and Development, 1952-87 233 The foreign financing of investments in state-owned enterprises started with the rehabilitation of the state-owned mining enterprise COMIBOL in the early 1960s. In the so-called Triangular Plan, Lhe gov- ernment obtained credits from the U.S. government, the Inter- American Development Bank, and a consortium of Gerrman banks. Two elements were significant in this deal: first, the heavy foreign financing of a state enterprise and, second, the use of an official US.. 'loan for a nationalized enterprise. The Triangular Plan to rehabilitate COMIBOL set a precedent for continued use of foreign financing of investment in state-owned firms. With the growth in world liquidity in the 1970s, the Bolivian govern- ment hoped that it could find development funds in the expanding international loan market. With access to large-scale loans from com- mercial banks, the process of indebted industrialization accelerated. In the second half of the 1970s the capacity of the tin smelters of the Empresa INacional de Fundiciones (ENAF) wias significantly expanded; mills were built by COMIBOL to concentrate low-grade ores as a precur- sor to the smelting process; a huge polymetallurgic mill was built in Karachipampa, in southern Bolivia; the oil refining installations of the. state petroleum company YPFB were increased; and new factories were expanded or built to process agricultural products such as sugar, soya, and milk. Foreign financing made up a substantial proportion of these. public investments. In some cases, the government contracted debts with foreign com- mercial banks to compensate foreign companies nationalized in 1969 and 1971. It is quite-ironic that once more, foreign savings contributed to the expansion of the public sector. The tin smelting company ENAF and the polymetallurgic smelter Complejo Minero Karachipampa (cMu) exemplify the troubles of pub- lic investment financed with foreign indebtedness. The creation of ENAF responded to a long-held Bolivian wish to reduce the transporta- tion costs and dependence on foreign oligopolistic smelting firms for. its main export, tin. But of course domestic refining could only make economic sense if the savings in transport and foreign refining costs per unit were greater than the expenditure on domestic smelting per unit of final output. Unfortunately, because of. technical problems, related in part to the low metal content of the ores and to poor design and management, smelting costs went well beyond expectations. The CMK case is even sadder. The project was based on the assump- tion that a sufficiedt supply of ores would be available, which turned out not to be the case. An ex post analysis revealed that the strategy of pushing the creation of value added on known mineral deposits at the expense of investing in exploration and development of new mines : was misguided.' 234 Juan-Antonio Morcies Public debt also served to finance the undertakings of the private sector. The private sector borrowed heavily abroad, with guarantees by the government. Also, loans contracted directly by the government were intended for and channeled to the private sector. The loans intermediated by the government to the private sector often con- tained a high subsidy element, and, worse, many of them were never repaid. In addition, many foreign loans to the private sector that were guaranteed by the government went unpaid, and the debt was shifted to the guarantor. The diversion of loans obtained from the govern- ment (or guaranteed by the government) from their intended invest- ment uses was common. Thus, we can assume that the foreign debt financed more consumption (and accumulation of assets abroad) than domestic capital formation. Consumption loans had to be repaid by a drop in consumption. Table 9-7 traces the- evolution of Bolivia's external debt, emphasiz- ing the public external debt. By 1982 Bolivia had accumulated a large foreign debt, amounting to 71.2 percent of the GDP. The debt was mostly incurred by the public sector or by the private sector with public guarantees. Given the interest rates in the mid-1970s, a growth strategy based on access to foreign loans was not unreasonable. The problem was not indebtedness per se but, as mentioned earlier, the extremely inefficient use of those external resources. By- 1980 Bolivia was already finding it hard to make principal repayments. The severe political problems of that time aggravated the situation. In 1982 the. drastic reduction in foreign inflows precipitated a severe domestic crisis. On average, between 1982 and 1985 Bolivia transferred in net terms around 3.5 percent of its CDP to foreign creditors. The overextension of the public sector in face of the growing financing difficulties once foreign inflows had dried up reached a climax during the crisis years of 1982-85. The appearance of hyper- inflation in those years can largely be attributed to this fiscal crisis and the attendant debt problems. Debt continued to accumulate rapidly, notwithstanding the lack of new commitments and disbursements. In other words, between. 1982 and 1985, debt accumulated without new resource inflows; indeed, resource inflows were negative in net terms. The debt growth resulted from arrears and the need -to resort to (involuntary) distress. financing. By the end of 1987, after a valiant struggle against inflation, Bolivia had started on the road to economic recovery. Today the country still faces formidable obstacles, however, such as the huge external debt. The overborrowing of the 1970s and, especially, the misuse of foreign credits are important causes of Bolivia's present difficulties. bble 9-7. External Debt, Bolivia, Selected Years, 1962-86 Debt 1962 197) 1978 1982 7985 1986a Total accerual debt --- 3,169.0 4,143.2 4,619.1 Long-term debt (millions of dollars) - 534.2 1,671.9 2,897.9 .3,744.9 4,077.6 Public and publicly guaranteed debt (millions of dollars) 174.0 527.0 1,649.0 2,769.1 3,189.9 3,522.6 Private nonguaranteed debt (millions of dollars) -7.2 22.9 128.8 555.0 555.0 Short-term debt (millions of dollars) --91.5 185.0 347.0 897.0 Public external debtb (peircentage of GDP) 35.5 42.8 43.7 71.2 109.6 122.2 Public external debtb (percentage of exports) 251.0 260.5 233.6 301.5 432.6 516.1 Public external debt scrviCeb 19.1 31.0 355.0 287.1 242.1 160.8 Interest payments (Millions of dollars) 5.5 9.0 84.0 181.0 100.0 86.7 rrincipal repayments (millions of dollars) 13.6 220 210 106.1 142.1 74.1 'Thtal debt service (as percentage of exports)c 27.6 11.9 49.6 31.3 32.8 23.6 Total debt service (spercentage of Grip) 3.9 2.5 9.4 7.4 6.4 4.3 Memorandum items Net resource transfers of public sector (millions of dollars) -65.0 17.0 -92.7 -131.9 138.4 Net resource transfers of public sector (as percentage Of GDP) -5.3 0.5 -2A -3.5 3.7 .Gross domestic product (GDP: Millionis of current dollars)d 490.2 1,232.5377. 3897 3,15 3795 -Not avalab a. Preliminary. b. Public and publicly guaran'teed debt. c. Goods and services. d. Derived from5P v'alues in current Bolivian pesos divided by a purchasing power parity (inp) exchangeTrare. Source: Inter-American Development Bank (1983) for 1960-78; World Bank (1987) for 1982-86. 236 Juan-Antanio Morales Export Policies Until 1982 the most important and systematic export policy was Boliv- ia's adherence to the international tin agreements (rrAs). During 1965- 80,-Bolivia also signed agreements for tungsten and antimony. Bolivia participated in all the ITAS except the last one signed in 1982, because the government thought that it could ride on the back of the agreement without actually signing it-that is, that it could benefit from the above-equilibrium prices without incurring the costs, just as Brazil was doing.2 As a high-cost producer, Bolivia systematically lobbied for a joint policy of high intervention prices. The other countries did not always follow Bolivia's demands for higher prices, because they feared that the upkeep of the buffer stock, the main instrument of price stabiliza- tion, would be too costly. More important, they feared that a long-run policy of above-equilibrium prices would backfire. Time proved them right. In the 1970s, in addition to the policy of price support through the ffAS, Bolivia systematically lobbied in the United States to forestall the tin sales of the General Services Administration. Despite the producers' efforts, the tin market collapsed when the London Metal Exchange ceased its tin operations (see table 9-8 and figure 9-2). Note that although tin prices (in real terms) fluctuated strongly between 1950,and 1987, the trend was upward until 1980. In that year a process of steady decline started that continued until 1985, when the price of tin took a nosedive. The production and export of tin and other minerals were severely affected by punitive taxes in the 1970s, by the exchange rate policy, and by the'decline of the Bolivian tin industry. Few major mines' had been opened since the Depression, and the"metal content of the ores of the old mines had been decreasing. By 1980 Bolivia had lost its. place as a major world tin supplier., Policies on the other significant Bolivian export, natural gas, were less clear-cut. In 1972 Bolivia signed a twenty-year treaty with Argen- tina for the sale of natural gas. The natural gas deal included provi- sions on the quantities to be sold and a fixed price, with rather vague clauses on the procedures for readjustments. Shortly after the treaty was signed, the first oil shock occu'rred and world energy prices sky- rocketed. Bolivia's natural gas exports did not benefit from this increase for quite some time. Until 1980 price increases for. natural gas trailed those for petroleum (see figure 9-3). Only after the second oil shock of 1979 did Bolivia and Argentina agree on important revisions of the-unit price of natural gas. After this second-shock, natural gas prices have either increased more rapidly or ' decreased more slowly than petroleum prices. The natural gas-nego- -~~dcrae mor sl ps nego:- E' Boliuian Dade and Development, 1952-87 237 hbIe 9-8. International Prices of lin and Natural Gas, Indexes, 1952-87 (1980 = 100) Tin Natural gas- In current hn 1980 In current in 1980 Year prices' prkesb prices pricshe 1952 15.8 58.0 - 1953 12.0 45.1 - - 1954' 11.8 45.4 - 1955 12.2 46.0 - 1956 12.9 47.2 - 1957 12.4 44.3 -. 1958 12.1 42.4 - 1959 12.9 45.9 - 1960 13.1 45.6 - 1961 14.6 50.0 - 1962 14.7 49.6- - 1963 14.9 51.2 - 1964 20.3 68.2 - 1965 23.2 77.4 1966 21.3 68.5 - 1967 19.9 63.2 - 1968 18.6 59.7 - - - 1969 20.4 62.3 - 1970 21.9 62.7 - 1971 20.9 56.7 - - 1972 22.5 56.2 - 1973 28.8 62.0 10.6 22.9 1974 48.9 86.5 17.4 30.9 1975 41.0 65.2 25.2 40.1 1976 45.2 71.0 32.3 50.7 .1977 64.2 91.6 37.6 53.8 1978 76.9 95.6 45.8 57.0 1979 92.1 101.0 56.2 61.6 1980- 100.0 100.0 100.0 100.0 1981 84.4 84.0 141.6 141.6 1982. 76.5 78.3 153A. 157.2 1983 77.4 80.3 156.8 162.7 1984 73.2 77.3 136.7 144.4 1985d 69.0 73.9 147.1c 154.0 1986d 38.8 34.3 128.1c 113.2 1987d.c 41.7 33.1 108.0c 85.7 - Not available. a. Standard, minimum 99.75 percent1 settlement price. a Current prices deflated by the World Bank Manufacturing Value Added Index. c. Price adjusted to the austral real exchange rate to take into account Argentine payments in goods and services. d. Preliminary estimates. e. June 1987. Source: WorldBank(1986);CentralBankofBolivia (1987TlImr(1987). 238 Juan-Antonio Morales Figare 9-2. Real World Price of Tioi, 1950-87 (1980 = .100) 110 100 90 - 80 60 50. 40 1950 1955. 1960 1965 1970 1975 1980 1985 Source. Central Bank of Bolivia (various issues).- tiations in the 1980s were extremely difficult because Argentina had itself found important deposits of this resource, which it had not-fully exploited because of its agreement with Bolivia. The steady fall in energy prices since 1982 has aggravated the prob- lem. Bolivian export'volumes and prices have fallen. Moreover, between 1983 and the last quarter of 1987, Argentina made. only 40 percent of its payments in dollars, the rest being paid in Argentine goods and services. Because of the frequent overvaluations of the Argentine peso and the austral, the effective price Bolivia received from its natural gas sales to Argentina was -substantially smaller than the price indicated by the dollir value. In addition, the frequent delays in the Argentine payments in the 1980s caused liquidity crunches, which- had strong repercussions in the domestic economy. In 1974 a far-reaching agreement -for sales of natural gas to Brazil was signed, but internal opposition in Bolivia killed the deal.' The promotion of nontraditional exports came relatively -late, in 1977. Generous incentives 'were provided: in addition to tax rebates, there were subsidies at varying rates depending on "embodied value'. Bolivian 7/ade and Development, 1952-87 239 Flzure 9-3. Real World Prices of Natural Gas and Petroleum, 19273-87 (1980 100) 160 140 1200 800 * 60 *: 40 20 1973 1975 1980 1985 Souce: Coentl Bank of Bolivia (vaiious issucs). added:" The concept of embodied value added, used to determine the base for export subsidies, included labor and capital services, as well as all domestic inputs. The base for the benefits was clearly incorrect, discriminated heavily against traditional exports, and was a source of corruption. For instance, Bolivian gold jewelry was priced intera- tionally entirely for.its gold content and not for its artistic features yet it benefited from the highest subsidy. The law pro'moting nontradi- tional exports was reformed in 1982, widening the benefits. With the policy reforms of August 1985, which we will discuss later, the varying rates of subsidy (or, more appropriately, tax rebates) were simplified to two rates: 10 percent for nontraditional exports and 5 percent for, traditional exports. In 1966 Bolivia joined the Latin American Free Trade Association, which later became the Latin American Integration. Association, and' in 1969 it joined the Andean Group. The aim was to promote manu- factured exports and, to a lesser extent, agricultural exports. Bolivia's participation in both agreements has been modest despite the govern- 240 Juan-Antonio Morales ment's considerable expectations when it signed the agreements. The scant benefits are mainly explained by Bolivia's initial conditions of extreme industrial underdevelopment and by the uneven quality of the economic policies followed in Bolivia and in the partner coun- tries. Although movements to freer trade may have benefited the country, Bolivia was completely unprepared to follow the joint indus- trial programming that was the goal of the Andean Group and to adopt a high common external tariff.- Import Policies Like its export policies, Bolivia's. import policies did not have well- defined economic purposes, at least until the policy reforms of August 1985. Domestic industry 'was protected from competitive imports by import tariffs, quantitative restrictions, and credit subsidies. The tariff structure had a number of objectives. It was intended to provide fiscal revenues, to substitute for devaluations in the face of transitory balance of payments difficulties, and to provide protection. Tariffs varied depending on whether the items were finished products, intermediate inputs, or capital goods (see table 9-9). The rates on finished goods (especially luxury goods) were frequently determined more by fiscal revenue considerations than by protection, although ultimately this created unintended protection for the domestic pro- duction of luxury goods. Many domestic industries were also pro- tected until 1985 by quantitative restrictions on competitive imports, especially by bans. Quotas and import licensing were rarely used. Both high tariffs and bans were difficult to enforce in practice, given the length of the Bolivian borders, which are sparsely popu- lated, and the extreme administrative weakness of the government. Bolivia's geographical features made control of smuggling difficult. The import-substituting industries often benefited from- two addi- tional forms of subsidy: first, the availability of credit at concessional interest rates, which were negative in real terms on occasion, and, second, preferential government purchases. The effective rate of sub- sidy was probably very high. The question then arises as to why Bolivia made so little progress in developing a manufacturing base, especially when compared with other countries of similar market size that fared substantially better. There are three tentative answers to this puzzle: * The prolonged periods of overvaluation of the Bolivian peso reduced the net effective subsidy rate significantly. * Macroeconomic and political uncertainty reduced 'long-term - investment. Industrialists were-more interested in- generating the B-olivian fade and Dovelopmen4 1952-87 241 Ihble 9-9. Effective Rates of Protection Yielded by LMriffs, Bolivia about 1982 (percent) Thrijon 7hrf on Effectiw nt ermediale finished rate Of Itm g60w - goodfs protections Products Agricultural products 2.0 28,1 31,9 Livestock products 0.0 0.0 -1.2 Forcstry products, gamc, f'ish 25.8 17.5 17.5 Processed meats 8.3 8.3 80.4 Dairy products 2.0 2.8 2.1 Mills and bakeries -10.0 17,5 65.0 Sugarandsugarproducts 26.3 41.7 114.5 Other foodstuffs 13.4 1 9.9 40.0 Beverages 13.3 60.2 88.4 Tobaccob 17.6 100.0 195.7 . Tcxtiles, clothing, leather, and leather products 38.7 72.2 169.1 Wood and wood products 35.4 62.1 145.8 Paper and paper products 10.8 35.0 76.0 Chemicals 11.7 37.2 64.8 Stone, earth, clay products 16.9 45.6 56.1 Metallic products, machinery and equipment 19.0 45.7 57.4 Others 29.9 43.3 49.9 Oilher inputs Crude petroleum and natural gas 4.0 n.a. n.a. Metallic and nonmetallic minerals 11.9 n.a. n.a. Petroleum products 4.5 n.a. n.a. Basic metals 12.6 n.a. n.a. Summary measuTresc Weighted mean for all sectors 27.6 44.4 Standard deviation for all sectors 20,1 43.9 .Variation coefficient for all sectors 72.7 99.1 -Weighted mean for manufacturingsector 47.7 94.1 Standard deviation for manufacturing sector 19.6 44.5 Variation coefficient for manufacturing sector 41.1 47.3 n.a. Not applicable. a. Intermediate goods imported into the sector. b. Quantitative restrictions applied in this sector, hence equivalent tariffs have been computed. c. Veighted by value added at internationial prices. Source: National Chamber of Industry (1987). 242 Juan-Anionio Morates fast, above-normal profits that the protection structure permitted than in making long-term investments. They seldom reinvested their high profits, because they feared political-turmoil and possi- bly expropriatory measures. The best example of the costs imposed by uncertainty appeared during the h oerinflation years, when, despite the high effective rates of subsidy resulting from such factors as negative real interest rates and foreign exchange controls, investment in the industrial sector was almost nil. Manufacturing value added actually declined a staggering 36 percentage points between 1981 and 1985. - * Excessive protection, especially during the 1970s, may have: had high long-run efficiency costs. Exchange Rate Pblicies After the high inflation of the 1950s and until the 1980s, Bolivia maintained a policy of fixed exchange rates. The government's fiscal policies were expansionary, but.as long as the fiscal deficits could be;. financed by borrowing abroad, they had no effect on the exchange rate. Throughout most of the period,= the parallel market premiums were very low.3 One of.the main tenets of economic.policy was indeed exchange rate stability, and the peso was devalued only twice between 1957 and 1981. Prompted by.the debt crisis of 1982, the peso went into free fall until the August 1985 stabilization program. ; An overvalued peso has been a permanent feature of Bolivia's econ- omy for the last-thirty years. Mild overvaluation hampered the coun- try's export potential in the agricultural and manufacturing sectors more than in the traditional exports, which benefited from. high inter- national prices during most of the 1960s and 1970s. The severe over- valuation of the official exchange rate grealy harmed: legal exports between 1982 and 1985 by: reducing the -profitability of the-enterprises that produced them. Contraband exports, however, benefited from the fiavorable parallel market exchange rate-and flourished.4 Although the evidence supports the notion that overvaluation was .present during most of. the 1960s and 1970s, the extent of the over-.- valuatCnn is difficult to assess. The noticeable improvement in the- terms of trade from 1960 to 1980 hampers the computation, of what: - should have been the right exchange rate. The.real exchange rates shown in table 9-10 indicate what happened, but the numbers are rough and other experts. may disagree with the methodology used., Export prices and real exchange rates- show. a strong negative- correla- tion. The sudden and relatively long-lastirg jump in export prices in the 1970s allowed the exchange rate to diverge from the higher long- run value needed forsustained development Bolivian l3ude and De-velopment, 1952-8 7 243 TAbble 9-10. Real Exchange Rate, Bolivia, 1960-69 (1980 - 100) Real Real Real Year exchange rate Year e%duarg rate Year exchange rate 1960 171.1 1970 110.5. 1980 100.0 1961 155.1 1971 114.2 1981 82.6 1962 142.4 1972 120.1 1982 98.6 1963' 142.0 1973 157.3 1983 95.3 .1964 131.4 1974 115.7 1984 66.9 .1965 134.0 1975 . 115.4 1985. 114.0 1966. 128.5 1976- 115.7 1986 128.0 1967 112.7 1977 115.6 1987M 123.5~ 1968 108.7 1978 l11l4 1969 112.9 1979- 107.2 Note: Official exchange rates dcflated by the ratio of Boliviain consumer pr'ices to US. wholesale prices and caverted to index. Exchange ratios and price indexes are annual averages. a. Preliminary estimate. Source: imF (1987), Central Bank of Bolivia (various issues). More important than a precise computation of real, exchange rates is the political economy problem of why no lobby of expo-rters and other interested grusapared to' push for exchange rate co'rrec- tion. Indeed, the lobbying by. tin producers was essentially aimed at. decreasing the tax, burden on their. output an,eventually, on their export sales. They sought to increase their effective'exchange rate through reduced taxatian rather than devaluation. The other main exporter, -the state-owned petroleum company YinB, made significant profits on its exports of petroleum and natural gas, even during the worst yars of overvaluation in the 1980s; therefore, it never lobbied for devaluation. A better strategy for. the'period would probably have been a more activist exc-hange rate- po1icy, such as. that followed by Indonesia, a country also dependent on 'resource-based exports. On the import side, overvaluation-was reflected in the relatively high'tariffs and the extent of. quantitative restriction Between 17 and .1982, except in 1980 when a favorable;-terms of trade shock occurred, the high tariffs and quantitative- restrictions were insuffi-: cient to restrain imports and prevent.deterioration ~of. the- current account of the balance of payments. Trade, deficits were financed with increased foreign indebtedness and from foreig exchange reserves.-- Overvaluat'ion waslalso. a factor. in'the huge volume of capital -flight- bewen 1976 and 1982. The' public's percption. that the exchange -rate was: below equilibrium prompted people to change their pe-sos. into dollar-s and to deposit their.dollar assets abroad. Capital flight accelerated during the militar-y governments of: 1980-82 (Ugarteche' 244 Juan-Antonio Morales 1986). With increased awareness of the difficulties Bolivia would face in obtaining new loans and servicing its old loans, speculators-and the public increased their assets.abroad. The main risks they were hedging against were future peso depreciation and limits on peso convert- ibility. The continuing capital flight between 1982 and 1985 was almost entirely the result of political uncertainty and macroeconomic instability. Between 1983 and 1985 exchange rate controls-restrained imports more effectively and limited the losses in reserves. The internal costs of the exchange rate controls, however, were huge in terms of effi- ciency in production and, especially, in the. fight against inflation.5 With the high black market premiums for foreign exchange prevalent * at the time, legal exporters often felt penalized, viewing the premium as a tax on their profits. As a consequence, their strategy was fre- quently aimed at minimizing the burden of this tax. For example, the mining enterprises would. accumulate stocks of their output and sell' them when thc premium was lower. They could finance retention of large stocks because of.their access to highly subsidized. domestic credit. Less scrupulous exporters would sell abroad and circumvent the exchange rate regulations entirely:' Foreign exchange users in, for example, the manufacturing sector * ': viewed the black market premium as a tax on their inputs and pressed for more access to the already scarce official foreign exchange reserves. . Complaints in the press about not having enough foreign exchange to.pay for imported inputs were frequent, but although complainants did not have access-to all the foreign exchange that they wanted at the official rate, the pamallel market could always -meet their demands. On the output end,. with few exceptions, manufacturers priced their products in dollars, although payable in domestic cur-: rency at the black market rate of exchange. Their complaints on the scarcity of foreigni exchange at the official exchange rate were actually * pleas for more subsidies. Producers' uncertainty about whether for- eign exchange would be available at the official rate to purchase imported inputs was one of the main factors reducing production. * In the high-inflation period of 1982-85, exchange rate management was indeed one of the government's most difficult tasks: Because of incomplete, or on some occasions defective, fiscal and monetary poli- cies, the government had to face virtually unsolvable dilemmas.. On, the one hand, rapid depreciation of the official exchange rate would have increased the inflation rate even more and raised the burden of the foreign debt,. as indeed happened. On the other hand, trying to repress inflation with the exchange rate was costly and ultimately inef- fective, because black market premiums surged, eventually leading to the needed official exchange'rate corrections and to more inflation. Bolivian Dade and Dewlpmen, 1952-87 245 Moreover, the policy of exchange controls led to a high degree of corruption. The Policy Reforms of August 1985 In August 1985 the newly elected government of Victor Paz-Estenssoro unveiled a package of extensive policy reforms. The package had two main components that proceeded simultaneously: a stabilization pro- gram and a trade liberalization program. The package was followed by a string of other measures collectively known as the New Economic Policy. Liberalization was intended to reinforce stabilization by pro- viding a natural check on domestic price increases, the initial situa- tion being one of disequilibrium in many markets. Liberaiization without first achieving stabilization would have probably been impos- sible, given the pressures on the current account of the balance of payments (see Sachs 1987a for a similar argument). The core of the stabilization package was a policy of stable and unified exchange rates supported by tight fiscal'and monetary policies (see Sachs 1986, 1987a, 1987b; Morales 1987a, 1987b; Morales and Sachs 1987 for a description of the stabilization program and its achievements, and table 9-11 for a summnary of the policy reforms). The August 1985 program and subsequent measures complemented exchange rate unification with other far-reaching liberalization mea- sures.'All quantitative restrictions on exports and imports were abol- ished except for those related to imports of sugar.6 In addition, import tariffs were lowered to a uniform rate of 20 percent except for imports by the petroleum enterprises and imports covered under standing economic integration agreements. The liberalization of international trade was accompanied by liber- alization of domestic markets. Price-ceilings: have been eliminated in most markets; only petroleum derivatives and a handful of public utilities, most of them state-owned, have regulated prices. The liberalization of the labor market may be especially relevant to development of the external sector. Job security clauses have been softened, making dismissal of workers easier; wage indexing has been . suppressed; and collective bargaining with arbitration by the Ministry of Labor has been greatly weakened, leaving most wages, at least in the' private sector, to be set by individual negotiations. The combination of the policy reforms and the collapse of export markets has caused the main export enterprises in both the public and.the private sectors to lay off a significant number of employees.. Liberalization of the current account of the balance of payments has proceeded simultaneously with the liberalization: of the capital account. No restrictions are imposed on capital inflows and outflows. 246 Juan-Antonio Morales Bible 9-11. Summary of Main Policy Reforms, Bolivia E-xchange rates Unified floating cxchange rates. Rates determined by auction. Exporters obliged to surrcnder 100 percent of export proceeds to the central bank. Foreign trade All quantitative restrictions on cxports and imports lifted. Uniform import tariff rate of 20 pcrcent. Foreign capital All restrictions on forcign capital inflows and outflows eliminated. Normalization of rclations with multilateral official creditors. Negotiations in Paris Club. Negotiations under way to buy back debt owed to foreign commercial banks with discounts given by secondary market Domestic market for Price controls eliminated in almost all markets. goods and services Price for gasoline and other oil derivatives fixed by government at border prices. Most public utility rates and urban transportation fares fixed by local governments after negotiations with concerned parties. Only a handful ofpublic utility rates set by the central government. Domestic financial Interest rate ceilings eliminated. Bolivian residents markets freely allowed to open deposits in foreign currencies in the domestic banking system. All contracts in foreign currencies.among residents authorized. Development loans channeled through private banks. Requirements on capital-debt ratios for private banks increased and more cx post supervision of banks portfolio enforced. Labor market Job protection clauses in legislation reduced in scope and the principle of marginal market adjustments stated. Wage indexation eliminated. - Fiscal reform 131x reform enacted with value added tax and wealth taxes as main taxes. Reorganization of state Several mines of state-owned mining enterprise enterprises COMIBOL closed and several thousand miners dismissed. Worker dismissals envisaged ir. the state-. owned petroleum company YPFB. Decentralization Enterprises of state-owned holding Corporaci6n Boliviana de Fomento ceded to regional development corporations. Assets of the National Transport Company transferred to municipalities. Boivian Trade and Dleelopmen4, 1952-87 247 In addition, the domestic capital market has been completely liber- alized by eliminating interest rate ceilings and floors and by permit- ting the public to open dollar and dollar-indexed demand and time deposit accounts in Bolivian banks. The result is that interest rates have soared, leading Bolivians to repatriate a significant amount c. capital-$350 million (or about 8 percent of GDP) in two years. Most of the capital, however, has gone into dollar or dollar-indexed time deposits with very short maturities (thirty to sixty days). The August 1985 program also called for divestment of certain state-owned enterprises. As a first step toward privatization, the enter- prises of the state-owned holding company the Corporaci6n Boliviana de Fomento and a small state-owned public transportation company have been transferred to the regional development corporations and the municipalities. Some other small state-owned enterprises have been dissolved. It is much too early to give a complete assessment of the accom- plishments of the policy reforms. Inflation was rapidly stabilized, how- ever, and in 1987 the country obtained a positive growth rate for the first time in six years. Normalization with official creditors, multi- lateral and bilateral, has also been achieved, implying significative debt relief and the granting of new loans. In the first two years after the policy reforms, consumer imports expanded and :he current account deficit of the balance of payments rose in both absolute terms and as a percentage of GDP. This growth in imports may, however, be a transitory. phenomenon experienced after. a process ofrapid liberalization. The reforms have eliminated most- policy-induced, anti-export biases, but obstacles of a more:structural character remain, including currency. overvaluation. The temptation to hold down inflation with overvaluation is a persistent danger (see Dornbusch 1986 for a discus- sion of this problem in several Latin American countries). Overvalua- tion is not yet severe, but it is steadily increasing. Also, interest rates are too high and will have to decrease for long-tern productive invest- ment to take place. Appropriate disbursement of the foreign loans' that Bolivia has contracted may help lower interest rates. Bolivia must reconstruct its external sector entirely. The collapse of iarkets for Bolivia's traditional exports (tin and natural gas) and the poor price prospects for most basic commodities that Bolivia could reasonably export in the near future make the situation challenging. In addition, the external debt burden continues to hinge on the coun- try's recovery and growth prospects, compounding the problem of weak markets for exports. The policy reforms are a step toward recovering previous export levels, but they are insufficient. The structura obstacles to reconstruc- 248 Juan-Antonio Morales tion of the external sector and achievement of export diversification are indeed formidable. If agricultural exports, including soya and other grains, are to take the place of tin and natural gas, unit transpor- tation costs will have to be lower than current levels. Well-designed investments in the transport infrastructure are a precondition for the resource shifts needed to rehabilitate the export sector. Investment in training workers and managers is also needed, given that the techno- logical content of exports, both agricultural and light manufacturing, is increasing. Finally, political stability is essential to Bolivia's development, espe- cially for its foreign trade sector. To attract foreign investment, Bolivia must provide not only legal guarantees against expropriation and punitive taxation but also a good social and political environment. In turn, durable political stability will be achieved only if the government finds-ways to close the wide -gaps in the distribution of income and wealth. This may prove to be a difficult task in view of its current fiscal - constraints. Conclusion *urinng the last,thirty-five years Bolivia. has endured two periods of -great economic distress, with substantial decreases in the GDP and investment rates. and with very high inflation. The- first economic arisis occurred in the 1950s during a period of profound social trans- form'ation. The second crisis, in the 1980s,- followed a long period of sustained growth- accompanied by an accumulation of major develop- ment problems. It was triggered by an interplay of external shocks and. domestic factors. One of the most important. internal factors was over- borrowing abroad, which was caused by a peublic sector that' was too large, a pattern of resource-based industrialization, and a misuse of investment funds. Bolivia's development hinges on the performance of its foreign sec- tor. Export earnilngs that depend on only a handful of basic products are subject to wide fluctuations. For example, more than 50 percent of * export proceeds depend on natural gas sales to Argentina. bThejump in export prices and the easy access to foreign debt in the 1970s allowed a substantial discrepancy between the short-run value of -the. exchange rate and the long-run value needed for sustained growth. The lasting overvaluation impeded the development of a more diversified export basket and a production structure better able to withstand external shocks. The public sector has had an important role in Bolivia's develop- ment as well as its problems. The extraordinary incursion of the pub- lic sector in investment is principally explained by. the government's Bolivian 71u and Deomt ,1952-87 249 inability to achieve broad taxation in the face of an ambitious redis- tributive agenda and the continuous reassessment of development goals. The search for new tax sources became tantamount to the search for new public investment opportunities. Taxation of a few natural resources is easier to implement than taxation on a broad base, and one of the main problems in Bolivia's. development has been overtaxa- tion of the natural resource sector. The state enterprises are also over- burdened with their role as the main providers of foreign exchange. In this role they have been frequent victims of. misguided macro-' economic policies that tried to subsidize consumption via the exchange rate. The public sector was overextended and weak at the same time. A good example of the state's weakness. is given by the regulations on foreign trade, many of which stayed on the books and could not be enforced. For example, high import tariffs were difficult to collect and only gave risc to more contraband. A similar comment can be made about the formrer regime of exchange rate controls. The Bolivian expe- rience clearly shows that a weak state cannot support a heavy burden of regulations. Most people simply ignore them and, what is worse, the state and the few who abide by them suffer. The political instability of some periods took a heavy toll on the economy. On several occasions, instability impeded timely recognition of the shocks affecting the economy and.thus delayed needed action. The uncertainty that went with political and macroeconomic insta- bility also imposed heavy costs in terms-of production and welfare. With a high degree of uncertainty, no system.of incentives can work well, and speculation thrives at the expense of productive, long-run investments. The policy.reforms introduced in August1985 are promising. The most immediate and visible outcomes have been the taming of infla-- . tion and the introduction of order in the fiscal budget. In the medium term following the reforms, the results on economnic growth are less encouraging than the results on inflation. But there are already signs of productivity growth and improvement in resource allocation,'which [it is hoped will lead- to. sustained economic development. Further- more, the reforms should help in. achieving an economy that is more diversified and open to foreign trade. The problem now is to persist in the reforms in face of severe obstacles and an increasing impatience in the public with the still-modest benefits in economic growth. Notes 1. The attempt at stabilization made by the civilian interim president Lydia Guciler stands out as an exception in this chaotic penod. 250 Juan-Antonio Moraoks 2. The ITAS were agreed on by the main producer and consumer countries, with some important exceptions. The main but not the only instrument for achieving price stabilization was a buffer stock of tin metal. Interventions by the International Tin Council (the governing body of the IrAs), with its buffer stock, and sometimes with the setting of production quotas, helped to achieve producers' price objectives until the third quarter of 1985. After 1981 the ITC sustained prices that were well above long-run market equilibrium prices. 3. The exchange rate was de facto pegged to the dollar during the whole period 1957-82, although some statutory provisions allowed for the possi- bility of floating and for the ratc being fixed in terms of gold. 4. The rapid deprecation of the peso in the parallel market during the period of high inflatic n led to exports of the most unlikely goods, even of highly perishable good'X such as bread, or of those with high transport costs, such as bottled soft dri iks. Even tin ores were smuggled out to Peru, which started to appear in the official statistics as a tin producer. - 5. The public's beliet:in the corruption of those persons who either admin- istered the allocation of the rationed foreign exchange or had access to it seems to have been a more important factor than the corruption itself in discrediting the system of controls. 6. Sugar quotas violate the general spirit of the liberalization measures, but given the highly manipulated nature of the international sugar market, Boliv- ia's regulation of sugar imports has to be viewed as an antidumping measure. Selected Bibliography Afcha, Gonzalo,'and Gualberto Huarachi. 1988. El programa econdmico boliviao: La nueva politica econdmica. La Paz: Unidad de Anilisis de Poifticas Econ6micas. Bolivia, Ministry of Planning. 1970. Revista de Plangicacdn. La Paz. Central Bank of Bolivia. 1987. Statistical Bulletin 259. La Paz. - Various issues. Statistical Builetin. La Paz. CEPAL (United Nations Economic Commission for Latin America). Various years. Statistical Yearbookfor Latin Amica and the Caribbewn 1964-85 eds. New Y York. - . 1987. Balance preliminarde La economia. Santiago, Chile.. Dornbusch, Rudiger. 1986. Inflation, &xchange Rates, and Stabilization. Essays in International Finance 165. Princeton, NJ.:. Princeton University. Frieden, Jeffrey A. 1981. "Third World Indebted Industrialization" Interna- tional Organiwion 35(3):407-31. IMF (International Monetary Fund). 1986. Government Fineoce Statisics. Vol. 10. Washington, D.C. . 1987. International Financial Statistics Yearbook. Washiington, D.C. Inter-American Development BanL 1983. Economic and Social Progress in Latin - . - America. Washington, D.C. - : . Malloy,James. 1970.Bolivia: The Uncompleted Revolution. Pittsburgh, Pa.. Univer- sity of Pittsburgh Press. - . Morales, Juan-Antonio. 1982. "The Bolivian External Sector after 1964r In Jerry R. Ladman, ed., Modern Day BoviaLegacy oftheRevolution and Prospects for the Future. Tempe, Ariz.: Arizona State University, Center for Latin Amer- ican Studics. Bolivian Trade and Developnent, 1952-87 25) . 1 987a. Precios, salariosypolitica econdmi adurant la alla inJlacidn boliviana de 1982 a 1985. La Paz: Instituto Latinoamericano de Investigaciones Sociales. 1 . 987b. InJflaion Stabilization in Bolivia. Documcnto de Trabajo 06187. La Paz: Universidad Cat6lica Boliviana. 1 1987c. "Estabilizacidn y nueva polftica econ6mica en Bolivia:' El Tri- uistre Eeondmico Special Number (September): 179-212. 1 1988. "Adjustment and Growth'in Hyperinflation: The Case of Boliv- ia' Universidad Cat6lica Boliviana, Department of Economics, La Paz. Morales,Juan-Antonio, andjeffrey Sachs. 1987. "The Economic Crisis in Bo- livia:t National Bureau of Economic Research, La Paz. National Chamber of Industry. 1987. Estudio integral del mrancel de importaciones. La Paz. Ram, Rati. 1987. "Exports and Economic Growth in Developing Countries: Evidence from Time-Series and Cross-Scction Data' Economic Developmet and Cultural COne 36(1):51-72. Sachs,Jeffrey. 1986. "The Bolivian Hyperinflation and Stabilization" Working Paper Series 2073. National Bureau of Economic Research, Cambridge, Mass. 1987a. "International Policy Coordination: The Case of the Develop- ing Counitry Debt Crisis:' Working Paper Series 2287. National Bureau of Economic Research, Cambridge, Mass. - 1987b. "The Bolivian Hyperinflation and Stabilization" American Eco- nomic Revieu4 Papers and Proceedinp (May):279-83. UDAPE (Unidad deAnalisis de Poiftica Econ6mica). 1987. Dossier de infornm-adn estaditca 1980-1987. La Paz, October. Ugarteche, Oscar. 1986. El estado deudor Eonomia politica de La deudar Pmery Bolivza 1968-1984. Lima: Instituto de Estudios Peruanos. World Bank. 1983. "Bolivia.- Structural Constraints and Development Pros- p :cts' Report 4194-BO. World Bank, Whshington, D.C.- 1986. "Commodity Trade and Price Trends:" Commodities and Export Projections Department, Development Policy Staff, Washington, D.C. - 1987. "Commodity Trade and Price Trends:' Commodities and Export Projections Department, Development Policy Staff, Washington, D.C.' .1988. WorldDebt TablesiFai naal Debt of DevelopingCounties. 1987-88 ed. Washington, D.C. Index Africa: capital flight from, 80; external 25; Nigeria and, 205,209, 210-11; trade capital and, 12;trade taxes and, 116.See liberalization and, 145,146 also names ofeonulries Banking system: in Israel, 24-28,33-36;. Agriculture: in Argentina, 102; in Bolivia, money supply role of, 19-24.37 nn3-6; 223, 227; exchange rate and, 102; in oligopolistic, 24, 28. Srealso CentraI Nigeria, 199-200 bank; Commercial banks Andean Group. Bolivia and, 239-40 Bank of Israie, 24,26,27,33-34,35,36 Argentina: agriculture in, 102; Bolivian Banks: credit allocation and, 49-52; exports and, 224; Bolivian natural gas supervision of, 53-54. See also Banking and, 236, 238; capital flight from. 80, system; Bank of Israel; Central bank; - 83,111; capital flows and, 55-56,61; Commercial banks economnic difficulties in, 8, 9, 10; export Banzer administration (Bolivia), 228, 231 duties and, 113; financial reform in, 13, Barrientos administration (Bolivia), 227, 39-40,46,48,53; financial repression 228 in, 41; inflation and, 65; interestrate Black market?,91,96,107-9,111-12,119I and, 10; money creation and, 78; quasi-- for dollars, 107-9; foreign exchange fiscal deficit and. 79; real product per losses and, 117-18; inflation and, 118; capita in, 8. 9; trade misinvoidngand, macroeconomiceffects of. 115-19; rev- 112 enue losses and, 116-17; taTiffrates Argentine Austral Plan, 33: and, 113; trade restriction and, 109; Se - Asia: budget deficits and, 74; capital flight also Tade misinvoicing from, 80; economic growth in, 9; eco-; Bolivia: agriculture in, 223,227; Andean nomic performance of, 65; external Group and, 239-40; Argentina and, capital and, 12; finance in, 64; inflation 224,236,238;Banzeradministration of, and, 73, 74; interest rates and, 69; trde 228,231; Barrientos administrationof, gaxes and, 116. See alo names ofcounitries 227,228; budget deficit and, 14-15; Assets. See Financial assets; Foreign- capital flight from, 24344; cocaine exchange-linked assets; Net foreign and, 223; debt and, 220,222,224,227, assets 228,229,232-34,248; economic crisis Assets market, exchange rate and, 92,93- in,220,222-24,227-34,236,238-40' 94 242-45,247-49,250 nn2-6r, ecoliomic growth in, 222-24,227,228,247,249; Babangida administration (Nigeria), 198, economic history of (1952-87), 220; 214-15 , 222-24,227-30; economic policies of, Balance of payments: exchange rate and, 220,222-24,227-34,236,238-40,242- 28; foreign exchange reseres and,23, 45,247 49,250 nn2-6, embodied value 253 a~~~~~~~~~~~~~~~~~~~ 254 Index added and, 238-39; cxchangc rate and, 84, 110; dollarization and, 80, 82-84; 242-45.248,250 nn3-5; exports and. inflaLion and. 118; from Latin America, 220,222-24.227,228,236,238-40,242, 80,83; launderingof assets and, I1; 243,245,247,248,249,250 n2; Foreign from Mexico, 57,80,85,86,87-88, 111, borrowing and, 220, 224, 227, 232-34, 115; from Peru, 80,82; seigniorage and, 248; Carcfa-Meza administration of, 83-84; trade misinvoicing and, 109-1 3, 229; cm' in, 220,222,223,229; govern- 115; from V'nezuela, 57 mcnt rinancing in, 232; Guciler admin- Capital flows: Argentina and, 55-56,61; istration and, 249 nl: imports and, 224, Chile and, 55,56-58, 61; fixed exchange 240,242,244; imp6rt substitution and, rate and, 23-24; interest rate and, 30, 224, 227, 240; indebted industrializa- 36; Latin America and, 55-51B, 59-60, tion and, 232-34; inflation and, 220, 61; Uruguay and, 55,56,61 222, 229-30, 244, 247; labor market Capital income, taxation of, 163-67, 170 and, 245; mining and mincrals and, nn1O-13 223,227,235,23Cr, money creation and, Capital market, 6, 7-8 14-15,78; natural gas and, 223-24,230. Carter administration (United States), 236,238,245,247,248; nontraditional Bolivia and, 228 exports and, 238-39: oil and, 223,227, Central bank. 18: loans sales by, 27; money 228, 233, 236,238. 243, 245: Ovando supply role of. 19-21, 37 nn3-5. See also administration of, 228,231; overvalua- Banking system tion in, 242, 243-44, 247, 248: Paz- Chile capital flows and. 55, 56-58, 61; Estenssoro administrations of. 227, financial reform in, 39-40,46.48, 53; 229-30,245.247-48; policy reforms in financial repression in, 41, trade liber- 229-30,245,247-48.249. 250 n6; popu- alization in, 14 lation in, 228; private sector and, 227, cMwc See Complejo M incro Karachipampa * . 234; production in, 223, 227; public Cocaine, Bolivian economy.and, 223 sector in, 220,229,230-32,248; revolu- COKIBOL. See Corporacidn Minera de tion of 1952 and, 227; Siles-Salinas Bolivia - administration of, 228; Siles-Zuazo Commercial banks, money supply role of,' administration of, 229; state capitalism 19-21,37nn3-5.SeealfoBankingsys- in, 230-32; state cnterprises in, 231-52, temn; Banks 233,247; taxation in. 113, 228,232, 249; Competition (in markets), 4, 6 terms oftrade and. 224,228,230; tin Competitiveness, exchange rate and, 6-7, and. 223.224,227,228,230,233,236, 94-95 247,248; Torrez administration of, 228; Complejo Minero Karachipampa (cMK) tradeand,220,222-24,227-34,236, (Bolivia), 2533 238-40,242-45, 247-49.250 nn2-6; Corporaci6n Minera de Bolivia-(comiBoL), trade liberalization and, 245; trade 231,233 partners of, 220; trade taxes in, 113 Corruption: exchange rate and, 106; pro- Bond prices, 27 tection policies and, 147'- Bond sales, 22,26,30 'Costa Rica, trade taxes in, 113 "Border tax adjustments:' 16B n3 Credit market financial liberalization Brazil, 71;capital flight from, 83,84, 110; and, 48-52,62 nn7, 8; financial repres external shocks and, 74, trade misin- sion and, 49-52, 62 nn7, 8; separation voicing and, I11 from money market of, 54-55 Budget deficit Asia and, 74; Bolivia and, Currency: in circulation, 19-20, 37 nn2-4; 14-15; inflation and, 7380, 82-84,85; nominal domestic, 17; unrestricted Israel and, 24-25,27-28,31-34; Nigeria. convertibility of; 55-56, 59,61, 96-97 and, 193 Buhari administration (Nigeria), 211-14 Debt, 12; Bolivia and, 220.222,224,227, 228,229, 232-34,248; indexing of, 60 Capital flight, 65;fromAfrica, 80;111; Debt service shock, 78 from Argentina, 80;83; from Asia, 80; Deficit, policies for, 7. See also Budget def- from Bolivia, 243-44; from Brazil, 83, icit; Quasi-fiscal deficit Index 255 Demand deposits, 19-20,21,37 nn2-5 European Economic Community (trZ), Denison't' "gucsstimate"' (1985),65 trade protection and, 129 Deposit insurancc, 54 Exchange controls, 137,138 Dcpreciation, or exchangc rate, 197-98, Exchange losscs,quasi-fiscal dericits and, 206,214,215 79 Deregulation (of financial system). 59 Exchange market: segmented, 95,96,97: Devaluation: exchlage rate and, 23,31- unified, 95-96,97. Sedalso Exchange 32,195-96, 215; in Israel, 35-36; in ratc Latin America, 6 Exchangc rate: alternative regimes for, Development banks, ill Lalin America, 52 91, 95-97; assets market and, 92,93- Disintermediation. Sec Financial 94; balance of paymcnts and, 28; black disintermediation rnarkers and, 91, 96, 107-9, 119- Boli- Dollarization,39: capital flight and,80, via and, 242-45, 248, 250 nn3-5; bud- 82-84: financial liberalization and, 55- getary problems and, 102, 106-07, 125 60; Mexico and, 84,86-87,88; Peru and, n4; commercial transactions and, 96, 82-83; scigniorage and, 83-84 97-107; 125 nn3, 4; competitiveness of, Duvalier,Jean-Claude (Haiti), 110 6-7, 94-95; corruption and, 106: crawl- ing, 23; cnwling peg, 25; definition of, * Economic development: financial factors 91; depreciation of, 197-98, 206, 214, in, 64-80,8249; financial repression 215; devaluation of 23, 31-32, 195-96, and, 66-73 215; dual rate systems, 91, 94 119-20, Economic growth: in Bolivia, 222-24,227, 123, 125. n6; exports and, 102, 103:. 228,247,249: financial deepeningand,. fixecd, 23-24, 97,123, 193, 242, 125 n2; 70-71; interest rates and, 71-72: in nexible, 91,96-97, 123-24; floating, *Korea, I; in Zambia, 1. Sc also Eco- 23, 25;-foreign borrowing and, 184-85, nomic performance; Economic policy 186, 191; France-and, 96; goods market Economic performance Asia and, 65; . and, 91-92, 93,94; impons and, 100-. economic policy and, 1-2,3, 8-10;. 101, 103-5, income distribution and, Latin America and, 65; real per capita 94-95; industrialization and, 100, 101, income and, 1-2. See awo Economic 102; interest rate and, 13-14; Israel growth and, 25, 29-50,31-32, 33, 34, 36-37; Economic policy: in Bolivia, 220,222-24, -. Italy and ,9&Japan and, 96; Korea and, . - 227-34,236,238-40.242-45,247-49, 96, Mexico and, 119, 120, 123; mone- 250 nn2-6: competition and, 4; eco- tary policy and, 17,28.37 nI: multiple, nomic performance and, 8-10; equity 13-14,91, 92, 93,95,%96 97-107, 125 ' in, 4; cxchange raLeand, 10,28-29,59- nn3,4; New York and, 119,120,123; 60,91-120, 123-26; institutional setting Nigeria and, 200, 205, 206,207,209, for. 3-4; macroeconomic, 6-8,91,92- . 214-15; pegged, 23, 29;-Peru and, 94; microeconomic, 4-6,91-92,94-95; . 97-99, 102, 107; policies for, 10, 28-29, objectives of, 2-9; property rights and, 59-60, 91-120, 123-26; prices and, 3,4,8; rules for, l5; terms of tradc .. 28-29,92,93,99; purchasing power of shock and, I92-98. See aLso Economic wages and, 100, 101; rcd tape and, 106, growth; Economic performance 125 n3; rentseeking and, 102,105-4, Economy, exchange rate rolc in, 91-95; resource allocation and, 91, 102-5; role S- calso Economic Policy; Nominal in economy of, 91-95; special regimes economy; Real economy - ff for, 100-102, 125 n3; stability of, 6-7; Education:-Nigeria and, 173-74,204 terms of trade and, 190, 195-96, 198; E-. See European Economic Community trade misinvoicing and, 91, 96, 107, Embodied value added,238-39 - .109.19; unified flexible, 91, I119, Ernpresa Nacional de Fundiciones (ENAF) 123-24; uniform, 96, 125; unrestricted (Bolivia), 233 convertibility arnd, 96-97; Venezuela - ENvv. See Emprcsa Nacional de Fundiciones and, 120. -See also Real exchange rate, Environmentalissues,tradepolicy and, -hde ' 5-6. Exchange rate overvaluation: Bolivia and, 256 Index 242,243-44, 247,248; Mexico and, 57, Financial resiriction: description of, 40- 96; results of, 3; Venezuela rnd, 57 4 1; In Korra, 40-41. 46: In Latin Amer- Export duties: Argentina and, 113; trade ica, 41; intvlwan, 40-41, 61 misinvoicing and, 109, 11-14 Financial sector, value added tax and, Lxports: Bolivia and, 220,222-24, 227, 154-55 228,286,238-40,242,243,245, 247, Financiems. See Nonbank financial 248,249,250 n2; exchange rate and, institutions 102, 108; Nigeria and, 172, 177,217,tnl IFiscal authorities, 18 Export subsidies, 1 5-56; tradc niisin- Fiscal contractlion, 22-23,24 voicing and, 109, 114 Fiscal expansion, 22-23,24 Extcrnal shocks: Brazil and, 74; indcxa- Fiscal policy: prudence in, 6, 7; terms of tion and, 75-77; Korca and, 74 &Sealso tradc shock and, 196-97. See also Fiscal Bolivia; Nigeria refonn Fiscal reform, 24-28 Financial assets, monetary framework Fixed-term deposits, 27 and, 18-21, 22-24.37 nn2-6 Foreign borrowing: Bolivia and, 220, 224, Financial deepening, 70-71 227; 232-34,248; exchange rate and, Financial disintermediation, 40,44,45, 62 184-85, 186, 191: Nigeria and, 205; n2 terms oftrade shock and, 174,175-76, Financial factors in economic develop- 177-79,180,181,182,184-8,186-87, ment 64-80,82-89 191 Financial liabilities, monetary framework Foreign currency assets, 25 and, 19,20-24, 87 nn3-6 Foreign exchange: black market and, 17- Financial liberalization: credit market 18; rationed, 95,97; sale of, 20; supply and, 48-52, 62 tin 7, 8; dollarization of, 95-97; unrestricted, 95-97. See also and, 55-60; domestic. 48-55, 61, 62 Exchange rate nn7,8; domestic markets and, 39; exter- Foreign-exchange-linked assets,26.29 nal, 39,55-60; financial reform and, Foreign-exchange-linked domestic 39-41,43-46,48-63; financial repres- accounts, 25,29-30, 32,34 sion paradigm and, 64-73,85; inflation Foreign exchange reserves: balance of and, 78-79; investment and, 45-46,48: payments and, 23,25; monetary frame- in 1980s, 12-13; savings and, 45-46,48. work and, 19,20-24,37 nn4, 5. Seealso See also Financial reform Net foreign a5sets Financial reform: in Argentina, 18,39-40, Foreign investment, 5, 12 46,48, 53; in Chile, 39-40,46,48,53; France, exchange rate and, 96 financial liberalization and, 39-41,43- Friedman, Milton, on policy, 1 46,48-63; in Korea,39, 40-41,46,52, 61, 64, 67-68, 70; in latin America, 39- Garcfa-Meza adnministration (Bolivia), 229 40,46,48-49; in Mexico, 85-88; in Gas. See Natural gas 1980s, 12-13; policy decisions for, 10; cATr. See General Agreement on laTiffs seigniorage and, 61; in Uruguay, 39-40, and Trade- 53. See aIso Deregulation; Financial lib- General Agreement on Tariffs and Trdce eralization; Fiscal reform; Regulation (cArn), 161, 162 Financial regulation. See Regulation Germany, tariffs in, 134,137 Financial repression, 12-13; inArgentina, Goods market, exchange rate and, 91-92, 41; in Chile, 41; credit market and, 49- 93,94 52,62 nn7, 8; description of,139-41,62 Growth, economic policy and, 1-28,-10, nn ,2; economic development and, 66- 60-61 73, effects of, 44-46, 48,62 n6; versus Guciler administration (Bolivia), 249 n I financial liberalization, 64-73,85; infla- tion and, 41,44-45,46,48; paradigm Health, Nigeria and, 173-74,204 of, 66-73; seigniorage and, 39,41, 41- 44,62 nn4, 5, in Uruguay, 13,41 iun. See International Monetary Fund : - - : . . . = . , - . \~~~~~~~~~~~~~~~~~W index 257 Import duties, 28; trade misinvoicing and, International capital markets, terms of 109,115 trade shocks and, 175 Importcd Inputs, tariffi on, 161-63, 170 International Monetary Fund (IMp): - n9 Nigeria and, 21 1; trade data and, 114- Imports: Bolivia and, 224, 240, 242, 244: 15 exchange rate and, 100-101, 103-5; Investment, 7-8; financial liberalization quotas on, 137-38,139-40, 143-44- and, 4546,48; interest rate and, 71 value added tax on, 108 nS 72-73; terms of Lrade shock and, 175, Import substitution: Bolivia and, 224, 227, 176, 187-88, 190. See alo Foreign 240: current status of, 10-1 1, value investment added tax and, 168nS Inward-looking economics, I 19, 145 Income, terms of tradc shock and, 174, Iran, revolution in, 172, 189, 198,207 176-77,180,181, 184-85, 188,191, 217 Israel: banking system in, 24-28,35-36; n2. See also Capital incomc; Real per budget deficit in, 24-25,27-28, 31-34; capita income i devaluation in, 55-36: exchange rte Income distribution: equitable policies and, 25,29-30, 31-32, 33, 4,36-37; fis- for, 2-3; exchange rate and, 94-95: in cal reform in, 24-28; incomes policies Latin America, 2; trade liberalization in, 34; inflation and, 25,27-28,32-33, and, 146847 35-36; intercst rates and, 25, 26-27,28, Indebted industrialization, Bolivia and, 30; monetary policy in, 18, 24-28,31- 232-34 37; reserve requirements in, 25; wnge Indexation: inflation and, 75-77; of tax policy in, 32 revenues, 77 . Italy: banking reforms in, 37 n8; exchange India: foreign currency black market in, rate and, 96. 170 nIO; protection and, 131 Industrialization: exchange rate and, 100, Japan: exchange rate and, 96; financial 101, 102: in Nigeria, 199 regulation in, 39; trade protection and, Infant industry protection, 134-35.1561 134 Inflation, 18-19; Argentina and, 65; Asia Joint-stock companies, 64 and, 73, 74: black market and, 118; Boi- via and, 220,222, 229-30,244,247; Korea, Republic of: exchangc rate and, 96; budget deficit and, 73-80,82-84, 85, external shocks and, 74; financial poli- capital flight and, 118; causes of, 77-79; cies in, 39,40-41, 46,52, 61, 64. 70; control of, 6, 10; debt service shock and, financial reforrn in, 67-68; real per cap- 78; financiaL liberalization and, 78-79; ita income in, 1; trade liberalization- in, financial repression and, 41, 44-45,46, 14; trade protection and. 142 48; indexation and, 75-77; Israel and, 25,27-28,32-33,35-36; Latin America Labor market: Bolivia and, 245: protec- and, 73-74, 77; monetary policy and, tionist policies and, 1334-4, 136. See 22; money creation and, 14-15, 75, 78; aLso Unemployment Nigeria and, 173, 200,205,206; Peru labor standards, trade policy and, 5-6 and, 65, 107; resource allocation and, Latin America: capital flight from, 80,83; 84 capital flows and, 55-58,59-60, 61; Insurance. See Deposit insurance devaluation in, 6; development banks Interest rate, 13; Argentina and, 10: Asia in, 52; econornic growth experiences of, and, 69; boond sales and, 22; capital 9; external capital in, 12; financial col- flows and, 3D, 36; economic growth and, lapses in, 53; financial reform in,'39-40, 71-72; exchange market and, 13-14; 46, 4849; financial re-striction in, 41; investment and, 71, 72-73; ISrael and; income distribution in, 2; inflation and, 25,26-27,28, 30; monetary policy and, 73-74,77; nonbank financial institu- 17,22,23,24, 30-31; savings rate and, tions (financieras) in, 56,61, 78-79; 69-70,71; seigniorage and, 43-44, 62 trade associations in, 239-40; trade mis- n5; United States and, 69 invoingin,117,118 trade prtection 258 rndex and, 129: trade taxes and, 113 , 1 16. See Moral hazard, macroeconomic instability atw names of countries and, 40, 53-55, 62 n I Lau ndering of assets, capital flight and, Multinational companies, taxation of, *l 111164-65 170 n13 "Lerncr theorem:' 169 nG Multiple exchange rate. See Exchange rate Liabilitics. See Financial liabilities Liberalization. See Financial liberaliza- Naira (Nigeria), 200,211 tlon;, adc liberalization Naturml gas, Bolivia and, 223-24, 230,236, "Luxury" taxes, 152 23, 243,247,248 NDc. See Net domestic credit Macroeconomic instability, moral hazard Nct domestic credit (Nbc), 30, 34-35, 37 and, 40, 53-55, 62 n I nIO Marcos, Ferdinand (Philippines), 1 10 Net foreign assets (NFA), 19,20-24,25,26, Markets: microcconomic policics and. 4- 28,29-30, 37 nn3, 4,9 6; protection and, 127,-129-30. See also New York exchange nate, 119, 120,123 Capital moarket; Goods market NFA. See Net foreign assets Mauritania, trade taxes and, 116 Nigeria: adjustment to shocks and, 212-. Mexico: capital night from, 57.80,85, 86, 15; agriculture in, 199-200; Babangida 87-88, 111, 115; doltarization and, 84, administration of, 198, 214-15; balance 86-87, 88; exchange rate and, 57,96, of payments and, 205, 209, 210-11: .119, 120, 123; financial reform in, 85- budget deficit and, 193: Bukari admin- 88; trade liberalization and, 14 istration of, 211-14; economic and Mining and minerals, Bolivia and, 223, financial background of, 198-215; cdu- -227,233 . cationand, 173-74.204: exchange rate Monctary authorities, 18, 19-24 and, 200,205,206,207,209,214-15; Monctary policy, 10,12; definition of, 17; expansionary policies in, 203-6,208- effects of, 17-18; exchange rate and, 17, 10: exports and, 172,217 nl; first oil.. 28, 37 nl; financial assets and, 18-21, boom (1972-78) in, 200-206, fiscal: 22-24,37 nn2-6, foreign exchange restraint in, 200,206-7,208,212-14; reserves and, 20-24, 57 nn4,5; frame- food production per capita in 173; for- work for, 18-24,37 nn2-6; growth and. eign borrowing and, 205; GNP per capita 60-61; inflation and, 22; institutional in, 172; health standards and, 173-74, structure and, 18; interest rates and, 17, 204; import quotas and, 21 1; industrial-', 22,23,24,30-31; in Israel, 18,24-28, ization in, A99: inflation and, 173,200, 31-37; open economy and, 17-38; 205,206; International Monetary Funa1 prices and, 17, 37 n 1; targets and tools and, 21 1; Iranian revolution and, 172, - of, 28-31. 37 nn9, 10; terms of trade 189,198,207; money creation in, 193; shock and, 192-98; timing of, 31, 61; money supply in' 193,205; oil and,. wage ratesand, 17,37 nl. See also Money 172-74,177,193,,198-215;217n1,218 supply - nnll, 13; output declines in, 173; per Money creation,7,10,13,20; Argentina capita investment in, 173; protection in, and, 78; Boliviaand, 14-15, 78; infla- - 199; second oil boom (1979-83) in,' tion and, 14-15,75, 78; Nigeria and, - 206-12; subsidies in, 199; ternm of trade 193. See aLso Seigniorage shock and, 14,172-74,177,1921 93, Money market, separation from credit 196,198-216,217 nnl, 8,218 nn9-13 market of, 54-55 - . Nominal economy, definition of, 17 - Money supply: banking system role in, Nominal exchange rate, 97 - - 19-24,37 nn3-6; control of,' 18; expan- Nonbank financial institutions: in Latin Sion of, 18-19; in Nigeria, 193, 205 America (financieras), 56,61,178-79; in Money velocity, 31 . - Mexico, 86, 87 Monopolization: market faiures and, 4; prices and. 130; protection and, 127, Oil: Bolivia and, 223,227,228,233,236, 130-31; rent seeking and, 131-32 243, 245; exchange'rate and, 102; Monopoly;profits, 165-67' Nigeria and, 172-74, 177, 193, 198-215, tindex 259 217 nl, 218 nnl 1, I3; Peru and, 102: and, 129; rent seekingand, 127,151-32: Vcnezucla and, 102. See also Oil shocks; resource allocation and, 128: Singapore Terms of trmde shock and, 142; "strategic" arguments and, Oil shocks: Bolivia and, 236, 238; Nigeria 135-36; Sweden and, 130; tarifli and, and, 172-74,177,198-215 137-38,159-40,141, 143,158-63; tradc Oligopolics: in banking systems, 24,28, liberalization and, 127,145-47, 158-61; 36; market failures and, 4 Lradc offsets and, 138, 139; United onse.c See Organization of Petroleum States and, 129, 134; wages and, 131. See Exporting Countries also Tariffs; ieade liberalization Organization of Petroleum Exporting Public sector, in Bolivia, 220, 229,230-52, Countries (oiEc), 172, 189 248 Outward-oricntecd economies, 144-45 Ovando administration (Bolivia), 228,231 Quasi-fiscal deficit: Argentina and, 79; exchange losses and, 79; Peru and, 79 Pakistan, trade taxes and, 116 Patam accounts. See Foreign-exchange- Real economy, definition -of, 17 linked domestic accounts Real exchange rate, 94, 97,125 n2 Paz-Estenssoro administrations (Bolivia), Real interest rates, 13 227, 229-30, 245, 247-48 Real per capita incomc economic policy Peru: capital flight from, 80, 82; dollariza- and, 1-2; in Korea, 1; in Zambia, I tion and, 82-83; exchange ratc and, 97- Real product per capita: in Argentina, 8, 99, 102, 107; inflation and, 65, 107; oil 9; in Spain, 8,9 - and, 102; quasi-fiscal deficit and, 79 Red tape, exchange rate and, 106, 125 n3 Petroleum. See Oil. Regulation (offinancial system): descrip- Policy, definition of, 1. .Se ae Economic tion of, 39-40,41,62 nir, 2; injapan, policy; Monetary policy 59; in Korea, 39, 46,61; in Taiwan, 39, Prices: exchange rate and, 28-29,92,93, 61; in the United States, S9,40 . 99; monetary policy and, 17,37 n1; Rent seeking: exchange rate and, 102, monopolization and, 130; wages and, 105-6; monopolization and, 131-32; 29. See also-Oil; Oil shocks protection and, 127,.131-32 Private sector, Bolivia and, 227,234 Required reserve ratios, 25 Privatization, deficits and, 7 Reserves, taxes on, 44, 62 n5 'Property rights, 3,4,8 Resource allocation: exchange rate and, Protection: arguments for, 132-36; con- 91, 102-5; inflation and, 84; protection tent restrictions on sales and, 138-59:. and, 128 corruption and, 147; cost-benefit analy- Retailing,yvalue added tax and, 155-56, . D. sis of, 141-42, 143-45; costs of, 127-32; 168n2 distortion of incentives and, 127, 128-! 29, 132; domestic distortions and, 132- Savings: financial liberalization and, 45- 34; European Economic Community 46,48; terms of trde shock and, 175, (EEc) and, 129; exchange controls and, 176,187-88,189,190 137,138; export subsidies and, 135-36; Savings rate, interest rate and, 69-70,71 extent of, 139-43; forms of, 136-39; Seigniorage: capital flight and, 83-84; Germany and, 134; import quotas and, dollarization and, 83-84; financial 137-38, 139-40, 143-44; India and, 131; reform and, 61; financial repression: infant industries and, 134-35,136; and, 39,41,43-44, 62 nn4, 5; interest. Japan and, 134; Korea and, 142; labor rate and, 43-44,62 n5; taxation of. market distortions and, 133-34,136; depositors and, 59-60. See also Money Latin America and, 129; market frag- . creation - mentation and, 127, 129-30;rmonopol- Senegal, trade taxes and, 116: ization and, 127, 130-31; Nigeria and, Siles-Salinas administration (Bolivia), 228 199: open economy and, 10-11; policies Siles4Zuazo administration (Bolivia), 229 for developing countries, 127-48; polit- . Simons's proposal, 54-55i ical economy of, 145-47; productivity - Singapore, trade protection and, 143 260 Index "Social market economy:' 4 Terms ortrade shock. adjustment to, 174- Spain, real product per capita in, 8,9 98,217 nn2-4, 6; consumption expen- Sri la nka. tmrde taxes and, H 6 diture and, 174, 175, 177-79,181-82, State capitalism, Bolivia and, 230-32 189; demand changes and, 175; eco- State enterprises, in Bolivia, 231-32,233. nomic policy and, 192-98; cxternal bal- 247 ance and, 195-98; financing versus Sumptuary taxes, 152 adjusting to, 177-79,182; foreign bor- Sweden, Lrade protection and, 150 rowing and, 174,175-76,177-79,180, Switzerland, governmentat intervention 181, 182, 184-85, 186-87, 191; income in,4 cffects of,174,176-77,180,181,184- 85,188,191,217 n2: international capi- ikiwan (China): financial regulation in, tal markets and, 175; invcstments and, .39, 6 1; financial restriction in, 40-41 175,176,187-88,190; monetary policy ThrifTs: black market and, 113 distortion and, 192-98; Nigeria and, 14,172-74,- of incentives and, 128; Germany and, 177,192.193,196,198-216,217nn1,8, 137; on imported inputs, 161-63,170 218 nn9-13;nonmonetaryaspectsof n9; labor market distortions and, 133; adjustmcnt to, 174-91,217 nn2-4,6; protection and, 137-38.139-40.141, production changes and, 175,178,179- 143,158-63: uniform, 152, 158-63, 169 82, 186,190-91;relativepricechanges nn4-8,170 n9; United States and, 137. and, 175,179, 180, 181,182-87,217 n3; See also Protection; Taxes; Tax reform savings policy and, 175, 176, 187-88, Taxamnesties, 168nl 18n, 90 Tlxes: Bolivia and, 113,228,232,249; cor- Tame deposits, 28 porate income, 164-67; equity promo- Tm, Bolivia and, 223,224,227,228,230, tion and, 4,7, 150-51;erosion of, 150, 23,3 236,247,248 151, 168 nnl,2; fraud and, 112-13,114; Torrez administration (Bolivia),228 indexation and, 77; multinational corn- Trade: Bolivia and, 220,222-24,227-34, panies and; 164-65, 170 nI3; reserves 236,238-40,242-45,247-49,.250 nn2- and,44, 62 n5. See also Tax reform;. 6: market policies and, 46. See also Trade taxes Exports; Imports; Markets; Protection; Tax incentives, 152 Tenns of trade shocks; Trade liberaliza- Taix reform, 11-12,14; capital income tax- tion;Trade restrictions; Trade taxes a:ion and, 163-67, 170 nnIO-13; Trade.liberalization, 14; balance of pay- enforcement of compliance and, 151; ments and, 145,146; Bgolivia and, 245;. "luxury" taxes and, 152; monopoly Chile and, 14; income distribution and, profits and, 165-67; moral dimensions 146-47; Korea and, 14; Mexico and, 14; of, 149-51, 168 ni; multinational comr- protection and, 127, 145-47, 158-61; panies and, 164-65, 170 n 13; policy transitional costs of; 145-46; Turkey consideTations for, 149-67,168-71 and, 14; unemployment and, 145,146. nn 1-13;sumptuary taxes and, 152; tax See alo Protection amnesties and, 168 nl; tax evasion and, Trade misinvoicing: aggregate costs of,. 150,151,152, 168 nnl, 2; tax incentives 118-19; Argentina and, 112; Brazil and, and,152; tax system design and,149- 111; capital flightand;-109-13,115; 50, 151-52; uniform tariff and, 152, exchangerateand,91,96,107 109-19; 158-63,169 nn4-8,170 n9; value added export duties and, 109, 113-14; export taxand,152-57,168nn2,3 subsidies'and, 109, 114; iniportduties -Tax systemn design, 149-50, 151-52 and, 109,1 13; Latin America and, 117, Terms of trade: Bolivia and, 224,228,230; 118. See also Black market exchange rate and, 190, 195-96, 198; Trade ofFsets, 138,139 relative policies and, 182-87. See also . Trade policy. See Protection; Tariffs; Trade; Tenns oftrade'shock -Trade liberalization Terms of trade disturbance. See Terns of Trade restrictions: black market and, I 09; trade shock . United States and, 65 tildex 261 lucle taxes: Africa and, 116; Asia and, financial scctor and, 154-55; on 116; Bolivia and, 113; Costa Rica and, imports, 168 n3; import substitutes and, 113; Latin America ancl, 1 13,116; 168 nS;retailing and, 155-56, 168 n2; Mauritania and, 116; Pakistan and, 116; strategy of, 157; Lax evasion and, 152; Senegal and, 116; Sri Lanka and,! 16 tax relurm and, 152-57, 168 nn2, 3 l'hansaction costs, 43; 62 n4 VAT. See Value added tax nlrkey, tradc Iileralization in, 14 Venezuela: capital flight and, 57; exchange rate and, 57, 120; oil and, Unemploymcnt: protcCLionist policics 102 and, 133-34;subsidicsand, 133, 134: trade liberalization and, 145,146 Wages: differcntials and, 133,134: in United States: financial regulation in, 39, Israel, 32; monetary policy and, 17, 37 40; governmental intervention in, 4; . n 1; prices and, 29; protection and, 131; interest ratc in, 69; tariffs and, 137; purchasing power of, 100, 101 trade protection and, 129, 134; trade restrictions in, 65 Yacimientos Patroifferos Fiscales Boli- Uruguay: capital flows and, 55, 56, 61; vianos (YrrB), 231,233, 243. financial reform in, 39-40, 46,48, 53; vrarz. See Yacimientos Pctrolfferos Fiscales financial repression in, 13, 41 Bolivianos Value added tax (vAt): coverage of, 153- 54; diffcrent lcvels for, 156, 168 n3; Zambia, real per capita income in, I