27549 tva? C,)~ ~ ~ ~ ~ ~~~~00 [i. m iSSi, . ..an GL and ass..2 ation I. , ! Oil Windfalls Blessing or Curse? A World Bank Research Publication Oil Windfalls Blessing or Curse? Alan Gelb and Associates C'TIOS5TE O9 Ecos4sI~J~lo u Published for The World Bank Oxford University Press Oxford University Press NEW YORK OXFORD LONDON GLASGOW TORONTO MELBOURNE WELLINGTON HONG KONG TOKYO KUALA LUMPUR SINGAPORE JAKARTA DELHI BOMBAY CALCUTTA MADRAS KARACHI NAIROBI DAR ES SALAAM CAPE TOWN I 1988 The International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Manufactured in the United States of America First printing October 1988 The findings, interpretations, and conclusions expressed in this study are the results of research supported by the World Bank, but they 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 Gelb, Alan H. Oil windfalls: blessing or curse? / Alan Gelb and associates. p. cm.-(A World Bank research publication) "Published for the World Bank." Bibliography: p. ISBN 0-19-520774-2 1. Petroleum industry and trade-Case studies. I. Title. II. Series. HD9560.5.G423 1988 338.2'7282-dcl9 88-25256 CIP Contents Preface ix Part I. Approach to Windfall Gains 1 Introduction 3 The Oil Windfalls: Expectation and Reality 7 Scope of the Study 10 Notes 13 2 Theoretical Approaches to the Analysis of Windfall Effects 14 Linkage Theory 14 Growth, Two-Gap, and Three-Gap Models 17 Export Instability Theories 18 Booming Sectors and the Dutch Disease 21 Notes 29 3 Absorbing Mineral Rents: The Record and the Policy Options 32 Do Minerals Promote Development? 32 Options for Absorbing Windfall Gains 37 Notes 45 Part II. Comparative Analysis 4 Introduction 49 Note 51 5 The Magnitude and Uses of the Oil Windfalls, 1974-81 56 A Simple Decomposition of Income and Spending 56 Size of the Windfalls 59 Use of the Windfalls: The Fiscal Response 61 Overall Absorption of the Windfall 70 v vi Contents Consumption and Permanent Income 71 The Financing of the Current Account 75 Notes 75 6 Sectoral Shifts, Real Exchange Rates, and the Dutch Disease 78 Real Effective Exchange Rates 79 Real Wages during the Boom and Relative Price Effects 83 Measuring the Dutch Disease 86 Notes 93 7 The Efficiency of Fiscal Linkage: Windfalls and Growth 94 The Composition of Public Investment Programs 94 Resource-Based Industry and Other Large Projects 107 Reasons for Poor Performance 109 The Drift toward Subsidies 117 The Growth Impact: Models and Performance through 1981 121 Notes 123 8 From Boom to Bust: The Downside of the Cycle and Asymmetric Macroeconomic Response 124 Adjustment Patterns 124 Modeling Growth and Adjustment during the Cycle 129 Notes 132 9 Summary of Findings 134 Note 144 Part III. Country Studies 10. Algeria: Windfalls in a Socialist Economy 147 with Patrick Conway Economic Development and State Control 148 Response to Higher Oil Prices 152 Adjustment of the Non-Oil Economy 159 Model: The Adjustment of a Fix-Price Economy to Oil Windfalls 163 Conclusions 166 Notes 168 11 Ecuador: Windfalls of a New Exporter 170 with Jorge Marshall-Silva Before the Windfall 171 The First Windfall 177 The Second Oil Boom 185 How Has Ecuador Benefited from the Windfalls? 191 Notes 195 Contents vii 12 Indonesia: Windfalls in a Poor Rural Economy 197 with Bruce Glassburner Politics and the Economy before 1973 199 Macroeconomic Responses to the Oil Windfalls, 1974-81 204 Adjustment in the Downturn, 1982-85 212 Avoiding the Dutch Disease 214 Distribution of Benefits 222 Conclusion 223 Notes 224 13 Nigeria: From Windfall Gains to Welfare Losses? 227 with Henry Bienen Nigeria before the Oil Boom 229 Policy Goals since Independence 235 Use of the Oil Windfalls, 1974-84 239 Some Effects of Oil Spending 249 Accounting for the Policy Response 256 Summing Up: Have the Windfalls Benefited Nigeria? 259 Notes 259 14 Trinidad and Tobago: Windfalls in a Small Parliamentary Democracy 262 with Richard Auty Before the First Boom 263 The First Oil Shock, 1974-78 266 The Second Oil Shock, 1979-81, and Its Aftermath 274 Reversing Oil Dependence: The Role of Gas 280 Assessment 286 Notes 288 15 Venezuela: Absorption without Growth 289 with Francois Bourguignon Economic and Political Background 290 The First Oil Windfall, 1973-78 293 The Second Oil Windfall, 1979-82 309 The Exchange Crisis of 1983 321 Conclusion 321 Notes 323 Bibliography 327 Index 349 Preface The research project underlying this book was started in 1981 and continued for several years, during which the temporary-even if quite long-lived-nature of the oil booms became more and more apparent. This required some changes in the focus of research, but the result is a more interesting study of the impact on economic development of the large export windfalls which accompanied the two oil shocks. The later chapters of this episode, however, are still to be written. Many more years will be needed before a final assess- ment, especially of the downside of the cycle, becomes possible. The study itself involved continuous interaction between the com- parative component at the core and the six country reports drafted by collaborators. A certain tension existed between the analysis of any one country and the comparisons made with other countries: aspects which appeared most illuminating in the comparative context were not always those which a country expert would choose to em- phasize. Therefore, although in most cases the country chapters fol- low the lines of the original reports, the author of the overall study cannot disclaim final responsibility for all chapters. Boubker Abisourour, Perry Beider, Shahrzad Gohari, Carlos Me- deiros, and Ann Meyendorff provided helpful assistance at various stages, and Lu Oropesa, Maria Ameal, and Raquel Luz processed the manuscripts beyond the call of duty. Arne Drud and Alex Mee- raus of the Development Research Department's Analytic Support Unit made a particular contribution in the setup and solution of the optimization model which underlies the results reported in chapter 8 and, in a broader sense, underpins the whole approach. Without the expert editorial help of Jeanne Rosen it would have been much harder to keep up writing momentum. In addition to the six experts who collaborated on the country chapters, Mauricio Garcia Araujo, Bela Balassa, Martha de Melo, Jack Duloy, Joan Nelson, Graham ix x Preface Pyatt, and two anonymous referees have contributed useful com- ments. So have many of the World Bank's country economists; their cooperation has been much appreciated. The collaborators on the country chapters deserve a special mention for their efforts and their patience over the period needed to put together this comparative volume. The book is dedicated to Robert Alexander Simon, who arrived just in time for galley proofs. PART I Approach to Windfall Gains I I Chapter 1 Introduction The dramatic Arab embargo of October 1973, which quadrupled world oil prices, set in motion a chain of events that was to have a great impact on the economies of virtually all countries, industrial- ized and developing, oil-importing and oil-exporting. Although some signs of strain in the global economic environment had emerged earlier that year-notably the monetary crisis in February, which led to the system of generalized exchange rate floating-the first oil shock signaled the end of a quarter century of high growth and relative economic stability. In dollar terms, oil prices slumped slightly in 1975-78. But they almost tripled in 1979 with the fall of the Iranian government and cutbacks in output by major producers. These two oil price shocks, charted in figure 1-1, have dwarfed all other terms of trade movements in the postwar period. Because oil represented the largest internationally traded commod- ity (having displaced coffee from this position in the late 1960s), the shocks caused unprecedented transfers of income between con- sumers and producers. Each of the abrupt increases in the price of oil brought about a shift from importing to exporting countries of $300 million a day.' Over a year, this represented the equivalent of one-and-a-half times all medium- and long-term debt owed by developing countries in 1973. The oil boom, moreover, has been of surprisingly long duration given the price movements usual for pri- mary products. Not until 1985 and 1986, when the market for oil from established producers contracted and the price of oil plunged briefly to about $10 a barrel before returning to about $16, did the purchasing power of these countries fall back toward the pre-1973 level. For all countries dependent on imported oil, the terms of trade shocks, followed by the "interest rate shock" and global recession, had severe adverse effects on inflation, employment, and productiv- 3 4 Approach to Windfall Gains Figure 1-1. Relative Prices: Oil, Manufactured Imports, and Non-Oil Exports of Developing Countries, 1960-84 700 - /' " 600 I oil - 500 i 400 - 300 - Developing countries' 200 - Developing countries' j manufactured importsa (100) non-oil exportsb ' j 0 1 1960 1964 1968 1972 1976 1980 1984 a. Unit value of manufactured exports from five industrial countries to developing countries. b. Unit value index for thirty-three commodities excluding oil, weighted by 1979-81 developing countries' export values. Source: World Bank data. ity. This set of issues has been extensively researched and will not be considered here.2 For countries with oil to export, the shocks and their effects were of course very different. It is the experience of the second group of countries-the apparent "winners" in the glo- bal oil drama-that is the subject of this study. How important were the oil windfalls to developing exporters? Did these countries merely consume their terms of trade gains? If not, what strategies did they formulate for using their new wealth to promote growth and development? Did the windfalls actually ac- celerate growth and lay the framework for sustained development? Did they enhance well-being and raise consumption? Or did the diffi- culties of managing producer economies through volatile, poorly pre- dicted terms of trade shifts nullify the potential gain, perhaps even turning it into a net loss? Is it in fact possible for a country receiv- ing a large windfall gain to end up less well-off than it might have Introduction 5 been without it? What lessons can be learned from these countries' experiences? Many more years must pass before economists can draw up a bal- ance sheet reflecting the full impact of the terms of trade windfalls on oil producers. Some of the projects funded by spending or borrow- ing against oil income cannot be expected to realize their intended benefits for a long time-as is true of development efforts in gener- al. Moreover, many projects that are now financial disasters could eventually prove to be profitable, if poorly timed, investments should the prices of energy and primary commodities regain lost ground. Nevertheless, enough time has elapsed to permit at least some preliminary answers to the questions raised above. If, upon careful analysis, the payoff to oil producers should appear to be small (espe- cially when judged by their own criteria for success), then the fluc- tuations in oil markets since 1973 may not simply have resulted in transfers between consumers and producers. They may also have engaged the global economy in a negative-sum game of monstrous proportions. In attempting to determine whether this is indeed what has hap- pened, this study considers the recent experience of six developing oil exporters-Algeria, Ecuador, Indonesia, Nigeria, Trinidad and To- bago, and Venezuela. These countries differ widely in per capita in- come, area and population, natural and human resources, political system, and the government's entrepreneurial and regulatory role in the economy. But they also share some important characteristics. For all of them, oil represents a large share of exports but a moder- ate share of gross domestic product (GDP). Oil income per capita is fairly low. Proven, low-cost oil reserves are high in value at aver- age post-1973 prices relative to the countries' estimated non-oil capi- tal stocks, yet are insufficient to bear the burden of financing develop- ment for more than another fifteen to twenty-five years unless real oil prices exhibit a rising secular trend. In 1974 the terms of trade oil windfalls for these countries totaled $22.5 billion, about 23 per- cent of their GDP (see table 1-1). Together, the six countries constitute the group of capital-deficit de- veloping oil exporters (also referred to as the high-absorbing or capital-importing oil exporters) that have long enough experience and comprehensive enough data to make a comparative study possi- ble. Mexico and Egypt, although sharing many features with this group, were not oil exporters in 1974. Iran is excluded because of in- adequate data after 1977. In contrast, the capital-surplus oil exporters such as Kuwait and Saudi Arabia have small populations, exceptionally underdeveloped 6 Approach to Windfall Gains Table 1-1. The 1974 Oil Windfall GDPa Windfallb Windfall (billions of (as percentage (billions of Country U.S. dollars) of GDP) U.S. dollars) Algeria 11.6 29.0 3.4 Ecuador 3.7 22.8 0.8 Indonesia 25.8 16.5 4.2 Nigeria 29.8 26.6 7.9 Trinidad and Tobago 2.1 30.4 0.6 Venezuela 26.2 21.3 5.6 All 99.2 22.7 22.5 a. 1974 GDP at market prices converted at annual average exchange rate. b. Windfall is measured according to the method in chapter 5. Source: World Bank, and International Monetary Fund (IMF) International Financial Statistics. non-oil economies (aside from activities financed by oil revenues), and very large, low-cost reserves that guarantee the comparative ad- vantage of oil for the foreseeable future. Their capacity to absorb reve- nue has been far lower than the maximum oil revenue they could extract, especially at the prices prevailing between 1974 and 1984. Their main medium-term economic problem, therefore, has been to choose an optimal asset portfolio of oil in the ground and financial and real assets abroad. The capital-deficit exporters also face a portfo- lio problem, as do all countries that can lend or borrow abroad. But in addition, their medium-run policy responses are crucial in determining the economic value of windfall gains and the effects of those gains on the non-oil economy. As indicated by statements of their governments, since at least the beginning of the 1970s these countries have been well aware that they could look forward to rela- tively few years of resource abundance. This study focuses on the medium-term management of windfall gains and reversals caused by changes in the terms of trade. It does not consider the implications of the discovery and exploitation of oil resources; all but one of the sample countries were well- established exporters by the time of the first oil shock. It also does not emphasize a third, and important, issue-the rate at which "ex- haustible" oil reserves should be depleted. The separation of the first and third issues requires that a distinc- tion be made between two concepts of economic rent. The first, whose magnitude, use, and impact is addressed in this study, is Ricardian rent, defined as the residual component of market price over production cost where all reproducible factors of production (labor, capital) are paid at market rates. This concept should be sepa- Introduction 7 rated from that of Hotelling rent, the equilibrium rise in the unit price of a nonrenewable resource at the real interest rate as the stock is exhausted. Ricardian rent is associated with resources of a superior quality (land particularly suited to the production of cer- tain crops, such as coffee or rubber), which may be renewable. Hotelling rent is associated with exhaustibility, and the Hotelling rule reflects arbitrage between holding the resource and holding interest-bearing assets. Given the geographical concentration of large, low-cost oil reserves and the increased dependence on a few suppliers that emerged during the 1960s and 1970s, it is not at all clear that price movements have reflected exhaustion on a global scale. It is equally plausible to explain them in terms of producer monopolies or backward-bending supply curves, as discussed in chapter 2. The Ricardian rent element in the price of low-cost oil has been massive since 1973-far in excess of the rent element in the price of any other major product. In the largest and cheapest fields, produc- tion and transport costs absorb as little as $1 a barrel, although the av- erage for the six countries in this study probably lies between $3 and $5 a barrel, with some extraction techniques raising costs to as much as $12 a barrel. In contrast, the average cost of extracting oil from Alaska's North Slope and delivering it to market is perhaps $8 to $12 a barrel, although the marginal cost is believed to be considera- bly lower. North Sea oil costs are still lower, but well above costs in the sample countries.3 The Oil Windfalls: Expectation and Reality Having long been the victims of an international division of labor which had relegated them to the classic dependency status of raw materials producers, the OPEC nations could now rede- fine their external relations as well as make fundamental changes in their internal models of capital accumulation. (Karl 1982, p. 3.) The 1973-74 oil price rise fundamentally altered the relation of the oil-producing countries to the international economic system. Oil producers were elevated to a distinct status among developing countries-in political influence, in trade matters, and in access to foreign capital. Expectations for their rapid economic development were little short of euphoric. Shah Mohammed Reza Pahlavi pro- claimed that Iran stood on the threshold of a "Great Civilization," while Carlos Andres Perez, president of Venezuela, announced his vi- sion of "La Gran Venezuela." And indeed, the windfalls relaxed three traditional constraints on economic growth: foreign exchange, domestic savings, and fiscal revenues. 8 Approach to Windfall Gains Not surprisingly, developing countries that lacked oil were envi- ous: The capital-short and aspiring Third-World planners have kept tell- ing themselves (and each other) that if only they had this black gold, the magical elan vital for their economic takeoff would be close at hand. (Amuzegar 1982, p. 814.) Yet as early as 1975, this optimistic view was being questioned: You think we are lucky. I don't think so. We are dying of indiges- tion . .. I call petroleum "the devil's excrement." It brings trouble. Look around you. Look at this locura-waste, corruption, consump- tion, our public services falling apart . .. And debt, debt we shall have for years. We are putting our grandchildren in debt. (Juan Pablo Perez Alfonso, one of the founders of the Organization of the Petroleum Exporting Countries [OPEC]; cited by Karl 1982, p. 18.) As time went on, prominent citizens of oil-exporting countries began to voice more and more doubts. History might show, they feared, that their countries "have gained the least, or lost the most, from the discovery and development of their resources" (Attiga 1981b, p. 7). The very titles of several studies are revealing: "The Years that the Locust Hath Eaten: Oil Policy and OPEC Develop- ment Prospects" (Levy 1978); "How Oil Revenues Can Destroy a Country" (Attiga 1981a); and "Oil Wealth: A Very Mixed Blessing" (Amuzegar 1982a).4 In his critical assessment of the Iranian experi- ence, Katouzian (1978) concluded that the main consequence of Iran's depleting its oil reserves was to squander its agricultural resources-a case of "dual depletion." Significantly, such views emerged well before the oil glut caused a steep drop in the reve- nues of the oil-producing economies. It is probably fair to say that those analysts taking the most criti- cal view of the uses and effects of the oil windfalls have empha- sized political and institutional factors more than is usual in quanti- tative neoclassical economic analyses. They have generally not attempted to quantify benefits and then assess them objectively in light of the potential offered by the windfalls, national priorities, and the desires of the recipients. To do so is important, however, since there is no a priori reason why certain outcomes usually de- plored (such as a drop in non-oil work productivity or the neglect of agriculture) might not be appropriate responses to windfall gains in certain circumstances. Some policymakers have regarded the ex- tent to which oil revenues have led to capital formation as virtually Introduction 9 the only criterion of success. Others have argued that the promo- tion or preservation of certain favored sectors of the economy (nota- bly industry) is an important criterion. The present study takes the view that in the long run the level of consumption and its distribution over time and across groups are the most important criteria for assessing the use of windfalls. Sectoral structure could be a factor in achieving a desirable consump- tion profile (for example, if there are learning-by-doing effects or if certain changes are hard to reverse) and is considered in later analysis. The level of investment is a relevant consideration in the short term, but it should not be considered irrespective of the ulti- mate payoff in terms of growth and, ultimately, consumption. In a quantitative neoclassical framework assuming perfect fore- sight and a consistent set of priorities, it is hard to turn windfall gains into overall losses without supposing very large divergences be- tween private and social benefits. Some analysts, accordingly, have adopted an intermediate position. On balance, they argue, it is diffi- cult to make the case that a revenue windfall, especially if fairly sus- tained, can have a negative overall value since it expands the op- tions available to the government and therefore to the entire country. This implies that the country can follow exactly the same policies as it would have done in the absence of the windfall and give the windfall away, for example, as foreign aid. Any different out- come will be, by definition, preferred. This argument is very power- ful, but it abstracts from changes in the international environment as- sociated with the windfall, which may make it impossible for the same policy to yield the same outcome as before. It also assumes that the windfall itself does not eliminate parts of the range of previ- ously feasible policies (for example, by changing expectations or insti- tutions), and in this respect abstraction comes at the expense of politi- cal reality. Further, certain sectors of the non-oil economy-especially those producing tradable goods-as well as certain groups heavily depen- dent on these sectors will be adversely affected by the windfall. Thus, in the absence of a costless, nondistorting system of compensa- tion, income distribution will be changed, perhaps in a way judged to be undesirable. Also, the adjustment costs of factor relocation, ac- celerated obsolescence of capital, and delays in regaining equilib- rium in labor markets may erase at least part of the potential bene- fit of the windfall. An intermediate view has probably become the dominant one, at least in those analyses restricted to a relatively narrow "economic" framework-windfall gains are "goods" not "bads," but with some adverse side effects. 10 Approach to Windfall Gains Scope of the Study In an effort to bridge the gap between institutionally based and quan- titative analysis, this study takes a broad look at the impact of wind- fall gains on six oil-producing countries. Emphasis will be on the size of the gains, the intentions of the governments, the way the gains were actually used and how this was affected by the coun- try's history and institutions, and the impact of increased spending on the non-oil economy-both on growth in general and on the per- formance of specific sectors. This approach requires some attention to macroeconomic, microeconomic, and political factors. It also re- quires some consideration of the strong association between the windfalls and changes in the global environment, notably the in- creased economic uncertainty in oil, commodity, and capital mar- kets since 1973. Windfall gains and rents have been approached from a number of analytical perspectives, and chapter 2 selectively reviews four strands of economic theory relevant in analyzing the exporters' expe- rience. "Linkage" theory stresses the nature of the demands engen- dered by a leading or booming sector, both those related to produc- tion (upstream and downstream linkages) and those related to final spending of income generated in the sector (consumption and fiscal linkages). More than the other theories, it leads to a consideration of the political and administrative factors that may reduce the effec- tiveness of the fiscal linkage that has been especially important for oil exporters. Neoclassical and related growth theory, in contrast, stresses the impact of the leading sector in augmenting the quan- tity of factors, both domestic and imported, as a determinant of growth. Export instability theories consider the effect of fluctuating and uncertain terms of trade on producer countries. As the empiri- cal chapters show, the stochastic and cyclical nature of oil windfalls has had important consequences. Prediction errors have led to seri- ous mistakes in the composition of investments across sectors and the location of industries among countries. Moreover, macroeco- nomic adjustment to rising and falling demand tends to be asymmet- rical, so that prediction errors in one direction are more costly than in the other. This can turn an expected windfall gain into an actual loss. "Ratchet effects", the irreversibility of policies initiated in the boom years, can have a similar result. Finally, "booming sector" the- ory stresses the impact of a booming sector on resource allocation in the rest of the economy. This approach raises the question of the appropriate degree of oil dependence and bears on strategic choices for absorbing windfall gains. In assessing the experience of the exporters, it is useful to have Introduction 11 in mind the long-standing and more general debate on the role of mineral resources and natural Ricardian rents in development-a de- bate to which this study makes a contribution. In most cases, it is not easy to distinguish the impact of mineral rents from the effects of other characteristics of mineral production. Moreover, certain nonmineral-producing activities may also generate sizable rents. The evidence on the impact of rents, which is surveyed in chapter 3, is somewhat ambivalent. This argues for approaching with an open mind the question of whether oil windfalls invariably prove ad- vantageous. Part I concludes with a summary, derived from the body of analyti- cal work, of the main strategic options for handling windfall gains. These options cover decisions on the primary uses of the windfall over time and across applications-consumption, saving abroad, do- mestic investment (private or public), and so on. They also include macro policies-such as trade restrictions, exchange rates, and con- trols on commodity and capital markets-that help to determine the ultimate uses of the windfall and the outcome. Part II is a comparative study of the evolution of the six oil-exporting economies both during and after the price shocks. In chapter 4, the introduction to part II, country characteristics are presented. A table also previews how the countries fit into the strategic options outlined in part I. The reader may wish to refer back to this table from the later empirical analysis. Chapter 5 de- velops and uses a simple methodology to estimate the size of the windfalls between 1974 and 1981 and how they have been used. Although the commonalities among the countries should not be overstated, in general most of the gains have been transformed into domestic capital in the public sector rather than consumed out- right or saved abroad. In fact, for a representative country, the split between consumption and other uses is not very different from what would be predicted by a simple permanent income model that allows for the depletion of proven oil reserves. This overall pattern of use reflects three important objectives of pro- ducer governments: * To accelerate growth of the non-oil economy * To modernize and diversify to reduce dependence on oil * To consolidate national control over the hydrocarbon sector and other industrial sectors The pattern also reflects a common desire, or political need, to spend rapidly. This precluded applications that would have required time-consuming institutional and social change; in fact, oil income may have reduced the ability of governments to perceive the 12 Approach to Windfall Gains urgency of such change. The result was a concentration on infra- structural and certain types of social investments and on large-scale industrial projects, since these are conducive to relatively quick dis- bursements. The following chapters examine the consequences of these wind- fall uses, using the above four theoretical approaches. Chapter 6 con- siders the impact on the structure of the countries' non-oil econo- mies and compares the outcomes with those expected from the "booming sector" theory. By and large, the predictions of this ap- proach are validated, although some notable exceptions point to the wide range of possible effects and to the importance of policy in influencing the outcome. Chapter 7 assesses the growth record of the six producers in rela- tion to both their own historical trends and trends in other coun- tries. Judged in terms of neoclassical growth and other, related theor- ies, the results are disappointing, even without considering the post-1981 period. Given the emphasis that the producer govern- ments have put on growth, this fact calls into question the effi- ciency of the crucial fiscal linkage between the oil sector and the rest of the producing economy. In particular, fiscal linkage has been ineffectual in developing competitive non-oil traded sectors. This critical weakness stemmed from a number of economic factors and management deficiencies. But the oil exporters were also in the difficult position of being impelled by their own internal political dy- namics to accelerate the implementation of long-range investment plans at a time when global uncertainty and volatility had greatly in- creased. The result was that large and vital industrial components of these investment plans performed far more poorly than ex- pected, not only because of poor implementation and management, but because market projections were off by wide margins. The fail- ure of these components to generate income and consequent expend- itures led, in turn, to the suboptimal use of new infrastructure. As a result of the interplay of these effects, capital formation did not ac- celerate growth beyond an initial demand-side stimulus, largely to construction. The slump in world oil markets after 1981 revealed just how frag- ile the development patterns of the exporters were in the face of un- expected fluctuations. Three of the countries studied experienced debt crises. Growth rates and demand fell sharply. And even though oil incomes were still far higher in real terms than before 1973, the oil exporters were propelled into stagflation. Chapter 8 re- views this period, for which data are still provisional, and analyzes the downside aspects of the oil cycle. It also summarizes the results of a simulation modeling exercise that combines elements of all Introduction 13 four theories-noted above. The results confirm that any upside bene- fits are easily negated by the costs of adjusting to falling revenues when macroeconomic behavior is asymmetric. They also suggest that incautious responses to overoptimistic oil price and export pro- jections can cost a great deal, particularly if inappropriate or reluc- tant adjustment policies are pursued on the downside. Chapter 9 summarizes the main findings of the study and tries to sketch an answer to the question: What should the developing oil ex- porters have done differently? Part III consists of six chapters, each a case study of one of the sam- ple countries. It provides sequential analyses to complement the issue-focused, cross-country treatment of part II. It also considers country-specific constraints on the deployment of rents. In princi- ple, natural rent is available for use in an "optimal" manner since it represents an income stream uncommitted to any reproducible fac- tor of production. Such revenues, therefore, should constitute the most highly valued income stream in the economy. One of the main findings of this study, however, is that oil incomes were com- mitted rapidly and allocated to uses determined by a country's preshock priorities and its political and economic institutions. In- deed, it is little exaggeration to say that most of the windfalls were committed before they were anticipated! To understand how and why this happened, it is necessary to study the experience of oil-producing countries in more depth; hence part III. Although the case studies follow a similar format, the treatment is not uniform. Differences in emphasis among chap- ters reflect important differences among the six countries in their re- sponses to windfalls and in their economic performance. To keep the study within manageable bounds, chapters 10-15 offer a fo- cused rather than a general economic history of each country dur- ing the years of the oil booms. Notes 1. Dollars are U.S. dollars unless otherwise specified. 2. In explaining the evolution of the main oil-importing economies after 1972, Bruno and Sachs (1985) attribute a more important role to supply price shocks than to reduced spending on investment or on research and de- velopment. 3. For an overview of extraction costs, see Adelman (1986a). Hughes and Singh (1978) discuss the incidence of rent in selected mineral extraction activi- ties. 4. The last-named paper also assembles critical commentaries and re- views of the performance of a number of oil economies over the boom period. Chapter 2 Theoretical Approaches to the Analysis of Windfall Effects Analysts disagree about whether growth and development are fur- thered or hindered by sectors with such an unusually strong compar- ative advantage that those who own the scarce factors of produc- tion garner large natural rents. In this debate, at least four theoretical approaches may be distinguished. Some understanding of these approaches is very useful when trying to assess the impact of oil windfall gains on an economy. Linkage Theory It has been observed that growth, particularly in its early stages, is likely to rely on a sequence of staple industries rather than just one. The "staple thesis," popular before the 1960s, attempted to show how a country's development could be shaped by a succes- sion of primary export products.' To answer the crucial questions of how, and under what circum- stances, "one thing would lead to another"-specifically, from a se- quence of staples based on natural resources to activities in which in- come would be generated by reproducible factors of production-a more structured theory was required. This led to the familiar but often misinterpreted concept of linkages.2 Although linkage theory, and especially attempts to quantify it, are commonly associated with input-output analysis, in its original form it emphasized the dynamic stimulus to entrepreneurship rather than a static framework of existing interrelations. Unlike the growth, two-gap, and "booming sector" theories surveyed below, it stressed that "development depends not so much on finding opti- mal combinations for given resources and factors of production as on calling forth and enlisting for development purposes resources and abilities that are hidden, scattered or badly utilized" (Hirsch- 14 Theoretical Approaches 15 man 1958, p. 5). According to this theory, in a given social, politi- cal, and economic context certain characteristics of the leading activ- ity are conducive to its providing such a stimulus. The effects of the interaction between the leading sector and other sectors are di- vided into production linkage, consumption linkage, and fiscal link- age. The first category, production linkage, is relevant to the proc- ess of increasing the density of the input-output matrix. It is further split into backward and forward linkages. Backward linkage refers to the potential for stimulating the growth of (upstream) sup- plying industries, forward linkage to the potential for stimulating the growth of (downstream) processing activities located in the same region or country. The importance of these backward and for- ward linkages (given the share of intermediates and processing in the final product price) depends on their technical requirements in re- lation to the endowment of the producing country. The strength of these linkages is usually considered to be inversely proportional to the "alienness" of the potential upstream or downstream activity to those who might be in a position to establish it. Hirschman (1981) also distinguishes between the stimuli provided to those engaged in the leading activity itself ("inside" linkage) and the stimuli of- fered to others ("outside" linkage). In contrast to other theories, linkage theory de-emphasizes comparative advantage and interna- tional trade and does not single out a few "fixed factors" as the main impediments to growth. Production linkages must, by definition, be potentially less signifi- cant for high-rent activities since intermediate inputs and returns to reproducible factors of production account for a smaller proportion of final value than in other products. The split between the poten- tial consumption and potential fiscal linkages then rests on the ex- tent to which surplus is taxed rather than appropriated by private owners. The actual stimulus to other activities reflects patterns of demand-which are influenced by the distribution of the surplus- and, in the longer run, the supply-side impact of investments made out of the surplus. Consumption linkages may be judged as favorable. or adverse from a developmental perspective. Adverse linkage is most often at- tributed to a propensity to import so high that it reduces global de- mand for the country's products and thus inhibits the development of other sectors. In a static analysis, this may not cause an imme- diate decrease in welfare. But even if some overall measure of well-being rises in the short run, it is possible for the consumption linkage to be considered unfavorable from a dynamic perspec- tive, usually because the shift in demand inhibits the growth of an infant sector.3 16 Approach to Windfall Gains Because oil windfalls since 1973 have overwhelmingly accrued to governments, the efficiency of fiscal linkage is the most important de- terminant of their ultimate benefits. If the objective is to consume windfalls directly, the focus is on the range of possible options for ex- panding public services and for transferring fiscal revenues to the pri- vate sector to enable private consumption to rise. Certain policy choices, such as negative indirect taxes and subsidies, tend to cre- ate allocative distortions. But direct transfers may be administra- tively impracticable, and it may not be possible to reach some groups at all at reasonable cost. If, however, the country attempts a supply-side development strat- egy funded by oil revenues, achieving an effective fiscal linkage is also likely to be difficult: The ability to tax must be combined with the ability to invest pro- ductively . .. In the case of [the more direct production and con- sumption linkages,] existing production lines or imports to be sub- stituted point to the tasks to be undertaken next, whereas no such guidance is forthcoming when a portion of the income stream earned in the enclave is siphoned off for the purposes of irri- gating other parts of the economy. Hence the possibility of either faulty investment, or a great deal of leakage on the way. (Hirschman 1981, p. 69.) For investments that are intended to produce commercially marketed (usually industrial) output, the problem is obvious. Governments are no more and are probably less likely to pick "1winners" than those whose livelihood depends on the outcome. But infrastructural investment, too, has its useful limits. Infrastruc- tural capital is usually created slowly and incrementally, and has normally been provided in response to demand emanating from other productive activities rather than in anticipation of demand in the hope of stimulating production.4 The task of usefully deploying windfall gains is arguably easier for the poorest countries because a wide range of physical infrastructure and human capital endowments is seriously deficient. But even when the state restricts itself to mak- ing good such deficiencies, the fiscal linkage at best permits, rather than stimulates, long-term growth. Fiscal linkage has another characteristic that is likely to weaken its efficiency. Surplus is concentrated in the hands of relatively few decisionmakers in the government, so that the constraint of inade- quate planning capacity rapidly becomes binding. As a result, the surplus tends to be used for large-scale projects. This may seem to minimize the administrative complexity of managing disburse- ments, but in fact it merely shifts the problem further down the Theoretical Approaches 17 line to the stage of project implementation. The labor/capital propor- tions required by large projects may also be inappropriate for the pro- ducer economy, since the criterion used is (immediate) administra- tive ease rather than such longer-run factor endowments. Several giant projects may then move a producing country, especially a small one, away from economic diversification toward dependence on a few key investments-leaving the country more vulnerable than before. In addition, public officials are not personally rewarded accord- ing to the economic soundness of their choices. This leaves open the possibility that a host of other considerations, political or finan- cial, will influence investment decisions. A large rent component in national income, if not rapidly and widely dispersed across the popu- lation, is liable to divert scarce entrepreneurial talent away from com- modity production into "rent-seeking" activities. This can manifest it- self in many ways. Public investments may be chosen not for any potential supply contribution, but to enrich contractors and politi- cians. Or there may be greater pressure for import protection to take advantage of the increasingly profitable home market. Incen- tives to maintain checks on the use of public funds by others may vanish entirely and be replaced by incentives to secure a share of the rent for oneself. Such rent-seeking activities, therefore, can be expected to lower the quality of domestic resource allocation. Com- mentaries on the experience of both high-income and developing oil-producing regions suggest that such "rentier economy" effects are indeed large in some of them although they are difficult to quantify.5 Finally, fluctuating fiscal revenues are likely to cause marked asym- metries in public decisionmaking. Investment programs started dur- ing a boom are hard to reverse; employees once hired are hard to dismiss. Asymmetric response is discussed below in the context of export instability, but here it should be noted that it weakens the effectiveness of fiscal linkage by lowering the average quality of public spending. Growth, Two-Gap, and Three-Gap Models Neoclassical growth theory characterizes output growth as a proc- ess of expanding a production possibility set, the frontier of which is set by the quantity (rather than the quality) of factors of produc- tion and by the efficiency of their allocation across activities. In the simplest formulation, growth is constrained by increases in the labor force and capital formation, hence by domestic savings. If labor and other domestic inputs are abundant, imports are an impor- 18 Approach to Windfall Gains tant complementary factor of production and, export revenues can- not easily be increased, foreign exchange can become a second bind- ing constraint on growth, as in the two-gap model. Further elaboration can lead to a three-gap model, in which development may also be constrained by a shortage of fiscal revenues. For this to happen, public funds must play a critical role-for example, in the process of capital formation, in ensuring access to foreign ex- change, or in relaxing bottlenecks to growth-and there must be con- straints on the ability to tax.6 Not much more needs to be said about these theories in the pres- ent context, except to observe that rent-intensive activities help to relax simultaneously all three types of constraint: domestic savings, foreign exchange, and fiscal revenues. To the extent that rents are taxed away and invested rather than consumed, these theories pre- dict a very favorable effect of windfalls on growth, especially if the do- mestic labor force is not a tightly binding constraint. Export Instability Theories This facet of analysis bears on the question of whether adverse ef- fects from the variability of oil income are likely to offset the bene- fits of temporarily high income. Typically, exports of developing coun- tries are more concentrated than those of developed countries and consist largely of primary agricultural and mineral commodities. The former, it is commonly argued, are price-inelastic in demand and, because of harvest fluctuations, are also subject to supply shocks which then induce large price swings. Mineral commodities are price-inelastic in both demand and supply, with demand very sensitive to economic activity in consuming regions. Cyclical de- mand fluctuations then induce large price and revenue shifts. Terms of trade variations, it is generally agreed, tend to be larger for developing than for developed economies,' and some studies sug- gest that mineral exporters are prone to wider fluctuations in ex- port prices and revenues than other developing countries. There is far less agreement, however, on the significance of such conclusions, and consequently little accord on whether it would be desirable to try and stabilize commodity markets, even if it were pos- sible to overcome the political difficulties involved in such an effort. There is also the perennial problem of how to identify the equilib- rium trend about which stabilization should occur. Newbery and Stiglitz (1981) regard the microeconomic arguments for stabili- zation, which center on the aversion of producers to risk, as quan- titatively insignificant. And indeed, no clear empirical relation has Theoretical Approaches 19 yet been found between growth rates and the degree of export stability.8 As conventionally measured, however, instability emphasizes short-run fluctuations, typically lasting one or two years. Possibly more important for the oil exporters have been the macroeconomic consequences of their far larger and slower fluctuations in terms of trade and fiscal revenues, and the effects of the resulting uncer- tainty on project selection in their public investment programs. Though very large, swings such as those in the oil market (as well as major shifts in steel, aluminum, and other commodity markets) are greatly attenuated by the filters used to compute conventional in- stability indices, which screen out oscillations of more than about four years' periodicity.9 For comparison, the periodicity of large oil market fluctuations since the early 1970s has been around ten years, with the overall boom of 1973-82 corresponding to a cycle of twenty years. These periods are long in relation to the fluctuations considered suitable for buffering, either by holding commodity stocks or by first accumulating and then drawing on international re- serves. But they are quite short for planning and executing major de- velopment projects and for carrying out large fiscal, sectoral, and macroeconomic adjustments to cope with huge swings in revenue and demand.10 There are also macroeconomic arguments in favor of stable ex- port revenues. These point to the well-recognized asymmetry of adjustment in response to fluctuations in demand." When domestic demand increases, supply is likely to hit capacity constraints. Infla- tion, real exchange rate appreciation, and rising imports then clear markets. But when demand decreases, unemployment is likely to rise either because of downward wage rigidity or because of sticky prices, with firms temporarily off their long-run supply curves.'2 Thus demand fluctuations (or intermediate price shocks) both raise average imports and lower average capacity use, output, and in- come. If savings and investment fall with income, this has an ad- verse impact on growth. If changes in oil revenue are manifested primarily in changes in the rate of public investment, a further growth argument for stabil- ity follows. There are costs in terms of quality-not easily measura- ble but apparently considerable-associated with very large shifts in the rhythm of investment. On the one hand, rapid growth of pub- lic spending is liable to reduce the quality of capital formation and raise costs because of more hasty planning, transport bottlenecks, and the need to use progressively more costly (or lower quality) fac- tors at higher growth rates. On the other hand, cutbacks mean 20 Approach to Windfall Gains costly postponement or cancellation, with partly completed ven- tures yielding no output. Even if they are later completed, delay will have reduced their rate of return. Finally, some policies and government programs put into place during the boom years may prove difficult to reverse as oil income falls. For example, restrictions on dismissing civil servants may in- duce a ratchet effect in the public wage bill. It may be politically diffi- cult to cut investments in the energy sector, even though declines in the world oil market reduce their profitability. Ratchet effects worsen the allocation of resources, and if sufficiently sustained they may prevent the reattainment of the pre-boom situation for a long time after the end of the windfall. The importance of considering changes in oil markets, rather than just levels of sales and prices, is heightened by the fact that oil mar- ket fluctuations have been very poorly predicted. Although the mar- ket had been tightening in the previous three years, the first oil price increase was not widely expected, and observers disagreed about its cause. At least four theories emerged, all with different im- plications for the future pattern of prices."3 The first argued that the increase reflected the gradual evolution since the 1960s of demand and supply conditions in favor of OPEC countries as reserve- output ratios fell, especially among the high-absorbing producers. The second stressed the shift of property rights from the oil multinationals to producer governments, and was sometimes in- voked to explain why the gradual evolution of demand and supply had not been reflected in a similarly gradual adjustment of prices. Both of these theories suggested that price increases would be perma- nent, but that prices would not continue to rise sharply once a new steady-state equilibrium was attained. Cremer and Salehi-Isfahani (1980) proposed a third theory based on the concept of "target revenue," which resulted in a backward- bending competitive oil supply curve. The higher world oil prices were, the lower the production levels a country needed to attain its revenue target. A shock to the system or a gradual outward shift in demand could then result in a jump from a low-price to a high- price equilibrium. Such an explanation points to a potentially vola- tile market with rapid price declines triggered by competition from new producers, which would increase the elasticity of OPEC'S de- mand curve.'4 A fourth theory was that the oil shock simply resulted from carteli- zation of the oil market; this allowed OPEC to institutionalize the sharp spot price increases that followed several significant political events threatening supplies. In this view, the path of oil prices would be set by the cohesiveness of the cartel. Some argued that Theoretical Approaches 21 OPEC would collapse fairly quickly like other cartels;'" but some, pointing to distinctive characteristics of oil production and of produc- ing countries, contended that OPEC could be sustained indefi- nitely. Theories of oil-price setting have continued to evolve. A recent one (see Adelman 1986b) departs from the proposition that oil export- ers tend to operate with short horizons and high discount rates be- cause their wealth portfolios are concentrated in a volatile oil sector rather than diversified. These high discount rates cause pricing deci- sions to be made on the basis of short-run, inelastic demand sched- ules for oil rather than the more elastic long-run schedule; this leads to a policy of "take the money and run" and to an inherently unsta- ble oil market. Whatever the theoretical views, by 1977 projections of prolonged scarcity were being taken seriously. It was asserted that by the mid-1980s OPEC production would have to rise by 50 percent, to 45 million barrels a day, or else real world prices would escalate fur- ther. The second oil price rise, which like the first was unexpected, was widely interpreted as reflecting the long-run depletion of an ex- haustible resource. Many official projections after 1979 foresaw a steady rise in real prices of about 3 percent a year for a long period rather than a ballooning for a relatively brief period,16 and some ex- porting countries factored even higher price increases into their na- tional plans. In the event, OPEC'S production in the mid-1980s was barely one-third of the level projected in 1977, and the price of oil fell to a third of its peak levels. Thus neither of the price increases was widely anticipated, nor was the oil glut of the 1980s. Nor were the policy responses of major consuming countries predicted-responses that significantly changed global scenarios between the first and second oil shocks. The stochastic nature of the windfalls and reversals, and the inade- quacy of predictions-not only of oil prices but of worldwide infla- tion, of interest and exchange rates, and of other commodity markets-must be borne in mind when accounting for the effects of the oil shocks. Booming Sectors and the Dutch Disease John Cairnes seems to have been the first to use this analytical ap- proach when, in 1857, he studied the effects of the 1851 Australian gold discoveries on other sectors of the economy.'7 The extensive the- oretical literature that followed has been surveyed by Corden (1984) and Neary and van Wijnbergen (1986); the treatment here is there- fore concise and selective. 22 Approach to Windfall Gains Unlike the three approaches already described, the "booming sec- tor" theory focuses on the sectoral reallocation of productive factors in response to a favorable shock emanating from either a resource dis- covery or an increase in the price of some commodity-typically an exportable one. If income is spent rather than saved abroad, the sum of the consequences includes a resource movement effect, which draws factors of production out of other activities and into the booming sector, and a spending effect, which draws factors of production out of activities producing traded commodities (to be sub- stituted by imports) and into nontraded sectors. The contraction or stagnation of the traded sectors is sometimes referred to as the "Dutch disease." 18 The key equilibrating variable in this process is the real exchange rate, the importance of which in the present context requires some elaboration. In some theoretical models of an economy, production is divided into "traded" and "nontraded" sectors where the former obey the so-called law of one price. This states that their domestic prices will equal their international prices because of the flow of goods across international boundaries, although tariffs may cause the prices to diverge by some specified amounts. The market for nontraded goods, however, must clear by domestic price move- ments. In this framework, the real exchange rate may be defined as the relative price of nontraded goods to traded goods, P,lPt. When domestic spending-power increases, an appreciation, or rise, in this ratio shifts production to the nontradables and demand to the tradables. This permits increased real income, in the form of higher absorption of nontradables and higher net imports of tradables. An alternative concept, sometimes termed the "real effective ex- change rate," is commonly used in empirical analysis because catego- ries of goods exhibit a continuum of tradability.19 It may be mea- sured by an index of domestic prices (for tradables and nontradables) relative to the prices of major trading partners con- verted at market exchange rates, pdlep*. If the price of traded goods pt is set by the level of foreign prices and a fixed tariff (so that pt = ep*[l + t]), the two concepts will clearly move together; real ex- change rate appreciation will imply real effective exchange rate appre- ciation.20 However, these two definitions of the real exchange rate may not al- ways point in the same direction. Suppose that domestic sectors pro- ducing tradable goods are heavily protected by tariffs or quotas that are then eliminated, at a fixed nominal exchange rate, in re- sponse to rising oil revenues. It is then possible for the domestic price level to fall (that is, the real effective exchange rate depreci- ates) even though the relative price of nontradables to tradables Theoretical Approaches 23 rises. When trade policy changes dramatically, caution is therefore needed in inferring the direction of domestic resource pulls from the real effective exchange rate. As discussed below, import liber- alization is, in fact, one of three policy options for limiting the infla- tionary impact of spending an oil windfall domestically. The others are nominal exchange rate revaluation and price controls with rationing."2 One other consequence of an appreciating real effective exchange rate is important if capital is internationally mobile. As domestic prices inflate at a constant nominal exchange rate, the real interest rate on foreign funds (that is, the nominal foreign interest rate de- flated by domestic prices) falls. This is likely to stimulate foreign bor- rowing for consumption and investment and further boost domes- tic absorption. Conversely, as oil prices and domestic spending fall and the real effective exchange rate depreciates, it is likely that capi- tal movements will reverse. It is more profitable to hold foreign as- sets than domestic ones, which experience the adverse impact of fall- ing demand on their returns. Alternatively, real exchange rate depreciation is expected to require nominal exchange rate devalua- tion and loss of value of assets denominated in domestic currency rel- ative to those in foreign currency. Unless domestic interest rates are allowed to rise sharply, capital outflows will, therefore, proba- bly accentuate the economic contraction. Such a pattern of capital movements, inflows on the upswing of the cycle and outflows in the downturn, accentuates the cyclical changes in demand owing to variation in oil revenues and is oppo- site to the pattern which would be desirable to stabilize the econ- omy over the cycle. Even if government saving abroad were to fol- low a stabilizing pattern, it is possible that this could be offset by procyclical private movements.22 A Simple Model Figures 2-1 and 2-2 present the essentials for analyzing the effects of a rise in the price of oil exports, followed by a fall. In figure 2-1, OWX is the domestic production set, the combinations of non-oil tradables and nontradables which the economy can produce. Equilib- rium is initially at A, with the real exchange rate (the price at which the tradables and nontradables can be exchanged) given by the slope of Po, where the collective indifference curve dd is tan- gent to the production set. A windfall equivalent to ZX in terms of traded goods shifts the ab- sorption possibility set outward to OWYZ. We assume away any fac- tor movements into the oil sector, and also abstract from growth.23 24 Approach to Windfall Gains Figure 2-1. The Oil Windfall Cycle P1 I0 w Doesi prdcin e PO W Nontraded goods With domestic production initially fixed at A, demand shifts to B and the real exchange rate appreciates sharply to P1, as the relative price of nontradables rises. This causes a reallocation of production toward the nontradables, and the economy shifts to production point E and demand point F. It is assumed that this shift is a smooth one along the boundary of the production set; we ignore pos- sible factor market rigidities which could cause it to move inside the set when demand is rising.24 The new real exchange rate is P2, Theoretical Approaches 25 Figure 2-2. The Oil Windfall Cycle with Government and Private Sector Distinguished P1 z ~1 BP2 A~ I F x Domestic production set \ 0 w Nontraded goods and at F the level of welfare is higher than at B, as shown by the indif- ference curve-the shift toward the nontraded sector is necessary to realize the full welfare gain. Now consider the adjustment problem posed by a collapse of oil prices that eliminates the windfall gain. Production must shift back from E to A and demand from F to A. During this phase of the oil cycle the demand for nontraded goods slumps, as does real in- come. Unless the economy responds flexibly through smooth real 26 Approach to Windfall Gains wage adjustment, and unless firms remain on their supply curves (so that relative prices absorb all the reduction in the demand), ad- justment will require a period of surplus capacity in the nontraded sectors such as along the path EJA. There may also be surplus capac- ity in the traded sector (point K) if, for example, imported inputs play a key role in production. Oil windfalls accrue mostly to governments, however, and this in- troduces some changes into the analysis, as shown in figure 2-2. In this two-agent model (government and the private sector) all of the windfall is assumed to accrue to government, which initially had zero demand. Government spends the windfall on traded and nontraded goods in the fixed proportion BD/CD. Initially, with pro- duction still at A, private consumers will be at C and the public de- mand vector is CB. When production has shifted to E private de- mand is at H and government demand is HF, larger than before because the reallocation of domestic factors toward nontradables has somewhat depreciated the real exchange rate from its peak (P1) and so reduced the cost of purchases by the public sector relative to oil prices. Note that the more the real exchange rate appreciates, the greater is the transfer of oil income to the private sector through the price mechanism, because the price index of its absorp- tion falls relative to the cost index of non-oil production.5 The fig- ure shows that the more intensive the nontraded good is in public demand, the greater is the extent of real appreciation. And typi- cally, real appreciation will be greater in the early stages, before the economy has been able to reallocate resources to the nontraded sectors.26 The overall value of the windfall is the integral of private welfare over the cycle ACHEJ, plus the value assigned to public absorption CB and HF, less the integral of private welfare at A over the cycle. It is likely to be lowest in the following circumstances: (a) public ab- sorption has a low value; (b) traded and nontraded goods are comple- mentary in private demand, and the private welfare function is strongly concave in income; (c) the windfall is transient and is fully absorbed, so that adjustment costs are raised; and (d) other policies also raise adjustment costs, especially in the downturn-so that the segment EJA is well inside the production set. Such policies could in- clude a reluctance to devalue in line with falling demand, which would make relative price adjustment more difficult between traded and nontraded goods. Some Extensions to the Model This simple model can be extended in a number of ways, toward a full simulation system of the type which underlies the analysis of Theoretical Approaches 27 chapter 8. Without going so far, some small extensions can make it more relevant to actual experiences discussed in parts II and III. CAPITAL FLOWS. As noted earlier, there may be a tendency for capi- tal inflows during the boom and capital outflows in the contraction to accentuate the cycle. The latter will be especially stimulated by in- terest rate controls, since the real exchange rate is seen to be unsup- portable. Capital flight when oil revenues are declining causes de- mand to fall below income, further decreases demand for domestic goods and accelerates and deepens the contraction along EJA. TRANSPORT AND DISTRIBUTION SERVICES. All imports require trans- port and distribution services before they can be absorbed domesti- cally, and these nontraded intermediate inputs may modify the con- clusions of the simple model. If the rate of growth of absorption is very high or if changes in its composition are large, congestion costs in transport and distribution systems are likely to rise sharply. This will be equivalent to a fall in the efficiency of the non- oil economy, which will be manifested in a rise of consumption and investment prices relative to non-oil production costs. This tends to offset the fall in absorption prices relative to production costs that usually results from real appreciation. Part or all of the oil income transfer effected through the real exchange rate is then dis- sipated through lower efficiency in transforming the new mix of do- mestic factors and foreign exchange into final output. INTERMEDIATE PRICE SHOCKS. So far, oil has been considered as a source of foreign exchange only, and not as an intermediate input. Once its latter role is included the sequence of intermediate price shocks will probably differ in oil-exporting and oil-importing coun- tries over the windfall cycle. On the upswing, given their ample fis- cal resources, the governments of oil-exporting countries are likely to resist raising domestic oil prices to world levels. To do so would shift resources from the private to the public sector at a time when the emphasis is on moving resources the other way. Holding domes- tic oil prices constant insulates these economies from the supply shock being experienced by oil-importing countries and reduces the tendency for the exporters' real effective exchange rates to appreci- ate against those of their trading partners. But on the downside of a cycle, falling oil export revenue will cause exporting governments to compensate by raising domestic oil prices. The non-oil economy of the oil exporter then experiences a supply shock and inflationary surge simultaneous with falling demand. Meanwhile, the opposite holds in oil-importing countries. The result may be a sudden real ef- 28 Approach to Windfall Gains fective appreciation of the exporters' exchange rate despite falling de- mand on the downside of the cycle. THE MONETARY DIMENSION. Finally, the model can be extended in the monetary dimension. Several studies, summarized in Neary and van Wijnbergen (1986), have analyzed the impact of resource booms in a monetary framework. In such studies, the supply of money normally enters, in a conventional way, as an independent policy variable. But the distinction between fiscal and monetary pol- icy is of limited value for developing oil exporters. Money is cre- ated when their governments run a "domestic deficit," defined as the excess of domestic expenditures over domestic (typically non- oil) revenues. In the boom years, domestic deficits are extremely large, and offsetting their effect on the money supply-by open- market operations or credit controls-is not feasible. Therefore the link between monetary and fiscal policies is close and inevitable. Nevertheless, monetary aspects can be important in several ways: * Before imports have risen sufficiently to offset the process of money creation initiated by the domestic deficit, the expan- sion of credit and money via the money multiplier may accentu- ate the boom if not choked off by higher reserve requirements or credit ceilings. * The government may crowd out private sector competition for scarce domestic factors of production through tightening cre- dit policy, if it considers that its own claims on the economy are paramount. * There is likely to be a fiscal impact from increased demand for real balances, which raises the inflation tax. This may be considerable because, in the boom years, oil economies are usu- ally inflating and rapidly monetizing their economies; sectors such as public services and other "modern" urban activities are growing fast, pulling labor from traditional, nonmonetized activities.27 COMBINING THE APPROACHES. These four strands of economic analysis-linkage, growth, instability, and "booming sector"-are not, of course, mutually exclusive. Indeed, they may be combined into a unified framework such as the one that underlies the simula- tions reported in chapter 8. They do, however, emphasize different aspects of the set of constraints, choices and policies available to countries faced with the difficult (if enviable) task of absorbing mas- sive windfall gains productively. Theoretical Approaches 29 Notes 1. Canadian economists, in particular, emphasized the staple theory. It originated with Harold Innes; see Watkins (1963). 2. Linkage theory was developed by Albert Hirschman; for a concise re- view see Hirschman (1981), especially chap. 4, "A Generalised Linkage Ap- proach to Development, with Special Reference to Staples." 3. A clear distinction between "absolutely adverse" consumption link- ages and those which are less favorable than they could be with lower im- port propensities is not always made. In a static model, real incomes could fall on account of the new activity if income distribution or other changes caused a large shift in aggregate demand away from goods produced by do- mestic factors of production. But such an effect may not be necessary for ad- verse consumption linkage. Although welfare may increase in the short run, factors of production could be drawn out of activities in which learning- by-doing externalities are important, possibly through the expansion of the nontraded sectors (see van Wijnbergen 1984a). One study which examines "displacement effects" is Resnick (1970). 4. The point is discussed by Hirschman (1981) who notes that Fishlow (1965) analyzes the growth of the U.S. railway system (a vital component of that country's development) as a response to demand from existing activi- ties. In constrast, Hunt (1973) argues that Peru's nineteenth-century guano boom failed to stimulate development because of inappropriate railway in- vestments, rather than because rents were dissipated through imports and remittances. 5. Rentier effects are not peculiar to developing countries. As oil reveenues rose in the state of Alaska, state representatives were inclined to give up monitoring each other's proposals for development projects in favor of lobbying for their own projects. The state program eventually frag- mented into numerous unassessed local programs. Rent-seeking and its costs are discussed more generally in Krueger (1974) and in Dervis, de Melo, and Robinson (1982, sec. 9.4). 6. The two-gap model was first elaborated by Chenery and Bruno (1962) and Chenery and Strout (1966). A three-gap model may be constructed along the lines of models sometimes used to analyze foreign borrowing; see Kharas (1984). 7. See, for example, Erb and Schiavo-Campo (1969). 8. Yotopolous and Nugent (1976, chap. 18) summarize the evidence then available. 9. Gelb (1979) analyzes the filters used in the construction of conven- tional instability indices. 10. Even the stabilization of short-term fluctuations may be academic. Davis (1983) analyzes the response of ten East African governments to export and fiscal fluctuations. His results confirm the impression that gov- ernments are ineffectual stabilizers because expenditures adjust rapidly to revenues. 11. For an evaluation of the importance of macroeconomic arguments for commodity price stabilization, see Kanbur (1984). 30 Approach to Windfall Gains 12. The model simulations in chapter 8 include the effects of these rigidi- ties. 13. Gately (1984) discusses these various theories of the evolution of the world oil market and describes their implications for the role of OPEC. 14. This was the process in operation in 1985 and early 1986 as oil prices plummeted from around $30 to $12 a barrel. 15. For example, see Milton Friedman, Newsweek, March 4, 1974, as cited in Gately (1984). 16. With hindsight, it is apparent that projections were unduly influenced by price increases in the immediately preceding period. See, for example, World Bank (1979, p. 35; 1980a, p. 6). 17. Bordo (1975) summarizes Cairnes's analysis. 18. Holland's claim to be associated with this theory followed debate on the impact of natural gas on the Dutch economy. The term "Dutch dis- ease" was apparently coined by the Economist on November 26, 1977. 19. See Dervis, de Melo, and Robinson (1982, chap. 7). 20. In some studies, the real effective exchange rate (or purchasing-power- parity-adjusted effective exchange rate) is measured as the price of foreign ex- change in domestic currency corrected for changes in domestic and foreign price levels, that is, the inverse of the present variable. The present mea- sure is preferred because it allows a natural association between the term "ap- preciation" and an upward move in the index. 21. Let Pd = a x pn + (1 - a) * pt be the domestic price level, where a is the share of expenditures on nontraded goods. The price of nontraded goods is pn and pt = e * p*(1 + t), where p* is the level of foreign prices of nontraded goods and e is the nominal exchange rate. Then, by substitu- tion, Pd = a pn, + (1 - a) * e p*(1 + t). If t and e are constant, Pd and p, move together. Assume that (a) there are no non-oil export sectors, (b) do- mestic oil prices are held constant, and (c) as world oil prices rise, t is re- duced. This introduces a divergence in the movements of Pd and pnIpt; the for- mer can fall while the latter rises. Reducing e can have a similar effect. For a discussion of price controls with rationing, see chapter 10. 22. Like other destabilizing speculative flows, this pattern of capital move- ments will probably not be profitable in the long run-unless the govern- ment absorbs losses of those who borrow abroad to finance domestic assets which later depreciate in value. 23. This analysis also disregards possible input shocks caused by rising domestic oil prices. In fact, most oil exporters stabilized domestic energy prices, at least until the early 1980s, so that input shocks were not impor- tant. 24. One such possibility, raised by van Wijnbergen (1984b), involves a downwardly rigid real wage and a consumption basket in which nontraded goods are heavily represented. In such a case, spending oil re- ceipts could lead to more labor being released in the traded sectors than is ab- sorbed in the nontraded sectors. This possibility is considered unlikely for cases in which real appreciation is actually caused by increased domestic spending, because it requires that investment and public consumption be so intensive in traded goods that wage earners derive none of the benefits Theoretical Approaches 31 of real appreciation. It is significant that this analysis was stimulated by the example of Britain, perhaps the one country where rising oil revenues co- incided with a government decision to cut back the role of the public sector in the economy. The sharp real effective appreciation of sterling, by 50 per- cent in 1976-80, has been attributed to tight monetary policy rather than to the spending effects from an oil bonus that represented, at best, about 5 per- cent of GDP; see Niehans (1981). 25. The private sector may be better or worse off at H than at C because the transfer effect and the gains from reallocating domestic production from A to E work in opposite directions. 26. This is still true in an intertemporal framework characterized by per- fect product and capital markets; see Neary and van Wijnbergen (1986). 27. A simple monetary extension of the basic model is as follows: let M, CG, CP, and CF represent the money stock, credit to the government and to the private sector, and net foreign assets of the banking system. For conve- nience, let the money multiplier be unity. There are no private capital flows. If d represents a change in a variable, dM = dCG + dCP + dCF = (G - T) - dO + dCP + (X - IM) + dO, where G is government spending, assumed to be all on domestic goods, T is non-oil (domestic) tax revenue, dO is an increment of oil revenue, which was previously zero. dO is exoge- nous and is assumed to accrue to the government, whose domestic deficit, G - T, together with credit to the private sector, are controls. Let P be the real effective exchange rate, that is, the price of domestic goods relative to im- ports. The private non-oil trade surplus is assumed to be a negative func- tion of dCP and P; we may then write: P = 1/a2 [d(G - T) + dCP(1 - a, - dM)] where a, is the net propensity of the private sector to import at fixed rela- tive prices, and a2 is the response of the private non-oil trade deficit to real ef- fective exchange appreciation. This equation indicates the dependence of the real exchange rate on fiscal policy, credit policy, import propensities (a1 and the extent of public spending on domestic goods), the substitutability of domestic and foreign goods (a2), and the willingness of the private sec- tor to hold money balances. It is possible to absorb the windfall with less real appreciation if trade is not restricted so that a2 is large, if public expendi- tures fall largely on imports, if private credit policy is tight, or if the pri- vate sector is willing to hold larger real balances. Chapter 3 Absorbing Mineral Rents: The Record and the Policy Options The composition of a country's output and exports changes continu- ally, so that it is not possible to label an economy definitively as "min- eral" or "nonmineral" on the basis of the share of minerals in GDP or trade. Despite this, categorization on the basis of general tenden- cies can be useful. Nankani (1979) finds that there were twenty- seven nations with a population of a million or more in which min- ing exceeded 10 percent of output and 40 percent of exports in the period 1967-75. These can, then, be considered "mineral econo- mies."' Twelve are producers of hard minerals, twelve are capital- deficit oil producers, and three (Libya, Kuwait, and Saudi Arabia) are capital-surplus oil producers. Do Minerals Promote Development? Mineral production is normally associated with the Ricardian rent which accrues to owners of reserves that are particularly concen- trated or that have low extraction costs. This is one of the main rea- sons for considering mineral economies as a distinct class of coun- tries. Under exceptionally favorable conditions, however, the production of some agricultural commodities is also able to gener- ate rents over and above normal returns to capital, labor, and land in alternative uses-sometimes only until the emergence of compet- ing producers. Any assessment of the role of rent in development should therefore be broadened to include countries other than min- eral exporters. At the same time, mineral production has some peculiar features: it is large-scale, enclave, and capital-intensive, usually with close links to multinational firms, often with high wages compared with the rest of the economy and with a high degree of uncertainty (Bosson and Varon 1977). These features may prove more impor- 32 Absorbing Mineral Rents 33 tant than the effect of spending out of mining rent in determining the evolution of the producing economy. The available evidence concerning the impact on development of rent in general, and of mineral rent in particular, can only be de- scribed as mixed. On the one hand, as described by the staple the- sis, rent-usually from a succession of activities rather than a single continuous one-has indeed played a leading role in the growth of such industrial economies as the United States, Canada, and Aus- tralia (Reynolds 1979). Profits from copper financed development in seventeenth-century Sweden (Heckscher 1963). An unusually strong comparative advantage in coffee provided funds and a source of de- mand to stimulate the growth of Brazil's industrial economy around Sao Paulo.2 The discovery of the gold fields of the Witwatersrand spurred the development of the modern South African economy, fi- nancing imports without which expansion would have been impossi- ble while creating demand for transport services, intermediate in- puts, and capital goods (Houghton 1972, chap. 5).3 On a more speculative note, the total of British foreign investments has been at- tributed to the windfall from Drake's circumnavigation in 1577-80, which paid a dividend of 4,700 percent: Indeed the booty brought back by Drake in the Golden Hind may fairly be considered the fountain and origin of British For- eign Investment. Elizabeth paid off out of the proceeds the whole of her foreign debt and invested a part of the balance (about £42,000) in the Levant Company; largely out of the profits of the Levant Company there was formed the East India Com- pany . . . If this is, on the average, a fair sample of what has been going on since 1580, the £42,000 invested by Elizabeth out of Drake's booty in 1580 would have accumulated by 1930 to ap- proximately the actual aggregate of our present foreign invest- ments, namely £4,200,000,000. (Keynes 1930, pp. 156-57.) Thus there is evidence that, at least in some cases, high-rent activi- ties (including piracy) have provided an important stimulus to growth. On the other hand, a positive causal link between high-rent activi- ties and development is certainly not inevitable. A lengthy litera- ture raises the possibility that the high-rent sector may inhibit the accumulation and upgrading of reproducible factors of production, and that in the long run this diversion of resources and attention can stultify growth. In the sixteenth century, discoveries of pre- cious metals in America had an adverse effect on Spain's domestic in- dustries by boosting wages above competitive levels (Keynes 1930) but a favorable effect on the industries of Holland-which was less Table 3-1. Investment and Growth Rates by Developing Country Group Hard-mineral Oil Other middle- Other low- exporters exporters income countries income countries Measure 1960-71 1971-83 1960-71 1971-83 1960-71 1971-83 1960-71 1971-83 Investment/GDP Mean 0.21 0.23 0.21 0.28 0.20 0.24 14.30 17.20 aT 0.06 0.05 0.10 0.10 0.05 0.05 4.10 6.10 Gross IOCR Mean 0.28 0.07 0.34 0.12 0.32 0.17 0.26 0.17 CT 0.05 0.02 0.06 0.05 0.02 0.01 0.03 0.04 Growth of GDP per capita (percent) Mean 2.5 -1.0 2.9 1.9 3.7 2.0 1.3 0.7 C 1.1 1.2 1.7 3.7 1.8 2.3 1.4 2.2 Number of countries 10 10 10 10 29 29 20 20 a. Incremental output/capital ratio. Source: World Bank data. Absorbing Mineral Rents 35 adventurous or perhaps less successful in its search for treasure (Wil- liams 1970). Before the Industrial Revolution, countries of northern and western Europe were considered to be poorly endowed with nat- ural resources compared with their neighbors in the south and east, who were soon to be outstripped economically (Southern 1952). And the most striking economic successes of the past several decades-Japan, the Republic of Korea, Singapore, and Taiwan- are all conspicuously poor in natural resources and rich in human re- sources. Historical evidence also provides some support for the thesis that mineral resources encourage countries to adopt highly leveraged, overextended strategies that render them more vulnerable to shocks: In seeking power and glory King Philip spent the combined pro- ceeds [of Spain's Peruvian and Mexican silver mines and reve- nues from the annexation of Portugal]-and more. He was the wealthiest monarch in Christendom, and went broke the most often. Accordingly, he paid increasingly high interest rates, be- cause his creditors realized that the risk had been transferred to them. (Adelman, 1986b, p. 25.) Table 3-1 considers the investment and growth performance, and the relationship between investment and growth, of four groups of developing countries: (a) the hard-mineral exporters, (b) oil export- ers, (c) other middle-income countries, and (d) other low-income countries. It covers two periods, 1960-71 and 1971-83. Capital- surplus oil exporters and countries with incomplete data are ex- cluded. The table indicates, first, the extent to which there has been a gen- eral structural shift in the relationship between investment and growth. In all groups, (unweighted) average investment rates rose be- tween the first and second periods. But this was more than offset by falls in the incremental output/capital ratios (IOCRS), so that aver- age growth rates deteriorated by a considerable margin. Second, the averages suggest that deterioration in the productivity of invest- ment (as suggested by the growth-investment relationship) has been more marked for the mineral exporting groups, which also ex- perienced the greatest variation in their terms of trade (see table 3-2). The hard-mineral exporters, whose terms of trade deteriorated sharply in the second period, almost stopped growing. The oil export- ers, whose terms of trade improved spectacularly and whose invest- ment rates rose the most, saw their average IOCR fall to barely one- third of its previous value. In the first period, the hypothesis that the mineral and nonmineral economies have identical IOCRS can- not be rejected; in the second period it can, at the 90 percent level. 36 Approach to Windfall Gains Table 3-2. Terms of Trade Indices Relative to Unit Value of Manufactures Imported by Developing Countries Metals and Period hard minerals Petroleum Agriculture 1960-62 100 100 100 1970-72 104 92 91 1980-82 78 636 84 Source: World Bank data. At first appearance, the mineral economies, whether facing favor- able or unfavorable terms of trade movements, have therefore experi- enced a more serious deterioration in the efficiency of domestic capi- tal formation than the nonmineral economies. In the hard-mineral economies this resulted in negligible growth. In the oil exporters, growth rates fell despite vastly greater investment. Mineral economies have also been alleged to exhibit certain unde- sirable structural characteristics. In the cross-section study of Adelman and Morris (1973) an abundance of natural resources, to- gether with the degree of dualism, was found to be an important ex- planation of the concentration of income. Labor endowments are far more evenly distributed than assets such as access to resource rents, so that an inverse long-run relation between resource abun- dance and income equality does not seem surprising.4 If the govern- ment is the major recipient of rent, much depends on the pattern of public spending-and this is often asserted to be urban-biased. One reason is political: mining and urban activities provide a natu- ral focus for efforts by organized labor to extract a share of rent.5 To complement this effect, mineral sectors enable countries to earn the foreign exchange they need to import food. Indeed, a charge fre- quently levied at mineral economies is that they neglect, if not sup- press, their (tradable) agricultural sectors, which suffer from overval- ued real exchange rates and public indifference. The long-run consequences for the distribution of income are predictable. Migra- tion to the urban or mining sectors is the only way to secure access to a share in the natural rent through protected employment, pub- lic services, or subsidies.6 More surprising, perhaps, is the observation that mineral econo- mies have not enjoyed particularly high levels of education and health care despite the strong positive association between a large mining industry and the ratio of tax revenue to GDP.7 Failure to allo- cate fiscal resources to social sectors may be related to the less cru- cial role played by reproducible factors of production in such econo- Absorbing Mineral Rents 37 mies. Mining income may lessen the pressure to ensure wide access to public goods and to develop human capital broadly across the population. The long-run impact of these tendencies on growth could be con- siderable. It is generally accepted that capital markets are not suffi- ciently perfect for individuals to fund their own education by borrow- ing against their expected future earnings. At the same time, microeconomic data often suggest that human capital formation has a high private rate of return, and there may also be large externali- ties in such areas. Neglecting the social sectors may therefore have an adverse long-run effect on growth that offsets any positive gains derived from the mineral sector itself. Options for Absorbing Windfall Gains The ultimate use and impact of windfall gains depend, first, on the policy choices adopted by the producer government. These include not only extraction rates and public spending, but also possible shifts in trade and other policies in response to changes in national in- come, which affect the distribution of the gains across the econ- omy. Second, they depend on how the macroeconomy responds to the surge in demand, which will create disequilibrium in some mar- kets at prewindfall output and prices. Table 3-3 presents a frame- work for analyzing these policy and macroeconomic factors, which is discussed in this section. Two underlying factors of a political and administrative nature should first be mentioned, since they affect the efficacy of any pol- icy choices that are made. The first is the "horizon" of public decisionmaking. This is likely to be closely related to the stability of the government. If leaders change often and in a discontinuous man- ner, it may be futile to attribute the policy responses of the country to any consistent set of objectives through the period of windfall gains. The second factor is the "unity" or centralization of public decisionmaking. When there are powerful, competing actors (such as autonomous public enterprises or an equally strong executive and legislature), inconsistent decisions can result in an outcome dif- ferent from, and inferior to, what would result if decisions were for- mulated by any single agent. Together, these factors largely determine whether a country can formulate and execute any coherent strategy at all. As described in part II of this book, the six countries in this study differed considera- bly in these two respects. The governments of Indonesia and Alge- ria had long horizons and centralized power. Those of Ecuador and Nigeria, however, were subject to frequent changes and powerful re- Table 3-3. Absorbing Windfalls: Policy Choices and Macro Variables Windfall Main subsi- Oil extraction Net savings absorption Public investment dies and rent Development rate abroad emphasis emphasis transfers emphasis (1) (2) (3) (4) (5) (6) Increase extraction Increase net savings Public spending Traded: hydrocar- Subsidies to firms Urban to sell more at abroad to reduce (consumption bons, other indus- high prices domestic demand investment) try, agriculture; ex- and build up for- porting, import- eign assets substituting Slow extraction to Increase net borrow- Transfers to private Nontraded: physi- Subsidies to house- Rural keep wealth in ing abroad against sector (subsidies, cal infrastructure, holds via direct the form of oil future oil income cuts in non-oil social capital (ed- transfers, subsi- reserves taxes) ucation, health) dized credit, and subsidies on com- modities (oil, food, etc.) Two-way flows Incentives to pri- Other considera- Regionally decentral- (public capital in- vate consumption, tions: scale, factor ized (possible role flows, private out- investment intensity of resource-based flows) industrial projects) Other Macro Variables and Policies Availability of Main macroeconomic Nominal exchange labor and skills Trade policy clearing method rate policy Large: labor supply can increase Open: clearing of markets Real exchange rate equilibrates In boom: revaluation to reduce in response to demand with- by allowing demand to spill nontradable markets (with inflation; constant nominal out large real wage increases over onto a wide range of or without nominal rate flexi- rate or depreciation imports bility) Small: labor force (or skills) is Closed: try to protect some Liberal import policy allows de- In downside: flexible rate to re- a bottleneck: growth of non- non-oil traded sectors at the mand to spill over onto for- duce recessionary impact and tradeds requires shrinking expense of others eign goods cushion budget; fixed rate traded sectors Labor and skills somewhat var- Price controls and rationing Large devaluations to try to iable through immigration to limit spending on domes- restore "real" oil income policy tic goods 40 Approach to Windfall Gains gional pressures. Trinidad and Tobago and Venezuela occupied an in- termediate position, though the factors involved were quite differ- ent in the two cases. Absorption Policy Choices For convenience, absorption choices may be arranged as a hierar- chy, even though in reality (and in an optimizing decision frame- work) some of these decisions would be made simultaneously. Such a choice hierarchy is set out in table 3-3. 1. The first policy choice concerns the rate of extraction. In re- sponse to a surge in world prices, should oil be kept in the ground or should reserves be depleted more rapidly? In view of the difficulty of abruptly changing the rate of depletion, this choice may be somewhat constrained. If the country can lend and borrow abroad as much as it desires at a given interest rate, oil will be left in the ground if the rate at which its value is expected to rise exceeds the interest rate. If, however, access to international funds is limited, the country might raise its extraction rate to bring forward consumption, even though the price of oil was expected to rise more rapidly than the interest rate. Optimal extraction then also depends on the opportunities for domestic investments. 2. The second choice concerns the proportion of revenues to be ab- sorbed at home as opposed to being saved abroad. Producers are likely to have a wide range of options, with excellent ac- cess to foreign loans as oil exports soar. To hold total domestic demand to the desired level, especially when the perceived cost of foreign funds falls with real effective exchange rate ap- preciation (as noted in chapter 2), the government may have to constrain borrowing abroad by public enterprises and the pri- vate sector. Large two-way capital flows are possible, either with government saving abroad and the rest of the economy borrowing, or the reverse. 3. The next policy choice is the broad type of absorption. Should oil income fund public investments, public consumption, or transfers to the private sector? The last of these could be imple- mented in many ways-by reducing non-oil taxes, granting sub- sidies to selected activities or groups, or subsidizing certain products. Most of the practical possibilities (such as a subsidy on oil for domestic use) are distortionary in that they cause prices to diverge from world levels or from costs. At first Absorbing Mineral Rents 41 glance the policy that seems to promise the least distortion is to reduce non-oil taxes, particularly on traded goods. Indeed, such a policy promises to decrease price distortions. The tax bur- den is borne by rent, a totally inelastic income source, but this leaves the budget dangerously dependent on volatile oil re- ceipts. It may also be possible to influence the composition of private demand through the method used to effect the trans- fer, so that the choice of how to absorb oil income extends be- yond the public sector. 4. A further set of choices relates to the thrust of public invest- ment to be funded (or encouraged) by oil revenues. Two broad strategies for "sowing the oil" to strengthen the non-oil economy can be identified. One approach is to emphasize the non-oil tradable sectors-conventionally, agriculture (both food and nonfood), manufacturing, and natural resource process- ing. This approach includes a wide range of suboptions: Is the focus to be on activities that are large-scale and capital- intensive or small-scale and labor-intensive, import-substitut- ing or exporting, urban or rural? A second approach is to em- phasize key infrastructural sectors that are generally nontraded. These include physical capital largely serving pro- duction (transport, communications, and power supply); physi- cal "social" capital (housing); and human capital (education, health). Such investments do not themselves create activities that can substitute for the foreign exchange provided by the oil sector. Their contribution to economic diversification and growth must lie in the stimulation and greater efficiency they promise for the future production of traded goods. The first strategy is therefore a more direct attempt to diversify out of oil, the second an indirect route that can be expected to take longer to end the dependence on oil. How the government balances these two types of domestic investment has other important repercussions as well. First, some investments-particularly in large-scale, resource-processing indus- try-have high import requirements, whereas others-generally in housing, infrastructure, and education-rely primarily on domes- tic resources.8 Second, certain investments-notably those in the social sectors-entail high recurrent costs and so impose future claims on fiscal revenue, whereas others-notably in the industrial sector-will, in theory at least, return a profit to the treasury.9 The spending pattern can therefore affect both the intensity of the spend- ing effect on domestic goods and the time profile of non-oil fiscal revenues and expenditures. To assess the appropriate pattern 42 Approach to Windfall Gains for a given country is a complex problem to which there are no easy answers. 5. The method of rent transfer to recipients has many potential dimensions. Table 3-3 notes two methods-direct transfer or commodity (or credit) subsidies-and two classes of potential recipients, firms and households. 6. A final aspect of absorption policy noted in the table is whether the emphasis is urban or rural, and whether regional decentralization (perhaps through the establishment of new industrial centers) is an important objective. Macroeconomic Impact and Tradeoffs The response of non-oil output to increased domestic spending (be- fore supply-side effects of higher investment increase capacity) de- pends mainly on the degree of slack in the economy. This is the first macro variable identified in table 3-3. If labor (with necessary skills) is in elastic supply at a given wage, non-oil growth will be vig- orous and driven by demand, and the real exchange rate will appreci- ate less because of high supply elasticity. Any contraction exerted on the non-oil traded sectors will be small because these do not need to give up labor to the booming sectors. But if labor markets are initially tight so that the non-oil economy is operating on its pro- duction frontier, the spending effect imparts less of a growth stimu- lus. Expansion of the nontraded sectors requires contraction (or, in a dynamic context, slow growth) of the non-oil traded sectors and a rise in the relative price of the nontradeds.10 This raises the problem of the sectoral tradeoffs posed by the ab- sorption of oil rent, a topic which has aroused considerable concern in producing countries. Is there a case for protecting some or all of the non-oil traded sectors from the adverse resource pulls caused by an oil-fueled spending effect? And, if this is desirable, how can it be accomplished? Whether a given sector expands or shrinks during the boom de- pends on three characteristics: (a) its degree of "tradability," (b) the absorption elasticity of demand for its output, and (c) its factor re- quirements relative to those of the most rapidly growing sectors (typi- cally construction and government services). Trade policy alone, the first and most obvious option (which avoids driving a wedge between domestic producer and consumer prices) cannot protect all sectors simultaneously. Higher barriers against manufactured im- ports, for example, will simply be reflected in greater appreciation of the real effective exchange rate since the range of traded prod- Absorbing Mineral Rents 43 ucts becomes narrower. Such a policy would further depress agricul- tural exports and therefore shift the burden of adjustment from one sector to another rather than protecting all traded sectors. The range of trade policy options is noted in table 3-3. A second option is "exchange rate protection," a possible strat- egy for protecting all sectors producing non-oil traded goods (Cor- den 1981c). This rests on saving abroad instead of spending domesti- cally at the rate that would be desirable were the non-oil tradables not a concern. A third option is to launch a public spending program that is more import-intensive than would otherwise be optimal. The cost of this and of the previous strategy lies in their otherwise subop- timal profiles of rent absorption, either over time or across uses. The cost must be offset against the presumed benefits to the non- oil traded sectors. A fourth option is to use oil income to subsidize the non-oil traded sectors. This could be done, for example, by imposing nega- tive indirect taxes on a specific product or its inputs. In addition to ad- ministrative complexities (such as those associated with a negative in- direct tax on food) this strategy has the limitation that the income injected into the economy through the subsidies itself causes a spend- ing effect which pulls resources toward the nontraded sectors. Offset- ting this entirely requires either taxation of the nontraded sectors or a cutback in public demand, to below the level before the wind- fall, to provide additional funding for the subsidy."1 The unavoidable conclusion is that it is inconsistent to try and im- plement a major development plan and at the same time protect all traded sectors. In the long run, however, these sectors may bene- fit from investments or productivity gains made possible by oil income. Now consider the case for using windfall gains to support some of the non-oil traded sectors. In general, permanent subsidization may be justified only if the social value of an activity diverges from the private value-reflecting, for example, distributional concerns or externalities due to some market failure. Can an oil windfall make a case, or strengthen an existing case, for subsidies based on such considerations? This question has been raised by van Wijn- bergen (1984a). Neary and van Wijnbergen (1986) distinguish two possibilities for a windfall whose time profile is known with cer- tainty. In the first, the country faces perfect capital markets. The time profile of absorption and its allocation between consumption and domestic investment are optimal in the sense that they maxi- mize some discounted sum of welfare over the entire period. If a learning-by-doing externality is introduced in the traded goods sec- 44 Approach to Windfall Gains tor, so that future productivity depends on current output levels, it can then be shown that the optimal subsidy is higher if, along the op- timal path, the real exchange rate depreciates in the future. Their second possibility, which may be more relevant empiri- cally, is that the country faces imperfect world capital markets. The government does not have the savings instruments to sterilize tran- sient windfall gains, and the private sector is not able to save, then dissave, abroad to offset changes in the level of public spending. The subsidy to tradables then represents a form of investment in rais- ing future levels of productivity. Its costs, notably those of distor- tions during the phase of high oil revenues, must be set against the benefits of increased future output. The social value of these bene- fits will typically be higher if real income subsequently falls, possi- bly due to a decrease in oil income. A third and perhaps even more relevant scenario, however, is that no real obstacle exists to saving the appropriate portion of oil in- come abroad at given (not necessarily constant) interest rates, yet the government faces political constraints to running large, cumula- tive budget surpluses. Suppose that the private sector cannot save or borrow abroad on a sufficient scale to offset the profile of public spending. Or, suppose that it could do so, but that individual agents do not have a sufficiently strong incentive to save abroad dur- ing the peak years of oil income because certain costs of rapid spend- ing are borne by society and are therefore external to the individ- ual."2 In either situation, any argument for subsidies that appeals to the windfall is second best; the optimal policy would be to smooth the absorption profile over time. The column headed "Main macroeconomic clearing method" in table 3-3 shows the strong interaction between trade policies and the way in which markets reach equilibrium in the booming pro- ducer economy. If imports are limited by policy or low natural substi- tutability, the real exchange rate plays a major role. Liberalizing trade can permit demand to spill over onto imports by broadening their coverage. Rationing mechanisms can also be used, as further ex- plained in the empirical chapters. Inflation, Trade, and Exchange Rate Policy Another set of choices is posed by the interaction of inflation, trade, and nominal exchange rate policy (the last column at the bot- tom of table 3-3). As noted earlier, with a fixed exchange rate and with trade policy unchanged, appreciation of the real exchange rate means appreciation of the real effective exchange rate, and is effected through domestic inflation. If price stability is an important policy ob- Absorbing Mineral Rents 45 jective, nominal exchange rate revaluation and trade liberalization are two ways of counteracting the inflationary effects of the boom. The more open the economy (in terms of larger import shares and fewer impediments to substitution between domestic and foreign goods) the lower will be the inflationary stimulus. Protectionist poli- cies to shield some traded sectors (at the expense of others) will there- fore also worsen inflation unless the currency is simultaneously reval- ued. Monetary policy can have some impact on inflation but, as noted earlier, its impact will be quite limited. On the downside of the oil cycle, the factors influencing the infla- tion rate are complex. In particular, they reflect the rate at which the government curbs expenditures relative to its falling oil revenues and foreign borrowing ability, and the extent to which additional revenue- raising measures entail higher domestic prices. The least inflationary combination of policies involves sharp cuts in public spending, in- creases in direct taxation, and moderate devaluation to stimu- late the traded sectors. The most inflationary choices include large, domestically financed budget deficits, tightened import controls, or massive devaluations intended to restore fully the domestic purchasing power of oil taxes. 13 As shown in later chapters, such responses can have a severely adverse effect on economic performance. Notes 1. Some very small economies are highly dependent on minerals. A fasci- nating account of the island of Nauru, where phosphate sales represented $25,000 for each of its 5,000 citizens, appears in the Wall Street Journal, Sep- tember 22, 1983. The Sultanate of Brunei is a good case of an oil ministate out- side the Gulf. 2. As indicated by high rates of taxation, coffee prices have frequently em- bodied a large component of natural rent. Furtado (1963, chaps. 30-32) de- scrihes the substantial role of coffee in Brazilian industrialization. A classic contrast is between the effects of coffee production and those of sugar pro- duction. Coffee, partly through its propensity to exhaust the soil and in- duce migration to new areas, stimulated commercialization, transport sys- tems, a processing industry, and eventually the development of the tenth largest industrial complex in the world in Sao Paulo. Sugar, geographically static and with less immediate links between production, processing, and commercialization, gave rise to a poor, densely populated region (Dean 1966). 3. Although gold has retained its key role since the early twentieth cen- tury, there has been a steady diversification of output and exports since the 1940s, at least until the world gold boom in 1981 raised the share of gold in exports back toward its 1946 level of 56 percent. 46 Approach to Windfall Gains 4. Nankani (1979) discusses the relation between mineral endowment and dualism. 5. One example is the powerful Venezuelan labor movement formed in the oil workers' camps in the 1920s and 1930s. 6. The "dual economy" case of Zambia, a mineral economy with good agri- cultural potential that is less exploited than that of its mineral-poor neigh- bor Malawi, is discussed by Baldwin (1972) and Barber (1961). 7. Nankani (1979) points to the lagging school enrollments in the mineral economies. Tait, Gratz, and Eichengreen (1979) and others who have made cross-country comparisons of fiscal effort note the overall strongly positive contribution of mineral revenues, even though this is often partly offset by lower tax rates on nonmineral sectors and incomes. 8. The direct and indirect import content of heavy industrial investment may be as high as 80 percent of total cost, whereas direct imports for a vari- ety of infrastructural projects may be close to zero. In the countries in- cluded in this study, over 60 percent of investment in the 1970s was of do- mestic origin, even though these countries lacked well-developed capital goods sectors other than construction. 9. Heller (1975) estimates the ratios of annual recurrent costs to capital costs. For some investments, these ratios may be up to 20 perent. 10. The economy will respond in a Keynesian manner to higher de- mand, if labor supply is elastic, rather than the neoclassical mode of fig- ures 2-1 and 2-2. 11. Using a computable general equilibrium model of Indonesia, Gelb (1985a) simulates policies to shield three traded sectors-export agriculture, food, and manufacturing-from the consequences of spending an oil wind- fall equivalent to 8 percent of non-oil GDP. The government maximizes pub- lic investment subject to the rate of return in those sectors not falling rela- tive to the mean rate of return in the nontraded sectors. This yields an optimal food subsidy of 19 percent, an agricultural export subsidy of 26 per- cent, and indirect taxation on manufactures of minus 5 percent. However, total public absorption has to be cut back to 78 percent of its prewindfall level. 12. Congestion costs in the upswing are one example of externality. On the downside, chapter 8 of this study confirms that an important argument for smoothing the time profile of absorption is likely to be the macroeco- nomic cost of adjustment, in terms of negative or slow growth. Such costs should be taken into account in setting the time profile of public spending, but they are external to individual private agents and therefore will not af- fect the private spending decision. 13. Devaluation may temporarily restore the purchasing power of oil reve- nues by cutting the price of nontraded goods relative to oil. Because the ag- gregate resource constraint has tightened, however, continued public spend- ing must be offset by cuts in private spending. The alternative to increasing non-oil taxes is therefore inflation. PART II Comparative Analysis li Chapter 4 Introduction Having reviewed some relevant theories, described the expectations raised by the windfalls, and provided some indication of the range of policies which may be adopted to absorb them and maintain mac- roeconomic equilibrium, we now proceed to compare the experi- ences of the six countries. In chapter 5 we develop a simple meth- od of decomposing changes in national accounts data so that the size of windfalls relative to nonmining GDP can be estimated. Then we consider the uses of the windfalls over the period 1974-81, first from the viewpoint of the fiscal response and then in terms of ab- sorption by the economy as a whole, using the same decompo- sition method. Chapter 6 analyzes the "booming sector" effects of absorbing the windfalls. It deals with real exchange rate adjust- ments and intersectoral resource shifts between non-oil tradables and nontradables. The supply of agricultural products in the oil exporters is also considered because of the sector's importance and vulnerability. Chapter 7 discusses the impact of the windfalls on growth through 1981. Closely related is the question of the efficiency of fiscal linkage: the composition of the massive public investment programs in each country and the reasons for their apparently low payoff are key elements of the story. As is apparent from the dis- cussion in chapter 3, the growth performance of the oil exporters was disappointing, even before the impact of the oil glut began to be felt. Data covering the oil glut years are still incomplete and subject to revisions. Therefore the period 1980-84 is treated separately in chap- ter 8 with only the broad outlines presented. In brief, the six oil ex- porters entered a phase of stagflation (much as the oil-consuming countries had when oil prices rose). This asymmetric response fur- 49 50 Comparative Analysis ther reduced their overall gain fronr the oil cycle. The volatility of oil prices is central to the analysis of this chapter. Chapter 9 draws together the main conclusions of the compara- tive study. Before proceeding, it is as well to indicate some limitations of the study. The sample size of six is far from adequate to embody all the economic, political, and policy differences among the widely scat- tered countries that nature has endowed with oil reserves. There- fore, simple unweighted averaging is used to provide indications of central tendencies, and individual countries are judged largely by their deviation from the averages. The cross-sectional approach of part II, which is organized by topics, is complemented by the case studies in part III, which provides a country-by-country treatment. Some indication of the diversity of the sample countries is pro- vided by table 4-1. In 1979 gross national product (GNP) per capita ranged from $370 in Indonesia to $3,910 in Trinidad and Tobago, while population ranged from 143.9 million in Indonesia down to 1.1 million in Trinidad and Tobago-which, being tiny, is by far the most densely populated country in the sample. Oil exports per per- son also varied widely, from $62 to $2,135 in 1979, as did the size and diversity of non-oil exports in 1973. In one important respect, however, there was surprisingly little variation. A very rough mea- sure of the ratio of oil wealth to the non-oil capital stock using 1975 re- serve estimates and prices lies within the range of 2.3-2.7 for all six countries. The share of oil in total wealth was therefore roughly simi- lar for the entire sample. Table 4-1 also summarizes features of the administrations of the countries over the period 1972-84, emphasizing the two aspects stressed in chapter 3: the degree to which decisionmaking was cen- tralized and the continuity, or "time horizon," of the government.' Again there has been a great deal of cross-country variation, but the sample can be grouped into three pairs. Two countries, Indone- sia and Algeria, are judged to have had political continuity and a sta- ble economic policy framework throughout the period, and also to have had a highly centralized process for determining the uses of oil income. These features are suggested by later analysis to have had some advantages for growth and to have helped to avoid a seri- ous foreign exchange crisis when oil income fell. Nigeria and Ecua- dor, in contrast, both experienced frequent changes of government of a discontinuous nature. In addition, their political systems were quite decentralized along regional lines. Both these countries were plunged into crisis by the fall in oil revenues, and their growth per- formance was not as good as that of the first-named pair. The third pair, Trinidad and Tobago and Venezuela, are judged to occupy an Introduction 51 intermediate position. Both had quite highly competitive, demo- cratic political systems throughout the period. The continuous gov- ernment of the former was, however, vulnerable to pressure groups which were well organized along sectoral and ethnic lines. This led to mushrooming subsidies and loss of control over public spending. The Venezuelan administration changed in the middle of the boom period, as did economic policy. Moreover, after 1978 poli- cies of the administration and of the legislature not infrequently diverged, and this had a substantial and adverse macroeconomic impact. In the previous chapter, table 3-3 set out a framework for the main policy decisions and macroeconomic variables affecting the ab- sorption of the windfall and its effects. To help link the compara- tive study with the theory, table 4-2 places the six countries into the framework. Because of the simultaneity of macroeconomic analy- sis, the link between policies and effects is not simple-for exam- ple, the rate at which windfalls are absorbed (discussed in chapter 5) bears on "booming sector" effects (chapter 6), growth (chapter 7), and the severity of the downturn (chapter 8). Therefore the reader might find it useful to refer back to table 4-2 at various points in the comparative study to ascertain how the response of a particular country compares with those of the others. Note 1. The following statements represent broad judgments in the comparative framework of the sample. They should not be taken to imply any absolute statement on the nature of the respective countries' governments and are of course subject to a variety of qualifications. More detailed discussion of each country is in part III. Table 4-1. Economic and Political Characteristics of the Oil Exporters GNPper Oil exports capita, Area per capita, Oil reservesl Main Political characteristics, 1979 Population, (thousands 1979 capital non-oil centralization, (U.S. 1979 of square (U.S. stock, exports, and continuity, Country dollars) (millions) kilometers) dollars) 1975a 19731 1972-84 Algeria 1,770 18.3 2,382 478 2.5 Wine, fruit Centralized socialist state; single politi- cal party with close ties to military. Change of president, 1979, but broad administrative and policy con- tinuity. Ecuador 1,110 8.1 284 127 2.3 Bananas, coffee, Regionally fragmented, with decen- cacao tralized public sector; executive presi- dent and legislature. Military coup, 1972; transitional military junta, 1976; elected government with popu- list leanings, 1979; death and replace- ment of president, 1982. Political dis- continuity. Indonesia 370 143.9 1,919 62 2.3 Timber, rubber, Centralized civilian government with coffee, tin, dominant political party (association animal products, of groups) and close ties to military. palm oil, coffee Exceptional political and adminis- tea, pepper trative continuity. Nigeria 910 82.6 924 201 2.7 Cacao, Federation of states subject to strong groundnuts centrifugal forces. Military govern- ment after civil war; coup, 1975; as- sassination of new president, 1976; ci- vilian government installed, 1979; military coup, 1983. Political disconti- nuity. Trinidad 3,910 1.1 5 2,135 2.3 Sugar Centralized Westminster-style govern- andTobago ment with dominant party through- out period facing strong interest groups. Death of prime minister and replacement, 1981. Political continu- ity. Venezuela 3,440 14.4 912 947 2.5 Iron ore Centralized administration but with power split between executive presi- dent and legislature. Election of AD (Social-Democrat) government, 1974; election of COPEI (Christian- Democrat) executive and AD house, 1979. Continuity subject to election cycle. Note: The entries in this table reflect a comparative judgment within the context of the sample. They are not a statement of the country's posi- tion in any absolute sense or relative to other developing countries. a. Estimated as (mineral value added x 0.80 x ratio of proven oil output) / (nonmining GDP X 2.5). b. Export values greater than $20 million. Source: World Bank. 54 Comparative Analysis Table 4-2. Absorption Choices and Macroeconomic Responses of the Oil Exporters Net Windfall Public Main subsi- Extraction savings absorption investment sidies and Country rate abroad emphasis emphasis rent transfers Algeria Slower after Negative Very strong Strongly T Subsidies 1979 (high ab- public in- (heavy in- from sorber) vestment dustry); Treasury bias some shift to public to N firms (housing) after 1979 Ecuador Slower than Negative Mixed Strongly N Large subsi- projected (infra- sidies on structure, energy, education) some on basic foods Indonesia On trend Slightly neg- Mixed Mixed: T Subsidies on ative; pos- (agricul- oil, ferti- itive after ture, in- lizer, rice 1979 dustry), N (infra- structure, education) Nigeria On trend Zero Public in- Strongly N Moderate vestment (infra- subsidies bias structure, on energy education, new capi- tal); after 1979 some T (indus- try, steel) Trinidad Increase Strongly Mixed: high Mixed: T Very large and from pre- positive consump- (hydrocar- subsidies: Tobago vious (low ab- tion em- bon-based sugar levels sorber) phasis industry); industry, N (roads, public electrifica- firms, tion) wages, consumer goods Venezuela Slowdown Balanced: Mixed: high Largely T Large sub- from public dis- consump- (particu- sidies: previous saving, tion em- larly steel, energy, levels private phasis aluminum) public en- saving terprises, consumer goods T = traded; N = nontraded Note: The entries in this table reflect a comparative judgment within the context of Introduction 55 Development Labor Trade and Main macro-clearing emphasis availability exchange policy variables Urban Very large, espe- Continued tight High import leak- cially relative to control of trade ages, household requirements of and capital flows; rationing, price capital-intensive pegged rates controls strategy Urban Large Some relaxation of Real effective ex- controls and cuts change rate in non-oil export changes taxes; pegged rates Mixed: major Very large; deep Some relaxation of Real effective ex- rural com- reservoir of labor trade controls, change rate ponent in Java and Bali free capital flows, changes rendered pegged but ad- more flexible by justable rates willingness to devalue Urban Supply inelastic, Trade policies Very high real supplemented by erratic, exchange effective ex- immigration rationing, pegged change rate ap- rates preciation with import controls Urban Large, due to Relatively open High import leak- initially high un- trade policy, ages, price employment, but pegged rates controls, later heavily exchange unionized appreciation Urban Supply inelastic, Relatively open Import liberaliza- supplemented by trade policy, tion with price immigration pegged rates controls, later exchange appre- ciation, massive capital flight the sample. They are not a statement of the country's position in any absolute sense or relative to other developing countries. Source: World Bank. Chapter 5 The Magnitude and Uses of the Oil Windfalls, 1974-81 The size and the uses of windfalls from higher oil prices can be esti- mated in a variety of ways. All involve measuring deviations from some counterfactual scenario setting out what would have tran- spired had there been no windfall. The most complete scenarios re- quire a full model of the economy, but this is not practical here be- cause of the large and complex differences among the countries. This chapter develops a simple decomposition of the changes in broad demand and supply aggregates to estimate the size of the windfalls and how they were absorbed. Chapter 6 uses a further sim- ple decomposition of the major sectoral components of the non-oil economy to assess the impact on the broad structure of production. A Simple Decomposition of Income and Spending In developing this approach, we begin by estimating the increase in domestic income which results from a larger current value of the oil sector relative to the aggregated non-oil sectors. To make cross- country comparisons easier, these computations are performed with the economy divided into mining and nonmining, rather than oil and non-oil, segments. Since non-oil mining is a very small part of GDP in countries without large hard-mineral sectors, this has only a minor effect on the results.' By the national income identity, the trade and nonfactor service balance (which is commonly called the resource balance) is the differ- ence between domestic output and total absorption, all at current prices. Domestic output is the sum of mining and nonmining out- put, and absorption is the sum of consumption and investment. Therefore we may write: (5-1) R = Y + Z - C - I 56 The Magnitude and Uses of Oil Windfalls 57 where R, Y, Z, C, and I represent the resource (trade and nonfactor service) balance, nonmining output, mining output, consumption, and investment, all at current market prices. Let p, pz, pc, and pi be the deflators for nonmining output, min- ing output, consumption, and investment respectively. In each per- iod t, the quantity and relative price structure of the economy may then be summarized by a state vector (r, z, c, i, qz, qc, qi). The respec- tive q's represent the mining (oil), consumption, and investment de- flators relative to p, the deflator of nonmining output. The ele- ments r, z, c, and i represent the (nominal) trade and nonfactor service balance, mining sector, consumption, and investment, re- spectively, relative to nonmining GDP. The state vector therefore "scales" the mining sector and the components of demand by nonmining GDP and also expresses their deflators relative to that of nonmining GDP. From equation 5-1: (5-2) rt = 1 + zt - (ct + it) A base period, denoted "0," is taken as the average for the per- iod 1970-72. We now construct a counterfactual scenario, denoted by h, using the following assumptions: * No changes in relative deflators from the base period 1970-72 * A constant mining sector share in total output * No change in the ratio of total absorption to total output. These assumptions imply that the counterfactual terms of trada: constant (if they were not, the consumption or investment deflators would have to change relative to output deflators), and also that the counterfactual ratio of the resource balance to GDP iS constant. This is consistent with a balanced growth scenario free from shocks.2 The composition of absorption, however, is known to change syste- matically with growth of income per capita, so that the balanced growth scenario must be modified. Therefore Cth and Ith (the super- script h denoting counterfactual values) are adjusted within the total absorption constraint in line with the projections of Chenery and Syrquin (1975), with real nonmining output per capita taken as the indicator of growth. In the counterfactual scenario: Ct h+ ih = Co + io so that (5-3) ro = 1 + zo - (cth + it h) 58 Comparative Analysis Adjustment of the exporter to the oil shock may then be written as the difference between the actual and the counterfactual scenarios, relative to nonmining output in the counterfactual scenario: (5-4) (Rt - Rth) / yth = (Zt - Zth) / yth - (Ct - Cth) / yth - (It - Ith) / Yt The righthand absorption terms may be further broken down into public and private components of consumption and investment. Next each of the resulting terms may be decomposed again into price and quantity effects. Since the terms express differences in cur- rent values-of mining output, consumption, and investment-and current value equals the product of price and quantity, the change in value may be expressed as the sum of the price change weighted by quantity and the quantity change weighted by price. In such ex- pressions, as is well known, use of the initial or final weights alone (here, counterfactual or actual weights) introduces a residual if the changes are not infinitesimal; therefore the average of the actual and counterfactual weights is used to minimize cross-effects and the residual terms.3 For example, scaling all base-year deflators to unity yields, for the oil windfall term: Zt - Z (5-5) t h t = (ztlqzt - zh) * (qzt + 1)/2 + (qzt - 1) * (ztIqzt + Zh)/2 The lefthand term is the "value windfall effect." It expresses the change in the contribution of the mining sector, relative to nonmining GDP, and is the sum of two effects. The first term on the righthand side of equation 5-5 is the "real windfall effect." It indi- cates the contribution to the windfall of changes in real mining out- put relative to real nonmining output. The second term is the "price windfall effect."4 It shows the contribution to the windfall made by relative price changes between mining and other sectors. If oil prices rise sharply but oil output falls relative to the output of other sectors, the real effect will be negative and the price effect will be positive; the windfall's value is less than it would have been had oil output grown as fast as output from other sectors. A similar breakdown of value changes into real and relative price effects can be made for consumption and investment terms. This can be useful. If, say, investment expenditures rise relative to nonmining output, is this due to greater real investment? Or does it merely reflect a rise in the relative cost of investment goods? If con- sumer prices fall because of cheaper imports, real consumption can grow more rapidly for given consumption expenditures. Needless to say, the use of deflator movements to indicate price changes is The Magnitude and Uses of Oil Windfalls 59 not ideal, especially given the residual nature of consumption in the national accounts, but no other indicator is available. Although the base period 1970-72 was one of reasonable stability for the world economy, it did not necessarily represent a period of stable, long-term equilibrium for each of the sample countries. There- fore economic structure in the base period is important in interpret- ing the results of such a decomposition. Table 5-1 provides an indica- tion of the base absorption structures of the six sample countries plus, for comparison, Iran. These are compared with the Chenery- Syrquin norms for countries at a similar level of nonmining income per capita. The table also shows the share of oil in commodity ex- ports in 1972 as well as the average current account, relative to nonmining GDP, in 1970-72.5 Iran, Algeria, and Venezuela-all with large, long-established oil sectors (and considerable remittance income in the second case)- had the highest initial ratios of absorption to nonmining GDP. Indone- sia, Nigeria, and Trinidad and Tobago occupied an intermediate posi- tion. Ecuador, which became an oil exporter most recently, was least dependent on oil in terms of absorption, but it already had a higher share of oil in total exports than Indonesia, which benefited from aid inflows and maintained a diverse portfolio of non-oil ex- ports reflecting its rich resource endowment. In the base period, the current account deficits of the sample coun- tries averaged 5.1 percent of nonmining GDP. This high average was due to the very large deficits of Trinidad and Tobago and of Ecuador, both of which were in a phase of rapid oil development largely fi- nanced by direct investment from abroad. Table 5-1 also shows the breakdown of extra absorption relative to patterns "normal" for countries at similar levels of income per ca- pita. Before the first oil shock, there were particularly strong invest- ment biases in Algeria, normal investment rates. Private consump- tion was high relative to public consumption in Ecuador, Nigeria, and Indonesia, reflecting a relatively small government role. Size of the Windfalls Tables 5-2 and 5-3 show the decomposition estimates of windfalls and their uses for 1974-78 and 1978-81 respectively. We first con- sider only the windfalls. The mean windfall ("value effect") was 22 percent of nonmining income during 1974-78 (24 percent including Iran) and 23 percent during 1979-81. Particularly after 1979, slower real growth of the oil sector reduced the value impact of the "price" windfall (32 percent of nonmining GDP) by about one-fourth on average and almost eliminated the windfall in Venezuela. Table Table 5-1. The Composition of Absorption in 1970-72 in Relation to Chenery-Syrquin Norms (percent) Trinidad Unweighted Item Algeria Ecuador Indonesia Nigeria and Tobago Venezuela mean Iran Private consumption Actual/nonmining GDP 64.5 75.5 83.5 80.2 80.9a 62.8 74.6 72.2 Norm: private consumption/GDP 66 68 75 70 78^ 62 69.8 64 Public consumption Actual/nonmining GDP 18 10.6 9.8 8.6 n.a. 16.4 12.7 23.4 Norm: public consumption/GDP 14 14 12 14 n.a. 15 13.8 14 Investment Actual/nonmining GDP 40.8 20.6 17.5 22.1 31 36.6 28.1 27 Norm: investment/GDP 20 19 15 17 22 23 19.3 22 Absorption Actual/nonmining GDP 123.3 106.7 110.9 111 111.9 115.8 113.3 122.7 Norm: absorption/GDP 100 101 102 101 100 100 100.7 101 Breakdown of extra absorption (percent) Private consumption -6 132 96 102 24a 5 n.a. 36 Public consumption 17 -60 -25 -54 n.a. 9 n.a. 41 Investment 83 28 28 51 76 86 n.a. 44 Percentage of oil in exports, 1972 79 77 51 82 78 91 76 91 Current account/ nonmining GDP, 1970-72 1.4 -6.9 -3.8 -3.3 -14.5 -0.6 -5.1 -3.2 n.a. Not available. a. Includes public consumption. Source: World Bank, World Tables data base. The Magnitude and Uses of Oil Windfalls 61 5-3 indicates that, at the time of the second oil price rise, oil sectors were relatively smaller than at the time of the first price rise (except in the case of Ecuador). This helped to mute the impact of the sec- ond boom on producing countries. Figure 5-1 indicates the (unweighted) average time profile of the windfalls in the period 1973-81, as measured above, for the six coun- tries. Again each year's windfall ("value effect") is expressed as a share of the nonmining economy. In 1974 the average windfall peaked at 33 percent of nonmining income, but by 1978 it had con- tracted to 15 percent. The second oil price increase raised the wind- fall to 27 percent of nonmining output in 1980 before declining sales began to reduce it. The combination of slumping prices and con- tracting sales in 1982-84 appears to have halved the windfall gain. Use of the Windfalls: The Fiscal Response With little direct linkage between the oil sector and the rest of a devel- oping economy, and with a rapid upward adjustment of tax and roy- alty rates to reduce the share of surplus accruing to the oil Figure 5-1. Oil Windfalls and Their Uses, 1973s81 Unweighted Average of Algeria, Ecuador, Indonesia, Nigeria, Trinidad and Tobago, and Venezuela 50 tp 401 xo 40 Trade and nonfactor services surplus t 301 /)/ Oil windfall Y Y / w~~~~~Tade and nonfactor services deficit 20 10 0 g<\ Consumption 1970-72 1973 1974 1975 1976 1977 1978 1979 1980 1981 Source: Tables 5-2 and 5-3. Table 5-2. Oil Windfalls and Their Uses, 1974-78 (percentage of nonmining GDP) Trinidad Unweighted Iran Windfalls and uses Algeria Ecuador Indonesia Nigeria and Tobago Venezuela mean (1974-77) Domestic oil windfall 27.1 16.8 15.9 22.8 38.9 10.8 22.1 36.7 Real -4.8 9.2 1.6 -2.3 2.3 -20.5 -2.4 -22.8 Price 31.9 7.6 14.3 25.1 36.6 31.3 24.5 59.5 Absorption effects Trade and nonfactor service -4.3 3.5 5.3 2.8 27.2 -1.0 5.5 17.0 Current balance -9.8 2.7 4.8 5.4 26.7 3.9 5.6 24.6 Non-oil growth effect -7.4 15.9 -2.4 -1.5 -7.8 5.9 0.5 12.2 Allocation effects Value Private consumption 3.6 -0.9 2.1 2.9 7.1b 1.9 2.1 -7.0 Public consumption 1.4 5.5 2.4 4.2 n.a. 1.6 3.2 10.2 Private investment n.a. 3.8 -1.7 -6.6 -2.6 3.3 0.6 6.7 Public investment 26.4a 4.8 7.9 19.5 7.3 4.9 11.8 10.0 Price Private consumption -0.9 -3.0 -1.1 1.2 -18.4 -10.1 -5.4 -6.2 Public consumption -0.2 0.3 0.7 -0.1 n.a. 3.5 0.7 2.3 Investment 4.1 4.3 -0.5 3.8 0.0 4.4 2.7 7.3 Real Private consumption 4.6 2.1 3.3 1.7 25.6c 12.1 7.5 0.8 Public consumption 1.6 5.2 1.7 4.3 n.a. 0.8 3.0 7.9 Private investment n.a. 0.4 -1.3 -9.5 -2.6 1.8 -1.8 2.5 Public investment 22.2a 3.9 8.0 18.6 7.3 2.6 10.4 6.9 Real allocation plus growth effects Private consumption 0.0 14.0 -1.5 0.5 19.4d 15.8 7.5 7.6 Public consumption 0.3 6.9 1.5 4.2 n.a. 1.8 3.0 13.2 Private investment n.a. 3.2 -3.4 -9.8 -4.6 3.3 -1.9 4.5 Public investment 19.1' 4.5 7.9 18.5 5.5 3.3 10.0 8.4 n.a. Not available. a. Adequate estimates are not available because of the small share of private investment. The private investment effect is taken as zero. b. Public consumption is included in private consumption. For computation of mean, total consumption effect is split in the average proportions for the other countries, 2.8 private and 4.3 public. c. Price effect is attributed to private consumption. For computation of mean, private consumption effect is taken as 21.2, public consumption ef- fect as 4.3. d. For means, this is split into 16.0 private and 3.4 public. Source: World Bank data. Table 5-3. Oil Windfalls and Their Uses, 1979-81 (percentage of nonmining GDP) Trinidad Unweighted Windfalls and uses Algeria Ecuador Indonesia Nigeria and Tobago Venezuela mean Mexico Domestic oil windfall 29.7 22.1 22.7 21.9 34.7 8.7 23.3 3.5 Real -17.6 10.8 -2.5 -6.1 -7.4 -28.0 -8.5 n.a. Price 47.3 11.2 25.2 28.0 42.1 36.6 31.7 n.a. Absorption effects Trade and nonfactor service 8.9 4.7 9.6 0.1 16.8 1.1 6.9 -1.8 Current balance -0.6 -1.2 6.1 3.9 19.2 7.0 5.7 -0.2 Non-oil growth effect -6.7 21.2 -3.5 -29.8 0.6 -6.6 -4.1 n.a. Allocation effects Value Private consumption 4.6 1.1 1.2 4.1 8.6a 9.4 4.1 -2.9 Public consumption 3.2 6.1 3.7 5.6 n.a. 0.7 3.9 2.3 Investment 12.9 10.3 8.1 12.1 9.3 -2.5 8.4 5.9 Price Private consumption -2.2 -4.0 -9.3 -4.1 -18.2b -14.7 -8.8 -12.7 Public consumption -0.1 0.1 0.3 0.6 n.a. 0.4 0.2 1.4 Investment 4.4 6.3 -0.5 4.8 -2.6 1.1 2.2 18.7 Real Private consumption 6.8 5.1 10.5 8.2 26.7c 24.1 12.9 -0.1 Public consumption 3.3 6.0 3.4 5.0 n.a. -0.3 3.7 1.0 Investment 8.6 4.0 8.6 7.3 12.0 -3.5 6.2 4.1 Real allocation plus growth effects Private consumption 2.7 21.7 7.7 -15.3 27.2d 20.0 9.8 n.a. Public consumption 2.5 6.8 3.0 2.3 n.a. -0.7 3.0 n.a. Investment 5.6 15.0 7.9 0.4 12.2 -6.0 5.8 n.a. n.a. Not available. a. Includes public consumption. For mean, this is split into 4.4 private and 4.2 public. b. Price effect is attributed to private consumption. c. For mean, this is split into 22.5 private and 4.2 public. d. For mean, this is split into 22.9 private and 4.3 public. Source: World Bank data. 66 Comparative Analysis multinationals, fluctuations in ml±iing value added were mainly re- flected in fiscal revenues. The impact is shown in tables 5-4, 5-5, and 5-6. For the six countries in the sample, the first oil price rise caused central government revenues to jump from 20 percent of nonmining GDP to 37 percent in 1974-78 (42 percent including Iran).6 This implies that, on average, about four-fifths of the wind- fall accrued to producer governments. Iran, for which data after 1977 are limited, experienced a particularly large windfall, 36.7 per- cent of nonmining GDP during 1974-77, which was also reflected in fiscal revenues. Central government revenues relative to nonmining GDP were similar in 1979-81 to what they were in 1974-78. This would be expected, since the windfall effect in figure 5-1 is similar for the two oil shocks. Although there were significant differences in non-oil fiscal per- formance among countries, almost all the increase in the ratio of fis- cal revenue to nonmining income is attributable to increased taxes and royalties on oil. The only exception is Algeria, where non-oil taxes, particularly on domestic goods and services, were unusually high at 25 percent of nonmining GDP and proved to be buoyant. Non- oil taxation remained low in the rest of the sample, on average little more than half what it was for typical countries at similar levels of GNP as indicated by cross-country regressions.7 Except for Algeria, only Trinidad and Tobago possessed a well-developed system of di- rect taxation. But as domestic prices rose, its brackets were ad- justed to compensate for progressivity so that collection was less buoyant than in Algeria. Tables 5-5 and 5-6 present estimates of the elasticity of non-oil taxes to growth of the non-oil economy during the first and second oil price shocks. Although most exporters relaxed import restric- tions and eased import tariffs and non-oil export taxes after 1974, the rapid increase in imports made possible a substantial rise in trade taxes in 1974-78. This increase in the tax base due to oil in- come was compensated (except in Algeria) by a cut in the taxes levied on domestic goods and services, as shown by less-than-unit elasticities. Underlying this policy was a widespread desire to moderate inflationary pressure caused by boosting public spending. On average, non-oil tax revenues remained fairly steady as a propor- tion of nonmining income during the first oil price rise. They de- creased slightly with the second price rise because import volumes and trade taxes did not continue to expand so rapidly. Producer governments therefore did not generally transfer their windfalls to the private sector by sharply cutting non-oil taxes, al- though in some cases, notably Ecuador, non-oil tax efforts slack- ened. Until the early 1980s they did try to insulate domestic consum- Table 5-4. Structure of Taxes and Government Expenditure, 1974-78 (percentage of nonmining GDP) Trinidad and Unweighted Item Algeria Ecuador Indonesia Nigeria Tobagoa Venezuela mean Iranb Total revenue 59.9 12.9 23.1 27.7 55.9 42.1 37.0 71.1 Non-oil direct taxes 6.2 1.4 1.8 0.1 5.2 1.3 2.6 0.5 Domestic indirect taxes 14.1 2.3 3.3 0.6 3.0 1.6 4.2 1.7 Trade taxes 4.2 5.8 2.5 3.9 4.4 2.6 3.9 4.3 Total taxes 24.5 9.5 7.6 4.6 12.6 5.5 10.7 6.5 Expenditure 71.5 14.4 25.0 24.2d 36.7 32.0 34.0 71.6 By function Administration, defense 6.5c 4.9 9.9 9.1 5.8 5.0 6.9 23.6 Social 9.8c 4.8 2.7 4.3 11.0 9.6 7.0 10.1 Economic 2.0c 3.7 8.2 8.8 15.6 10.6 8.2 20.5 By economic classification Current 27.0 n.a. 14.5 13.6 21.1 20.5 19.3 49.3 (Wages and salaries) (10.0) n.a. (4.1) (4.1) (10.9) (11.2) (8.1) (16.4) Subsidies and transfers 10.5 n.a. 5.1 3.4 3.1 5.8 5.6 13.1 Investment 44.5 n.a. 10.6 12.4 15.6 11.9 19.0 22.3 Net lending 29.0 n.a. 1.8 7.2 6.0 13.8 11.6 n.a. Total revenue 1970-72 32.6 14.2 15.6 12.3 19.9 25.2 20.0 31.7 Total revenue 1979-81 57.4 14.2 30.9 n.a. 57.2 36.3 39.2 n.a. n.a. Not available. c. Current expenditures. a. 1976-77. d. 1975-78. b. 1974-77. Sources: IMF Government Finance Statistics data base; World Bank data. Table 5-5. Elasticity of Taxes and Government Expenditure with Respect to Nonmining Output, 1974-78/1970-72 Trinidad Unweighted Item Algeria Ecuadora Indonesia Nigeriab and Tobago Venezuela mean Iranc Total revenue 1.84 0.91 1.48 2.25 2.81 1.68 1.82 2.24 Non-oil direct taxes 1.15 1.00 1.25 -_ 1.48 1.11 1.19 1.00 Domestic indirect taxes 1.25 0.85 0.92 0.20 0.71 0.90 0.81 0.72 Trade taxes 1.56 0.78 0.92 1.78 1.26 1.57 1.30 0.90 Total taxes 1.26 0.83 0.99 0.83 1.13 1.20 1.04 0.85 Expenditure 1.78 1.02 1.34 2.249 1.70 1.29 1.57 1.90 By function Administration, defense 1.18d 1.32 1.25 1.44 0.89 1.16 1.21 1.92 Social 1.05d 1.04 1.69 4.78 1.23 1.09 1.81 1.53 Economic _ 0.93 1.44 4.19 3.71 1.20 1.91 1.60 By economic classification Current 1.36 n.a. 1.43 1.63 1.34 1.11 1.37 1.98 (Wages and salaries) (0.98) n.a. (1.11) (1.11) (1.10) (1.07) (1.07) (1.24) Subsidies and transfers 2.19 n.a. 1.02 3.78 n.a. 1.38 2.09 2.67 Investment 2.15 n.a. 1.53 4.96 2.89 1.83 2.68 1.74 Net lending 2.74 n.a. 1.28 3.00 n.a. 7.26 3.57 n.a. n.a. Not available. e. Negligible in 1970-72. a. 1974-78/1973 f. Negligible in 1972. b. 1974-78/1972 g. 1975-78/1972. c. 1974-77/1970-72 Sources: IMF Government Finance Statistics data base; World Bank data. d. Current expenditures only. Table 5-6. Elasticity of Taxes and Government Expenditures with Respect to Nonmining Output, 1979-81/1974-78 Trinidad Unweighted Item Algeria Ecuador Indonesia and Tobago Venezuela mean Total revenue 0.95 1.10 1.34 1.02 0.86 1.05 Non-oil direct taxes 1.11 1.29 1.00 1.63 0.92 1.19 Domestic indirect taxes 0.80 1.17 0.78 0.80 0.96 0.90 Trade taxes 0.69 0.76 0.88 0.86 0.93 0.82 Total taxes 0.86 0.94 0.87 1.16 0.93 0.95 Expenditure 1.01 1.19 1.28 1.18 0.97 1.13 By function Administration, defense 0.89 0.94 1.46 1.90 0.94 1.23 Social 1.07 1.46 1.26 0.92 1.05 1.15 Economic 0.90 0.91 1.22 0.99 0.72 0.95 By economic classification Current 0.99 n.a. 1.15 1.23 1.13 1.13 (Wages and salaries) (1.20) n.a. (1.10) (1.11) 1.03 (1.11) Subsidies and transfers 0.80 n.a. 1.42 3.10 1.16 1.62 Investment 1.04 n.a. 1.46 1.10 0.68 1.07 Net lending 0.99 n.a. 1.06 1.50 0.50 1.01 n.a. Not available. Note: Data were not available for Nigeria. Sources: IMF Government Finance Statistics data base; World Bank data. 70 Comparative Analysis ers from the effects of higher oil prices by holding domestic oil prices at or near former levels.8 Except for Algeria, however, govern- ments did not attempt to restrain private competition for scarce do- mestic resources by increasing the non-oil tax burden. The question of whether a government should raise or lower non- oil taxes in response to windfall gains has no simple answer. Much depends on the range of saving and spending options available to the government. But improving the non-oil tax system (broadening the tax base and improving tax administration as opposed to rais- ing tax rates) is one way of preparing for the post-oil-boom period. Only one of the countries with weak fiscal systems, Indonesia, took advantage of the respite from revenue pressures to set in place a sys- tem to improve the efficiency of non-oil taxation. Indonesia's tax re- form was implemented in 1983. As demonstrated by the high elasticity of expenditure, especially during the first oil shock, the decade of the windfalls was marked by an unparalleled growth in the weight of the state and diversifica- tion of its role. Between 1974 and 1978, net lending by government and public investment outlays grew more than twice as fast as the non-oil economies of the respective countries (see table 5-5). Cur- rent expenditures grew less rapidly, with the notable exception of subsidies and transfers. This category, which will be discussed fur- ther below, continued to expand after 1979, whereas the relative growth of investment and net lending slowed. The growth of eco- nomic spending exceeded that of social spending during the first per- iod, but this was reversed during the second period as other commit- ments began to eat into the proportion of the surplus available for economic functions. Overall Absorption of the Windfall Returning to figure 5-1 and tables 5-2 and 5-3, we now consider the overall absorption of the windfalls. This includes the impact of both first-round and subsequent (multiplier) rounds of expenditures. The analysis of windfall use is restricted here to the evolution of value shares. Discussion of the breakdown between real and price ef- fects is deferred until after that of movements in the real exchange rate in chapter 6. As noted above, in 1970-72 current account deficits in the six coun- tries had averaged 5.1 percent of nonmining GDP, with especially high deficits in Trinidad and Tobago (14.5 percent) and in Ecuador (6.9 percent) because of the large investments needed to develop the oil industry and because of some spending in anticipation of in- The Magnitude and Uses of Oil Windfalls 71 creased revenues. As indicated by figure 5-1 and table 5-2, during 1974-78 about one-fourth of the windfall (5.5 percent of nonmining GDP went to reduce these trade (and current) deficits. Another one- fourth (5.3 percent of nonmining GDP) was consumed. Slightly over half of this increase in consumption was public, slightly under half was private. The remainder of the windfall-almost half-was used for domestic investment. Although non-oil private investment (which is estimated as a residual) boomed in certain countries, nota- bly in Venezuela in 1976-78, the increased investment outlays were overwhelmingly public. Consumption and Permanent Income Estimates of proven oil reserves are notoriously uncertain. They are also not fixed, but depend on prices and on exploration and develop- ment activity. At the extraction rates of the 1970s, the countries in the sample could have expected, on average, about fifteen years of low-cost (that is, high-rent) oil output. The equivalent flow of perma- nent income may be defined as that permanent flow which has the same discounted present value as the windfall. It depends on price projections and on the rate of discounting, denoted by i. If ex- pected future prices are set to equal current prices in real terms (a ran- dom walk hypothesis for prices), projected permanent oil income, zp, is given by: 15 zp = if e-it zdt 0 If i is taken to be 3 percent, permanent income is 36 percent of cur- rent oil income z. If instead the Hotelling rule is assumed, so that real oil rents rise with the rate of interest, permanent income will sim- ply equal 15 if zdt 0 or 45 percent of current oil income. Now suppose that the long-run sustainable current account defi- cit was zero (the average sample country was almost in current ac- count balance in 1974-78 and 1979-81; see table 5-7). Suppose also that consumption out of windfall gains was set by a constant long- run propensity to consume (say, 80 percent) out of permanent in- come. Of the windfall not used to reduce the unsustainable current deficit, between 29 percent and 36 percent should then have been Table 5-7. Current Accounts and Their Financing, 1970-83 (percentage of nonmining GDP) Item 1970-72 1974-78 1979-81 1982-83 Current account Algeria 1.4 -11.4 -2.0 -0.4 Ecuador -6.9 -4.1 -8.1 -5.0 Indonesia -3.8 -1.6 2.3 -10.1 Nigeria -3.3 2.3 0.6 -9.0 Trinidad and Tobago -14.5 12.7 4.7 -15.5 Venezuela -0.6 3.1 6.3 1.2 Unweighted average -5.1 0.2 0.6 -6.6 Financing Algeria Portfolio + long-term capital 1.7 13.6 4.7 -2.7 (Direct investment) (-0.4) (1.3) (0.5) (-0.2) Short-term capital + errors -0.5 -0.8 -0.2 1.1 Reserve change and other balance items 0.2 -1.4 -2.5 2.0 Ecuador Portfolio + long-term capital 8.9 6.8 9.0 11.7 (Direct investment) (3.4) (1.0) (1.0) (1.0) Short-term capital + errors -0.8 -1.1 -1.0 -6.5 Reserve change and other balance items 1.1 -1.6 0.0 0.1 Indonesia Portfolio + long-term capital 4.3 4.2 3.5 8.8 (Direct investment) (1.0) (1.2) (0.9) (1.4) Short-term capital + errors 0.6 -2.6 -3.3 0.4 Reserve change and other balance items -1.0 0.0 -2.5 0.8 Nigeria Portfolio + long-term capital 2.5 1.3 1.7 3.0 (Direct investment) (1.1) (0.7) (0.1) (0.8) Short-term capital + errors 1.4 0.0 -0.4 0.0 Reserve change and other balance items -0.6 -3.5 -0.2 0.8 Trinidad and Tobago Portfolio + long-term capital 11.3 5.8 8.9 5.2 (Direct investment) (9.6) (8.9) (10.4) (10.9) Short-term capital + errors 2.3 0.0 -0.7 1.5 Reserve change and other balance items -0.9 -18.5 -12.8 4.9 Venezuela Portfolio + long-term capital 0.8 4.4 3.3 2.5 (Direct investment) (-0.5) (-0.9) (0.5) (0.7) Short-term capital + errors 1.9 2.0 -3.3 -9.2 Reserve change and other balance items -2.0 -9.5 -6.3 5.5 Unweighted average Portfolio + long-term capital 4.9 6.2 5.2 4.7 (Direct investment) (2.4) (-0.9) (2.2) (2.4) Short-term capital + errors 0.8 -0.4 -1.5 -2.1 Reserve change and other balance items -0.6 -5.7 -4.3 4.0 Source: IMF International Financial Statistics. 74 Comparative Analysis consumed, depending on whether future prices followed the ran- dom walk or the Hotelling rule. The actual proportion consumed in 1974-78, 33 percent, falls within this range. The average use of the windfall in 1979-81 was similar, except that private consumption somewhat increased its value share at the expense of domestic investment, which accounted for only about one-third of the total. The average consumption propensity of the countries was thus not out of line with that predictable from a permanent income expla- nation. But there were significant absorption differences among coun- tries. In 1974-78 the use of windfalls for domestic investment was lower than average in Trinidad and Tobago and in Iran, both of which ran huge current account surpluses for very different rea- sons. The investment drive was far more intensive in Algeria. Tak- ing advantage of its improved creditworthiness, Algeria absorbed the entire first windfall in investment and borrowed abroad against future oil and gas income to finance a small rise in public and pri- vate consumption relative to non-oil output and to compensate for a decline in remittance income. In 1974-78 the Algerian current ac- count deficit was 11.4 percent of nonmining GDP. At its peak in 1977, Algerian investment reached the remarkable figure of 73 per- cent of nonmining GDP. Virtually all of this was public, with much going to develop heavy industry. The other country to run substantial current account deficits in 1974-81 was Ecuador. The story is very different from that of Alge- ria, whose foreign borrowing was undertaken as an essential compo- nent of a long-term strategy pursued with great continuity and (possi- bly excessive) dedication. Ecuador's deficits, in contrast, stemmed from weak and fragmented fiscal control, exacerbated by political in- stability and the decentralized nature of its public sector, and from its policy of holding domestic interest rates low and expanding Cen- tral Bank credit to the private sector. Ecuador's fiscal control deterio- rated seriously in 1979-81, and its current deficit rose to 8.1 percent of nonmining GDP. For comparison, in table 5-3 estimates based on the same methodol- ogy are also presented for Mexico. An exporter of more recent vin- tage, Mexico experienced only a small oil windfall in 1979-81, just 3.5 percent of its non-oil GDP. But, rather like Algeria in 1974-78 or Ec- uador later, it borrowed abroad against future oil earnings to boost ex- penditures by a further 1.8 percent of non-oil GDP. The increased Mexican expenditures, like those of the other oil exporters, mostly took the form of public investment. Development expenditure for the Mexican oil sector itself accounted for 2 percent of non-oil GDP. The Magnitude and Uses of Oil Windfalls 75 The Financing of the Current Account Table 5-7 also indicates how current deficits and surpluses were fi- nanced. It is notable that throughout the period 1974-81 all coun- tries, whether in surplus or deficit, maintained sizable long-term capi- tal inflows. In some cases, especially Indonesia and Trinidad and Tobago, it was considered advantageous to use excellent credit rat- ings to consolidate relationships with foreign lenders. Direct invest- ment flows, however, were disappointing in 1974-78 and fell below 1970-72 levels, partly because of the stagnation of private invest- ment and partly because of the end of the oil development phases in Ecuador and Trinidad and Tobago. Foreign borrowing was used to accumulate reserves and to fi- nance short-term and other capital outflows. Much of these were un- official and unrecorded. Although in 1974-78 the sum of dollar cur- rent accounts and direct investment flows was -$12.2 billion for the six countries, the increase in their medium- and long-term debt less reserves was $16.7 billion, as shown in table 5-8.9 Further, in 1979-81 when current accounts plus direct investment flows totaled a surplus of $10.3 billion, medium- and long-term debt less reserves increased by $4.8 billion. Borrowing facilitated by oil income thus permitted a considerable degree of asset diversification abroad by residents of the exporting countries, and registered gross foreign debt rose over the whole per- iod 1974-81 although current accounts were, on average, not far off balance. Notes 1. It is not possible to obtain national accounts data which are consistent both over time and across countries and which isolate the oil sector. Inclu- sion of non-oil mining affects the results only if the sector is large and if its deflator changes greatly relative to that of nonmining GDP. 2. This projection should not be taken to imply that countries would actu- ally have followed such a pattern in the absence of terms of trade changes. For example, Venezuelan constant-price mining output would certainly have declined relative to the non-oil economy during 1972-78 for technical reasons. Rather, it provides a common basis for the comparative assess- ment of the impact and use of windfalls. No non-oil growth effect is yet attrib- uted to the windfalls, for reasons described below. 3. For an analysis of this (Tornquist-Theil) index as a discrete approxima- tion to a divisia index, see Diewert (1976). 4. Residuals are small and have been subsumed into the price term. Since the actual and projected non-oil economy are identical, real apprecia- Table 5-8. Net Borrowing and Indebtedness, 1974-83 (billions of U.S. dollars) Trinidad Item Algeria Ecuador Indonesia Nigeria and Tobago Venezuela Total 1974-78 Current account plus direct investment -8.2 -1.2 -2.9 -0.2 1.1 -0.7 -12.2 Change in MLT debt less reserves 9.4 1.5 6.1 -0.1 -1.5 1.3 16.7 Short-term debt, 1978 2.0 1.2 1.8 2.4 0.1 8.0 15.5 1979-81 Current account plus direct investment -1.3 -2.3 3.3 0.9 0.6 9.1 10.3 Change in MLT debt less reserves 0.2 2.0 0.3 1.5 -1.3 2.3 4.8 Change in short-term debt 0.3 0.8 1.5 2.0 0.1 9.0 13.7 1982-83 Current account plus direct investment -0.3 -1.3 -11.7 -11.4 -1.9 0.2 -26.3 Change in MLT debt less reserves -0.6 2.0 7.1 8.8 1.5 2.1 20.9 Change in short-term debt -0.4 -0.8 1.4 2.3 0.2 -2.4 0.3 Note: MLT debt is medium- and long-term debt. Sources: IMF International Financial Statistics; World Bank, World Debt Tables. The Magnitude and Uses of Oil Windfalls 77 tion leads to a reduction in the price effect for a sector such as oil whose prices are determined internationally. 5. Some of the following tables appeared in Gelb (1986a); the tables here correct errors in the earlier version and therefore differ in some places. 6. The unusual decrease for Ecuador is explained by (a) the fact that oil rev- enues also accrue to special funds outside government as defined here, and (b) the easing of certain non-oil taxes after 1974. In some cases na- tional oil companies also received part of the windfall. 7. For cross-country tax norms and indications of the relation of non-oil taxes in oil countries to these, see Tait, Gratz, and Eichengreen (1979). 8. Domestic oil subsidies have been mostly implicit because oil produc- tion allocated to domestic use is valued at domestic rather than world prices. 9. It is possible for a country to register a negative total dollar current ac- count over a number of years and still have a positive average ratio for the current account to GDP if the dollar value of GDP changes. For this reason the sign patterns of tables 5-7 and 5-8 may diverge, as they do for Nigeria in 1974-78. Chapter 6 Sectoral Shifts, Real Exchange Rates, and the Dutch Disease The standard "booming sector" three-sector model reviewed in chap- ter 3 predicts two main consequences of oil-led domestic expendi- tures. Real exchange rates will tend to appreciate as the relative price of nontraded sectors rises. This tendency will be associated with a production shift toward the nontraded sectors, which will lead to greater dependence on oil for foreign exchange. These ef- fects are quantified in this chapter for the sample countries. Predic- tions of neoclassical or two-gap growth theory-that investing the windfalls domestically will cause the growth rate to increase-are in- vestigated in chapter 7 together with the efficiency of fiscal linkage. For various reasons, as noted in chapter 3, actual economies may deviate from the stylized patterns predicted by a simple model. Al- though the real exchange rate (defined as the price of nontraded goods relative to traded goods) may appreciate, trade liberalization could cause the more empirically useful measure, the real effective ex- change rate (defined as the level of domestic prices relative to for- eign prices), to move in the other direction. Measurement of the real exchange rate itself is complicated by the continuous rather than discrete nature of tradability. Also, services have large weight in the nontraded sector. Many of the most easily measurable serv- ice prices are those of public utilities controlled by the government. Rationing in such sectors could make the underlying pressure of de- mand difficult to measure. Policies which dampen multiplier effects from increased government spending can constrain resource pulls to- ward the nontraded sectors. If traded sectors can respond strongly to investments financed by oil revenues, product market pulls to- ward the nontraded sectors may be counterbalanced in the medium to long run. According to the analysis of table 3-3 this is particu- larly likely if labor markets are slack; then nontraded sectors do not draw labor from the traded sectors, and the economy is able to 78 Sectoral Shifts, Real Exchange Rates, and the Dutch Disease 79 enter a phase of rapid, demand-led growth. The tendency of a partic- ular sector to expand or contract depends on more than its tradabil- ity. It also depends on the extent to which its factor demands com- pete with those of the most rapidly growing sectors-notably construction-which will be heavily affected by the strong empha- sis on investment that characterizes rising public demand. Real Effective Exchange Rates Estimates of real effective exchange rates for 1970-84 for the six ex- porters are shown in table 6-1. The trade-weighted real effective exchange rate is defined as the ratio of the domestic consumer price level of the oil exporter to the geometrically trade-weighted price indi- ces of its major trading partners converted at average exchange rates. Averages for 1970-72 are taken as the base, 100; an increase in- dicates appreciation or loss of competitiveness. The unweighted mean (excluding Iran) indicates average real appreciation of 12.9 per- cent in 1974-78, 19.2 percent in 1979-81, 34.7 percent in 1982-83, and 46.3 percent in 1984. During the 1970s, however, the unit value of manufactures im- ported by developing countries (Muv index) rose relative to the Table 6-1. Real Effective Exchange Rate Movements, 1970-84 (1970-72 = 100) Average trade-weighted real effective exchange rate Country 1974-78 1979-81 1982-83 1984 Algeria 96.3 104.4 114.1 123.3 Ecuador 120.6 127.9 131.9 119.2 Indonesia 133.0 104.4 108.5 91.5 Nigeria 129.4 162.6 194.2 287.3 Trinidad and Tobago 105.6 115.8 139.5 169.8 Venezuela 92.9 100.1 119.8 86.9 Unweighted mean 112.9 119.2 134.7 146.3 United States 92.3 93.4 105.2 n.a. Iran 100.4 119.2 194.2 n.a. MUV index, 139 138 110 106 n.a. Not available. a. Unit value index of manufactures imported by developing countries relative to U.S. consumer price index. The 1985 value is 111. Sources: IMF International Financial Statistics; World Bank and United Nations, Monthly Bulletin of Statistics, various issues. 80 Comparative Analysis price levels of most of the trading partners of the sample countries. This rise was both because of oil and other primary intermediate price shocks, and because bursts of investment, largely in the devel- oping oil exporters themselves, boosted costs of imported plant and equipment (an effect discussed at greater length in chapter 7). This relative price shift reduced the extent of average effective appre- ciation relative to the MUv index and was one factor limiting any fall in consumption and investment prices relative to the cost of do- mestically produced non-oil commodities. Such a fall would other- wise have been expected to result from real currency appreciation. The oil exporters were, in fact, cushioned from increased import price shocks by their own real exchange appreciation. Because of low, controlled domestic energy prices, until the early 1980s they were also spared the inflationary supply-side shock felt by their trad- ing partners. The Cost of Foreign Borrowing Real exchange rates appreciated sharply relative to the dollar which, as shown in table 6-1, weakened during the first oil shock. The dollar began to recover strongly during the second shock with the adoption of a tight monetary policy after 1979. This implied that, in terms of domestic prices, real interest rates on dollar debts and returns to dollar assets were substantially negative for the sam- ple until 1979, averaging -6.8 percent in 1974-78. Real interest rates were even lower in terms of the cost of investment goods, partic- ularly construction prices, which tended to lead domestic inflation in the sample countries. The highly negative returns on foreign financial assets increased the urgency of domestic spending. Countries scrambled to acceler- ate investment projects to avoid increases in costs and to make heavy industrial plants inflation-proof relative to competing invest- ments which other countries were expected to make in the future. This introduced a destabilizing feedback into public decisions dur- ing the first oil cycle. The more rapidly spending accelerated, the greater was the tendency toward real appreciation and cost infla- tion, the more negative were real foreign interest rates, and the stronger was the incentive to absorb real resources. This speculative bubble in real assets was abruptly punctured in 1979, when the combination of the appreciating dollar and higher real U.S. interest rates resulted in a switch to a positive real foreign in- terest rate, averaging 3.6 percent for the sample in 1979-81. The wide- ranging consequences of these interest rate fluctuations for the qual- Sectoral Shifts, Real Exchange Rates, and the Dutch Disease 81 ity of large industrial programs and their returns are reviewed in chapter 7. The Policy Impact Table 6-1 indicates large differences, however, in the extent and tim- ing of real effective exchange rate movements. To a great extent, these reflected differences in the policies of the various countries. By far the greatest real appreciation was experienced by Nigeria, whose effective exchange rate had doubled by the end of the sec- ond oil shock. Nigerian ports, roads, and distribution systems were particularly inadequate to handle the flood of imports and the shifts in domestic trade patterns which followed rapid urbanization. This resulted, in turn, from the urban-biased pattern of public spend- ing. Congestion provoked initially rapid real appreciation. This was then increasingly sustained through tightened quantitative import controls as revenues and the expenditure stimulus diminished. Partly because of its restrictive and increasingly inefficient system of foreign exchange rationing, Nigeria experienced a severe decelera- tion in growth. The average size of its nonmining economy in 1979-81 was 29 percent lower than it would have been had the coun- try managed to sustain its preshock growth trend.' Indonesia's experience provides an interesting contrast. Its initial exchange appreciation in 1974-78 was, if anything, more violent than that of Nigeria. In November 1978, however, fearing the im- pact of domestic costs on labor-intensive non-oil export sectors and believing that the oil boom was at an end, Indonesia devalued by 50 percent, from 415 rupiah to the dollar to 625. As the dollar strengthened against other major currencies, the rupiah was al- lowed to drift downward, to 700 to the dollar by March 1983. It was then again devalued to 970 rupiah to the dollar. After 1979 the government also began to implement a more conservative fiscal pol- icy. A substantial part of the second windfall was saved abroad. This was in spite of an official requirement that the budget be bal- anced rather than in surplus.2 Thus, in tables 5-2 and 5-3, Indo- nesia's trade and nonfactor service effect increased from 5.3 percent of nonmining GDP in 1974-78 to 9.6 percent in 1979-81. Relative to the base period, it saved 42 percent of the second windfall abroad. The corresponding trade and nonfactor service effects for Nigeria were only 2.8 percent and 0.1 percent. Indonesia was, in fact, the only country in the sample to imple- ment a determined policy of expenditure reduction and exchange rate realignment before the fading of the second oil boom. In this 82 Comparative Analysis sense, its macroeconomic management in 1973-81 could be consid- ered as the most prudent in the sample. As is clear from table 6-1, it was fairly successful in avoiding prolonged exchange overvalu- ation and was quick, relative to other countries, to adopt a real effec- tive depreciation in response to falling oil receipts. Indonesia was also quick to adjust other policies to changes in the availability of oil revenues (see chapter 12). At least until 1981, two notable exceptions to the rule of ex- change appreciation were Algeria and Venezuela. These countries ac- tually managed to depreciate their effective exchange rates during the first boom period. Although price controls played a central role in both cases, the macroeconomic clearing mechanisms which substi- tuted for real appreciation were completely different, as noted in table 4-2. Venezuela had long since completely adapted to the sta- tus of an oil monoexporter, so that virtually all non-oil sectors re- ceived some protection. After 1973 there was particularly great scope for the relaxation of tariffs and the quantitative import con- trols that had rendered the non-oil economy essentially noncompeti- tive with imports. Venezuela had historically been a low-inflation country, and the response to the inflationary burst which accompa- nied the start of the first oil boom was, indeed, to strengthen price controls on a wide set of wage goods and to satisfy demand, where needed, through imports. Accordingly, there was a tendency for con- sumer prices to rise less than the cost of non-oil output. This was re- flected in the substantial real consumption effect shown in table 5-2. In addition, domestic oil prices were held constant, which shielded the economy from supply shocks and their inflationary im- pact, and the labor force was augmented by Colombians. Algeria, in contrast, had a highly planned, centralized, socialist economy. Virtually all prices and wages were controlled. Through- out the windfall period, foreign trade and capital flows were tightly regulated. The public investment program was very heavily ori- ented toward hydrocarbon development and heavy industry. Conse- quently, it had a high direct import content. Labor was in elastic sup- ply because of chronic underemployment. These factors helped to relieve pressure on the real exchange rate. In addition, private pur- chasing power, derived from the multiplier effects of domestic pub- lic spending, was drawn off in two distinct ways. First, non-oil taxes were heavy and buoyant in comparison with those of other ex- porters (see tables 5-4 and 5-5). Second, cash balances were accumu- lated on a massive scale. In 1973-79 the seigniorage extracted by the Central Bank averaged 7.5 percent of nonmining GDP, three times that of the other countries in the sample. This constituted a third leakage from the oil income stream (the other two being the Sectoral Shifts, Real Exchange Rates, and the Dutch Disease 83 high import propensity of public investment and high non-oil taxa- tion). The holding of money balances appears to have been associ- ated with rationing in markets for wage goods and housing (see chap- ter 10). The Algerian strategy therefore involved deferring consumption to allow for a period of exceptionally high investment. In table 5-2 Algeria's value consumption effects total only 19 percent of the value investment effects, whereas Venezuela's consumption effects total 43 percent of investment effects. The Algerian private absorp- tion effect is only 11 percent of total absorption, whereas the Vene- zuelan private sector accounted for 44 percent of the total absorp- tion effect. As the Algerian strategy shifted after 1978 toward liberalizing markets and investing in housing (which has a higher do- mestic resource content), the real exchange rate began to appreciate (see table 6-1), and the share of private absorption began to rise (see table 5-3). Real Wages during the Boom and Relative Price Effects Information on the evolution of real wage rates is in some cases sketchy, but indicators summarized in table 6-2 suggest certain broad patterns. Algerian real wage gains were initially moderate. They became substantial, particularly in agriculture after 1976, as the heavy industrial investment strategy began to be moderated. Wage levels rose slowly at first in Ecuador, but advanced rapidly after 1979 as exchange overvaluation and fiscal and current deficits became more pronounced. Real wages in Indonesia increased after recovery in 1976 from the Pertamina crisis. There has been much controversy over the evolution of real wages in Indonesia, but the table suggests that they have, at least to some extent, reflected the country's favorable economic performance, both urban and rural. At the same time, however, the breaking down of traditional ten- ure and cropping patterns threatened to displace rural labor. Urban Iranian wages (not shown in the table) rose particularly strongly until 1977. This may have resulted from a policy of holding down food prices (as discussed below), which had an adverse impact on rural areas. Real wage levels appear to have been almost static in Ni- geria over the boom periods, with large nominal increases (such as those granted by the Udoji civil service awards in 1975) rapidly being eroded by inflation. This was apparently caused by an acute shortage of wage goods, especially food, as domestic factors shifted toward satisfying the demands of public investment programs and imports were rationed.3 Real wage levels rose sharply in Trinidad and Tobago, where trade union pressure, especially among the Table 6-2. Real Wage Trends, 1974-84 (1975 = 100) Country 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 Algeria Agriculturea 108 100 114 109 133 142 157 171 n.a. n.a. n.a. Nonagriculture, 108 100 106 101 121 127 143 144 n.a. n.a. n.a. Deflated unit labor cose 91 100 104 107 122 124 128 129 127 142 n.a. Ecuadorc Agriculture 105 100 111 98 87 110 180 n.a. n.a. n.a. n.a. Construction 104 100 119 105 95 116 172 n.a. n.a. n.a. n.a. Indonesia Urband 101 100 105 118 123 135 153 n.a. n.a. n.a. n.a. Rural (1977 = 100)e n.a. n.a. n.a. 100 n.a. n.a. n.a. n.a. n.a. 115 n.a. Nigeria' Government 82 100 83 73 63 57 n.a. n.a. n.a. n.a. n.a. Construction 131 100 79 70 57 82 74 62 58 63 44 Trinidad and Tobagog Textiles 95 100 111 113 122 109 113 109 157 n.a. n.a. Sugar n.a. 100 193 198 205 218 246 295 313 n.a. n.a. Venezuela' Blue-collar workers 106 100 94 129 135 106 106 103 n.a. n.a. n.a. White-collar workers 107 100 98 117 112 110 97 97 n.a. n.a. n.a. n.a. Not available. a. Until 1979, national minimum wage indices for unskilled workers in agriculture and nonagriculture; thereafter, Salaire National Global. Source: IMF. b. Labor cost index from World Bank, deflated by consumer price index. c. Legal minimum wage in agriculture and for construction artisans. Source: IMF. d. Until 1975, average daily wages for skilled and unskilled workers in construction, Jakarta, Source: World Bank. After 1975, average real wages in manufacturing. Source: Warr (1986). e. Average increase of daily wages in hoeing, planting, unskilled and skilled construction deflated by cost of living deflators for six villages in rural Java. Source: World Bank data. f. Government minimum wages; construction market wages, Benin. Source: World Bank data. g. Average weekly earnings in sugar, and of production workers in textiles. Source: World Bank data. h. Average wages, blue-collar workers (obreros) and white collar workers (empleados). Source: World Bank data. 86 Comparative Analysis sugar workers, was strong. Real wages also increased in Venezuela during the booms, although they stabilized and then began to fall back as the economy stagnated. These data suggest that a combination of demand-led growth, which tightened labor markets, and the price shifts discussed next, which usually favored private consumers, should have resulted in considerable increases in private consumption through the 1970s. A possible exception is Nigeria. The price effects in tables 5-2 and 5-3 indicate the extent to which real consumption may have been increased or decreased because of relative deflator movements. Their unweighted average confirms a tendency for private consumption deflators to rise less rapidly than the deflator for nonmining GDP, and for investment deflators to rise more rapidly. This is largely because of the tendency for relative con- struction costs to increase with the investment boom. But except in Venezuela and Trinidad and Tobago (both of which implemented tight price controls on wage goods) and in Ecuador, the negative price effects were small. Nigerian consumers, in particular, appear not to have gained from the cheapening of imports in 1974-78, which one would expect to have accompanied the appreciating naira. Their gains in 1979-81 are more than offset by the invest- ment price effect, a result at first sight inconsistent with the degree of real effective exchange rate appreciation observed in table 6-1. Problems of data quality affect this measurement in particular be- cause consumption tends to be determined residually in national ac- counts. Other explanations for the initial failure of real consump- tion and wages to increase more rapidly in Nigeria probably lie in the congestion effects already noted, which overwhelmed the distri- bution system. Such effects were particularly serious in the poorest countries because of their shortage of basic infrastructure.4 The de- clining efficiency of resource allocation caused by the tightening of quantitative restrictions could have produced a similar effect after the first oil boom. Part of the windfall gain was then simply ab- sorbed by the lower efficiency of Nigeria's non-oil economy, which raised prices relative to factor costs and so offset the relative cheapen- ing of non-oil imports. Measuring the Dutch Disease Many observers have noted a tendency for agriculture to lag be- hind other sectors in oil-exporting countries, but care needs to be taken in interpreting this phenomenon. Cross-country studies such as that of Chenery and Syrquin (1975) indicate that growth of in- come per capita is normally associated with a considerable shift in Sectoral Shifts, Real Exchange Rates, and the Dutch Disease 87 the pattern of output, first from primary production to industry and later to services. These trends imply a "normal" shift from the non-oil sectors conventionally considered as tradable-agriculture (broadly defined to include fisheries and forestry) and manufactur- ing-toward services and construction. Although there are exports and imports of transport, financial, construction, and other ser- vices, these are not usually considered to be sectors where demand and supply can be easily equilibrated through net imports. The tendency for the nontraded sectors to grow faster than the traded sectors might be weaker if measured in constant prices, how- ever, or might not exist at all. This is because growth of income per head is also associated with an increase in the price of nontraded goods relative to those of traded goods, the so-called Balassa effect. This effect causes the currency of poorer countries to be underval- ued, relative to those of richer countries, by the criterion of purchas- ing power parity.' To assess the impact of the windfalls on economic structure, non- oil economies are decomposed into four major sectors: agriculture, manufacturing, construction, and services. Norms for shares in out- put (SNi) are derived for each sector i.6 So are changes in these norm shares owing to increased real income per capita, which is rep- resented by nonmining output per capita. These are compared with constant-price shares (Si) and changes in shares for the sample after 1972.7 The Dutch disease index, DD, is then defined as DD = (SNag + SNma) - (Sag + Sma) It measures the shortfall in the share of tradables (agriculture and manufacturing) in the non-oil economy relative to their "normal levels." Changes in DD indicate the direction in which the con- stant-price composition of output is evolving relative to a "normal" pattern. The results are given in table 6-3. As the initial conditions indi- cate, Algeria, Trinidad and Tobago, and Venezuela had severely skewed non-oil economies before the oil price rise. The exceptional in- itial Algerian DD value reflects the disruption of agriculture and man- ufacturing by the eight-year revolutionary war, which was followed by independence and the subsequent departure of French colonial proprietors. The government attempted to overcome this problem- with poor results, especially in the agricultural sector-by the strat- egy of autogestion or workers' self-management. The weakness of the Algerian traded sectors could be sustained by the foreign ex- change from remittances and oil. Trinidad and Tobago and Vene- zuela had long since adapted to the economic structure of an oil Table 6-3. Sectoral Structure and the Dutch Disease Algeria Ecuador Indonesia Iran Nigeria Trinidad Venezuela Sector Actual Norm Actual Norm Actual Norm Actual Norm Actual Norm Actual Norm Actual Norm Sectoral shares in nonmining GDP and modified Chenery/Syrquin norms, 1972 (percent) Agriculture 11.Ob 25 23.0 26 45.1 46 21.8 18 39.2 38 5.8a 16 7.6a 14 Manufacturing 15.8a 20 20.4 19 11.0 11 18.5a 25 5.lb 15 22.0 26 19.5a 27 Construction 13.4b 5 4.8 5 4.3a 3 6.3 6 10.7b 4 8.0O 6 6.0 7 Services 59.8 50 51.8 50 39.7 40 53.3 51 45.0 43 64.2 52 66.9a 52 Mining 17.9 4 2.0 4 12.1 6 28.5 3 15.1 5 8.8 2 20.5 2 Average annual change in share at constant prices, 1972-81 Agriculture -0.38 -0.82 -0.91 -0.77 -1.50 -1.31 -1.44 -1.02 -1.90 -0.67 -0.31 -0.49 -0.10 -0.32 Manufacturing 0.52 0.28 0.37 0.27 0.77 0.34 0.34 0.49 0.48 0.11 -1.03 0.20 -0.04 0.06 Construction 0.82 0.09 -0.12 0.07 0.26 0.10 0.13 0.06 0.53 0.03 0.37 0.05 0.02 0.03 Services -0.98 0.45 0.68 0.44 0.48 0.86 1.04 0.40 0.88 0.53 0.99 0.35 0.13 0.24 Dutch disease index 1972 18.2 1.6 0.9 2.7 8.7 14.2 13.9 Change, 1972-81 -7.0 0.5 -2.4 5.7 8.5 10.5 -1.3 Change in current prices, 1972-81 -3.5 12.5 -0.4 12.7 2.4 4.7 -3.2 a. More than one standard deviation from norm. b. More than two standard deviations from norm. Following results of Chenery and Syrquin (1975), standard deviation is approximated by 0.25 norm. Source: World Bank, World Tables data base. Sectoral Shifts, Real Exchange Rates, and the Dutch Disease 89 monoexporter. Indonesia and Ecuador, in contrast, had sectoral structures that were almost normal-and so, surprisingly, did Iran. Table 6-3 also shows the average annual changes in norms and ac- tual shares during 1972-81 and the implied change in the Dutch dis- ease index over the decade. A negative value for the change in DD in- dicates a strengthening of the non-oil tradables relative to their expected share, whereas a positive value indicates a shrinking of the non-oil tradables relative to the norm. Table 6-4 provides some additional indicators of sectoral evolu- tion: indices of agricultural and food output per capita for 1974-83, with the period 1969-72 taken as the base. To facilitate comparison, Table 6-4. The Performance of Agriculture, 1974-83 (1969-72 = 100) Index of output per capita Country 1974-78 1979-81 1982-83 Algeria Agriculture 89 83 70 Food 88 82 69 Ecuador Agriculture 109 119 113 Food 109 120 114 Indonesia Agriculture 109 123 127 Food 112 127 133 Nigeria Agriculture 91 91 84 Food 92 92 85 Trinidad and Tobago Agriculture 95 86 81 Food 96 87 81 Venezuela Agriculture 104 102 101 Food 105 104 104 Unweighted average Agriculture 100 101 Food 100 102 98 World Agriculture 103 104 105 Food 104 106 106 Developing world Agriculture 103 105 104 Food 104 106 106 Source: U.S. Department of Agriculture. 90 Comparative Analysis the table also shows the corresponding indices for the world and for all developing countries. These tables are complemented by table 6-5, which gives the price of food relative to other products in the sample countries and an index of world food prices relative to a price index of thirty-three nonenergy commodities. Table 6-3 indicates that Algeria, Venezuela, and Indonesia man- aged to strengthen their non-oil tradable sectors over the decade. The first two, however, had very small initial non-oil traded sectors and, as noted above, were able to limit domestic market forces by price controls, import liberalization (Venezuela), and constraints on private spending (Algeria). When agricultural sectors are already small, it is less likely that they will continue to reduce their share in the "normal" manner. The exceptional case is Indonesia, which managed to sustain a rather strong performance in the non-oil traded sectors through the period of the oil price booms. This was due to a combination of good fortune and good policies. The develop- ment of high-yielding and disease-resistant rice varieties had a great impact on agricultural performance. But their dissemination would not have been possible without the unusually broad develop- ment strategy followed by the government. Input subsidies may also have played an important role; in addition, the measures taken to prevent the real exchange rate from moving too far out of line after 1978 stimulated non-oil exports (see chapter 12). Table 6-5. The Relative Price of Food, 1974-83 (1970-72 = 1.00) Ratio of food price index to general price index Country 1974-78 1979-81 1982-83 Algeria 1.15 1.26 n.a. Ecuador 1.15 1.13 1.20 Indonesia 1.12 1.11 1.09 Nigeria 1.09 1.09 1.11 Trinidad and Tobago 1.09 1.05 1.09 Venezuela 1.13 1.30 1.36 Unweighted average 1.12 1.16 1.17 World, 1.12 1.04 0.99 Iran 0.95 n.a. n.a. n.a. Not available. a. Index of food prices relative to index of thirty-three commodities excluding en- ergy; weighting by developing countries' export values, 1977-79. Source: World Bank data. Sectoral Shifts, Real Exchange Rates, and the Dutch Disease 91 Ecuador avoided a substantial increase in its Dutch disease index. In this, it was assisted by the rapid growth of the new seafood indus- try. Private manufacturing and agribusinesses were stimulated by generous tax rebates (partly responsible for poor non-oil fiscal per- formance), other incentives, and credit extended by the govern- ment from its borrowing abroad (see chapter 11). The most marked shifts to the nontraded sectors occurred in Trini- dad and Tobago and Nigeria. Trinidadian industry was adversely af- fected by the decline of petroleum refining with the falloff in world crude exports. Its agriculture (particularly sugar) contracted despite mounting subsidies as labor moved off the land into construction and other public programs. These were introduced to alleviate the chronic problem of unemployment but paid wages far above those available in rural activities (see chapter 14). Nigeria's agricultural stagnation reflected a number of factors. Its development strategy was strongly urban-biased and emphasized the nontraded sectors, especially road construction and the spread of primary education, then secondary and university education. Unlike the rural econo- mies of Java and Bali, the Nigerian rural sector did not have a labor surplus. The manpower demands of the public construction pro- grams had an immediate impact on the supply of farm labor. The government lacked the institutions and perhaps the will to raise smallholder agricultural productivity, and the performance of large- scale irrigation schemes was poor. As noted in chapter 13, some observers consider that much might have been done to improve Nigeria's agriculture, even though there was no revolution in dryland farming comparable to the Green Revolution in rice. With a static technology and a shrinking labor force, agricultural supply grew little. As table 6-4 indicates, Ecuador, Indonesia, and Venezuela man- aged to raise domestic food and agricultural output per capita dur- ing the 1970s, although Venezuela began from an extremely small base. The table confirms the superior performance of Indonesia, both in relation to the rest of the sample and to non-oil countries. By the end of the period Indonesia's food output per capita was one- fourth larger than it would have been had supply increased at the worldwide average rate. In contrast, the performance of agriculture in Algeria, Nigeria, and Trinidad and Tobago is seen to be poor. Iran, too, experienced a rapid decline in the share of agricultural products in output over the period for which data are available. A common criticism leveled at the policies of mineral-exporting countries is that they tend to keep food cheap because of pressures from groups in the urban areas, and that cheap food policies are sus- tained by spending mineral rent on imports. But table 6-5 suggests 92 Comparative Analysis a different story for the sample in the decade of the oil windfalls. In 1974-78 world food prices were higher, relative to other non- energy commodities, than in 1970-72. This was paralleled by price movements within the sample.8 After 1978 relative food prices fell worldwide but rose further in the oil exporters, so that the aver- age relative increase was greater in the sample despite increased imports. These relative price movements point to the importance of factor market pulls and of potential import competition as influences on sec- toral evolution. In all the countries, income per capita was considera- bly lower in agriculture than in other sectors.9 As public investment and construction expanded, labor was therefore drawn largely from agriculture, so that its output fell short of growing domestic de- mand. Although food imports by the six countries in the sample grew rapidly, this growth usually reflected the need to fill the widen- ing food gap rather than a deliberate policy of holding down the price of food relative to, say, manufactured goods. Thus in terms of output and net import trends, Nigerian agriculture resembled a traded sector in the simple model of chapter 3 because imports in- creasingly satisfied demand. But in terms of relative price, it be- haved more like a nontraded sector, mainly because its labor de- mands competed with those of construction. One interesting exception to this pattern was Iran. As indicated in table 6-5, the relative price of Iranian food was lower in 1974-77 than in 1970-72, although domestic output had fallen short of increas- ing demand and imports had increased. Iran did attempt to hold down food prices through imports, and this may have been one cause of the very large increase in real urban wages indicated by offi- cial data. But at the same time, the policy limited the transmission of multiplier effects from oil spending to the rural areas, with unfortu- nate consequences. ° Table 6-3 also indicates changes in the Dutch disease index on the basis of current-price sectoral shares (rather than constant-price shares). As expected, there is a more pronounced shift toward the construction and service sectors in four of the seven countries, mainly because of higher construction and lower manufacturing de- flators. The sign pattern of resource allocation shifts, as measured by DD, remains unchanged across countries. The conclusion that the non-oil traded sectors remained weak or tended to weaken except in Indonesia (and possibly Ecuador) is con- firmed by trade data. Total export volumes fell, on average, by 1.7 percent a year in 1972-81. Nonhydrocarbon export volumes con- tracted except in Indonesia (which maintained a fairly strong non- oil export performance across a wide range of primary and manufac- Sectoral Shifts, Real Exchange Rates, and the Dutch Disease 93 tured commodities) and Ecuador (which shifted toward processed products but experienced little overall increase). As discussed in the country chapters, Nigeria did extremely poorly in this regard, and some of its traditional exports disappeared. Trinidad and To- bago also did poorly, while Venezuela was unable to fill its coffee quota and saw a decline in cocoa exports as well. On the whole, the countries made little or no progress toward reducing the degree of dependence on oil. Notes 1. As described in chapter 13, Nigeria entered into a vicious circle of infla- tion, trade restriction, and fiscal deficits. 2. Indonesia's balanced budget policy regards foreign aid and loans for de- velopment purposes as revenue; for details see chapter 12. 3. As noted in chapter 13, there is some divergence of views on Nigeria's agricultural and consumption growth. Data in this area are not strong. 4. Lagos, it was sometimes claimed, had the world's only concrete-bot- tomed harbor, created when ships jettisoned cement that had hardened dur- ing the delays in unloading. 5. Changes in the patterns of output with income per capita and other vari- ables are analyzed only in current prices by Chenery and Syrquin (1975). For estimates of the Balassa effect, see Balassa (1964) and Kravis, Heston, and Summers (1978). 6. Chenery and Syrquin do not separate out construction. To do so has re- quired additional computations using the World Bank World Tables data base to derive the share of construction in industrial output. The Chenery- Syrquin equations are used to project primary, service, and industrial shares. Industry is then split into construction and manufacturing. 7. There are no cross-country constant price norms for sectoral struc- tures. Constant-price data are preferable for the sample because of the untyp- ical impact of oil booms on relative prices, particularly of construction. Be- cause of the Balassa effect the bias introduced by comparing constant-price shares and current-price norms makes it less likely that oil-exporting econo- mies will appear skewed toward nontradeds. 8. Although energy is included in the country price indices in table 6-5, its prices were typically stabilized below world levels. 9. Some studies of villages in Java have suggested that wages are lower in certain small-scale service and manufacturing activities, but these results do not seem borne out by estimates based on economywide data. For more discussion, see Gelb (1985a). 10. The potential effects of Iran's food policy had been noted by a report of the International Labour Organisation; see Pyatt and others (1972). Chapter 7 The Efficiency of Fiscal Linkage: Windfalls and Growth As is clear from chapter 5, public expenditures, and in particular a vast expansion in the size and scope of public investments, consti- tuted the main use of the windfalls. Growth, modernization, and the extension of national control over production were the main goals of these expenditures, although countries also had various sub- sidiary objectives. This chapter first considers the nature of public in- vestments and the factors that affected the performance of this vital fiscal linkage. It then notes the drift toward subsidies, which as- sumed mounting importance in most of the sample countries. Fi- nally, in the light of the investment drive, the growth performance of the sample is compared with that which might have resulted from the application of a simple neoclassical model. The Composition of Public Investment Programs It is difficult to compare the composition of public investments across countries because there is no uniform system of classifica- tion. Nevertheless, table 7-1 provides a rough indication of the broad distribution of public investments in the sample countries. It is intended simply to provide background in considering the invest- ment strategies followed by the various governments. Levels of development and income per capita differ greatly among the six producers. So does the role of government in the econ- omy and the weight of the public sector. All of the countries saw ex- tensive and growing public involvement in the hydrocarbon indus- try in the 1970s. The industry was nationalized in Venezuela in 1976 and almost totally nationalized in Algeria. Disputes in Ecua- dor with foreign oil companies over taxes and concessions resulted in CEPE, the state oil company, assuming a leading role, also in 94 The Efficiency of Fiscal Linkage 95 1976. Moves to consolidate national control over the oil industry forged ahead in Nigeria and Trinidad and Tobago, especially during 1974-79. By the end of the 1970s national control had essentially been achieved in all of the countries. Salient features of this proc- ess of extending national control are summarized in table 7-2 to facili- tate a comparison. In all countries this process preceded the first sharp oil price in- crease by many years. Even in the new producer, Ecuador, state own- ership of all mineral deposits had been asserted in the Constitution of 1945. The oil shocks were therefore not directly responsible for moves to nationalize the oil industry. In some cases, however, the price increases did accelerate the process. By providing more reve- nues to producer governments, they enabled foreign interests to be bought out at relatively little sacrifice. In Ecuador, for example, pay- ments to multinational oil companies for CEPE'S share of the indus- try came to $160 million between 1973 and 1979. That sum repre- sented only 26 percent of the country's 1976 oil exports. The extent of public involvement in sectors other than oil has dif- fered widely. At one extreme, in Algeria the central government and public enterprises are estimated to have accounted for over 90 percent of domestic investment. At the other extreme, in Ecuador the role of the public sector outside the traditional functions of admin- istration, defense, and provision of physical and human infrastruc- ture has been quite limited. Differences in the public role result both from the ideological tendencies of governments and from his- torical accidents. For example, both the extensive involvement in agri- culture of Algeria's socialist state and the considerable holdings in plantation crops and timber of Indonesia's relatively conservative government stemmed from the departure of colonial proprietors, French and Dutch respectively; they left at independence and, in the case of Indonesia, also at the time of the Irian Jaya dispute. In cer- tain countries, notably Venezuela, public oil income had led the state to expand its role in production over a long period, even when the government was basically oriented toward the private sec- tor. Maintaining a private economy is difficult when a government disposes of most of the investable surplus.' As would be expected, all the oil exporters experienced unparal- leled growth of the public sector's share in the economy after 1973. In most the public sector also extended its role as it moved toward direct participation in production, especially in the industrial sec- tor. Although virtually all governments expanded their activities in virtually all directions, the emphases differed. Some governments, taking a direct approach, allocated public capital primarily to the trad- Table 7-1. Sectoral Distribution of Public Investment (percent) Economic Social infrastructure infrastructure Hydro- Agriculture (transport and (education, Administration Country Period Industry carbons and fisheries communications) housing, health) and defense Other Algeria 1970-73 53.2 25.0a 8.5 9.5 21.7 7.1 n.a. 1974-77 53.5 25.9a 5.5 10.3 25.3 5.4 n.a. 1978-79 56.5 26.9a 3.3 8.9 26.5 4.8 n.a. 1980-84 38.6 15.7a 6.0 13.6 36.2 5.6 n.a. Ecuador 1973-74 6.0 6.0b 7.6 55.9 13.5 n.a. 17.0 1975-78 7.2 6.9b 8.5 53.9 28.1 n.a. 2.3 1979-81 13.2 13.Ob 3.9 54.0 26.9 n.a. 2.1 Indonesia 1969-72 7.6 3.3' 22.4 46.2 12.5 2.1 8.9 1974-78 19.2 8.8c 12.8 42.8 15.4 6.7 3.1 1979-81 17.0 8.6' 10.0 31.7 22.5 12.2 6.7 1982-84 17.4 14.7' 8.5 32.0 24.7 6.7 10.6 Nigeria 1970-74 10.5 4.7d 12.9 33.3 27.9 14.5 0.9 1975-80 13.7 5.0 7.2 37.5 24.3 17.3 n.a. 1981-85e 18.8 7.7 12.6 26.0 33.7 8.8 n.a. Trinidad and Tobago 1974-79 33.5 25.6 2.8 34.2 27.9 0.2 1.7 Venezuela 1970-74 22.5 6.3' 8.1 32.1 29.2 n.a. 8.0 1976-80 41.3 20.6b 7.4 36.6 14.7 n.a. n.a. 1981-85 31.8 24.9b 6.4 31.2 28.5 n.a. 2.1 n.a. Not available. a. Hydrocarbon sector. b. Petroleum sector. c. Mining. d. Mining, fuel, power. e. Projected. f. Mining and hydroelectricity. Sources: World Bank data; Trinidad and Tobago, Ministry of Finance (1980); Marshall-Silva (1984). Table 7-2. Hydrocarbon Industries in the Oil Exporters Gas reserves Domestic Country Start of production Major companies Extension of national control and development energy pricing Algeria Oil discovered by Total, ERAP After 1965 a determined drive toward com- Major emphasis French in com- (French) plus plete state control of hydrocarbon sec- on gas exports in mercial quanti- Mobil, Phillips, tor and replacement of foreign nation- industrial sector ties in 1956. Shell. als by Algerians. 1963, SONATRACH, the plans. First LNG state hydrocarbon company, established. and later piped; 1971, nationalization of all non-French for- SONATRACH world eign operations and assumption of 51% leader in LNG public share in French interests. SONATRACH capacity (exports controlled 70% of crude oil production, approximately 100% of exploration, refining, and market- 1 tcf a year) ing. Foreign companies could participate in although pricing exploration programs. 1980, $3 per barrel disputes held exploration fee levied on contract crude oil output well below customers. capacity levels in some periods. Ecuador Minor production Anglo-Ecuadorian 1945, Constitution asserted state ownership Estimated gas re- State electricity since 1911; large Oilfields (BP sub- of all minerals. 1971, Hydrocarbon Law serves 4 tcf. Not company (INECEL) deposits found by sidiary) until 1963; created state oil company CEPE. 1972, mili- a major focus of heavily subsidized. Texaco in 1967. Texaco-Gulf Con- tary government of General Rodriguez Lara development. Gasoline prices sortium became extended state control; Anglo-Ecuadorian among lowest in major producer in interests transferred to CEPE. 1974, CEPE world but raised 1964. purchased 25% of consortium; 1976, CEPE from $0.18 to bought out Ecuadorian Gulf's 37.5% share. $0.60 per gallon in Exploration diminishes; 1978, Hydrocarbon 1981. About 20% Law amended to increase incentives for of locally sold gas- private investment. oline may have been exported illegally. Indonesia Oil discovered in Royal Dutch/Shell, 1968, state oil company Pertamina formed Estimated gas re- Domestic energy 1883; commercial later Caltex. by merger of three locally owned compa- serves 24 tcf. Out- use, notably kero- production by nies. Largely autonomous until 1975. For- put expanded sene and diesel 1890. eign companies still produced 90% of oil, rapidly to 0.5 tcf fuels, heavily under contract of work, production sharing, in 1980. Major subsidized. and technical assistance programs. Al- LNG investments though 1971 Pertamina Law made the com- together with pany responsible for all petroleum activities, Japan are an im- control was pragmatic and oriented toward portant part of hy- raising reserves. drocarbon devel- opment. Iran Oil discovered by BP sole developer 1951, Prime Minister Mosadegh nationalized Gas reserves 400 BP in 1908. until 1951; con- all petroleum resources and established the tcf. Ambitious sortium after 1954 National Iranian Oil Company (NIOC). 1954, program to export also included consortium deemed owner of fixed produc- gas to Soviet Shell, Gulf, Mo- tion assets and Iran owner of oil reserves; Union and Europe bil, Exxon. 50% tax principle established. 1960s, a via IGAT I and number of joint ventures on 50-50 basis IGAT II pipelines between NIOC and foreign companies. 1973, in 1970s; curtailed NIOC took over all operations but the con- after 1978. sortium was allocated a marketing function. After 1978, trend to reduce foreign partici- pation in energy industries. Nigeria Oil discovered in Shell/BP partner- Mid-1960s, political and economic signifi- Gas reserves about Some subsidiza- 1956; exports be- ship and others. cance of oil recognized. 1969, Petroleum 75 tcf. Plans for tion of petroleum gan in 1958. Decree established state's option to part major LNG facility products and ownership of hydrocarbon industry at 51%. at Bonny Island other energy 1971, National Oil Company (NNoC) estab- for gas use in forms. lished. 1973, ultimate objective announced steel production to be a nationalized industry; NNOC began and for power production-sharing. 1974, public equity generation. (Table continues on following page.) Table 7-2 continued Gas reserves Domestic Country Start of production Major companies Extension of national control and development energy pricing share in all oil operations increased to 55%; 1979 to 60%, and BP's crude oil and mar- keting interests nationalized. 1977, new incentive package to accelerate exploration. Trinidad and Oil discovered in Shell, Textrin, 1969 Petroleum Act and 1970 Petroleum Estimated gas re- Domestic petro- Tobago 1857; Shell became and others. Regulations laid regulatory foundation. serves: 12 tcf leum prices sub- first producer in 1970, Trinidad-Tesoro (50.1% government- probable, 21 tcf sidized; domestic 1918. owned) established. Bought out Shell; 1974, possible. Output sales of natural established National Petroleum Marketing about 0.5 tcf. Shift gas underpriced. Company. 1976, bought out Texaco's mar- to gas-based in- keting operations. 1979, National Energy dustry a central Corporation (NEC) established. 1980 on, part of develop- drive to increase public control of hydro- ment strategy, carbons, but 1981 Petroleum Act amended which implies to increase private incentives to production continuing good and exploration. relations with private capital and expertise. Venezuela By 1928 leading Standard Oil of Gradual forty-year process of increasing do- Proven gas re- Domestic petro- exporter and sec- New Jersey (Ex- mestic control and knowledge of industry; serves 42.5 tcf. leum and energy ond largest pro- XON), Shell, Gulf, culminated in nationalization of oil in 1976. Nationalized in prices heavily ducer in the Texaco, Mobil, 1935-48, extension of regulation and estab- 1971. Gas for do- subsidized world. Sun, and others. lishment of the 50% tax principle. 1960, mestic use plays throughout period. state-owned Venezuelan Petroleum Corp. a role in the heavy (cvp) established to receive all future oil industrial strategy concessions, with foreign investment developed after through service contracts. 1971, Hydrocar- 1974. bon Reversion Law called for reversion of existing concessions to state ownership in early 1980s. 1974, President Carlos Andres Perez called for early reversion. 1976, total nationalization under Petroleos de Venezuela (PDVSA, formerly cvp). tcf = trillion cubic feet. Sources: U.S. Department of Energy (1981); U.S. Department of Energy, Office of International Affairs (1977); U.S. Embassy, Jakarta (1982); World Bank data. 102 Comparative Analysis able sectors to ensure a source of non-oil export revenues and to sub- stitute for imports, if and when oil revenues faltered. Others adopted the indirect strategy of strengthening the nontraded sectors -transport, communications, housing-and augmenting human capital. It might be argued that in the long run such infrastructural in- vestments are necessary to develop efficient non-oil traded sectors. The strategies of the sample countries, which have been outlined in table 4-2, are summarized here. The use of the term "strategy" is not intended to obscure the fact that the outcome did not always re- flect a comprehensive, consistent plan. Algeria possessed limited oil reserves but the fourth largest gas res- ervoirs in the world. It placed high priority on the development of the gas industry, mainly for sale in primary form, both liquefied (LNG) and piped under the Mediterranean through Sicily to Europe; Algeria's LNG capacity rapidly became the world's largest. The cost of Algeria's gas investments, at about $18 billion, was roughly equal to its gross foreign debt at the end of 1981. The gigantic Alge- rian public investment program displayed by far the strongest bias in the sample toward heavy industry: together with hydrocarbons, heavy industry accounted for almost half of all investment. Little in- vestment was allocated to agriculture. But the share going to social in- frastructure, particularly housing, rose markedly during the second windfall years, which also saw a gradual deceleration of the total in- vestment effort. This shift represented an effective expenditure- switching policy. The domestic resource content of investment was increased as the total level was cut, which sustained domestic de- mand and growth despite a rapid reduction in the current account deficit. In its public programs Ecuador emphasized the building of eco- nomic infrastructure. This was mostly urban, with particular atten- tion to transport, electrification, and water supply. Education, too, re- ceived a large share of the oil revenues. (Ecuador's uniquely complex system of revenue earmarking effectively reserved por- tions of the windfall for specified institutions.) Public industrial in- vestment went mainly to the oil sector and refining. Although the government had inherited a number of public manufacturing firms at the time of the first oil price increase, the state's role in manufactur- ing was actually cut back in the windfall periods, a tendency unique in the sample. Instead, the private manufacturing and agricul- tural processing industries were encouraged by generous credit, tax and tariff rebates, and subsidies. This policy resulted in slackened non-oil tax collection but did stem the proliferation of loss-making public industrial plants that plagued some of the other countries. The Ecuadorian government initially intended to use oil revenues The Efficiency of Fiscal Linkage 103 to fund land reform. But only limited initiatives were carried out, and these involved the colonization of new lands rather than the pur- chase and redistribution of existing farms. Indonesia pursued a strategy finely balanced between physical infrastructure, education, agricultural development (both smallhold- ers and estates) and capital-intensive industry. A $20 billion pipe- line of industrial projects including fertilizer, LNG, and steel plants was in place by the early 1980s. But half of this pipeline was speed- ily postponed or canceled with the onset of the oil glut. As de- scribed in chapter 12, Indonesia was fortunate in that the 1974 Pertamina crisis helped to delay the buildup of an irreversible invest- ment momentum. More than any other exporter, Indonesia directed a high propor- tion of its development spending to rural areas for irrigation works, roads, schools, and other small-scale infrastructural improvements. This distinctive emphasis followed from the critical role assigned to rural reconstruction and to self-sufficiency in rice when the Suharto government came to power in the mid-1960s (Timmer 1975). Indone- sia's rural investments were labor-intensive. The successful INPRES programs funded productive public works at the kabupaten (county) level, in some cases supplying only materials and relying on self- help for labor. Although the administrative problems involved in this wide distribution of investments should not be glossed over, In- donesia seems to have had a relatively effective rural administra- tion. An extensive dual civilian-military system had previously been established at the level of the 373 counties and towns and in many of the 60,000 villages as well. Another notable use of oil income was for heavy subsidies on ferti- lizer.2 Indonesia's gas reserves, used for fertilizer plants, were a key component of its rural strategy. An econometric assessment of these subsidies by Timmer (1985) finds that they were effective in pro- moting the use of fertilizers and raising yields, and concludes that their benefit-cost ratio was high (see chapter 12). Although INPRES loan programs later ran into difficulties, oil income was thus used to raise productivity and output in a key non-oil traded sector. The public investment strategy of Nigeria, which also started the period with a large agricultural sector, was very different from that of Indonesia. A substantial part of public investment was decentral- ized to the states, but relatively little was spent to promote agricul- ture. Moreover, agricultural spending was directed to large-scale, capital-intensive projects with a low rate of return.3 Bienen (1983) has estimated that actual capital expenditures in agriculture made by the federal government under the third plan (1975-78) were only 2.5 percent of the total, compared with a budgeted 6.5 per- Table 7-3. Social Indicators in the Oil Exporters Trinidad Unweighted Indicator Algeria Ecuador Indonesia Nigeria and Tobago Venezuela mean Comparator Primary enrollment rate 1970 76 97 77 37 94 94 79 88 1982 93 114 100 98 99 105 102 102, Secondary enrollment rate 1970 11 22 15 4 42 33 21 25a 1982 36 56 33 16 61 40 40 42' Radio receivers per thousand 1970 63 290 22 19 226 155 129 1982 211 357 131 73 281 359 235 251b TV receivers per thousand 1970 8 26 1 1 58 72 28 1982 66 69 23 5 266 111 90 86 Population per nursing person 1970 2,730 1,580 7,680 5,070 330 560 2,992 1982 740 570 2,300 3,010 380 380 1,230 1,636b Infant mortality rate 1970 144 108 121 141 43 59 103 1982 107 76 101 113 28 38 77 62b a. Middle-income countries. b. Unweighted mean: middle-income countries in Asia and the Pacific and in Latin America and the Caribbean. Sources: World Bank, World Development Reports and Social Indicator Data Sheets. The Efficiency of Fiscal Linkage 105 cent. Disbursement was a serious problem because of the lack of an effective rural administration.4 The bulk of Nigerian investment went to the nontraded sectors for roads, ports, and the extension of universal primary education and later secondary and higher levels. In June 1977 a master plan for a new capital city at Abuja was commissioned; Abuja was partly built when the revenue constraints of the oil glut halted construc- tion.5 At the end of the 1970s the country committed itself to an inte- grated iron and steel industry against the advice of feasibility stud- ies which projected-correctly-low returns, especially since the industry was to be geographically dispersed for political reasons. These commitments began to have an impact on budgets by the end of the 1970s. The investment programs of Trinidad and Tobago, like those of Algeria, laid heavy emphasis on the exploitation of natural gas re- serves. However, Trinidad chose to establish gas-based industries- ammonia, methanol, urea, and steel-rather than to sell gas in pri- mary form. (Although the option of exporting LNG was assessed at various stages, it was never adopted.) Trinidad and Tobago's invest- ment programs also placed considerable emphasis on physical infra- structure, notably road building and transportation. Agriculture was not a major focus for investments (as opposed to subsidies) despite a concern, frequently expressed, about the poor performance of both the export and food subsectors. Public investment in Venezuela emphasized the metals indus- tries. Studies of effective protection in the early 1970s had con- firmed the view that the country's non-oil comparative advantage was greatest in such industries. Public manufacturing investments centered on steel and aluminum plants and their associated infra- structural needs, particularly the massive hydroelectric capacity of the Guri dam. Investment in the Venezuelan oil industry also picked up sharply after nationalization. This was needed to make up for a period of rundown which had occurred as the years of for- eign ownership had drawn to a close. Physical and social infrastruc- ture, too, were well funded,6 and there was also an attempt to re- verse the decline in the output share of the smallholder agricultural sector. Overall, some 60 to 70 percent of the public investment spending of the exporters appears to have been on physical and social infra- structure. The effect of this on the efficiency of transport, distribu- tion, and communications systems is not easily quantifiable, but in some countries it was substantial. As table 7-3 shows, although in 1970 the sample countries had been considerably behind middle- income comparators in education, by 1982 they had caught up in Table 7-4. Macroprojects in Oil-Exporting Countries Average Total Total Project sector (percent) Number of Total cost cost cost! costl Rank among projects (billions of (millions of 1980 1980 oil developing Hydro- Other Infra- Country included U.S. dollars) U.S. dollars) GNP windfall countriesa carbons Metals industry structure Iran 108 119.6 1,107 1.57b 10.2b 2 30 7 9 54 Algeria 69 38.7 561 1.07 4.2 5 36 7 33 23 Venezuela 27 27.4 1,015 0.51 5.4 10 33 41 7 19 Mexico 59 26.0 441 0.18 5.1 12 46 17 12 25 Nigeria 19 14.4 758 0.17 0.9 15 26 11 16 47 Indonesia 44 14.4 327 0.23 1.1 16 41 18 16 25 Trinidad and Tobagoc 7 6.9 983 1.35 4.5 - 57 43 - - -Country not ranked and projects not included in the study. Note: Only projects costing more than $100 million are included. For this reason and for lack of data, Ecuador is excluded from the table. a. Ranked by total investment on projects exceeding $100 million each. b. 1977 GNP and oil windfall. c. Gas-based industrial projects only; includes Tenneco-Midcon LNG project proposed for 1988. Sources: Murphy (1983), table 2.5; Auty (1986a). The Efficiency of Fiscal Linkage 107 primary enrollment rates and had made large gains in secondary enrollment rates. The number of nursing staff available per person more than doubled, and infant mortality rates were cut by a quarter. These data suggest that part of the windfall was transformed into social consumption. Most of the remaining investments were concentrated in a rela- tively small set of resource-based industries, especially oil develop- ment and refining, gas, hydrocarbon-based chemicals, and metals processing. These were the main initiatives in promoting the traded- goods sectors. This pattern of investment proved to be a significant determinant of future events. The high proportion of infrastructural spending im- plied that the payoff to much of the public capital stock depended on the continued growth of the economy; the investment itself did not lead directly to income-generating activity. It also meant that in most countries a few key projects serving a few critical markets would determine the returns to that component of investment capa- ble of directly generating net foreign exchange and fiscal revenue. If oil incomes should decline and these projects should fail to re- turn profits (or worse still, fail to generate sufficient revenue to serv- ice their debts), fiscal revenue would suffer and the demand-led en- gine of growth would slow down. This would reduce the value of much infrastructural investment and have serious consequences for the rest of the non-oil economy. Resource-Based Industry and Other Large Projects As investment rates rose in the developing countries between the 1960s and 1970s (see table 3-1), so did the number, size, and complex- ity of "macroprojects," defined by Murphy (1983) as projects with capital costs greater than $100 million. Although some large oil- importing countries (such as Brazil) undertook many such projects, of the nineteen developing countries that invested the most in proj- ects exceeding $100 million each, all but five were oil exporters. The number and cost of large projects in five of the sample coun- tries plus Iran and Mexico are shown in table 7-4; total cost is also compared with 1980 GNP and with the oil windfall in 1980. Ecua- dor, whose public industrial programs were far more modest than those of the other countries, is excluded. Only actual and some pro- jected investments in gas-based industry are included for Trinidad and Tobago. The table shows that Iran (occupying second place worldwide after Saudi Arabia) initiated or planned 108 projects averaging more than $1 billion each, with a total capital cost ten times its 1977 oil 108 Comparative Analysis windfall. The large projects identified in Algeria, Venezuela, Mex- ico, and Trinidad and Tobago represented four or five times the 1980 oil windfall of these countries. The poorest producers, Indone- sia and Nigeria, were somewhat less inclined to mortgage oil for large projects and slower to start, but their investments of this type were still considerable. The large projects identified in table 7-4 repre- sent roughly four and a half years of average oil revenue during 1974-81. For a representative country (not including Trinidad and Tobago) about 35 percent of the projects were related to hydrocarbons. An- other 19 percent were in metals (more in Venezuela) so that at least 54 percent of the projects were in the area of resource-based indus- tries (RBI). Another 16 percent was in other industry, and 32 per- cent in infrastructure. In terms of numbers of large projects, there was therefore quite heavy emphasis on the traded sectors and RBI, though some of the infrastructural projects were very large. Data compiled by Murphy (1983) suggest that the larger projects, which were also the more complex and technically demanding, had a greater tendency to overrun initial cost and time budgets than did smaller projects. By 1980 one-third of the largest projects had ex- perienced cost overruns averaging 109 percent. Overruns on the smaller projects were less frequent and more modest, averaging 30 percent. Delays of between one and two years plagued half the trou- bled projects; a further 25 percent experienced delays of three to four years. Yet these estimates greatly understate the true extent of cost and time overruns. Many projects were not completed by 1980, and a number may never reach completion because of the ef- fects of the oil glut and world recession. Unfortunately no later sur- vey of the experience of such large projects is available so that the ro- bustness of Murphy's conclusions (such as the relationship between scale and overruns) is open to question. As is clear from table 7-4, resource-based industrialization was a key element in the growth plans of the oil exporters. Auty (1986a, b) has studied the RBI strategies of eight oil-exporting countries and their outcomes. His sample includes four of the countries in this study: Trinidad and Tobago, Venezuela, Indonesia, and Nigeria.7 The following discussion draws heavily on his analysis. Resource-based industry was a dominant part of the develop- ment strategy of Trinidad and Tobago and Venezuela, which were among the first developing countries to invest in RBI. Trinidad and To- bago opted for a portfolio of gas-based investments that was fairly di- versified, especially in view of the small size of the country. Venezuela's portfolio, in contrast, was more heavily concentrated in the steel and aluminum industries. RBI investments in Indonesia and Nigeria were more spread out over time. As noted previously, The Efficiency of Fiscal Linkage 109 many Indonesian RBI projects (encompassing a variety of sectors) were fortuitously delayed by the Pertamina crisis, which meant that they could be canceled or rescheduled when the second oil boom ended. Nigeria's main RBI effort, its steel complex, was not initi- ated until the end of the 1970s. Why was RBI stressed so heavily? In the eyes of its promoters it promised two important economic benefits: (a) public revenues from taxes on natural resource rents and returns on public equity, and (b) foreign exchange earnings and access to foreign suppliers' credits at attractive interest rates to finance investments. RBI was also believed to offer (c) economic diversification, (d) an opportun- ity to enhance the skills of nationals, and (e) production linkages to more labor-intensive downstream activities, which, it was held, would compensate for the capital-intensive nature of RBI. On the po- litical side it offered (f) high visibility, (g) a focus for patronage (em- ployment for supporters, sometimes kickbacks from contractors), (h) gains from regional development, (i) the ability to absorb wind- falls rapidly in projects with a high import content, (j) reduced depen- dence on multinational firms, and (k) a shift in status from raw- material exporter to industrial power. With regard to the last benefit, RBI satisfied the urge to create an exporting industrial sec- tor without the need to take the politically difficult steps required to reform existing import-substituting, protected industry. Although it has not been possible to make a detailed quantitative assessment of rates of return, the overall performance of the RBI proj- ects using this measure seems to have been very poor. True, some projects (such as Indonesia's fertilizer plants) have been successful. But some public manufacturing projects have been unable to cover even their wage bills, let alone the costs of other inputs and a mar- ket return to investment. Many plants have survived only through the imposition of taxes, in the form of import barriers, on domestic users of their output. Loss-making public investments have left gov- ernments facing a difficult choice-to subsidize directly, to protect and so shift costs to other sectors, or to close plants and write off the losses.8 Reasons for Poor Performance In considering this disappointing outcome, the first issue which arises is the process by which projects were evaluated and selected. Project Selection It is sometimes alleged that many projects were implemented with- out an assessment of their returns or in spite of assessments which in- 110 Comparative Analysis dicated low returns, but this is an oversimplified and misleading view of what actually occurred. It is true that, in some cases, govern- ments proceeded against the recommendations of feasibility stud- ies. But according to Auty (1986a), much of the problem stemmed from the assessments themselves. Three general shortcomings may be noted. 1. Overemphasis on natural resources. Proponents of RBI usually as- sume that access to cheap raw materials or energy is the major fac- tor in global competitiveness. The weight given to this factor in proj- ect assessments is far too great relative to the weights given to capital and recurrent costs, product prices, and likely capacity use. Energy costs, for example, constitute only about 22 percent of the total cost of producing steel in the United States and Japan. Power ac- counted for only 23 percent of total costs in a 1982 hydroelectric alu- minum smelter in the United States.9 Figure 7-1 summarizes sensitiv- ity analyses carried out for two plants, one producing olefins and the other low-density polyethylene. Internal-rate-of-return lines are shown for plant size, cost of raw materials, capacity use, product price, and construction cost. For plants of a given production capac- ity, these lines indicate that the midpoint elasticities of rates of re- turn with respect to price, capacity use, and construction cost are at least three times those with respect to the cost of raw materials. Cost overruns, increases in interest rates, or deterioration in prod- uct markets could therefore cut severely into rents and profits ex- pected from the plants. For reasons outlined below, adequate ac- count was not taken of the possibility of large variations in these dimensions in the feasibility studies. 2. Consensus and moral hazard in appraisal. Auty argues that consult- ing firms appraising projects are of two types. The first type has no apparent financial interest in the project (though other observers argue that there are few truly independent project appraisers). Such a firm has little incentive when assessing project performance to depart from the "range of consensus" that prevails regarding the outlook for the product, energy costs, interest rates, and other varia- bles. If events turn out to be very different and the consensus changes, the firm can justify its assessment with the argument that "everyone else was wrong too." Risk is therefore not properly taken into account in the appraisal, and sensitivity analyses typi- cally do not address the performance of the project under radically different market conditions. The implication is that appraisals do not evaluate what the accepta- ble rate of return needs to be-if the project and its markets evolve as projected-to compensate for the probability of a major down- The Efficiency of Fiscal Linkage 111 Figure 7-1. Sensitivity Analysis: Indonesian Olefins Complex and Low-Density Polyethylene Plant Olefins Complex 30- 20 20 = 20 <~~~~~~~~~~~ 1~~ Plant size 2 Cost of raw materials 3 Capacity use 4 Exfactory price 5 Construction cost ,103 iO 85 100 3 Capacity use (percent) (200) (300) (350) (450) (600) 1 Plant size (103 MTA) 90 100110 2,4,5 Cost/Price (percent) Low-Density Polyethylene (LDPE) 30 2 t 20 5 70 85 100 3 Capacity use (percent) 80 120 141 180 240 1 Plant size, LDPE (103 MTA) (200) (300) (350) (450) (600) Plant size, Ethylene 90 110 2,4,5, Cost/Price (percent) 100 MTA = metric tons per annum Source: UNICO International Corp., "Republic of Indonesia Survey Report on Petrochemical Industry Development," vols. 1-6 (Tokyo, 1974), as cited in Auty (1986a). 112 Comparative Analysis turn in markets (involving price cuts or protection abroad) or a large cost or time overrun. The second type of appraising firm has a substantial potential inter- est in the project, possibly as a manager or a supplier of equip- ment. It may be prepared to take a small equity stake to encourage the project to go forward. This it can easily recoup through fees and sales-especially in the event of cost overruns, when it de- clines to put in more equity so that its stake shrinks while its sales to the plant increase. This situation introduces a double element of moral hazard into both project selection and implementation. The appraiser-supplier has a strong stake in having the project go forward, whether or not it is economically justified, and little interest in containing costs once it has started. 3. Sizing up the competition. Project appraisals rarely considered the cost structures of current competitors. Some older plants, although perhaps less efficient, are fully depreciated and therefore poten- tially very competitive with new plants. But, because of overoptimis- tic scenarios concerning market growth, appraisals have assumed that product prices would be set by the costs of new capacity, that is, of future plants. Old, depreciated plants presumably would then earn abnormal profits. In the first spurt of industrial investment in 1974-76, it was argued that new plants would be far more expen- sive because of rapid inflation in plant costs and negative real inter- est rates. And in the second spurt of industrial investment in 1980-81, overoptimistic projections of energy prices swamped all other considerations. As a result of these three factors, sensitivity analyses typically in- cluded only minor variations in market and cost variables, so that the risk inherent in devoting substantial resources to large, inflexi- ble investment projects was downplayed. Once a decision had been made in principle to go ahead with a project, it was not always easy to find foreign partners prepared to share risk by taking a sub- stantial equity position or, equally important, by contracting to pur- chase part of the output. And in the cases of more doubtful feasibil- ity studies, supporters of the projects could always appeal to the less tangible of the expected benefits to support what was gener- ally, in the last analysis, a political judgment. Market Trends The most important factor in the performance of the large RBI invest- ments has been the unexpected severity of the downturn in world markets for many of their products after 1980. By the early 1980s, The Efficiency of Fiscal Linkage 113 world demand for copper, lead, zinc, aluminum, and nickel was be- tween 30 and 50 percent lower than what extrapolating the trend of the 1970s suggested. The relation between output and energy use in developed countries changed from an elasticity of unity to an elas- ticity of less than 0.7. In 1980 the Organization for Economic Co- operation and Development (OECD) forecast a doubling of world steel demand to 1,400 million tons by the year 2000; later forecasts projected only a 20 percent rise to 900 million tons. Price projec- tions were revised downward, sometimes by 30 to 40 percent in real terms. In table 7-5 projections made in 1980 for oil, urea, and alu- minum prices for 1985 and 1990 are compared with similar projec- tions made in 1984. In addition to their effects on price, adverse market develop- ments affected rates of capacity use, since the domestic markets of even the larger countries in the sample were insufficient to absorb the full output of world-scale plants in sectors such as aluminum and steel. In 1979-81, for example, Venezuela's steel industry oper- ated at less than 50 percent of capacity, while its aluminum indus- try operated at 40 percent of capacity. Even at the peak of the sec- ond oil boom domestic steel demand in Indonesia, Nigeria, and Venezuela averaged only 3.5 million tons a year, which is about the output of one world-scale integrated steel complex. Further, part of this demand was for steel that had to be imported, and part was for specialized steel products that had to be fabricated abroad. Such a plant would still have needed to export to achieve full-capacity oper- ation. The domestic markets for aluminum in Indonesia and Vene- zuela were similar in size and averaged only 78,000 metric tons a year. This was far less than one world-scale smelter's output of 300,000 metric tons. Domestic demand for metals then contracted sharply with the end of the second boom, falling further below the output of efficient plants. 1 But marketing was a problem for many countries that built RBI plants far in excess of domestic demand. With the slowdown in world demand, many countries protected their own producers. Lack- ing foreign partners and long-term marketing arrangements, plants in the oil-exporting countries were forced to sell on residual world markets, where price was set by the globally most efficient or most heavily subsidized competitors. In some cases they faced trade barri- ers when they attempted to export to major markets in the devel- oped countries. For example, ISCOTT, the steel company of Trinidad and Tobago, had production costs 50 percent higher than those of the efficient U.S. minimills, and antidumping charges were success- fully brought against the company when it tried to export to the vital U.S. market. The involvement of foreign partners sometimes Table 7-5. Commodity Price Trends and Projections (constant 1980 U.S. dollars) Commodity 1975 1976 1977 1978 1979 1980 1985 1990 Petroleum (dollars per barrel) 17.3 18.3 18.6 16.1 22.1 30.5 1980 projections 33.0 38.4 1984 projections 23.7 27.0 Urea (dollars per metric ton) 315 175 184 181 191 222 1980 projections 256 270 1984 projections 155 229 Aluminum (dollars per metric ton) 1,394 1,528 1,641 1,453 1,446 1,730 1980 projections 1,795 1,851 1984 projections 1,268 1,361 Source: World Bank data. The Efficiency of Fiscal Linkage 115 helped to secure markets; thus Indonesia's LNG plants, constructed with heavy foreign financial and technical involvement, benefited from marketing linkages abroad. In contrast, Algeria's larger LNG in- stallations, which are wholly state-owned, ran at less than half of ca- pacity for lack of markets, so that Algeria's LNG exports in the mid-1980s fell below those of Indonesia. Plant Costs Plant costs were typically higher in the oil exporters than in devel- oped countries. For chemical plants built on new sites (greenfields), it was estimated that construction costs in Mexico, Saudi Arabia, and the Gulf exceeded those in the United States by 25 percent, 50 percent, and 100 percent respectively."' The cost of steel capacity in some of the oil exporters has been estimated at $2,000 per metric ton, compared with $800 per metric ton for integrated mills in the Re- public of Korea. Venezuela's aluminum smelters experienced cost overruns of more than 70 percent, its refinery of 100 percent. Costs were boosted by the necessary infrastructural require- ments, by congestion in domestic construction industries, and by the overloading of suppliers in developed countries during the two intensive investment periods, 1974-75 and 1980-81. The sensitivity of plant costs to the timing of investments is indicated in figure 7-2. The first period represents the speculative real asset boom noted in chapter 6. The second represents the investment boom that resulted largely from widespread projections of secularly rising real energy prices, which in turn translated into large expected prof- its on RBI investments. Operating Costs In addition to high capital costs, high operating costs often contrib- uted to lower returns. Some industries were overmanned; Nigeria's steel plants were estimated to have three times the necessary num- ber of operatives before production even started. In some cases, oper- ating costs were raised by technical errors caused by limited links with foreign suppliers. Such errors were costly in the early phases of the Venezuelan aluminum industry. Some degree of technical error may be considered to be part of the learning process of estab- lishing new industries. But the public bodies in overall charge of the investment program often had little expertise in the particular in- dustry concerned. They were therefore slow to recognize and cor- rect emerging technical and management problems, which in- creased the costliness of such problems. 116 Comparative Analysis Figure 7-2. Plant Cost Index for a Japanese Hydrocarbon Process Plant Supplier 250 - 230 - Plant cost 210 -.- N 190 / 170 -' / _ 150 Wholesale price index 130 - 110 _ 1970 1972 1974 1976 1978 1980 1982 Source: H. Tanaka, "An Analysis of Elements Affecting Project Profitability Re- lated to Contract Marketing and Contracting," paper presented to Pertamina Sympo- sium, Jakarta, February-March 1985, as cited in Auty (1986a). The worst performers among the RBI projects have probably been the metals projects. Very few, if any, of these appear to have been profitable. The gas-based chemical and fertilizer industries seem to have done better, though their returns have been lower than origi- nally projected. Such plants could operate efficiently on a smaller scale and respond better to the needs of the domestic market. Oil re- fineries producing for export did less well, partly because transport costs for small batches of refined products are higher than for large volumes of crude oil. For the same reason, a refining industry serv- ing the domestic market was a more attractive investment. But a chronic surplus of capacity worldwide had squeezed margins by the 1980s, with refineries in many OECD countries operating at less than 60 percent of capacity. And some domestic markets were too small, given the need for a variety of products, to support world- scale refineries. The poor returns of many RBI plants, then, reflected a variety of fac- tors. None was peculiar to oil-exporting countries, but all were accen- tuated by the scope and pace of their investment growth and ren- The Efficiency of Fiscal Linkage 117 dered more important by the weight of RBI in their development strategies.'2 The exporters chose to convert enormous investable sur- pluses into fixed assets in a period when the range of uncertainty over key global economic variables and structural relationships was unusually large. Their strategies for developing the traded goods sec- tors emphasized large-scale projects in a limited set of industries. This emphasis reduced their economic flexibility at a time when just the opposite was needed. Risk was downplayed in project assess- ments. Projects were hastily and in some cases inadequately pre- pared. Foreign partners with large equity stakes were sometimes avoided for reasons of nationalism. In other cases they were sought but, being more sensitive to risk, could not be attracted on terms con- sidered acceptable by producer governments. Despite these difficulties, such large projects were probably the only domestic investments capable of satisfying the political urge to absorb the windfalls so rapidly; other potential options, such as land reform, labor-intensive industry, or smallholder agriculture faced administrative barriers and posed political and other problems which could not be addressed speedily, if at all. For all these reasons, the acceleration of public investments lowered the quality of public capital formation and increased risk. Many of the RBI investments made in the sample countries were internationally uncompetitive, and some were doomed from the start to make negative rates of economic return. At the same time, it is hard to avoid the conclusion that the poor performance of other key RBI investments has been mainly caused by the unexpected evolution of world markets and interest rates, which has also had a serious impact on the same types of industries in other developing and in developed countries. In this sense, factors external to any individual country, but closely related to the oil price shocks by their impact on the global economy, have reduced the returns from the oil windfalls. Some projects that now show a loss may become profitable as losses are written off with restructuring (most projects have a heavily front-loaded capital structure), espe- cially if world markets for their products improve. Even so, their unlucky timing makes it unlikely that the overall returns to invested capital will be satisfactory. The outlook is bleaker still if one accepts the proposition that a substantial part of the natural resource endow- ment of some countries is committed to specific product mixes, which curtails flexibility in the future use of resources. The Drift toward Subsidies An additional claim on public purses came from rapidly growing pro- grams of subsidies and transfers. Consumer subsidies were usually 118 Comparative Analysis directed to holding down the rate of inflation, producer subsidies to supporting loss-making firms. Between 1970-72 and 1974-78 fis- cal subsidies and transfers expanded, on average, twice as rapidly as nonmining GDP; between 1974-78 and 1980-81 they rose 1.6 times as rapidly. Recorded fiscal subsidies rose sharply in Trinidad and Tobago to 7 percent of total GDP or 11 percent of nonmining GDP by 1981. By 1983 it was estimated that the production costs of Caroni Sugar in Trinidad were five times those of efficient produc- ers, despite the fact that some of the latter had unit labor costs sev- eral times higher. The state enterprise deficit in Venezuela reached 8.3 percent of GDP in 1982. For several reasons, the fiscal accounts of the sample countries greatly understate the true magnitude of the subsidies. Producer gov- ernments were reluctant to raise domestic oil prices to world levels in 1974 and 1979 and chose to pass on this part of the windfall di- rectly to domestic consumers. In at least three countries domestic oil prices were set at the cost of production, so the governments de- rived no revenue from oil consumed at home. Some producers, nota- bly Ecuador and Indonesia, raised domestic prices of oil derivatives in the early 1980s, but on the whole their prices still remained well below world levels. As their domestic oil consumption (which some- times included illicit exports) grew more rapidly than their non-oil economies, the implicit fiscal burden on the state increased. In 1980 energy subsidies were estimated to be equivalent to almost 10 per- cent of household income in Ecuador, where electricity was also very heavily subsidized. Subsidies amounted to 4-6 percent of GDP in some of the other countries. Studies of their distributional im- pact have been limited, but there is no indication that they were tar- geted to the poorest groups. Fiscal data also do not include the subsidies implicit in loans made to loss-making (usually public) firms, nominally to finance in- vestment, and in the government guarantees which enabled them to obtain commercial loans. It is difficult to estimate these subsidies (since many such firms would probably have been unable to bor- row at any interest rate without a government guarantee), but they must have been considerable since they were accorded to some ex- tremely unprofitable firms. The RBI industries in particular were heav- ily subsidized. In the steel industry, for example, losses of $400 mil- lion by P. T. Krakatau in 1977-80 and $1.8 billion by SIDOR in Venezuela had to be covered by the state. Despite subsidies of $59 million in 1982-83, ISCOTT in Trinidad and Tobago recorded losses of $201 million. Some governments stimulated employment directly through pub- lic works programs. In certain cases these programs made a useful The Efficiency of Fiscal Linkage 119 contribution to capital formation; in others, they represented a dis- guised subsidy to labor. Consider the contrast between the INPRES and Bantuan Desa (village-level) rural construction programs in Indo- nesia and the Special Works (DEWD) infrastructure programs in Trini- dad and Tobago. Each is estimated to have provided at least part- time work to some 2.5 percent of the country's labor force. INPRES created slack-season jobs at wages generally not above what was available in agriculture. Projects had to demonstrate results (peri- odic photographs of work in progress were one requirement) and to be completed within one year. The Bantuan Desa projects financed only materials and were conditional on local self-help in the form of unpaid labor inputs. The effect of such programs in labor- surplus Java appears to have been positive. In contrast, the Trini- dad programs offered daily pay at least twice that available in agricul- ture. Since the average workday was light-about two hours- these programs really represented a method of transferring oil rents to the poorer sections of society. It is not clear that the pro- grams had a major impact on infrastructure. But they accelerated the shift of labor off the land, which exacerbated the drop in agricul- tural output. An indistinct yet interesting pattern emerges across countries con- cerning the use of oil rents to fund consumer subsidies and sup- port price controls on consumer goods. Such measures to distribute oil income to consumers seem to have been more pronounced in the higher-income countries in the sample than in the poorest ones. This may have been because political life was marked by more electoral competition in Trinidad and Tobago and Venezuela than in Indonesia and Nigeria. It may also have reflected greater ad- ministrative reach in the richer countries. Yet another important fac- tor may have been the more pressing need for basic infrastructural in- vestments in the poorer countries, which argued for having public rather than private agents spend the first round of oil income. The apparently low returns to much public investment raises a key distributional issue. Are subsidies and direct transfers of oil in- come from the state to private citizens irresponsible or are they sound policy? Some high-income oil-rich regions, such as Alaska and Alberta, have made direct distributions of oil wealth to their resi- dents, but this option appears not to have been considered in the capital-importing developing producers. Yet, from the perspective of second best, such an approach may be preferable to a program of public investments which yields little or no benefit on the sup- ply side, and is even more preferable to investments with a negative rate of economic return-such as steel plants requiring indefinite pro- tection (or taxes on users) to cover costs. Simulations with a gen- Table 7-6. Growth Trends in the Oil Exporters, 1967-81 (percent) Nonmining Domestic Goods and nonfactor service GDP investment Exports Imports Country 1967-72 1972-81 1967-72 1972-81 1967-72 1972-81 1967-72 1972-81 Algeria 9.5 8.6 16.7 10.8 5.7 -1.0 11.6 10.8 Ecuador 4.7 7.6 3.2 10.2 15.9 6.0 6.0 9.7 Indonesia 8.5 8.2 24.3 13.0 15.7 4.3 16.7 19.1 Iran, 10.1 13.3 10.2 21.1 12.9 -0.3 17.7 23.7 Nigeria 9.2 5.3 __b 8.7 __b -4.2 __b 15.3 Trinidad and Tobago 5.3 5.4 6.1 9.3 2.5 -6.5 6.6 8.4 Venezuela 6.5 5.1 11.9 3.5 -1.3 -8.7 7.7 12.8 Unweighted mean (excluding Iran) 7.3 6.7 12.4 9.3 7.7 -1.7 9.7 12.7 Middle-income oil importers, 5.8d 5.ld 8.2 5.6 6.7 4.0 7.4 1.5 a. 1967-72 and 1972-77. b. Deflated data unreliable before 1970. c. 1960-70 and 1970-82. d. GDP. Source: World Bank data. The Efficiency of Fiscal Linkage 121 eral equilibrium model suggest that, without a supply-side impact, only about one-third of investment expenditures will be reflected in private consumption through the induced relative price and demand- multiplier effects (see Gelb 1985b). Low-return public investments are therefore a very inefficient way to distribute oil wealth to the citi- zens of a producing country. Subsidies tend to become institutiona- lized, however (see, for example, chapter 14), and this can also be shown to have a detrimental effect on efficiency and growth should oil revenue fall. The Growth Impact: Models and Performance through 1981 Most theories of economic growth are based on stable relationships between factor inputs and output; they would therefore predict a sub- stantial acceleration of growth following an investment program on the scale of those implemented by the oil exporters after 1974. Con- sider the simplest aggregate neoclassical production function: y = A KaL(l - a) where y is output, K is the capital stock, L is labor, and a is the share of capital in output. The corresponding growth model is: = A + aK + (1 - a)L where y represents the GDP growth rate, A is the rate of (Hicks- neutral) technical change, K is the rate of capital growth, and L is the rate of growth of labor. Suppose that, for a typical country, A = 0.02, a = 0.30, K = 0.10, L = 0.03, the depreciation rate is 5 per- cent a year, and the capital-output ratio is 2.0. This is compatible with a gross savings rate of 30 percent, a growth rate of 7.1 per- cent, and a growth of income per capita of 4.1 percent a year."3 In such a stylized economy, with a constant quality of capital for- mation, an extra investment effort on the scale of the real invest- ment effect of tables 5-2 and 5-3 would add 1.4 percentage points to the growth rate, even if it is assumed that (a) no immigration oc- curred to increase the labor supply (in fact, immigration was consid- erable in Venezuela and Nigeria over the boom years) and (b) techni- cal change was totally disembodied. If growth was theorized to be constrained by the capital stock only, or by the availability of for- eign exchange or fiscal revenues, the expected impact of the wind- falls would be far larger. The projected addition to growth is there- fore quite conservative.14 Table 7-6 summarizes growth trends in the six oil exporters dur- ing 1967-72 and 1972-81 and compares these with the growth rates of middle-income oil-importing developing countries in the same pe- 122 Comparative Analysis riods. The cutoff date, 1981, is chosen so that the period of the emerg- ing oil glut does not affect results. Despite this, the impact of invest- ment on growth was at first sight very disappointing. Only Ecuador significantly accelerated the growth of its nonmining econ- omy in 1972-81 relative to its low level in 1967-72.15 On average, in 1974-78 non-oil economies were almost exactly the size they would have been had they maintained their 1967-72 growth trends, and in 1979-81 they were 4.1 percent smaller than along this trend. These comparisons yield the real growth terms in tables 5-2 and 5-3. Combining the real allocation and the growth terms in these ta- bles shows how the components of absorption have deviated from a scenario which combines (a) nonmining growth at the rates of 1967-72 and (b) the hypothesized structures of relative prices and composition of demand. On average, real private and public con- sumption and public investment were considerably larger than these scenario values during 1974-78. The averages in table 5-3 sug- gest a further rise in consumption (except in Nigeria) but a substan- tial shortfall in investment in 1979-81 relative to the first boom per- iod, on the basis of the new hypothetical scenario. On closer examination, the growth record through 1981, though still disappointing, is less adverse. Before 1972 Indonesia, Nigeria, and Algeria had all been in recovery phases, the first two from se- vere internal disturbances and the third from a protracted war of in- dependence. The economies of Ecuador and Trinidad and Tobago had been stimulated by spending on the oil industry. The sample's non-oil growth rate of 7.3 percent in 1967-72 was exceptionally high, 1.5 percent above the average growth rate of middle-income de- veloping countries in the 1960s. Average non-oil growth after 1972 was still 0.9 percent above this level. However, growth in the sample countries was demand-led rather than supply-driven. As is clear from the growth effects of ta- bles 5-2 and 5-3, growth responded to increased absorption after 1974 but slowed after 1978 despite the expectation that large invest- ments undertaken in 1975-78 would begin to contribute to output. An outstanding case is Venezuela, which simply stopped growing in 1979 despite the largest investment program in its history. Al- though the spending boom of 1974-77 raised the growth rate above its previous trend, by 1983 the nonmining Venezuelan economy was one-third smaller than it would have been had it continued to grow at the pre-1973 rate. The poor growth response is important in assessing the benefits of the windfalls. Growth was a major priority for all of the coun- tries, and as noted in chapter 5 it cannot be argued that the export- ers directly consumed an excessively large share of their windfalls rela- The Efficiency of Fiscal Linkage 123 tive to the predictions of a "permanent income" model. The widespread ineffectiveness of fiscal linkages, and especially of pub- lic investment programs, has thus been a major feature in the evolu- tion of the sample countries. Notes 1. Tugwell (1975), Bigler (1980), and Karl (1982) analyze the role of the Ven- ezuelan state and the "third sector" in the economy. 2. In this study, fertilizer subsidies are included in development spend- ing but excluded from measures of public investment. 3. Katouzian (1978) describes a similar agricultural investment policy in Iran and its similar results. 4. It is not possible to form a picture of the expenditure patterns of the states. The general lack of data raises some doubts, however, about the effi- ciency of state investments. 5. In 1987 the population of Abuja was estimated at 30,000 following a 2 billion naira investment. The original target was for a population of 3.1 mil- lion by the turn of the century. 6. The Caracas metro system was reputed to be the most costly, per meter, in the world. 7. Other countries considered by Auty are Saudi Arabia, Bahrain, Came- roon, and Malaysia. See Auty (1986a), (1987), and related works for a more detailed analysis. 8. Recent unpublished estimates by Auty suggest capital/output ratios of over 10 for most RBI projects. 9. Auty (1986a), tables 10 and 5. For a 1972 Japanese oil-powered plant, however, power represented 52 percent of total costs. 10. In the mid 1980s steel demand in Indonesia, Nigeria, and Venezuela averaged only about 1.5 million tons. Aluminum demand was also about half of its boom-year levels. 11. United Nations Industrial Development Organization (UNIDo) esti- mates as cited in Auty (1986a). 12. As shown by the $8 billion cost of the trans-Alaskan oil pipeline ver- sus its $900 million original budget, cost overruns can be large in devel- oped countries also. Their peculiar significance for the oil exporters is due to the weight of large projects relative to the size of their economies. 13. With gross savings of 30 percent and a capital/output ratio of 2, gross investment is 15 percent of the capital stock so that net investment is 10 percent of capital. The gross incremental output/capital ratio on such a growth path is 0.24. This is slightly below average for developing countries before 1971 (see table 3-1). 14. Embodied technical change would also cause the impact of new invest- ment to be greater. 15. Data are limited for Iran. Chapter 8 From Boom to Bust: The Downside of the Cycle and Asymmetric Macroeconomic Response As world oil markets turned downward after 1981, a combination of slumping prices and cuts in sales reduced the windfall gains of the oil exporters. With 1981 taken as a new base period, the wind- fall effect for the six countries in 1982-84 averaged -8.1 percent of nonmining GDP, as shown in table 8-1. Only Ecuador, which man- aged to raise oil output considerably and so compensate for falling prices, avoided a substantial reduction in windfall gains. Adjustment Patterns The exporters reacted to this adverse shift in a variety of ways but typically with a lag of at least a year. They had previously found that the momentum of public spending was hard to curb as reve- nues fell, particularly since an increasing proportion consisted of re- current expenditures to service the capital investments of earlier years and of subsidies to keep loss-making ventures in operation. In 1978, at the close of the first boom, government deficits had aver- aged 4.1 percent of nonmining GDP; current account deficits had aver- aged 11.8 percent of nonmining GDP (excluding Trinidad and To- bago where expenditure was more cautious in this period). Some countries had taken steps to slow the absorption of goods and ser- vices, and Indonesia had devalued by 30 percent in November 1978. These contractionary moves were interrupted by the second oil shock, which resulted in a current account surplus of $11.8 bil- lion for the six countries in 1980, a year widely regarded as signal- ing the start of an era of scarce and costly energy. But the second windfall fell off more abruptly than the first. In 1980-82 commodity exports, in current dollars, shrank by 21.6 per- cent as imports rose by 22.3 percent, so that current account bal- ances shifted to a deficit of $19.6 billion by 1982. More than half 124 From Boom to Bust 125 the current account deterioration was due to the unexpected fall in export revenues. In 1983 the deficit dropped to $7 billion as con- tractionary policies were implemented. These policies included sharp reductions in public spending and cuts in subsidies. Some ex- porters took steps to raise domestic energy prices closer to world lev- els. In addition, countries postponed or canceled projects in their pipeline; Indonesia, for example, delayed or cut over half its planned heavy industrial spending, for an estimated import saving of $11 billion. The adjustment patterns of the oil exporters in 1981-84 are pre- sented in table 8-1, which uses the methodology developed in chap- ter 4 but indicates only the value effects.I About two-thirds of the nec- essary decrease in absorption relative to nonmining output was borne by investment-more in Nigeria and Venezuela and less in In- donesia, where private consumption adjusted to offset the relative fall in oil revenue. Ecuador's continuing windfall gains went to in- crease the share of consumption, not investment. In assessing these responses one should bear in mind that, when the second oil boom ended in 1981, the sample countries found them- selves in quite different situations. Trinidad and Tobago and Vene- zuela had run up large trade and nonfactor surpluses (in Venezuela this surplus was partly offset by capital outflow), whereas Nigeria had run up a large trade and nonfactor service deficit. Nigeria's in- vestment was therefore cut sharply to achieve external balance in the face of a substantial worsening of its terms of trade rather than to attain a surplus position. Indonesia's conservative fiscal policy dur- ing the second oil boom had allowed it to build a reserve cushion to smooth adjustment. This made possible a negative trade and nonfactor service effect of -3.4 percent of nonmining GDP in 1982-84. Economic management through booms and slumps was rendered more difficult because access to foreign savings tended to be cycli- cal, increasing with the actual levels of world oil prices, since these were strongly related to future price expectations and perceived creditworthiness, but reducing when oil prices fall. Until world oil markets deteriorated, countries such as Mexico and Venezuela were able to cushion the impact of growing outflows of private capital by public borrowing abroad. When it became obvious that exchange rates could not be sustained, capital outflows accelerated, especially in countries such as Venezuela with open capital accounts and inter- est rate controls. In 1982 that country may have experienced an out- flow as large as 10 percent of GDP. The reasons for this outflow included interest rate ceilings in the face of rising foreign rates, a reluctance to adjust the exchange rate Table 8-1. Adjustment and Growth, 1981-84 Trinidad Measure Algeria Ecuador Indonesia Nigeria and Tobago Venezuela Mean Windfall effect, -10.4 2.8 -8.7 -10.2 -13.6 -8.3 -8.1 Trade and nonfactor service effect, -1.2 0.5 -3.4 10.7 n.a. 1.0 1.5b Consumption effecta -3.8 4.5 -9.2 -4.2 n.a. -1.1 -2.8b Private -4.6 5.9 -8.6 -2.4 n.a. 1.9 -1.6b Public 0.8 -1.4 -0.6 -1.8 n.a. -2.6 -1.1b Investment effecta -5.4 -2.2 3.9 -16.7 n.a. -8.2 -5.7b Non-oil growth rate for 1981-84 7.0 -1.0 4.8 -5.6 -2.9 -2.0 0.0 Ratio of trade and nonfactor service surplus to nonmining GDP in 1981 1.4 1.1 2.8 -10.6 23.4 7.8 4.3 n.a. Not available. a. Effect, measured as percentage of nonmining GDP, is the average for 1982-84, with 1981 as the base year. b. Mean for five countries. From Boom to Bust 127 to widely perceived trends in the oil market, and contradictory poli- cies initiated by the Venezuelan administration and its opposing legis- lature, which squeezed domestic profit rates while opening industry to greater foreign competition (see chapter 14). The capital outflow resulted in a prolonged decline in private investment, which was the main contributor to the large negative investment effect recorded in table 8-1. The investment slowdown caused negative growth in the period after 1978, which in turn contributed to the slowdown through the "negative accelerator" effect of depressed yields on the domestic capital stock. Increases in Venezuela's labor force and capi- tal stock should have ensured growth of at least 4 percent a year even with zero productivity gains. In fact, the average country experienced virtually no non-oil growth at all in 1981-84, as shown in table 8-1. Only Algeria and In- donesia managed to maintain positive growth rates in this period. Algeria was aided by growing exports of natural gas, LNG, and conden- sate (a liquid by-product of gas extraction) and by the implementa- tion of an effective expenditure-switching policy, which shifted the composition of public investments from import-intensive manufac- turing to social capital, notably housing. Indonesia, once again, was quick to devalue as oil income fell and was able to prevent the real effective exchange rate from appreciating for extended periods. That country's recovery was helped by a strong performance in agri- culture and in miscellaneous (mostly manufactured) exports, which compensated for the depressed markets facing non-oil primary com- modities. The average growth and inflation rates of the oil exporters from 1976 to 1984 are summarized in figure 8-1. The period may be di- vided into five phases: * 1970-72 marked the end of recovery from the domestic turbu- lence of the 1960s in Algeria, Nigeria, and Indonesia and of in- tensive oil development phases in Ecuador and Trinidad and To- bago. The average nonmining GDP growth rate fell from more than 10 percent in 1970 to 5.3 percent in 1971 and 1972. * 1972-77 was the upswing of the first oil boom. Domestic prices rose sharply as real exchange rates appreciated and import costs rose. As resources were reallocated to the construction and service sectors, growth peaked in 1976 and 1977. In this pe- riod, construction sectors grew an average of 13.9 percent a year. * 1977-79 was the downside of the first boom. As income and spending slackened, growth rates of real output fell sharply but inflation actually accelerated slightly. This period provides 128 Comparative Analysis Figure 8-1. Boom and Stagflation during the Oil Cycle, 1970-84 (unweighted averages, six countries) 20 18 1974 1975 1984 16 1983 < 1979 1980 14 1981 1976 12 -1978 1977 10 1982 1972 8 1970 6 1971 4 2 * I I ~I I I I1I ' 1 I 1 1 -2 0 2 4 6 8 10 Growth of nonmining output (percent) Source: World Bank data. From Boom to Bust 129 the first experience of asymmetric macroeconomic adjustment; by the start of the second boom in 1979 inflation was higher and growth lower than in 1972 and 1973. * 1979-81 was the second boom. It was smaller and more abrupt than the first one. Average inflation and growth rates varied only slightly in these years. * 1981-84 marked the downturn of the second boom. Both infla- tion and non-oil output growth fell sharply in 1982, but the large current account deficits were not sustainable. Govern- ments therefore turned to domestic markets for financing, re- duced subsidies, and scaled back other measures which had held down domestic prices. Rather like the oil importers in the upswing of the cycle, the exporters moved into stagflation on the downside, with negative growth in 1983 and 1984.2 * Ecuador, Nigeria, and Venezuela experienced foreign exchange crises. Indonesia and Algeria, as noted above, adjusted more smoothly, although their policies were very different. Trinidad and Tobago's ample reserves enabled it to postpone adjust- ment until 1985, although poor agricultural performance and a decline in the oil-refining industry resulted in negative growth in 1981-84. Inclusion of the fifth period reduces average non-oil growth rates to only 4.7 percent during 1974-84. This is 1.1 percent below the aver- age growth rate for developing countries in the 1960s. Despite the emerging oil glut, adverse terms of trade for oil were not responsi- ble; even in 1984 the purchasing power of oil exports was far larger than it had been before 1974. The massive investment programs of the previous decade were failing to stimulate output. Modeling Growth and Adjustment during the Cycle To obtain a better quantitative picture of gains and losses during oil cycles and of the effects of policies and macroeconomic responses, Gelb (1985a, 1985b) investigates the impact of suboptimal and opti- mal policies using a general equilibrium model. The rest of this chap- ter outlines the model, its results, and the implications of those results for the sample countries. The exercises use the model to com- pare the evolution of an economy under alternative conditions: an oil price cycle and constant real world oil prices. This enables the im- pact of the cycle to be assessed under various policy and macroeco- nomic assumptions. The model itself is a dualistic six-sector computable general equilib- rium (CGE) representation of an economy similar to that of Indone- 130 Comparative Analysis sia, with oil, manufacturing, construction, services, export, and food agriculture.3 "Modern" and "traditional" sectors are distin- guished (in the latter, labor may be paid its average, rather than mar- ginal, product), and manufacturing and service sectors are split into modern and traditional components. There are two economic agents, government and the private sector. The latter is treated from the demand side as a single household whose reaction func- tions are taken into account in setting policy. The model is simu- lated or optimized over a run of ten periods, each representing two years. The outcomes are then assessed according to a welfare func- tion in two parts: the discounted sum of the logarithms of consump- tion per capita during the twenty-year period times population, and a terminal condition on the output of year 20 to represent the dis- counted value of the posthorizon steady-state consumption stream. The base-year social accounting matrix of the model is for 1975 and is based on Downey and Keuning (1983). Oil output, world prices, and interest rates are exogenous sequences of variables. The model includes a richer set of interperiod links than most CGES. In addition to the conventional updating of capital stocks and labor force through investment, population growth, and migra- tion, government consumption levels are set by the need to service public capital. This introduces a ratchet effect of investment booms on public spending. Public infrastructural investment raises produc- tivity and output, rather than adding to capital stocks in produc- tion functions as does private investment. Rapid acceleration or de- celeration in the rate of investment lowers the quality of capital formation, so that unstable investment paths are less productive than stable ones. In the basic version of the model, prices and wages adjust smoothly to equilibrate demands and supplies, and private domes- tic investment is determined by private savings. But the model also includes three optional variants, which allow simulations to incorpo- rate other important interactions during the cycle. One option is to in- troduce downward wage rigidity in the modern sector. The result is unemployment rather than real wage cuts to clear the modern labor market when labor demand falls. A second option is that of "sticky prices"; faced with falling demand, firms move off their long- run supply curves and cut output to slow the rate of price decline. This results in a phase of operating at less than full capacity in re- sponse to sharply falling demand. Third, private agents may be given the option of exporting their savings in addition to investing domestically. The capital-flight decision is modeled as sensitive to sharp depreciations in the real effective exchange rate, which in the real world would require large nominal devaluations.4 From Boom to Bust 131 Certain features of this model, such as the treatment of labor mar- kets, are rather specific to Indonesia. But for the most part its dy- namic behavior reflects more common characteristics. The exercises described below should be interpreted as delineating oil exporters in general rather than any single country. A first base run for steady world oil prices is constructed, and then a second base run for a windfall scenario such as that of 1978-84. In these base model runs, government is constrained; it nei- ther lends abroad nor borrows in the windfall period-a policy repre- sentative of average behavior for the sample of exporters. Windfall gains feed public investment, and the real exchange rate appreci- ates with the spending effect. Private demand rises to absorb about 30 percent of the windfall during the boom because of the demand- led effects of public spending on relative prices and activity levels. The base value of the windfall-denoted here by V1-is given by the difference between the valuation functions of the two simula- tion runs. V1 is the windfall's value under empirically representa- tive, not optimal, policies and with flexible wages and prices in the economy. The next step is to optimize the time profiles of public absorp- tion, both with and without the windfall, to determine the optimal path of saving and borrowing abroad. With perfect foresight, about 70 percent of the windfall is optimally saved abroad during the boom, to be used up gradually as oil prices fall. This raises the value of the windfall by over 60 percent from its base level Vl. The op- timized value of the windfall is denoted by V2. The gains from cautious spending are far larger, however, if the producing economy cannot adjust smoothly in the downturn through flexible prices and wages. Under the fiscal assumptions of the base run, sticky prices and wages reduce the value of the wind- fall by between 20 and 40 percent of V1 (to a value denoted by V3) be- cause of unemployment of labor and capital on the downside of the cycle. In optimized runs with perfect foresight, the model economy never runs into such a recession because rigidities in factor and prod- uct markets are very costly in terms of lost output, investment, and growth. The value of the windfall in the optimized run, V2, is there- fore more than twice that in the base run if the economy cannot ad- just smoothly; that is, V2 is more then twice V3.5 The simulations also show that capital flight is extremely costly. Be- cause it accelerates on the downside of the cycle, it exacerbates the problems caused by falling levels of public spending and deepens the recession by cutting investment further. The economy may now be considered to have two private agents. "Speculators" own fi- nancial assets, the more internationally mobile factor of production. 132 Comparative Analysis By moving funds abroad as the oil market weakens, they throw the burden of adjustment onto "workers" who respond by defending real wages at the high cost of unemployment. Optimized runs with perfect foresight always avoid situations likely to trigger such capi- tal flight. Finally, the model is used to analyze the impact of imperfect fore- sight. How important has poor prediction been in determining the macroeconomic outcome of the oil shocks? Incorporating errors in predicting oil prices requires a two-stage process. First, the model is optimized subject to a path of oil prices which is predicted to rise at 6 percent in real terms after the peak of the boom. Second, the solu- tion to that problem is used to create a new starting point for the model at the peak of the boom, for a reoptimization subject to the "correct" price expectations for the rest of the period. This process simulates the sharp downward revisions in oil market projections which took place at the peak of the second oil boom in 1981. The results of this exercise suggest that the costs of such predic- tion errors are likely to be enormous. The model economy borrows against expected future oil income to increase absorption by a fur- ther 60 percent of the windfall during the boom.6 The peak of the boom finds the economy with a strongly appreciated real exchange rate, unsustainable wage levels, and large debts (which must be re- paid by the end of the twenty-year horizon). Under a wide range of parametric assumptions, the windfall value is negative because of the severity of subsequent recession and stagnation. In absolute mag- nitude the "cost" of the windfall, V4, is often more than three times its "benefit" V, in the base run or more than four times its "ben- efit" V3 with the base fiscal assumptions and sticky downward adjust- ment of wages and prices. One major implication of these simulations is that the oil export- ers should have saved a far higher proportion of the windfall abroad than they actually did. Another is that the increase in global volatility and uncertainty after 1973 was a very important aspect of the oil shocks. Given the magnitude of the resulting prediction er- rors, the shocks could certainly have left producing economies worse off than before the windfalls. This is because downside macro- economic costs are potentially far larger than upside absorption benefits. Notes 1. Deflators are not sufficiently reliable for enough countries to warrant presenting real and price effects. The value effects compare value shares and projected value shares and therefore combine real and price effects. From Boom to Bust 133 2. If the exporters had held domestic oil prices at world levels, they would have experienced the supply shock on the upside of the cycle as the oil importers did. But because domestic oil prices were raised on the downside to increase fiscal revenues, the pattern of spending and supply shocks was the mirror image of that in the oil importers. 3. CGE models are described extensively in Dervis, de Melo, and Robin- son (1982). They capture the economy on the production side and on the de- mand side through production and demand functions and represent the whole as a set of interrelated markets which (usually) clear by adjustments in relative factor and product prices. 4. This model, like other CGEs, does not have a nominal price level; the price of imported manufactured goods is taken as the numeraire. 5. As noted in chapter 2, adjustment costs on the upswing and the downside of the cycle are largely external to individuals. The optimized run therefore cannot be identified with the outcome of a policy of distribut- ing the windfall to private agents. This is one of the main reasons why the private sector is not modeled as an optimizing agent. 6. Mexican projections were for real oil prices to rise by 6 percent a year, and the increased borrowing of the model corresponds exactly to the ex- tent of increased Mexican borrowing during 1979-81; see table 5-3. This is a pure, yet interesting, coincidence. Chapter 9 Summary of Findings Although commodity price fluctuations have long been common- place, in a number of respects the oil shocks were unprecedented. Oil is a key input into production and distribution. Even before 1973 it was the commodity with the largest value in international trade. The periodicity of the oil cycle-or cycles, if considered as two events-was far longer than that characteristic of fluctuations in most commodity prices, and the amplitudes of the oil cycles were larger. Also, no other commodity has embodied so great a share of national rent in its price. The shocks therefore had a large impact on the global economy. They affected trade flows and capital balances between importing and exporting countries and changed the incentives for activities that consume and produce energy. In energy-importing countries, the shocks had profound regional effects. For example, the crisis of financial institutions in Texas in the mid-1980s highlights the disequilibriums and investment losses induced by oil price swings. In the producing countries, cumulated rent flows concentrated the surplus in the hands of relatively few economic decisionmakers. Oil price trends and the policy responses of major industrial countries to the shocks were poorly predicted. Consequently, so were real interest rates, inflation, exchange rates, and the evolution of world markets for other important commodities after 1973. One result of the shocks has therefore been greater global volatility and uncer- tainty. This book considers the impact of these events on a set of appar- ent winners in the global equation-the capital-deficit, developing oil exporters. The six countries studied differ along many political and economic dimensions, and this raises difficulties for compara- tive study because of the small sample. But the countries share 134 Summary of Findings 135 some common characteristics as well. The sizes of their oil sectors rel- ative to their non-oil economies were similar during the windfall dec- ade; so were the proportions of their total wealth made up of proven oil reserves. All desired to use their oil revenue to spur growth, modernization, and economic diversification and to in- crease the degree of national control over key economic sectors. None overtly sought to consume most of the windfall gains or to move toward a rentier state through prolonged savings abroad. These priorities account for the perspective of this study. The ef- fects of the oil booms on income distribution, for example, have not been given greater emphasis here not only because of severe data problems but also because this was not the main thrust of pol- icy in these countries. It is important to recognize the extent to which continuity shaped the use of the oil windfalls. Social and political institutions changed more slowly than the availability of resources. Indeed, in some cases the abundance of oil revenues may have reduced pres- sure for institutional change by funding the inefficiency of existing in- stitutions. Although in principle the windfalls represented uncom- mitted income which could be allocated to those activities deemed to have the greatest social value, in practice they tended to fund fairly conventional, though disparate, activities reflecting priorities and administrative capabilities that had evolved before 1973. At the broadest level, continuity of policy dictated that the rents be translated into domestic investments, the traditional use of surplus funds. The proportion of the windfalls invested domestically was indeed large. On average, it was roughly what an economist would have recommended in 1975 on the basis of a "permanent income" model, allowing for depletion of reserves and projecting constant real prices, but abstracting from two important features: constraints on absorption capacity and uncertainty. It was above the levels which might have followed from assuming that real resource rents would increase at or above a reasonable rate of discount-an assumption which would have argued for still higher consumption. Increased investment was overwhelmingly public; for a variety of reasons pri- vate investment was unresponsive in most of the countries. The six countries in this study all took significant steps to bring their oil and hydrocarbon sectors under national control. Some also took wide-ranging steps toward indigenization in other sectors. The oil windfalls were not responsible for such policies, which pre- dated the shocks-often by many years. But the windfalls did acceler- ate the process by providing producer governments with the finan- cial resources that helped them to buy out foreign interests and to compensate for any adverse effects on output. Progress toward the 136 Comparative Analysis objective of increasing national control over key economic sectors was therefore quite strong. Progress toward achieving economic diversification was far less im- pressive, however. Only Indonesia and possibly Ecuador managed to strengthen and diversify the nonhydrocarbon traded sectors dur- ing the windfall decade. The other producers either began and ended the period with uncompetitive manufacturing and agricul- tural sectors, which contributed, in total, a smaller than "normal" share to nonmining GDP (Venezuela and, especially in agriculture, Algeria), or they experienced a dramatic relative or absolute decline in these activities (Nigeria, Trinidad and Tobago) and moved to still greater dependence on oil. As for growth, the only favorable impact of the windfalls on aver- age performance has been the powerful demand-led expansion of 1974-77. From 1974 to 1981 average growth rates were well below what would have been predicted by a simple neoclassical model, given the size of the investment boom in relation to either the past ex- perience of the individual countries or the performance of develop- ing countries in the 1960s. Growth rates were even further below what would have been predicted by theories of capital- or foreign- exchange-constrained growth. Extending the period through 1984 lowers the average growth rate of the non-oil economies of the sam- ple countries to only 4.7 percent, considerably below the growth rates registered by developing countries in the 1960s. Terms of trade were still far more favorable for oil exporters in 1984 than in 1972, so the oil glut cannot be blamed. The theories reviewed in chapter 2 may be grouped into two sets. Booming-sector and neoclassical growth theories emphasize the allo- cative consequences of windfall gains. They highlight the spending ef- fects on the relative prices of traded and nontraded goods and conse- quently on factor allocation, and they also highlight changes in growth rates caused by an increase in the quantity of capital funded by oil income. Both theories characterize the economy as being on the frontier of a production set. Although the declining effi- ciency of accelerating investment (and similar effects) can be incorpo- rated, this involves adding exogenous assumptions rather than at- tempting to explain such a fall in efficiency. The linkage and macro-instability theories, in contrast, empha- size more the quality, nature, and degree of use of windfall gains and of domestic factors of production during the oil cycle. By analogy to the theory of the firm, they could be considered as X-efficiency ap- proaches since they do not suppose that economies are on the bound- ary of a conventional production set. Both types of theory are important in accounting for the lack of eco- Summary of Findings 137 nomic diversification and the poor growth in the sample countries. As spending accelerated after the first oil price rise, real exchange rates appreciated. So did real effective exchange rates, except ini- tially in Venezuela and Algeria, which liberalized trade or applied price controls. Domestic factors of production shifted toward the nontraded sectors-particularly construction, which boomed with the rise in public investment, especially in 1974-77. As real effec- tive exchange rates appreciated, returns on foreign financial invest- ments fell further below increases in the prices of domestic invest- ment goods, an effect which was exacerbated by the impact of the investment boom on foreign suppliers of capital goods. This fed the speculative boom in real capital formation in the oil exporters; the more rapidly they spent, the higher domestic costs rose and the greater seemed the need for investment to make projects inflation-proof and render plants more competitive than later mar- ket entrants. This cycle was briefly punctured by the fall in reve- nues in 1978 and by the adoption of tight monetary policies in oil- consuming countries, which caused the dollar to appreciate after 1979 and raised interest rates relative to the price of capital goods. But predictions of rising real energy prices in 1979 and 1980 caused a resurgence of the investment boom since they promised ever grow- ing revenues and increased the anticipated returns to energy-based industrial projects. Because of the desire to spend rapidly, public investment favored large projects that minimized decision time and did not require labori- ous and controversial institutional and political changes. In the repre- sentative oil-producing developing country, about two-thirds of investment was directed toward either infrastructure or human capi- tal formation. This part of investment therefore went to strengthen the nontraded sectors. The impact on transport, communications, and housing was in most cases considerable, and education and health indicators rose markedly-although they also improved con- siderably in non-oil developing countries in this period. Most of the remaining third of public investment went to the hydrocarbon sectors and to other large RBI projects, the major initiatives in the traded sectors. In many cases these projects were inadequately planned and en- countered substantial cost and time overruns; sometimes decisions did not properly allow for comparative advantage. But more gener- ally, feasibility studies did not account properly for uncertainty. Risk was high because of the very specific nature of investments; the scale of plants, which was far larger than warranted by the size of guaranteed domestic markets; and the correlation of returns to a wide set of plants similarly affected by the level of global activity. 138 Comparative Analysis Therefore, when oil, petrochemical, and metals markets slumped after 1981 and interest rates simultaneously rose, the oil exporters were vulnerable. Their non-oil traded sectors were weakened, they had become more dependent on imported intermediates to sustain industrial production, and in most cases they had sizable gross inter- national debts. Falling oil revenues and poor returns from resource- based industrial projects reduced income flows. Inflation surged as governments were forced to raise administered prices, notably of oil derivatives, and to devalue in an attempt to maintain fiscal balance by boosting the domestic buying power of oil income. Rather like the oil-consuming countries in the upswing of the oil cycles, the producers entered a phase of stagflation on the downside. Slow, and in four out of six cases negative, growth reduced the benefits of much infrastructural spending. As revenues fell, partially com- pleted projects were postponed or canceled, and the result was that returns on the investment program were further reduced. To date, the overall payoff to most public investment programs appears to have been very low when judged by incremental output and financial results. In most countries, labor seems to have shared in the windfalls to a considerable extent. The spending effect tightened labor markets and raised employment. The combination of price controls and con- struction's weight in the nontraded sectors caused consumer prices to fall relative to overall domestic prices and production costs. This al- lowed a considerable rise in real wages, and consumption increased as a share of non-oil GDP by about one-fourth of the windfall. Ni- geria may have been an exception because a chronic shortage of wage goods followed the movement of labor out of agriculture. The poor growth performance of the sample suggests, however, that la- bor's gains were only temporary. Indeed, by the early 1980s there was evidence in several countries of a fall in real wages at the end of the boom. Given the diversity of the countries, there was naturally consider- able divergence from this representative picture. Some of the country-specific effects have been noted in earlier chapters, and they are discussed at greater length in the chapters that follow, so only a concise overview is offered here. It is not easy to discern a relationship between the type of politi- cal system and the efficiency with which oil income was used. Never- theless, two broad conclusions about the interaction between politi- cal structure and oil seem to emerge. First, the higher-income countries, whose political systems were more subject to electoral com- petition throughout the period, had a greater tendency to use oil in- come for subsidies than did the poorest countries. This may also Summary of Findings 139 have been due to the more obvious infrastructural needs of the latter, which strengthened the case for initial spending of the windfalls by the public rather than the private sector. Second, the time frame of governments and the degree to which political power is centralized seem to be important determinants of the outcome of a windfall cycle. In all three countries with either fre- quent and discontinuous changes in government or (as in the case of Venezuela) divergent executive and legislative policies, there were foreign exchange crises when oil revenues declined. Indonesia and Algeria, which had the most centralized and continuous policy- making structure in the sample, managed to turn in the highest growth rates, especially later in the period. They responded cau- tiously to the second oil boom and adjusted relatively smoothly (though not painlessly) when it ended. Trinidad and Tobago also had a centralized, continuous government during the period. But its experience shows how hard it is to handle oil rents cautiously for a sustained period in the face of powerful interest groups. Organ- ized along sectoral lines, these groups generated irresistible pressure for raising subsidies. A fragmented or decentralized decisionmaking structure reduces the ability to spend windfall gains cautiously. First, as noted in ear- lier chapters and again below, some of the costs of a volatile spend- ing sequence, both on the upswing and on the downside, are exter- nal to individual agents or small groups. If each group assumes that its actions will not influence those of other groups, spending pressure will not be adequately contained. Second, the windfall it- self reduces incentives to maintain a system of checks and balances among agents. The costs to an individual agent of such a system are normally balanced by the benefits, in the form of resources made available by constraining the spending of others. Windfall gains re- duce these benefits, making the task of monitoring others less attract- ive and shifting incentives toward obtaining as large a share of the gains as possible for the use of one's own constituency. Highly cen- tralized decisionmaking can, however, also reduce the efficiency of investment. What should the oil exporters have done differently? This is not an easy question to answer, partly because a final judgment about the consequences of their actual choices will not be possible for many years. From the perspective of the individual countries, how- ever, the most important recommendation to emerge from this study is that spending levels should have been adjusted to sharp rises in oil income far more cautiously than they actually were. The general equilibrium mod- eling exercise described in chapter 8 suggests that about two-thirds of a windfall comparable in size and duration to that received by 140 Comparative Analysis the oil exporters should be saved abroad. This would maintain the quality of capital formation and avoid a severe recession as well as losses in output and growth because of below-capacity production at the end of the boom. But the optimal degree of saving-or dissaving-abroad appears to be very sensitive to projections of oil prices after the peak of the boom. If oil prices are firmly expected to keep rising in real terms-contrary to the historical record that commodity price trends reverse themselves-it may indeed make more sense to borrow against future revenues and run larger current account deficits than to save abroad. Even if public investment is increased to the point at which marginal expenditures have zero or negative supply-side impact, the (demand-side) spending effect raises private income, im- ports, and consumption and so advances the enjoyment of future oil income, even if very inefficiently. This may create a dilemma for oil exporters. Producers have usu- ally justified higher oil prices by citing the depletion of resources even when the immediate reasons for large price jumps lay else- where. It may be difficult to persuade a domestic constituency to ac- cept a spending policy based on the likelihood of lower prices and demand in the future, while at the same time arguing for high cur- rent prices in international fora by referring to the inevitability of still higher prices in the future. The solution to this problem is to rec- ognize that the costs of overoptimistic projections are far greater than the costs of overcautious ones, so that a spending policy based on less than the ex- pected price trend is to be preferred.' At the global level the recommendation to spend more slowly raises, of course, the recycling problem: how to channel large sur- pluses to oil-importing countries. The sample countries were not net borrowers during the booms, but they did recycle their own surpluses and, through their residents' considerable degree of asset diversifica- tion, they enabled portfolios in other countries to be diversified. (For example, some Venezuelans may have financed Brazil while Saudi Arabia may have lent to Venezuela through financial intermediaries.) Slower spending would have made the recycling problem more prounounced, although it might have eased debt problems of the mid-1980s by reducing the indebtedness of oil exporters. Had the urge to spend rapidly not been so great, public projects might have been subject to more rigorous appraisals that took risk more fully into account. The experience of the sample countries suggests three important roles for nonresident equity participation in resource- based, export-oriented projects: (a) to check on the realism of feasibil- ity studies by scrutinizing the assessment of rate of return and risk; (b) to facilitate construction, start up, and operation; and (c) to re- Summary of Findings 141 duce risk by securing markets and lowering the government's stake. To obtain such benefits, foreign equity participation should be considerable-at least, say, about 30 percent. The inefficiency of low-yielding public investments as a conduit for oil revenue raises another question: should the exporters have made a greater attempt to disburse revenues directly to their popula- tions? Such payments did take place in Alaska and Alberta. Had the countries in the sample emphasized the direct distribution of oil income (as Ecuador did in part through credit policy and lower non-oil taxes), spending booms might have taken place out of pri- vate income, especially with imperfect capital markets. The composi- tion of demand would have differed from actual patterns. But some of the costs of a spending boom-notably congestion effects when de- mand rises and surplus capacity when it falls-are external to individ- ual agents and are not adequately taken into account when individu- als make spending and saving decisions. It is not certain, therefore, that direct distribution would have resulted in a much better macro- economic outcome. The scale of loss-making investments would surely have been smaller; spending cuts on the downside might have been made more promptly (or might not, if transfer programs had become institutionalized); and transient consumption gains would have been larger. The use of oil income to fund improvements in public administra- tion does seem to be a good option. For example, it can facilitate tax reform to broaden the basis of non-oil tax collection by permit- ting a period of low tax rates. Such options were not taken up by the sample countries except, to an extent, by Indonesia. Because the costs of a volatile spending path are largely external to individual agents, there is an argument for centralizing control of the oil windfall in order to moderate spending, and also possibly for limiting foreign borrowing by public enterprises and private agents who desire to anticipate future windfall income. The main prob- lem is then to render longer-run saving abroad politically acceptable. This may not be easy; indeed, it may be impossible for a country whose government faces (or consists of) powerful groups competing for a share of the rent. The experience of Indonesia and Algeria in the sec- ond boom period and of Trinidad and Tobago in the first suggests that sterilization might be possible for some countries for a number of years. But for countries such as Ecuador and Nigeria with short- lived governments and regionally disparate power groups, restraint will always be very difficult. Whether the "Dutch disease," as manifested by poorly perform- ing non-oil traded sectors, is inevitable in the longer run will be deter- mined by (a) static effects-that is, the impact of the spending 142 Comparative Analysis boom on the non-oil traded sectors through its effect on factor and product markets; and (b) dynamic effects-that is, the possible uses of oil revenue to raise productivity growth in the non-oil traded sec- tors. A particularly interesting contrast is the strong performance of the agricultural sector in Indonesia, which was far above world aver- ages, and the poor performance in Nigeria and Trinidad and To- bago, which was far below world averages. Indonesia had a far lower ratio of oil income per capita than the other two countries, and a large reservoir of labor in its traditional agriculture and serv- ice sectors. The labor pull out of agriculture caused by oil income was therefore less, especially as a large share of public spending went to rural areas where it could fund the slack-season use of labor. In addition, a combination of new crop varieties, investments in irrigation and infrastructure, and the production, subsidization, and distribution of fertilizer seems to have been effective in raising productivity, especially in rice. In contrast, productivity fell in Trini- dadian agriculture, especially in the nationalized sugar industry, which was increasingly subsidized. In Nigeria, smallholder produc- tion technology stood still and large-scale public projects had disap- pointing results. In both countries, public programs drew significant amounts of labor from the farm sector. In addition to the static and dynamic differences, the greater will- ingness of Indonesia to devalue the nominal exchange rate may have affected agricultural performance and that of non-oil exports in gen- eral. Devaluation enabled the real exchange rate to depreciate rapidly as the spending effect fell off, and so non-oil exports were protected. The contrast with Nigerian exchange rate policy is notable. Although the results of the decomposition in chapter 4 do not prove the case, they do support the thesis that tightening foreign exchange controls in Nigeria led to a progressive fall in the efficiency of the non-oil economy, which largely offset the benefits of the oil windfalls. Initial conditions and policies are important determinants of the impact of windfall gains on the structure of the non-oil economy and can modify the predictions of simple models. In some respects the empirical results suggest the need to reas- sess simple stereotypes. For example, the poor agricultural perform- ance in most of the sample countries resulted from agriculture's sensi- tivity to labor market pulls rather than simply from its status as a traded sector. Instead of becoming relatively cheaper, therefore, food tended to become more costly during the boom period in most of the countries despite rising imports. This, in turn, may have helped to distribute at least a part of the windfall gains to rural sectors. Summary of Findings 143 Because of the association of the windfalls with greater global uncer- tainty (which also affected importing countries), it is indeed possible to make the case that oil exporters ended the period worse off than they would have been with a far lower, more predictable rate of increase in oil prices or, indeed, with constant real oil prices. The simulation results of chap- ter 8 suggest that such an outcome is not unlikely. Volatility plus poor prediction translates, on average, into a poorer use of re- sources during the cycle, which may more than offset the incre- ment to resources from the windfall. This conclusion may not apply to all the countries because of real consumption gains during the windfalls in most countries, because of the possible long-run im- pact on growth of infrastructural and educational investments in some countries, and because some residents of some countries have built up sizable foreign assets. It is possible to conclude, how- ever, that a great proportion of the potential gains to the exporters in this study was nullified by a combination of the changes induced in the global economy by the oil shocks and the poor economic poli- cies of the exporters themselves during the period. Together, these seriously reduced the efficiency with which the countries used their resources, as judged by their own criteria of diversification and growth. To some extent, policy errors reflect a generally incautious approach to greater global uncertainty, but the abundance of oil wealth seems to have encouraged a deterioration over and above this. How much of the deterioration in the investment-growth relation- ship (shown for all groups of developing countries in chapter 3) can be attributed to the direct and indirect effects of the oil shocks is not clear. If none, then the only cost to the oil importers would have been the real resources transferred to the exporters. If it is sup- posed that the extra capital formation in the exporters was, say, one- third as efficient as "normal" (as suggested by table 3-1), the dead- weight global loss on this account would have been about one- third of the transfer, which is considerable. But if even a minor part of the deterioration in economic performance of oil importers may be attributed to the oil shocks and their associated effects (as sug- gested by other studies), these shocks would have constituted a far more serious loss. The windfall transfers could then have involved a cost much greater than the transfers themselves while providing uncertain and small benefits, and the global deadweight loss could have far exceeded the transfer. But even without such an effect, the conclusions of this study indeed support the view that the decade of the oil windfalls has involved the global economy in a massive, negative-sum game. 144 Comparative Analysis Note 1. For such reasons the state of Alaska instituted a rule of budgeting on the 30th percentile of the distribution of future revenue rather than on its expected value. PART III Country Studies Chapter 10 Algeria: Windfalls in a Socialist Economy with Patrick Conway Algeria is unique in this study of oil exporters. Whereas the other five economies relied mainly on the price system to clear markets, Al- geria is a socialist economy in which product, factor, and foreign ex- change markets are heavily controlled. Alone among the sample countries, Algeria responded to the first oil windfall by strongly increasing its foreign borrowing, although its response to the sec- ond windfall was far more cautious. Its patterns of relative price changes and sectoral evolution were also unusual. How has Algeria's strategy for using its oil revenues differed from the strategies of more market-oriented economies? How has the non-oil economy adjusted, in the absence of price flexibility, to changing levels of public spending? Have controls helped it to avoid the so-called Dutch disease with its symptoms of appreciat- ing real exchange rates, contracting non-oil tradable sectors, large shifts of resources toward construction and services, and unsustaina- ble consumption? If so, at what cost in efficiency? And have con- trols helped to avert an economic crisis in the downturn of the oil cycle? Since independence in 1962, Algeria has taken an approach to de- velopment that has been both firm and consistent. The first section below outlines this approach, as well as the structure of the econ- omy before the first oil price increase. The distinctive elements of Algeria's response to the first and second oil windfalls are explored in the following section; both the priorities adopted by the govern- ment in the 1950s and 1960s and the natural resource endowment of the country shaped that response in important ways. The sec- tion "Adjustment of the Non-Oil Economy" analyzes the evolution Patrick Conway is assistant professor of economics at the University of North Carolina. 147 148 Country Studies of the economy after the oil price increases and notes some un- usual features, including the rationing of consumer goods and the ac- cumulation of cash balances on an exceptional scale. The findings imply that, especially in the first boom, Algeria followed a strategy of deferred consumption, with the windfall and the foreign borrow- ing it facilitated going into creating a heavy industrial base, largely in hydrocarbons. Only with the realignment of investment priori- ties after 1978 did consumption, as well as industries producing con- sumer goods and housing, receive greater attention. The next sec- tion outlines the main features of a "fix-price" equilibrium model of the Algerian economy. This model is used to investigate the impor- tance of rationing and of cash balances in explaining the Algerian pat- tern. The final section offers an assessment of the Algerian strategy for absorbing windfall gains. Economic Development and State Control Algeria's development strategy was first enunciated during the eight- year revolution that led to independence. Its substance was shaped by the bitter and costly nature of that war. Despite factional dis- putes among Algerian political leaders, stemming in part from differ- ences in generation and the role played in the revolutionary strug- gle, all groups seem to have accepted three basic principles of development:' * Stimulating growth and creating jobs are important objectives. But Algeria should develop from a socialist perspective and should promote income equalization among both regions and in- dividuals. * Economic policy should promote self-reliance and minimize long-run dependence on foreign financial and technical assist- ance. * Economic planning is the responsibility of the state but should include some participation by workers. The economic as well as political history of Algeria since indepen- dence can be divided into three periods. The constitution of Septem- ber 1962 established Algeria as a one-party socialist republic with Ahmed Ben Bella as president. After three years he was toppled in a bloodless coup that placed Houari Boumedienne in the presi- dency. In 1978 Boumedienne died and was succeeded by Bendjedid Chadli. Each of these presidents emphasized different aspects of the three basic principles of development. Algeria 149 Development Strategy As the first president of independent Algeria, Ben Bella took a prag- matic approach to reconstituting the economy. In the wake of a mas- sive exodus of French colons during and after the revolution, he fo- cused on the promotion of light industry, especially consumer goods, and on agrarian reform. Management of the abandoned in- dustrial and agricultural firms was entrusted to workers (the autogestion program). One assessment of Ben Bella's program sug- gests that economic performance was disappointing and that Algeria's chronic problem of a rural and urban labor surplus per- sisted (Mallarde 1975).2 This period marked the emergence of la nouvelle classe-a middle class of public bureaucrats, technocrats, and managers-which was to play an increasingly important role in future years. Boumedienne's accession in June 1965 marked an important shift in Algerian development policy. Influenced by the prominent econo- mists Francois Perroux and G. Destanne de Bernis, whose work was in the spirit of the linkage theory of Albert Hirschman, the new president formulated a threefold strategy of industrialization, in- tegration of domestic industry within the nation, and rejection of for- eign influences.3 "Industrializing" industries-those with strong backward and forward linkages-were thought to be the key to devel- oping the economy. And of these, heavy producer-goods industries such as steel, fertilizers, and petrochemicals were believed to prom- ise the strongest linkages. The role of the hydrocarbon sector was considered particularly crucial; a "premier industrializing indus- try," it would provide a market for capital goods, supply raw materi- als for downstream processing activities, and generate resource rents to fund industrialization. Perroux's study of "growth poles" suggested that regional integration could best be achieved by estab- lishing new industrial centers in three locations-Arzew, Skikda, and Annaba-which at the time had only small industrial bases. Fi- nally, the decision to become a closed economy was heavily influ- enced by the work of Samir Amin. Amin identified a dependent econ- omy as one in which export goods and luxury consumer goods were dominant sectors, and a self-reliant economy as one in which producer goods and necessities were dominant sectors. The implica- tions of such a classification for the Algerian strategy were clear. To implement this revised development policy, the government strengthened the planning mechanism, initiated a sustained public investment drive, and held down private consumption to finance cap- ital accumulation. In 1967-72 investment rose from 21.7 percent to 150 Country Studies 25.5 percent of GDP, while private consumption fell from 58.9 per- cent to 55.2 percent. In 1970-72 parastatal investment (representing the public sector's productive role) rose by 60 percent in nominal terms, and direct Treasury investment (representing the public sec- tor's infrastructural and social role) rose by 50 percent. In the same period, the small component of private investment-estimated as a residual-was reduced by half. The public investment drive thus began some years before the first oil price rise and accelerated after 1973. Chadli became president in 1978 just as the first oil windfall per- iod was ending. Under Chadli capital accumulation and central plan- ning have been given less emphasis than earlier. This movement ac- tually began in Boumedienne's last years, as la nouvelle classe of technocrats gained greater power, but Chadli's accession acceler- ated the process. More recent economic policy, as exemplified by the five-year plan of 1980-84, has put greater stress on developing consumer-goods industries and has introduced market prices for some agricultural products. Some shift, then, has occurred in Algeria's development strategy during the years of the oil windfalls. But the broad lines of the strat- egy have been well defined and stable, more so than for any other oil exporter in this study, with the possible exception of Indonesia. Controlled Economy Algeria has been almost unique among developing countries be- cause of the highly controlled nature of its socialist economic sys- tem. All prices and formal wage scales were administratively deter- mined until a very partial liberalization in the 1980s. Treasury subsidies to public enterprises essentially removed the constraint even of administered market forces, so that firms faced a "soft" bud- get constraint like that in the Eastern European economies, which the Algerian economy resembled.4 Until recently, private enterprise was marginal except in some agricultural and a few minor urban ac- tivities. The state had a virtual monopoly on the allocation of investa- ble resources through its regulation of the (public) banking system, and the Treasury also acted as a major financial intermediary. Trade took place through official channels only, imports were controlled, and foreign exchange was tightly allocated. An obvious question is the extent to which control has been more a facade than a reality. In Algeria the institutions of control ap- pear to have been powerful during the period under consideration. Black markets undoubtedly existed, but by all accounts they were, and probably still are, minor. The Algerian bureaucracy appears to Algeria 151 be conspicuously free of the corruption endemic in so many other oil- exporting countries. Despite moves after 1978 to liberalize and decen- tralize production and free certain markets (notably for agricultural products), Algeria is still aptly described as a socialist, centrally planned "fix-price" economy. The fix-price economy with gradual adjustment is the analytical framework that has been adopted in this discussion. "Fix-price," in the Algerian context, means not that prices are immutably fixed, but that price changes are not an important mechanism for clearing markets.5 Relative prices play a major equilibrating role when a mar- ket economy increases absorption in response to windfall gains. This is particularly true of the real exchange rate, which appreciates to shift factors of production to nontraded sectors and to deflect de- mand to non-oil exportables and imports. Such standard mecha- nisms cannot operate in a fix-price economy, which must instead sub- stitute various rationing systems (on either the demand or supply side of individual markets, depending on where the scarcity is). If these systems are enforced, there is no reason to presume that the ev- olution of key price and resource indicators will follow the free- market paradigm. But the cost of enforcing a particular relative price structure may rise, or the effectiveness of the controls decline, as the administered price structure becomes less and less appropri- ate. Then, to retain control, the state will probably make some adjustments-albeit reluctantly and with a lag-in the setting of wages and prices and in the allocation of investable resources. Even before the first oil price increase, the Algerian economy dis- played distinctive structural characteristics because of the ambition of its development plan. As shown in table 5-1, total absorption of goods and services as a share of nonmining GDP was high (123 per- cent) by international standards. The excess was financed mainly by oil exports, but foreign assistance and remittances of Algerians working abroad also contributed significantly to the current ac- count. Private consumption represented only 65 percent of nominal nonmining GDP, compared with a norm of 66 percent of total GDP for a typical developing country at Algeria's level of income per ca- pita.6 Public consumption at 18 percent only slightly exceeded the norm of 14 percent. The high absorption ratio, therefore, was due al- most entirely to investment, which at 41 percent was twice as high as the norm. It is estimated that more than 85 percent of the total in- vestment in the years just before the first oil shock was attributable ei- ther to the central government or to public enterprises. In the structure of production, too, Algeria's economy was distinc- tive, as table 6-3 makes clear. Mining, equivalent to 18 percent of nonmining output in 1972, was of course a dominant sector. Agricul- 152 Country Studies ture provided only 11 percent of nonmining output (the norm being 25 percent) and manufacturing 16 percent (the norm being 20 percent). In contrast, the construction sector was much larger than the norm (13 percent compared with 5 percent) and the service sec- tor somewhat larger. Overall, the non-oil tradable sectors (agricul- ture and manufacturing) were smaller by some 18 percent of non- mining output than the norm for a typical non-oil country at Algeria's level of development. This reflected the disruption of agri- cultural and industrial production by the exodus of the French, high investment rates in the middle of the first four-year develop- ment plan (1970-73), and the financing of imports from oil, aid, and remittances rather than from non-oil exports. In this sense, Alge- ria suffered severely from certain symptoms of Dutch disease be- fore 1974-even more so than Venezuela and Nigeria, which had non- oil tradable sectors that were only 14 and 9 percentage points, respectively, below the "normal" levels for non-oil economies. Response to Higher Oil Prices Half of Algeria's exports in the early 1960s came from oil. Following the prolonged war of independence, which severely disrupted the non-oil economy, this proportion rose to three-quarters by the early 1970s. In 1972 oil exports were equivalent to 16 percent of GDP. The quadrupling of world oil prices in 1973-74 and their redoubling in 1979-80 conferred large windfall gains on Algeria as it did on the other oil exporters. In 1974-78 Algeria's domestic oil windfall aver- aged 22 percent of nonmining GDP, and in 1979-81 it averaged 30 per- cent. In comparison, the unweighted averages for the other five oil- exporters in this study were 21 percent and 22 percent for the first and second periods. With both price and volume changes taken into account, the windfall-as shown in figure 10-1-peaked in 1974 at the equivalent of 42 percent of nomining GDP. It fell to 26 per- cent in 1975 and to 18 percent in 1978 as world oil prices eased and as the volume of Algeria's hydrocarbon output grew more slowly than its non-oil economy. With the second oil price rise, the ratio of oil revenue to nonmining GDP rebounded to 34 percent. Use of the Windfalls Although the relative size of Algeria's first oil windfall was not atypi- cal for a capital-importing, oil-exporting developing country, the pro- pensity to augment that windfall by borrowing abroad was distinc- tive. As indicated in figure 10-1, the ratio of the trade and nonfactor service deficit to nonmining GDP increased in 1974-78, rela- Algeria 153 Figure 10-1. The Oil Windfall and Its Use, 1973-81 Algeria 50 - 40 -Trade and nonfactor 40 A services deficit Oil windfall OS 30 20- 10 Consumption 1970-72 1973 1974 1975 1976 1977 1978 1979 1980 1981 Unweighted Mean of Comparators: Ecuador, Indonesia, Nigeria, Trinidad and Tobago, 50 - and Venezuela 40 - 30 30 E ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~i _w 0 indlfal 20 Trae and nonfactor services surplus 100\ Consumption 1970-72 1973 1974 1975 1976 1977 1978 1979 1980 1981 Source: World Bank data. 154 Country Studies tive to its base in 1970-72, by the equivalent of 4.3 percentage points. The ratio of the current balance to nonmining GDP deterio- rated by 9.8 percentage points, and net borrowing totaled $8.2 bil- lion. Figure 10-1 also illustrates that the comparators reduced their ratio of trade and current deficits to nonmining GDP by 7.5 and 8.7 percentage points respectively, thereby cutting the impact of the windfall on absorption by one-third. Algeria, in effect, reacted to the first oil shock as an extremely resource-constrained economy; it took advantage of its improved creditworthiness to boost absorp- tion almost twice as much as the average for the comparators. Figure 10-1 reveals another unusual feature of the Algerian re- sponse. During 1974-78 only 17 percent of the increased absorption made possible by oil revenues and borrowing went to private and public consumption (5 percent to the latter). The remaining 83 per- cent went to raise investment further, to the remarkable level of 72 percent of nonmining GDP by 1977. Investment programs exploded in all the oil-exporting countries after 1974, but the Algerian re- sponse was extreme even by the standards of the sample. To a considerable extent this strategy was moderated in the sec- ond windfall period, which began just after Chadli became presi- dent. Trade deficits were reduced, private and public consumption was increased, and investment was cut by the equivalent of 13 per- cent of nonmining GDP. As a result, the share of consumption in ab- sorption began to rise. In 1980 and 1981, the current account was ac- tually in surplus despite the interest burden from past borrowing. The relatively restrained response to the second oil price rise was far more in line with that of the other oil exporters. With an 8.9 per- centage point fall in the ratio of absorption to nonmining GDP from the base period 1970-72, the contraction in absorption relative to the high-spending years was equivalent to 13.2 percent of non- mining GDP. The converse to the question of how the Algerian econ- omy, without price flexibility, adjusted to the colossal rise in absorp- tion during the first windfall is therefore: How did the economy respond to the contraction of demand-from very high levels, of course-during the second windfall? Nature of Public Investment Public investment in Algeria has two distinct components. The largely infrastructural component directly financed by the Treasury (on average, about one-third of the total) has remained relatively steady over time. The component going to public enterprises has been more volatile. It is financed by (a) long-term loans from the Alge- rian Development Bank and similar domestic financial institutions, Algeria 155 (b) medium-term credits from domestic commercial banks, and (c) ex- ternal credits from foreign banks and suppliers with approval of the Finance Ministry. In the early 1970s, Treasury savings financed about 55 percent of the public investment program. Much of the rest was financed by the excess of domestic private savings over pri- vate investment; foreign savings also financed a small share (Con- way 1982, p. 52, table 14). The increase in oil taxes permitted the Trea- sury to fund 65 percent of the expanded program during 1974-78. Nevertheless, between 1973 and 1979 the surplus of private savings channeled to the public investment program (estimated as a resid- ual from the public budget and the foreign accounts) represented an unweighted average of 7.6 percent of nonmining GDP. More- over, in this period the estimated share of public in total invest- ment rose further, from 86 percent to 92 percent. An indication of the composition of public investment by sector is given in tables 10-1 and 10-2. These tables indicate the remark- able extent to which investment was concentrated in the capital- intensive industrial sectors, notably in hydrocarbon exploitation. The public investment program was at least twice as large, relative to the non-oil economy, in Algeria as in the comparator countries. At the same time, the weight of the capital-intensive industrial sec- tors in that program was two to three times greater in Algeria than in the other countries. Relative to the non-oil economy, therefore, the Algerian investment in these sectors was at least four to six times as large as in the other countries, which themselves imple- mented some of the most ambitious heavy industrial programs in the world. Roughly $18 billion was invested in liquefied natural gas (LNG); this compares with Algeria's gross foreign debt (short-, me- dium-, and long-term) of $17.6 billion at the end of 1981. Algeria, in other words, took out a mortgage, secured by future oil reve- nues, to diversify its industrial base, primarily toward non-oil hydro- carbons. With investment went moves to consolidate national control. These also had a fiscal dimension. In 1963 the share of oil company profits paid to the government was positively associated with the ex- tent of public ownership and negatively associated with multina- tional ownership, with French ownership occupying an intermedi- ate position (Mazri 1975).7 In the same year, a public enterprise, SONATRACH, was established as a channel for state expenditures in the hydrocarbon industry. Its scope was steadily extended. By 1970 it controlled roughly half of all activity in the hydrocarbon sector and had a dominant role in exploration and domestic distribution, but only a minor role in oil, LNG, and petrochemical production. By 1980 all oil production and all but 5 percent of exploration activity Table 10-1. Algeria: Sectoral Composition of Public Investment (percent) Four-year plan, 1974-77 Interplan, 1978-79 Five-year plan, 1980-84 Sector 1975 1978 1982 Agriculture 7.7 3.1 4.0 Water development 2.9 1.5 4.0 Hydrocarbons 26.4 34.9 14.8 Other industries 36.2 28.4 16.4 Housing 5.3 8.2 13.1 Construction' 10.3 14.3 32.7 Education and training 5.4 6.0 9.4 Servicesb 5.8 3.5 5.6 a. Includes economic, administrative, and social infrastructure, construction enterprises, commune plans, and special programs. b. Includes tourism, transport, commerce, and distribution. Source: World Bank data. Table 10-2. Algeria: Structure of Industrial Investment (percent) Three-year Four-year Four-year Five-year plan plan, plan, plan, Interplan, Planned, Actual, Sector 1967-69 1970-73 1974-77 1978-79 1980-84 1980-82 Heavy industry 57.3 56.0 44.0 50.8 43.0 47.7 Iron and steel 43.7 30.4 21.5 21.5 16.4 27.4 Other 23.6 25.6 22.5 29.3 26.6 20.3 Light industry 42.7 44.0 56.0 49.2 57.0 52.6 Chemicals 23.1 10.1 16.0 11.8 10.4 9.5 Food and tobacco 7.3 7.0 5.1 8.8 14.5 13.2 Textiles and leather 7.9 5.7 8.2 12.6 10.6 15.4 Construction materials 1.5 12.3 20.1 9.5 11.7 5.1 Miscellaneousa 2.9 8.9 6.5 6.6 9.7 9.5 a. Mainly wood and paper. Source: World Bank data. 158 Country Studies was under the control of SONATRACH. This was a rapid transition from the 1963 situation when foreign administrators and techni- cians in the hydrocarbon sector had outnumbered Algerians by nearly four to one. Between 1975 and 1980 SONATRACH was host to twenty-seven macroprojects (each with a capital cost of more than $100 million), a greater number than any other single institution in the develop- ing world (Murphy 1983). By the early 1980s Algeria had installed the world's largest LNG export capacity, and gas began to be piped to Europe under the Mediterranean via the Straits of Messina in 1983. Two factors account for the high priority accorded hydrocarbon in- vestment. First, as noted above, the hydrocarbon subsector was con- sidered a "premier industrializing industry." Second, Algeria's oil was expected to be rapidly depleted, and the proceeds were ear- marked for the creation of a viable natural gas export industry.8 Algeria's natural gas fields were estimated as the fourth largest in the world, in contrast to its limited proven oil reserves, which were expected to be exhausted by the end of the century. By 1980 proven oil reserves represented twenty-four years' output and proven natural gas reserves eighty-one years' output. Natural gas ex- traction also produced condensate, a liquid crude oil by-product, so that the expansion of gas output effectively increased oil produc- tion as well. The strategy of hydrocarbon diversification began to have a real im- pact on exports in the early 1980s. In 1980 total hydrocarbon ex- ports were $13.6 billion, of which crude oil-at $9.3 billion- constituted 68 percent. By 1984 crude exports had decreased to only $3.1 billion, but total hydrocarbon revenues had remained al- most steady at $12.4 billion, with roughly equal shares for gas, crude, condensate, and refined products. Algeria was therefore cush- ioned, at least for a time, from the impact of the oil glut. Although it is beyond the scope of this study to attempt to esti- mate the payoff to Algeria's hydrocarbon investments or to predict their returns in the future, the case of LNG is illustrative. In 1983-84 LNG exports averaged $2.26 billion. Before taking into account cur- rent costs such as labor, intermediate use, and depreciation, this was only a 12.5 percent return on the estimated capital costs-a fig- ure roughly equivalent to the interest rate on foreign debt. The impli- cation is that the payoff to such investments depends critically on the evolution of gas prices and demand relative to world interest rates.9 It also depends on the level of capacity use; unlike Indone- sia, Algeria was not able to operate its plants at full capacity be- cause of marketing constraints (see below). By affecting the feasibility of increasing the various components Algeria 159 of the public investment program, past priorities shaped the use of windfalls. In general, the concentration of investment in the heavy in- dustrial sector was facilitated by the capability of this sector to imple- ment projects. This capability, in turn, resulted from the priority that the sector had enjoyed since 1965 in the planning process. In ac- cordance with the development strategy, manufacturing investment initially focused on heavy industry (producer goods) at the expense of light industry (consumer goods), although the share of light indus- try increased with the shift in the Algerian development strategy to- ward greater consumption after 1978. Social and infrastructural in- vestment also rose after that date; housing plus infrastructure accounted for only 15 percent of total public investment in 1975 but for 46 percent in 1982. In contrast, agriculture received little emphasis in the public invest- ment program. In the plan periods 1970-74 and 1974-77, agricul- ture had been budgeted for 11 percent of public investment. Actual disbursements in these two periods, however, were only an esti- mated 8.3 percent and 4.8 percent respectively, with a low of 3 per- cent in 1978. Agriculture did undergo many institutional changes in the 1970s, including land redistribution as part of the agrarian revolu- tion of 1971. But the sector continued to perform poorly, and much food had to be imported. In 1980-83 food imports were equivalent to 80 percent of value added in agriculture. Adjustment of the Non-Oil Economy The size of the Algerian investment program raises three questions for the adjustment of the non-oil economy. First, how could so mas- sive a program be reconciled with the scarcity of domestic re- sources relative to foreign exchange windfalls and borrowing? Sec- ond, how did the government manage to extract so large a share of domestic savings to finance public accumulation? Third, what has been the impact of the investment patterns on Algerian growth and the sectoral structure of the non-oil economy? The following sec- tion considers the first two questions; the third is taken up later. Reconciling Aggregate Demand and Supply The first factor that made possible the extraordinary amount of in- vestment, especially in the first windfall period, was its high im- port intensity. This, in turn, followed from the sectoral distribution of investment. Roughly two-thirds of total investment in 1974 is esti- mated to have consisted of imported capital goods, and close to half in 1978 (Conway 1982, p. 47, table 11). 1 The high import compo- 160 Country Studies nent lessened the demand for domestic goods. This moderated pres- sure on the domestic economy; it also facilitated borrowing through suppliers' credits, which represented 41 percent of public outstand- ing and disbursed debt by 1977. The same method of adjustment, put into reverse, was used to cushion the effect of slowing investment rates on the Algerian econ- omy during and after the second oil boom. The reorientation of in- vestment toward light industry and the social sectors increased the domestic content of investment. As a result, the elasticity of im- ports of goods and nonfactor services with respect to GDP fell from 1.4 in 1973-79 to only 0.8 in 1980-84, and the level of domestic activ- ity was maintained."1 Despite the high import content of Algeria's investment program (in 1983 imports of capital goods still accounted for 19 percent of capi- tal formation), the magnitude of the program had important domes- tic effects. The construction sector expanded its constant-price share in nonmining value added from 13.4 percent to 20.8 percent be- tween 1972 and 1981. So massive a shift toward investment goods would normally be associated with substantial appreciation of the real exchange rate. This, in turn, would be mirrored in appreciation of the real effective exchange rate-unless trade policy shifted dra- matically toward import liberalization. This shift did not occur in Al- geria. Yet the real effective exchange rate actually depreciated from an index of 100 in 1970-72 to an estimated average of 96 in 1974-78. This contrasts with a real appreciation of 20-40 percent in such market-oriented economies as Ecuador, Indonesia, and Ni- geria (see table 6-1). Real effective depreciation in Algeria was associ- ated with low increases in administered prices, which-though higher than their rates in 1970-72-rarely exceeded 11 percent. Im- port prices thus tended to rise relative to domestic prices, and the ad- ministered domestic price structure moved rather uniformly across major categories of production and demand. By maintaining the pur- chasing power of oil revenues over domestic goods and services, de- preciation of the real effective exchange rate facilitated the large pub- lic investment program. The second factor that enabled the government to implement a massive investment program was an exceptional propensity for Alge- rians to hold cash, which also kept inflation low. In common with other developing countries, and especially with oil exporters experi- encing an accelerated transformation toward a cash economy, Alge- ria manifested a high income elasticity of demand for money. For the period 1967-80 the elasticity of Ml stood at about 2.0 with re- spect to nonmining income. Even in 1970-72 an unusually high share of Ml was composed of currency circulating outside banks. Algeria 161 The share of currency tended to rise over time rather than to de- cline as is usual with deepening financial intermediation, so that be- tween 1970 and 1981 currency holdings increased from 25 percent to 42 percent of nonmining GDP. The bulk of these holdings is known to have been in the hands of the private sector. This im- plied large seigniorage gains to the Central Bank, as shown in fig- ure 10-2. Seigniorage averaged 7.5 percent of nonmining GDP dur- ing 1973-79, virtually equal to the internal public financing gap as residually estimated above. The propensity to hold cash absorbed about one-third of the domestic savings pool generated outside the Treasury during 1974-79, almost three times the corresponding ratio for the five other oil exporters, notwithstanding their gener- ally higher inflation rates.12 This caused a second leakage (after im- ported intermediate and capital goods) from the multiplier process normally initiated by large increases in public investment out of oil revenue. Algeria's high propensity to hold money balances appears to have resulted from rationing in markets for both private consump- tion and investment goods, and to have been facilitated by the mod- erate upward adjustment of administered prices and the low inter- est rates (2-4 percent) available on alternative financial assets. The strength of the public accumulation crowded out private consump- tion and the accumulation of real assets. Therefore Algeria did not at- tempt to contain further the already moderate rise in wage rates rela- Figure 10-2. Algeria: Currency, Inflation, and Seigniorage, 1970-81 50 -- Currency in …_--- circulation/Ml 40 _____ ~ -~~~~~ _.. .Currency in circulation/ Nonmining GDP Increase in 20 nonmining output deflator Seigniorage/ 0 _ _ ~ Nonmining GDP 1970-72 1973 1974 1975 1976 1977 1978 1979 1980 1981 Source: IMF, International Financial Statistics, and World Bank data. 162 Country Studies tive to product prices (see table 6-2) or to increase sharply the rate of non-oil taxation-which was high and buoyant by the standards of oil exporters (see tables 5-4 to 5-6; non-oil taxes represented about 25 percent of nonmining GDP). Instead, the Algerian govern- ment opted for a strategy of deferring private consumption. Private agents built up financial claims against the state, and the real unit value of these claims decreased only gradually through inflation. This strategy is captured in a fix-price macroeconomic model esti- mated for Algeria and outlined below. A third factor, which limited the inflationary impact of the invest- ment boom, was the elastic labor supply. Algeria's rate of popula- tion growth was high at nearly 3.3 percent, and its development strat- egy was capital-intensive. The rural population, which in 1963 was 63 percent of the resident population, continued to increase at about 1.8 percent a year despite rural-urban migration. Although the nonagricultural labor force rose by 5.0 percent a year in the dec- ade 1973-83, Algeria therefore still had large labor reserves. As time went on, however, it proved harder to keep consump- tion low and prices stable. From 1978 on, an increasing share of the windfall went to consumption (see figure 10-1). Inflation continued at just below 10 percent (measured by the Grand-Algiers consumer price index). With the nominal exchange rate pegged (causing a mod- est devaluation against the rising dollar after 1980) and inflation fall- ing abroad, the real effective exchange rate began to appreciate. From a base of 100 in 1970-72 it reached an average of 104 in 1979-81, 114 in 1982-83, and 123 in 1984. The degree of rationing tended to decrease in consumer-goods markets as gradual liberaliza- tion proceeded. Though still extensively regulated, Algeria's econ- omy was beginning to adjust more after the pattern of the other oil ex- porters. Growth and Sectoral Structure In 1972-81 Algeria attained the highest nonmining GDP growth in the sample: 8.6 percent. The investment surge did not, however, ac- tually accelerate growth. The preceding years, 1967-72, had repre- sented a period of recovery from the disruption of the revolution, so growth rates had been even higher-9.5 percent-before the oil boom. Incremental capital-output ratios (ICORS) were also high. Across all sectors, at constant 1974 prices, the gross ICOR was 6.8 in 1973-77, 4.4 in 1978-79, and 10.1 in 1980-82. In non-oil mining, man- ufacturing, and energy it was 8.5 for 1973-77, 4.7 for 1978-79, and 5.4 for 1980-82. In the hydrocarbon sector, real output remained con- stant between 1973 and 1977 despite the sector's receiving one- Algeria 163 quarter of public investment. Output expanded by 11 percent be- tween 1977 and 1979 but contracted sharply after 1979, even though LNG and piped gas export capacity came on stream, because of demand constraints. Both supply and demand constraints played a role in low capital productivity, which was frequently associated with the existence of surplus capacity. In 1976 Algeria's cement industry was operating at 56 percent of capacity, the metal frame industry at 52 percent, and the ceramic tile industry at 17 percent (World Bank estimates). These figures reflect supply-side constraints-bureaucratization, in- adequate training, and the absence of competition-rather than low demand, since imports of building materials were high. In contrast, the LNG industry has been operating at 40 percent of capacity in the 1980s because demand has been less than projected. The sectoral evolution of Algeria's nonmining economy after 1972 followed a distinctive pattern. The constant-price share of construc- tion in nonmining GDP rose by 0.82 percent a year during 1972-81. Yet because of the rapid expansion of manufacturing, growth in the share of the tradable sectors exceeded the norm for an economy at Algeria's level of per capita income and rate of growth by 0.68 per- cent.'3 By 1981 Algeria's Dutch disease index- defined as the imbal- ance between nonmining tradables and nontradables relative to the size of the nonmining economy-had almost halved, falling below the corresponding indexes for Nigeria, Trinidad and Tobago, and Ven- ezuela (see table 6-3). To some extent this was because the pre- shock Algerian economy had been markedly skewed against non- oil tradables. Although the pervasive system of state controls contributed to the less efficient use of labor and capital, it played some part in limiting further skewing of the economy by reducing multiplier effects emanating from public investments. Model: The Adjustment of a Fix-Price Economy to Oil Windfalls To appreciate the distinctive nature of the Algerian adjustment proc- ess it is useful to consider an aggregated fix-price model of the re- sponse to windfalls of a controlled economy. The essentials of such a model and its application to Algeria are outlined in this section."4 First, the labor market. Suppose that maximum non-oil value added is a function of capital and labor, which are combined in fixed proportions given by the development strategy selected. With overall capital per worker in the economy below the full-em- ployment level k*, unemployment must occur whether or not produc- tion is at full capacity. With capital per worker greater than k*, labor would be in excess demand if production were near full capac- 164 Country Studies ity; but there might still be unemployment if demand is less than at full-capacity output and if it determines actual output and labor use. Next, the goods market. Non-oil value added is split into wages and profits (which accrue to the public sector). Households that re- ceive wage income either spend it on consumer goods, or they build up money balances to some desired level. Thus, given out- put, employment, and wages, desired consumption is higher the greater is m, the level of real money balances per person. Depend- ing on the level of demand relative to the potential availability of con- sumer goods, households could have their purchases rationed (that is, m is above its equilibrium), or else their demands could be insuffi- cient to sustain full-capacity output (that is, m is below its equilib- rium). The equilibrium level of m for output corresponding to full- capacity production with capital k* is denoted by m*. Finally, the oil windfall. The size of the public investment pro- gram is determined by the level of oil income (assumed to be exoge- nous) and by the extent to which k is below k*; the greater struc- tural unemployment is, the more strongly investment responds to oil income and the borrowing opportunities it creates. In the me- dium run, therefore, an increase in oil income has two offsetting ef- fects on the potential supply of consumer goods. On the one hand, if it stimulates a very large increase in the public investment pro- gram or other public spending, it may reduce the availability of con- sumer goods as factors of production are switched to producer- goods sectors."5 On the other hand, it offers the prospect of increasing imports of consumer goods, thereby augmenting their supply.', The four possible configurations for such an economy are shown in figure 10-3. At the point WAL, both the labor and goods markets are in unrationed (Walrasian) equilibrium. In region REP, the econ- omy is in a state of repressed inflation with excess demand for goods and labor because of high k and m. In KEY the economy is in a state of Keynesian unemployment with output constrained by lack of demand. In CLA the economy is in a state of classical unem- ployment: the capital stock is too small, given the chosen degree of capital intensity, to employ all workers, and there is excess demand for consumer goods as indicated by high money balances. It may be shown that the boundary between CLA and REP on the one hand and KEY on the other is an upward sloping line as depicted in the fig- ure. Consider the impact of an oil windfall on the economy. If the capi- tal stock is near or above k*, the availability of labor, not capital, con- strains output. There will be little urge to accumulate further, so con- sumption can increase by importing more consumer goods. Higher Algeria 165 Figure 10-3. Fix-Price and Walrasian Equilibria m REP /~ // m* - ~-~ - - - 6Y - - / | ~~~~KEY / ki k* k CLA = Classical unemployment KEY = Keynesian unemployment REP = Repressed inflation WAL = Walrasian equilibrium consumption requires higher m; the dividing line that brings the goods market into equilibrium therefore shifts upward, as shown in the figure. But if the rate of structural unemployment is high (that is, if k is well below some level kl), a large investment response can crowd out private consumption demand so that the dividing line shifts downward. In such a case, the windfall will cause rationing in the goods market to increase-or it could shift a previously unrationed goods market into a state of rationing, with an involun- tary accumulation of real balances. The data are not sufficient to test econometrically which region Al- geria occupied during the windfall years. But there are strong a pri- ori grounds for believing that the Algerian economy fell within CLA- notably the persistence of the employment problem and shortages of both housing and diverse consumer goods. The fitting of alterna- tive econometric models in Conway and Gelb (1988) provides sup- 166 Country Studies port for this proposition."7 The model can then be estimated, assum- ing that the economy is in CLA, by specifying a process of adjustment for the administered price system. The results of estimating the model for 1967-84 suggest that the level of rationing increased considerably during the first oil wind- fall, from the equivalent of 3 percent of actual consumption to about 10 percent in 1973-78. Households accumulated money bal- ances, and the economy shifted further into CLA, an indication that it had originally been at a point such as A in figure 10-3. The re- duced intensity of investment and the shifts in its allocation after 1978 are reflected in a change in the response of public spending to oil income and borrowing, and in some decrease in the rationing coef- ficient. 8 These results appear to confirm that the restraint on consump- tion and the increases in cash balances significantly affected Al- geria's response to its oil windfalls, particularly during the first oil boom. Venezuela, the only other country in this study whose real ef- fective exchange rate depreciated during the first oil boom, ad- justed by controlling prices and liberalizing imports, which had been tightly constrained before the boom. As a consequence-and in contrast to Algeria-the share of Venezuela's windfall that went to increase real private consumption was high, as discussed in chap- ter 15. Conclusions The Algerian pattern of development since the early 1960s has dif- fered from that of the other oil exporters in important ways. First, Algeria's economy has been centralized and state-controlled, with the public sector playing a crucial entrepreneurial role. Second, alone among the countries in this study but like Mexico in 1979-81, Algeria borrowed heavily against its anticipated future oil income during the first boom. It channeled both its oil revenues and its bor- rowed funds into investment, especially in a limited range of trada- ble sectors, and followed a strategy of deferring large increases in con- sumption. During the second boom, the strategy shifted toward consumer, social, and infrastructural goods. This raised the domestic content of investment and permitted the momentum of growth to be main- tained. In 1981-85 nonhydrocarbon GDP grew at nearly 7 percent. Large current account deficits were avoided after 1979. Total exter- nal debt, net of reserves, actually fell by $1.0 billion between 1981 and 1984. Algeria was also cushioned from the oil glut, at least through 1985, by its diversified hydrocarbon industrial base. Algeria 167 The remarkable investment build-up was effected without soar- ing inflation, without real effective exchange appreciation (for a time), and without further shrinkage in the relative size of the non- oil tradable sectors. Instead, accommodation seems to have de- pended on holding down consumption out of the windfall pro- ceeds until the second boom, at which time investment was scaled back. During the period of the oil shocks, Algeria appears to have been in the classical unemployment variant of fix-price equilibrium, with excess demand for goods, excess supply of labor, and involun- tary holdings of money balances resulting from rationing in the goods market. Econometric results support this view of the Algerian adjustment process. They suggest that without rationing the Algerian price level would have had to increase in 1974-78 by some 40 percent more than its historical trend to hold consumption at pre-1973 lev- els, even under the favorable assumption that desired real balances were unaffected by inflation. And if desired real balances had been sensitive to inflation, the Algerian strategy of the 1970s would not have been feasible without far tighter wage controls or higher non- oil taxes; otherwise the state could not have secured so large a share of purchasing power. Have the distinctive features of Algeria's controlled economy been an advantage or a liability in dealing with the changes in its terms of trade since 1973? On the negative side, there are indica- tions that Algeria's investments have not been as productive as planned, either in the hydrocarbon or nonhydrocarbon sectors, and that its strategy has failed to stimulate nonhydrocarbon exports. But it is not clear how much this can be attributed to its planned econ- omy since the same might be said for some other oil exporters as well. A serious problem, which more flexible planning might have al- leviated, however, has been the capital intensity of the public invest- ment program in the face of widespread underemployment. The strategy of deferring consumption has also had a cost in the short to medium run. On the positive side, Algeria has avoided foreign exchange cri- ses, partly because of the extent of control over international transac- tions and partly because of its more cautious approach to spending during the second oil boom. High growth has been maintained even through the oil glut. An excessive shift to nontraded goods was avoided. Moreover, major steps have been taken to diversify away from oil, a resource that is rapidly depleting, to natural gas, a resource that has a far longer expected lifetime. It will be possible to evaluate the Algerian strategy fully only when the benefits of rapid accumulation, as well as the costs- 168 Country Studies including forgone consumption and the accumulation of foreign debt-become clear. The outcome will depend on three factors, two external and one internal: developments in gas markets, the level of world interest rates, and the efficiency with which nonhydro- carbon investment and domestic labor can be used. If the external fac- tors foster a high return to hydrocarbon investments (that is, if the ratio of gas price increases to interest rates is favorable) and if domes- tic policies encourage efficiency, the transformation of oil income and borrowing into domestic capital will permit substantial growth of real consumption in the future. If not, alternative strategies, such as saving abroad or increasing domestic consumption more di- rectly through relaxing import restrictions and easing non-oil taxa- tion, will appear in retrospect to have offered greater promise of im- proving the well-being of the Algerian population. Notes 1. For analyses of the political differences, see Quandt (1971) and Zartman (1975). The documents enumerating the Algerian principles of de- velopment are: Program of Tripoli, May 1962; Constitution, September 1963; Charter of Algiers, April 1964; National Charter, June 1976; Constitu- tion, November 1976; Three-Year Plan (1967-69); Four-Year Plan (1970-73); Four-Year Plan (1974-77); Five-Year Plan (1980-84). 2. Estimates made by the Ministry of Planning suggest that during Ben Bella's rule perhaps two-thirds of the rural and one-fourth of the urban labor force were either unemployed or underemployed. 3. The intellectual origins of the Algerian strategy are discussed by Raffinot and Jacquemot (1977). 4. Kornai (1978) analyzes the "soft" budget constraint of the production sector in the context of East European socialist countries. Among the devel- oping countries only pre-reform China would seem to rival Algeria's de- gree of economic control. 5. The terminology is from Hicks (1965). 6. The norm is derived from regressions of Chenery and Syrquin (1975). 7. It is estimated that in 1963, whereas 44 percent of the gross profits earned by public oil companies accrued to the Treasury, only 17 percent of profits earned by privately owned companies accrued as tax or royalty (Conway 1982, p. 40, table 8). 8. This strategy was outlined in "Hydrocarbon Development Plan: 1976-2000," prepared for Algeria by the Bechtel Corporation. 9. Inclusion of condensate exports with LNG would double the gross re- turn to gas investments as computed here. But it is not clear how to value condensate exports because their real contribution appears to have been to permit crude oil output to be reduced to levels judged more consistent with the objective of preserving the long-run production capacity of the geo- logically complex Algerian fields. If the cost of attempting to accelerate Alge- Algeria 169 rian crude oil output is indeed high, the value of condensate exports should be at least partly included in the calculation of gross return to gas in- vestments. 10. Large hydrocarbon projects frequently have a direct import compo- nent of 80 percent, and the domestic component consists largely of construc- tion. Value added in construction is also typically less than half of gross out- put. Algeria's construction sector is known to import a high proportion of intermediates. Although transport costs for such imports are often high, gross construction expenditures can overstate the direct and indirect domes- tic resource content of investment. 11. A similar realignment of public investment away from projects heav- ily dependent on foreign exchange was used to cushion adjustment in Indo- nesia to the downside of the second oil boom; see chapter 12. 12. Seigniorage (the rent accruing to government for its monopoly to issue fiat money) normally accounts for about 1.5-2.5 percent of GDP in devel- oping countries. Morgan (1979a, b) has noted the tendency for money stocks to rise unusually rapidly in the oil exporters with the rapid monetiza- tion of their economies. 13. The tradable sectors-agriculture and manufacturing-would have re- duced their aggregate share in the nonmining economy by 0.54 percent a year according to the Chenery-Syrquin norm. Instead, in Algeria they ex- panded their share by 0.14 percent a year. 14. For a complete description of the model and its estimation, see Conway and Gelb (1988). 15. Such an effect can also dominate in a flexible-price economy in which public investment draws resources from wage-goods sectors where import- ing is constrained; see chapter 13 on Nigeria. 16. In the long run, the domestic supply of consumer goods could of course rise, depending on the investment profile. 17. As noted above, some firms did operate at less than full capacity be- cause of limited demand. But these were in hydrocarbons and producer goods, not in housing and consumer goods. 18. Estimates of the unrationed consumption and money demand func- tions imply propensities to consume and to accumulate real balances that are similar to those that have been estimated for the United States. Algeria's response is thus due more to rationing than to distinctive private preferences. Chapter 11 Ecuador: Windfalls of a New Exporter with Jorge Marshall-Silva Unlike the other oil producers considered in this study, Ecuador was a new exporter at the time of the first oil price increase. Be- cause the country's oil output during 1974-78 was only half of what had been projected, the rise in price largely compensated for the shortfall in volume, leaving a relatively small first windfall. Nevertheless, it is harder to discern a coherent strategy for absorb- ing oil income in Ecuador than in the other five countries. This is partly because of political instability during the period and partly be- cause its public sector is highly decentralized. Earmarking of oil in- come as well as other revenues was extensive, so that the feasible range of choice was more limited than in the other countries. Per- haps as a result of this constraint, the Ecuadorian government did not use its oil revenues to implement ambitious, large-scale public industrial programs. Instead, it channeled most of the oil income received into "traditional" areas such as social and physical infra- structure. It also returned much of its potential revenue gain to the private sector through subsidies, tax rebates, and the subsidized on- lending of foreign borrowing. In some respects this pattern of use may have turned out to be an advantage, given the performance of large-scale industrial invest- ments in other oil-producing countries. Yet the same political and in- stitutional conditions also weakened the government's control over public spending. In contrast to its earlier tendency (and that of the other exporters except Algeria), Ecuador ran persistent and sizable current account deficits in the 1970s and 1980s. Non-oil taxation slack- Jorge Marshall-Silva is professor of economics at the University of Santiago, Chile. 170 Ecuador 171 ened. Public consumption, public investment, military expendi- tures, and fuel, power, and food subsidies increased. Credit policy was expansive and coupled with restrictive controls on domestic in- terest rates. And despite domestic inflation, the exchange rate was fixed from 1970 until the foreign exchange crisis in 1982. Nevertheless, Ecuador's record was quite favorable compared with that of some other oil exporters. Non-oil growth increased from its past levels. Gains in infrastructure, education, and health were impressive. The country acquired considerable private indus- trial capacity, although primarily in protected sectors. Agricultural growth was moderately good though far less impressive than in Indo- nesia. But oil wealth made less of an impact on poverty than it might have because of the capital-intensive pattern of industrial de- velopment encouraged by public policies. And the mid-1980s found the country facing falling world oil prices with a far larger external debt than before the first boom and with an insufficiently diversi- fied export portfolio. The danger, therefore, is that gains will prove to be transient since the oil windfalls were not used to correct long- standing socioeconomic problems. The first section of this chapter outlines Ecuador's economic and political structures and the nature of its public sector before the wind- fall, features that influenced the country's subsequent evolution. The following sections consider the use of revenue from the first windfall of 1974-78 and then the second windfall, which ended in a foreign exchange crisis. The final section sums up with the ques- tion: How has Ecuador benefited from the windfall? Before the Windfall Before 1972 Ecuador was one of the poorest countries in Latin Amer- ica, with great disparities among the three main geographical re- gions: the Coast, where plantation agriculture predominated; the Sierra, where subsistence farming predominated; and the undevel- oped and sparsely populated Orient, or Amazonian region, where oil was to be found. The people of the Coast (called montevios) are a mixture of Indians, whites, and blacks while the people of the Si- erra are 80 to 90 percent Indian or mestizo. The traditional way of life of much of the populace, the nature of social and political institu- tions, and great inequalities in income distribution reflected the country's economic backwardness. Vast differences between the Coast and the Sierra-in climate, pattern of landownership, type of farming, and course of economic development-were exacerbated by difficulties of communication. Table 11-1. Ecuador: Equivalent Exchange Rates (sucres per dollar) Item 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 a. Imports 31.2 33.3 31.1 29.0 29.8 30.8 30.8 30.3 30.0 30.6 b. Traditional exports 23.1 22.8 21.0 21.5 23.0 21.6 23.3 22.2 22.3 23.3 c. Nontraditional exports 20.2 21.5 24.6 25.5 26.6 27.9 30.5 30.0 27.3 26.9 (a-b)/b (percent) 35.1 46.0 48.1 34.9 29.6 42.6 32.2 36.5 34.5 31.3 (a-c)/c (percent) 54.5 54.8 26.4 13.7 12.0 10.4 1.0 1.0 9.9 13.7 Note: Entries are nominal rates adjusted for tariffs and subsidies. Source: World Bank data. Ecuador 173 Economic Structure Much of the political conflict between the dominant classes of the Coast and the Sierra and the consequent political instability can be traced to the socioeconomic differences that reflect regional dual- ism. In the Sierra output was mainly determined by domestic demand for food; industry, other than handicrafts, was almost non- existent, and economic surpluses were spent largely on imported lux- uries. Foreigners dominated internal trade. The best land had long been in the hands of big landowners (hacendados), while small proprie- tors eked out a subsistence by farming their hilly and often eroded land and by working part-time for the hacendados or in nearby towns and villages. The tropical climate of the Coast favored the cultivation of export crops. Cacao was the main engine of Ecuador's economy from the late nineteenth century until the 1920s. It led to the development of plantations and made the port city of Guayaquil a center of com- merce and finance. A sharp decline in cacao production from 1920 to about 1950 thrust the country into a long period of economic stag- nation. This ended with the introduction of bananas in the 1950s, which became the main export crop until superseded by oil in 1973. The banana boom of the 1950s and 1960s spurred industrial growth, which rose from a very small base at 8 percent a year in the 1950s and 10 percent a year in the 1960s. Especially in Quito a new bourgeoisie began to emerge, the re- sult first of industrialization and then of oil. Industrial growth pro- ceeded within the framework of import substitution, effected by very generous concessions from the mid-1950s. By the early 1970s, duties and taxes raised the average cost of foreign exchange for im- ports by more than 40 percent above that received by exporters, as table 11-1 indicates. Many industrial activities added little domestic value so that effective protection rates varied widely, ranging up to 300 percent. The Ecuadorian market was small and quantitative im- port controls promoted oligopolistic and uncompetitive production. Meanwhile, efficient and competitive coastal agrobusiness provided nearly all of Ecuador's foreign exchange earnings until 1970. Ecuador has produced oil in small quantities since 1911. But in 1967 Texaco made major new finds, and in the early 1970s oil develop- ment began on a large scale. From 1970 to 1976 cumulative invest- ment in the exploration, extraction, and export of oil represented nearly half of Ecuador's average annual GDP. Current account defi- cits of nearly 7 percent of GDP in the early 1970s were mostly fi- nanced by direct foreign investment in the oil industry. Growth was rapid in this phase, and the ratio of total investment to GDP 174 Country Studies soared from only 6.2 percent in 1965-66 to 23.8 percent in 1970-72. By the end of 1973 foreign exchange reserves represented 70 per- cent of external public debt as the oil sector shifted from its extrac- tive to its high-rent phase. Political Structure Ecuador's government is modeled on that of the United States. The executive branch is headed by a president who is elected for a sin- gle four-year term, and legislative power resides in a one-chamber parliament. But there is a plethora of political parties, which fall into three broad groups: conservative, centrist-popular, and marx- ist. Socioeconomic differences and rivalries between the dominant groups of the Coast and the Sierra help explain the country's his- tory of political instability. Following the cacao crisis of the 1920s, the hacendados of the Si- erra and the smaller farmers of the Coast joined in unstable alli- ances, interrupted periodically by armed intervention. For almost forty years, political life centered around a strong populist caudillo, Jose Maria Velasco Ibarra, five times elected to the presidency and four times toppled by the armed forces. Most of Ibarra's Coast constit- uency shifted to another populist movement, headed in the late 1960s by Assad Bucaram, former mayor of Guayaquil. The likeli- hood of Bucaram's winning the presidency in June 1972 was the main reason for the military coup in February of that year. During the decade of oil windfalls that followed, several more shifts in government occurred. The military dictatorship of General Rodriguez Lara that was established in 1972 was strongly nationalis- tic and attempted to eliminate economic dependency on foreign pow- ers, to promote self-sufficient economic development, and to carry out a radical agrarian reform. The government gradually moderated its revolutionary stance and compromised with the dominant classes, particularly on agrarian reform. In January 1976 Lara was re- placed by a military junta which was to draft a new constitution and effect a transition back to civilian rule within two years. This was finally achieved in 1979, with the election of Jaime Roldos as pres- ident. In 1981 the death of Roldos and a border conflict with Peru again precipitated political instability. Dr. Osvaldo Hurtado became president until 1984, when power passed to a conservative coali- tion. These frequent shifts in government shortened the horizons of poli- cymakers and so affected the way Ecuador used its oil windfalls. Ecuador 175 The Ecuadorian military resembled the Peruvian military then in power, although it was less radical. Its anti-imperialistic program included defense of the country's national interests and natural resources, solidarity with Third World nations, Andean integration, active participation of workers and peasants in political life, and agrarian reform. The Lara government moved gradually to a more centrist stance. The junta (1976-79) was more conservative, al- though its transitional character, its collective leadership, and its preoccupation with constitutional and political problems hampered economic decisionmaking. The constitutional government (1979-84) was left of center, with Roldos (1979-81) concerned more with urban issues and Hurtado (1981-84) more with rural poverty. How- ever, the exchange crisis and the recession which beset the econ- omy during Hurtado's administration undermined efforts to reduce poverty. In this highly political situation, the country could implement few long-term plans for allocating its oil revenues. Political continu- ity was provided mainly by the public bureaucracy (which is esti- mated to have increased from 6.4 percent to 8.7 percent of total employment between 1975 and 1981), the camaras (organizations of private businessmen that wielded a good deal of political influ- ence), and the armed forces.' The Public Sector Ecuador's public sector had grown through the establishment of au- tonomous institutions and public enterprises, mostly financed from specifically earmarked sources of revenue. When account is taken of other contractual payments, such as interest on debt, some three- quarters of budgetary revenues were so earmarked. This decentral- ized structure was anotherimportantfeature thatlimited the government's power of fiscal control. In the context of the oil windfalls, the most important public enter- prises were CEPE (the State Petroleum Corporation) and INECEL (the Ecuadorian Electricity Institute).2 In the early 1970s the government set up, acquired, or participated in a variety of other productive enter- prises. Between 1970 and 1973, forty-two public industrial enter- prises were established with an investment roughly equivalent to $160 million. This process was discontinued in later years. And al- though it was proposed to create a petrochemical industry with the gas of the Guayaquil Gulf, this possibility faded with the onset of the oil glut. Ecuador's government therefore played by far the small- est entrepreneurial role of any government in the sample. Table 11-2. Ecuador: Oil Output, Consumption, and Net Exports, 1972-83 (millions of barrels) Item 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 Production of crude 28.6 76.2 64.6 58.8 68.4 67.0 73.7 78.2 74.8 77.1 77.1 86.3 Consumption 9.6 10.7 12.0 14.4 14.4 19.6 22.6 25.3 27.7 29.9 30.4 27.5 Net exports 19.0 65.5 52.6 44.4 54.0 47.4 51.1 52.9 47.1 47.2 46.7 58.8 Value (millions of U.S. dollars) 45 260 702 526 648 611 629 1,194 1,571 1,589 1,344 1,623 Index of domestic use/ nonmining GDP 100 103 103 112 104 130 142 151 155 162 161 155 Source: Marshall-Silva (1984); World Bank data. Ecuador 177 The First Windfall This section considers the use of Ecuador's oil revenues in 1974-78, even though the discussion occasionally touches on trends over the entire period 1974-82. Because the first oil price increase coincided with the start of oil exports, the whole of the net oil earnings may be considered as a windfall. Table 11-2 summarizes oil output and use for 1972-83, and table 11-3 computes the oil bonus. Net exports in table 11-2 are the difference between production and internal con- sumption. Net export revenues in table 11-3 are the value of net ex- ports less profits of the foreign oil companies and less imports by the oil industry. Two adjustments are then made. Part of the bonus accruing to Ecuador was passed on to domestic users of oil deriva- tives, which continued to be sold at their pre-1974 prices. As a result of this policy, domestic sales (which were used for illicit exports as well as domestic consumption) rose rapidly; their ratio to nonmining GDP increased by 62 percent during 1972-81. The first ad- justment made in table 11-3 is that the value of this implicit sub- sidy, which peaked in 1980 at 7.3 percent of GDP, is added to the net export revenues. Second, payments for CEPE'S shares in the oil production consortium are deducted. The windfall, as thus mea- sured, averaged 12.3 percent of GDP in 1974-78 and 16.3 percent in 1979-80. Even by the standards of oil exporters Ecuador's domestic energy prices were kept low. The subsidy to domestic users accounted for a sizable part of the total windfall, particularly after the second oil price rise in 1979. Oil for the home market was priced at or near the cost of production plus transport (estimated at $1.00 a barrel in 1973 and about $2.75 in the early 1980s). The government therefore derived almost no revenue from oil consumed at home. Reduction in Oil Output Annual production, as table 11-2 shows, was lower in 1974-78 than in 1973. Although the government's revenue (in real terms) per bar- rel of oil exported was 2.6 times that projected before 1973, output was only half of the projected 400,000 barrels a day. This can be attrib- uted to the nationalistic orientation of the Lara government, which retroactively applied a 1971 law limiting the concessions granted to foreign oil companies, raised oil taxes and royalties, and increased the reference price for oil exports on which taxes were based. In mid-1974 conflicts between the government and oil companies para- lyzed exploration activities, and in 1976 the government decided to compute CEPE'S quarter share in the oil consortium on the basis of Table 11-3. Ecuador: The Oil Windfalls, 1973-83 (millions of U.S. dollars) Item 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 Net exports 305 702 526 648 611 629 1,194 1,571 1,589 1,344 1,623 Profits of foreign companiesa 106 157 11 25 18 15 16 20 28 46 n.a. Inputs from abroad 60 58 57 99 83 130 183 170 240 263 n.a. Net export revenues 139 487 458 524 510 484 995 1,399 1,321 1,035 n.a. Subsidy to domestic oil use 26 127 129 148 198 212 519 857 625 590 290 Subsidy/GDP (percent) 1.1 3.4 3.0 2.8 3.0 2.8 5.6 7.3 4.5 4.4 2.3 Payment for CEPE'S shares 43 0 0 0 82 0 35 0 0 0 0 Total windfall/GDP (percent) 4.9 16.5 13.6 12.7 9.4 9.1 15.8 19.2 14.0 12.0 n.a. n.a. Not available. a. Oil company profits for 1974 and perhaps for 1973 are probably overstated because the reference price for oil at which the companies had to surren- der foreign exchange to the Central Bank was higher than the actual price the companies were getting for their exports. Source: Marshall-Silva (1984). Ecuador 179 maximum legal production rather than actual production. In 1977 Gulf withdrew, selling its share in the consortium to CEPE. One use of the windfall, therefore, was to finance national control of the oil in- dustry and to conserve reserves. Intention versus Outcome: The 1973-77 Plan How did Ecuador plan to use its oil revenues? And did it follow its plan? In 1972, at the beginning of the oil era, the Planning Board (junta de Planificaci6n) drafted a detailed "Plan of Transformation and Development" (PIDT). The plan noted that in the past "the peri- ods of relative bonanza [export windfalls] were translated in a short time into economic instability manifested in balance of payments problems and a fiscal deficit of even greater magnitude than pre- vailed prior to the period of prosperity" (PIDT, pp. 3-4). A main objec- tive was to avoid a similar occurrence in the future. Ecuador's oil wealth was to be used to transform its productive structure and in particular: * To strengthen and integrate the country and reaffirm national sovereignty * To improve living conditions, especially for the poor * To strengthen the productive sectors so that they could absorb more labor at rising productivity. To achieve these goals, the plan called for three basic institutional re- forms. Agrarian reform was essential to eliminate rural poverty. Tax re- form would provide the state with domestic resources needed for the public investment program. And reform of public administration would allow the state to participate effectively in transforming the economic structure. In practice, however, these structural reforms were essentially ig- nored. Agrarian reform fell far short of its goals, and the land that was made available to peasants came from colonization of new areas rather than from redistribution. There was no tax reform and virtually no reform of public administration. Growth of the public sector was even more rapid than planned. This was especially true of consumption since the recurrent expendi- tures required by investments in hospitals, schools, and other facili- ties had been underestimated. Public consumption absorbed a third of the windfall as measured in table 5-2, the highest ratio in the sam- ple. Again, fiscal restraint failed. After 1973 and 1974, two years of surpluses, public spending outran revenue; deficits averaged 4.4 per- cent of nonmining GDP in 1975-78. The distribution of public expendi- tures in 1974-78 also differed considerably from that of the plan. In Table 11-4. Ecuador: Nonfinancial Public Sector (percentage of nonmining GDP) Change from 1973 Item 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1974-78 1979-82 1974-78 1979-82 Current income 29.9 35.8 31.5 29.7 29.9 27.7 33.1 32.1 31.2 29.9 30.9 31.6 1.0 1.7 Petroleum taxes 4.3 10.5 6.9 7.0 5.5 4.8 6.9 8.5 4.1 10.6 6.9 7.5 2.6 3.2 Other taxes 15.8 14.1 12.6 11.1 11.1 11.5 11.2 10.6 15.2 9.1 12.1 11.5 -3.7 -4.3 Current expenditures 18.4 22.8 23.7 22.8 23.9 22.4 23.3 25.9 26.0 27.4 23.1 25.7 4.7 7.3 Government consumption 11.4 15.2 16.4 15.7 16.4 14.9 14.6 16.5 16.3 15.7 15.7 15.8 4.3 4.4 Remuneration 9.4 10.3 11.1 10.9 10.0 9.5 9.3 11.3 11.2 11.1 10.4 10.7 1.0 1.3 Interest 0.9 1.3 1.0 1.3 1.8 1.8 2.7 2.9 2.5 5.6 1.4 3.4 0.5 2.5 Capital expenditures 8.1 12.1 10.3 10.6 12.0 10.6 10.4 11.4 11.8 11.4 11.1 11.2 3.0 3.1 Balance 3.5 1.0 -2.4 -3.7 -6.0 -5.4 -0.6 -5.1 -6.6 -9.0 -3.3 -5.3 -6.8 -8.8 Current account/ -0.5 - - - - - - - - - -4.1 -8.5 -3.6 -8.0 nonmining GDP (percent) Source: World Bank data. Ecuador 181 particular, a considerably lower share was allocated to industrial de- velopment than had been projected. These departures from the plan resulted from Ecuador's planning and decisionmaking processes. The Planning Board had only lim- ited influence. Many decisions were made on the spur of the moment in the face of political pressures. All major state institu- tions participated in policy decisions related to oil. Most of the oil revenues were already earmarked for specific recipients. Notwith- standing the weakness of the Planning Board and the constraints im- posed by earmarking, however, some of the goals of the 1973-77 plan-particularly those related to growth-were exceeded. This stands in stark contrast to the outcome of the second oil boom; none of the economic goals of the 1980-84 plan were achieved, al- though that plan laid greater stress on improving social conditions than had the previous one. The National Income Department of the Central Bank of Ecuador has attempted to establish the destination of oil revenue by eco- nomic activity for 1974-78, but this is not easy since it requires an as- sessment of spending trends in the absence of such revenue. For the period as a whole, expenditures by economic sector were: Administration, research, and planning: 20.7 percent Agriculture and fishing: 7.0 percent Industry, mining, and petroleum: 9.6 percent Electricity, roads, and transportation: 20.5 percent Education, health, and housing: 19.4 percent Other, including defense and the public debt: 22.8 percent This distribution of the oil windfall signified a substantial shift to- ward economic and social infrastructure. Reduction in Non-Oil Tax Revenues An alternative fiscal classification, based on the evolution of the pub- lic sector budget relative to the nonmining economy, is presented in table 11-4. (Unfortunately, comparable data are not available be- fore 1973.) Comparing 1973 with 1974-78, the table indicates a rise in oil revenues equivalent to 2.6 percent of nonmining GDP, but this was more than offset by a fall in non-oil taxes of 3.7 percent of nonmining GDP. Current spending increased its share by 4.7 per- cent, and capital spending its share by 3.0 percent. The result was a swing from a surplus of 3.5 percent in 1973 to a deficit of 3.3 per- cent for 1974-78, an evolution which was reflected in the current account. Almost half of Ecuador's non-oil tax revenues came from taxes Table 11-5. Ecuador: Subsidies, Exchange Rate Effects, and the Relative Price of Consumer Goods, 1973-82 (percentage of GDP) Item 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 Subsidy Internal oil use 1.1 3.4 3.0 2.8 3.0 2.8 5.6 7.3 4.5 4.4 Credit (interest) 0.5 0.5 0.5 0.5 0.5 0.5 0.6 0.6 0.7 0.7 Wheat 0.6 0.6 0.5 0.6 0.6 0.5 0.5 0.5 0.4 n.a. Milk and other 0.3 1.0 0.4 0.2 0.1 0.0 0.2 0.2 0.2 n.a. Electricity n.a. n.a. n.a. 0.2 n.a. 0.2 0.3 0.1 0.2 0.5 Total subsidies 2.5 5.5 4.4 4.3 4.1 4.0 7.1 8.7 6.0 5.4 Effect of cheaper imports from real exchange appreciation n.a. 1.4 3.8 3.8 4.4 4.5 4.5 3.7 5.3 2.5 Total 2.5 6.9 8.2 8.1 8.5 8.5 11.6 12.4 11.3 7.9 Consumer price index/ nonmining GDP deflator (1970-72 = 100) 96 97 98 93 90 92 91 88 89 87 Consumer deflator/ nonmining GDP deflator (1970-72 = 100) 98 98 99 94 91 91 93 90 93 93 Minimum real wage index (1970 = 100) Paid by employers 95 122 132 148 131 119 156 247 215 199a Received by employees 95 126 138 153 136 124 165 254 221 201b n.a. Not available. Note: Columns may not add up to totals because of rounding. a. For 1983 the figure was 173. b. For 1983 the figure was 183. Sources: Marshall-Silva (1984); World Bank data. Ecuador 183 on international trade, about 20 percent from income taxes, and 30 to 40 percent from sales and other excise taxes. The 1973-77 plan (and later plans) envisaged fiscal reforms to increase non-oil tax reve- nues and to make the distribution of the tax burden more equitable. Tax concessions and exemptions were granted, mainly to industry. Average nominal tariff rates were cut and export taxes were some- what reduced to compensate for the overvaluation of the sucre, ex- cept for taxes on coffee during the boom of 1976-78. Nontraditional exports received incentives through negotiable export tax certifi- cates (CATS). As a result of these policies, the degree of anti-export bias implicit in the system of trade taxes was reduced sharply for non- traditional exports after 1974 (see table 11-1), although all traded sec- tors were affected by the real effective appreciation of the sucre in this period-from 100 in 1970-72 to 121 in 1974-78. One use of oil in- come, therefore, was to cut non-oil taxes. But little was achieved in terms of broadening the base and strengthening the tax system. CURRENT SPENDING. Public consumption increased sharply relative to nonmining GDP in 1974-75, then stabilized, although because of growing interest payments the share going to recurrent spending continued to increase. Subsidization of credit, wheat, milk, and other goods boosted current spending. Electricity was also subsi- dized: INECEL's return to its investments was more or less zero for 1974-78 (it turned negative after 1981) even though it benefited from fuel oil provided at roughly the cost of production. Estimates of the magnitude of these subsidies are given in table 11-5. PUBLIC INVESTMENT. Without the oil bonus, public investment would have been about 5 percent of GDP (a figure normal before 1973). In fact, it increased its share of nonmining GDP from 8.1 per- cent in 1973 to 11.1 percent in 1974-78 (see table 11-4). If 5 percent is taken as a base level, then roughly half of the oil windfalls (plus the net external borrowing that they facilitated) that was not used for defense went to public investment, while the other half went to current expenditures.3 Table 11-6 shows the distribution of public investments by eco- nomic category. Physical and social infrastructure and electrifica- tion account for the overwhelming bulk of investment after 1973. Their combined share declined from 1975 to 1978, whereas invest- ment in the petroleum sector increased strongly. About one-fourth of public investment was used to provide drinking water, sewage fa- cilities, hospitals, medical centers, and schools and to improve liv- ing conditions in the cities. More than two-thirds went for energy, 184 Country Studies Table 11-6. Ecuador: Sectoral Composition of Public Investment, 1973-81 (percent) Economic category 1973 1974 1975 1976 1977 1978 1979 1980 1981 Production sectors n.a. n.a. 8.4 9.0 10.2 7.4 1.5 4.6 6.1 Agriculture 7.2 7.9 8.0 8.8 10.0 7.1 1.3 4.4 5.9 Other n.a. n.a. 0.4 0.2 0.2 0.3 0.2 0.2 0.2 Natural resources n.a. n.a. 12.2 20.3 22.0 24.7 34.2 34.7 30.3 Petroleum 2.3 9.7 2.8 7.0 8.6 9.2 12.2 11.3 15.4 Electricity 18.0 12.0 9.2 13.0 12.9 15.4 21.9 23.1 14.6 Other n.a. n.a. 0.2 0.3 0.5 0.1 0.1 0.3 0.3 Physical infrastructure n.a. n.a. 48.4 39.5 37.7 38.5 40.4 34.1 27.1 Roads 39.4 42.5 38.0 31.3 27.0 24.1 20.7 21.8 18.2 Other transport n.a. n.a. 4.2 3.8 5.9 6.5 11.3 6.3 2.8 Hydraulic n.a. n.a. 4.3 3.0 3.4 4.0 3.7 2.3 4.8 Other n.a. n.a. 1.9 1.4 1.4 3.9 4.7 3.7 1.3 Social infrastructure n.a. n.a. 28.3 28.6 28.2 27.6 22.2 23.9 34.5 Water supply n.a. n.a. 10.6 7.0 4.5 4.2 5.6 4.5 7.2 Sewage n.a. n.a. 1.6 3.3 2.7 2.2 1.8 2.9 4.1 Urban equipment n.a. n.a. 5.7 4.9 5.9 6.5 5.0 5.6 12.3 Health 4.2 2.0 5.2 6.7 7.6 8.3 5.5 3.9 3.1 Education 12.5 8.2 5.0 6.4 5.8 6.4 4.3 6.3 6.0 Other n.a. n.a. 0.2 0.3 1.7 n.a. n.a. 0.7 1.8 Other investments n.a. n.a. 2.7 2.6 1.9 1.8 1.7 2.7 2.0 n.a. Not available. Note: Totals do not include investments by public finance institutions. Sources: 1973 and 1974, U.N. Economic Commission for Latin America (ECLA), Re- port on Ecuador (1979); 1975-81, CONADE (1983). roads, and transport. Investment in the non-oil tradable sectors was low, with practically all of it directed to agriculture. BORROWING AGAINST FUTURE OIL INCOME. Ecuador's external debt at the end of 1976 was only 13 percent of GDP, but from 1976 on the government started to borrow heavily from commercial banks. Access was easy because of the country's status as an oil exporter, and real interest rates appeared to be strongly negative because of real effective appreciation of the sucre. Remarkably, public foreign borrowing financed more than the pub- lic sector deficit. As shown in table 11-7, loans to the private sector rose far more rapidly than deposits by the private sector, and at rates well above the growth rate of output. This was not surprising given the interest rate policy. Real rates on savings deposits aver- Ecuador 185 aged -7.6 percent in 1971-83, and real lending rates were also negative, especially those from official development banks. Adminis- tered interest rates had been set at levels that took account of Ecuador's tradition as a low-inflation country. They were main- tained through 1982, even though inflation accelerated from its his- torical trend of 5 percent to 14.7 percent in 1977-78, as measured by the consumer price index. Far from crowding out the private sec- tor, easy access to foreign savings permitted the coexistence of a growing deficit and an expansionary credit policy until 1982. Ecuador's government therefore reduced non-oil taxes, borrowed abroad against future oil income, and passed some of the proceeds on cheaply to the private sector. The Second Oil Boom Fortunately for Ecuador, as oil income fell in 1977-78 non-oil ex- ports rose. The second oil boom, in contrast, was accompanied by a decline in non-oil exports-especially coffee and cacao, which fell both in price and in value. Although the rise in the oil price cush- ioned the drop in non-oil export earnings, it had less impact on the economy than the earlier price increase. Use of the Windfall The second oil boom coincided with a political transition in August 1979 from a military dictatorship to an elected government with pop- ulist leanings. Seeking to satisfy its constituency-and reassured by predictions of a sustained global energy shortage and rising real oil prices-the government used its expanding revenues to increase cur- rent expenditures further, by 2 percent of nonmining GDP. Salaries of public employees rose sharply in 1980. Although table 11-4 sug- gests only a moderate growth in the ratio of the public wage bill to GDP, other estimates suggest considerable growth of public employ- ment between 1973 and 1982, with a doubling of employment in au- tonomous entities and in the central government during this period (World Bank 1985, p. 34). Capital expenditures rose somewhat, with the greatest increases going to petroleum and agriculture and to upgrading the living and sanitary conditions in the main cities. Spending on electrification maintained its earlier level; spending on roads and transport declined. Defense spending increased in re- sponse to the tensions with Peru, which led to a brief war early in 1981. It accounted for an estimated 21 percent of oil revenues in 1979-82. Table 11-7. Ecuador: Consolidated Banking System Assets and liabilities 1972 1975 1978 1981 1982 As percentage of GDP Assets Public sector (net) 3.3 -1.9 -3.5 -6.7 -4.6 Private sectora 25.5 25.6 27.2 28.7 34.1 Liabilities Private sector 30.3 24.9 27.1 24.9 27.6 As percentage of total assets Assets Public sector (net) 9.5 -6.6 -11.4 -23.6 -14.6 Private sector 74.7 91.8 90.0 101.0 82.0 Liabilities Private sector 88.7 89.4 89.5 87.7 87.4 Note: All balances are at end of period. a. Includes other assets and float. b. Includes Ml and other liabilities. Source: World Bank data. Ecuador 187 The Public Deficit Despite higher oil prices, the public budget swung further into defi- cit, which averaged 5.3 percent of nonmining GDP in 1979-82 (see table 11-4). Almost one-third of the deterioration after 1973 was the result of rising interest costs. The expansive credit policy was contin- ued, and the current account deficit widened to 8.5 percent of nonmining GDP in 1979-82. Table 11-5 suggests that subsidies were a major factor in the pub- lic deficit. They peaked in 1980 at the equivalent of 12.4 percent of GDP, then fell abruptly from 1981 to 1982, primarily because domes- tic oil prices tripled. The most important subsidies during this pe- riod were those on the domestic consumption of energy, both oil and electricity. The subsidy on oil derivatives for the home market peaked in 1979 and 1980; that on electricity, in 1982. Higher domes- tic prices for oil derivatives would have meant a smaller increase in internal sales and higher export revenues. The subsidy to oil deriva- tives may have appeased urban consumers, but most of the bene- fits accrued to the middle and upper classes.4 The subsidy on imports included in table 11-5 was implicitly granted to importers through real appreciation of the exchange rate. Part of the cost of the appreciated exchange rate was borne by non-oil exporters and part by the government, which saw the domes- tic purchasing power of its oil income and borrowing decrease. The consumption of wheat, milk, and other selected goods was also subsi- dized, and their prices held down by controls. From tables 11-4 and 11-5 it is apparent that either of two options could have eliminated the public deficit during the period of the windfalls: pricing domestic oil at world levels or maintaining non- oil tax ratios and cutting subsidies to food and electricity. The buildup of Ecuador's foreign debt may therefore be attributed en- tirely to its domestic transfer and credit programs. However, the policy of holding down domestic oil prices and subsi- dizing other goods, together with the appreciating real exchange rate-which went from a base of 100 in 1970-72 to 121 in 1974-78, 128 in 1979-81, and 132 in 1982-83-had a substantial impact on real consumption levels. From a base of 100 in 1970-72, consumer prices declined by 13 percentage points relative to the nonmining out- put deflator (see table 11-5). This sparked a substantial rise in consumption, in real terms, compared with the relation between con- sumption and nonmining GDP that is expected for a country with Ecuador's rate of per capita income growth. The large real consump- tion effect is notable in tables 5-2 and 5-3, as is the corresponding neg- ative price effect. Table 11-8. Oil and Non-Oil Export Performance Growth rates at 1975 prices (percent) Export 1970-73 1973-79 1979-82c 1983c Bananas (tons) 3.2 0.2 -3.1 -27.9 Coffee (tons) 13.0 1.5 -20.6 83.0 Cocoa' 1.3 17.8 1.7 -58.1 Industrial goodsb 33.6 16.5 19.7 -66.4 All non-oil exports 7.6 2.0 -4.8 -31.8 Crude and fuel oil 464.8 -7.4 -1.7 38.4 Total 49.3 -3.3 -3.3 4.2 a. Includes processed cocoa. b. Excludes processed cocoa. c. Provisional. Source: World Bank data. Ecuador 189 One consequence of extending incentives to nontraditional im- ports and of reducing tariffs and export taxes was to reduce some- what the degree of anti-export bias, particularly for nontraditional exports, as shown in table 11-1. This did not offset the effect of real ex- change rate appreciation, however, and traditional exporters were still subject to considerable tax-induced disincentives. Ecuador's non- oil export volumes stagnated during 1973-82 before plunging in 1983, as shown in table 11-8, while industrial exports, though small, grew quite rapidly until 1983. In factor markets, duty-free im- port of capital goods and real appreciation of the exchange rate meant a growing disincentive to absorb labor. Debt, the Crisis, and Stabilization From August 1970 until 1982, Ecuador maintained a fixed exchange rate of 25 sucres to the dollar. A fluctuating market rate applied to perhaps one-third of transactions, mostly invisibles and certain capi- tal movements. The spread between the free rate and the fixed rate widened from less than 2 percent before 1975 to about 11 percent in 1976-80, before jumping to 20-33 percent in 1981. The existence of such a free market permits diversification of assets among curren- cies without clear indications of capital flight, so that arriving at reli- able estimates of assets held abroad by Ecuadorian residents is not easy. Marshall-Silva (1984) suggests that unofficial capital move- ments might have amounted to about $1 billion between 1973 and 1982. The Bank for International Settlements (Bis) has estimated that reporting banks' liabilities to Ecuadorian residents were $1.3 billion at the end of 1983, although some of this may have repre- sented reserves held with these banks (total reserves were $800 mil- lion).5 Gross debt less reserves at the end of 1983, at $7.3 billion, was equivalent to 3.3 years' merchandise exports, or 70 percent of GDP. If the unofficial outflow of $1 billion were deducted, net debt might have been equivalent to about 2.8 years' exports, or 60 per- cent of GDP. In 1982 the dollar value of oil exports fell by 16 percent, and the cur- rent account deficit reached 9 percent of nonmining GDP. In the middle of the year, foreign lending dried up. In May the sucre was devalued to 33 per dollar, the domestic price of gasoline was dou- bled, and measures were taken to limit imports. Then in early 1983, as the basis for an agreement with the International Monetary Fund (IMF), further steps were taken to strengthen the financial posi- tion of the public sector. Taxes and import surcharges were im- posed, duty exemptions were reduced, and the domestic price of oil derivatives was increased again. Electricity, telephone, and 190 Country Studies water rates were revised upward, and control was tightened on ex- penditures. In March 1983 the sucre was devalued again to 42 per dollar, and a crawling peg was established by which the price of the dollar went up 0.05 sucre a day. Interest rates were also raised and their structure simplified. The program of the IMF went into effect in July 1983 after a resched- uling arrangement had been reached with the foreign commercial banks. The main points of the program were: * To reduce the public sector deficit to 4.2 percent of GDP in 1983 and to limit its financing by the Central Bank * To lower inflationary pressures by imposing a limit on the expan- sion of net domestic assets of the Central Bank * To continue the flexible exchange rate policy; certain export pro- ceeds and import payments were to be transferred to the free market, and arrears in import payments were to be eliminated * To place a limit on the outstanding public sector external debt, in- cluding the debt guaranteed by the government. It is difficult to disentangle the impact of the stabilization program from that of El Nifno, the catastrophic rains and floods during the win- ter of 1982-83, especially in the coastal region. The effects of El Nifno included damage to property, livestock, and the transport net- work of about $400 million, loss of banana and cacao exports of about $115 million, and a rise in imports of about $150 million necessi- tated by reduced food output. Food shortages were superimposed on the elimination of subsidies, the adjustment of prices of public goods and services, and the devaluation of the currency. The con- sumer price index thus catapulted from an annual growth rate of 28 percent in January to 63 percent in September, while real GDP de- clined about 3 percent, real per capita consumption 7 percent, and real investment 28 percent. But tightened monetary and fiscal pol- icy and an improvement in weather conditions caused the rate of inflation to fall to less than 30 percent by mid-1984. Because of in- creased inflation, devaluation only partly offset the real apprecia- tion of previous years; the real effective exchange rate based on the 1970-72 index fell to 119 in 1984. All the targets of the stabilization program were fulfilled by Ecuador except the elimination of the im- port arrears, for which a special agreement was reached with the for- eign banks.6 External adjustment was rapid. The current account deficit plunged from $1.2 billion in 1982 to $104 million in 1983, largely through a drastic drop in imports and a large rise in oil output. Ecua- dor was the only country in this study with a positive windfall ef- fect in 1982-84 compared with 1981; see table 8-1. On the domestic Ecuador 191 front, Ecuador has no reliable statistics on employment and unem- ployment. But a sample survey of 282 companies showed a drop in employment in the manufacturing sector of 2.9 percent from 1981 to 1982 and an additional drop of 3.6 percent from mid-1982 to mid-1983. The limitations on public spending and the decline in con- struction also had a negative impact on employment. It is esti- mated that by mid-1983 the minimum real wage had dropped about 15 percent from its level at the end of 1982. And relative to 1980 the decline was 24 percent (see table 11-5). The non-oil growth rate in 1981-84 averaged -1 percent. How Has Ecuador Benefited from the Windfalls? Despite both the short political horizons of its successive govern- ments and the bureaucratic and decentralized nature of its public sector, Ecuador appears to have benefited considerably from its wind- fall gains. In certain respects, these apparently adverse features turned out to be advantages. Alone among the sample, Ecuador used its oil revenues and the borrowed funds they facilitated for tradi- tional public functions and transfers to the private sector. It is al- most certainly the only country in the sample to have ended the decade of windfalls without a large, unprofitable steel industry.7 Economic Effects GROWTH. The expansion of government spending, reinforced by vari- ous incentive policies, raised the rate of economic growth to a level far higher than anything Ecuador had experienced earlier in this cen- tury. Although growth slowed after 1978 and was negative after 1981, Ecuador's performance surpassed that of most other small and medium-size Latin American economies. Higher growth plus the effects of price controls, subsidies, and real appreciation re- sulted in a substantial rise in private consumption from projections based on past nonmining growth rates and constant relative prices. Private consumption was about 14 percent higher than such a trend would predict in 1974-78, and 22 percent higher in 1979-81; see tables 5-2 and 5-3. The danger, however, as borne out by experi- ence in 1982-84, was that accumulated debt, the lack of an adequately diversified export portfolio, and the dependence of con- sumption levels on transfers out of oil income would render such gains unsustainable. SECTORAL STRUCTURE. As shown in table 6-3, from 1972 through 1981 the sectoral structure of Ecuador's nonmining economy did 192 Country Studies not deviate too much from a "normal" evolution, although indus- trial growth was stronger and agricultural growth a little weaker than normal. Several factors account for Ecuador's fairly rapid indus- trial growth, which actually began before the oil era. Strong de- mand from the public sector and from the growing middle class was important, as was Ecuador's participation in the Andean Group. Industry enjoyed low interest rates, low energy prices, and ample labor available at reasonably low product wages partly be- cause of the subsidies on fuel, food, and utilities. But the main instru- ments for industrial development were the Promotion Laws applied by the Ministry of Industries, Commerce, and Integration. Under these laws, industries fall into various categories of priority. Incen- tives differ by category, and specific exemptions and privileges are granted by the ministry case by case. Included are exemptions from part or all of taxes-for specified periods or without any limit-and exemptions from custom duties for imports of machinery, spare parts, or raw materials not produced in Ecuador. Enterprises could also claim tax deductions for new investments in fixed capital, contri- butions to capital, or transfers of property. Industrial imports exempted from duties rose from 683 million sucres ($27 million) in 1971 to 8,412 million sucres ($336 million) in 1980. Industry enjoyed more than half the exemptions granted to the private sector, with most of the benefits going to the largest man- ufacturers. The structure of custom duties after 1971 was also an in- centive to industrial development. Whereas duties on nondurables were cut by half between 1971 and 1980, and those on inputs and capital goods by 64 percent and 55 percent respectively, duties on du- rable consumer goods were raised by more than 40 percent. Al- though average nominal protection dropped sharply from 59 to 21 percent, effective protection for much of the industry increased. The pattern of incentives was biased toward capital-intensive in- dustries. It reinforced the effect of exchange rate appreciation, which raised labor costs relative to those of machinery imports and investment credits. It has been estimated that the ratio of the cost of imported machines to that of industrial labor fell by 78 percent dur- ing 1971-78. Almost all employment generated in the industrial sec- tor during the first oil boom is believed to have been provided by incentive-induced projects. By 1978 the investment cost for each job had risen to $30,000 in medium- and large-scale industry and to $6,500 in small-scale industry. Consequently, manufacturing employ- ment grew less than 4 percent a year, although output increased 9 percent a year during 1973-82. As for the other traded sector, even in 1982 agriculture employed about half the working population and provided most of the non- Ecuador 193 oil exports. Ecuador has great agricultural potential. During the decade of the oil windfalls this sector did less well than other activi- ties, although far better than in Algeria, Nigeria, and Trinidad and Tobago. Agriculture was doubly disadvantaged-by real effective ap- preciation and by trade policy. Duty exemptions on its import re- quirements were relatively few, and many domestic inputs had to be obtained at high protected prices. Its products were subject to ex- port taxes and real appreciation of the sucre, while wheat and milk had to compete with subsidized imports. Government extension pro- grams did not always reach the farm level. Public spending on agri- culture declined markedly after 1977 as credit made available by the Banco Nacional de Fomento-which directs its resources mostly to the farm sector and which also provides technical assistance- represented a declining share of total credit, with its resources fall- ing in real terms after 1974.8 The sector's performance was not uniform: farm production lagged, but cattle raising did quite well, and forestry and fishing experienced an extraordinary expansion. It is not possible to discuss Ecuadorian agriculture without men- tioning land distribution. According to a 1968 agricultural survey, 85 percent of agricultural properties occupied only 17 percent of the farm area and averaged less than 10 hectares (about 25 acres), while 1.5 percent of properties occupied more than 47 percent of the farm area and averaged 100 hectares. Large properties had the best land, yet they often included areas that either were not ex- ploited or were exploited inefficiently. Because of the small size and bad quality of their land and their lack of capital, technical skills, and credit, smallholders were never assured of the means of subsisting. Despite heated debate over agrarian reform at the begin- ning of the 1970s, the land tenure problem remains basically as it was. The intention to use oil revenues to promote land reform came to little, other than some relatively minor colonization of new lands. Strong social and political opposition was-and is-a major obstacle. Social Effects The effects of growth and modernization on the distribution of in- come, among both individuals and regions, are more difficult to as- certain. Whereas some observers believe that oil wealth increased inequalities in income and regional distribution and thereby aggra- vated the social problems that existed before oil was discovered, oth- ers feel that the whole country benefited from the boom of the 1970s. Two general effects are fairly clear. First, there was a broadening and strengthening of the middle class (bureaucrats, technicians, 194 Country Studies and professionals), which had previously been quite small and weak. Second, the oil revenues did not manage to solve-and proba- bly only slightly mitigated-the problems of poverty and marginal- ity that affect between 30 and 40 percent of the population. Some subsidies, such as those on domestic oil products and electricity, were regressive, while those that were intended to benefit the lower income groups had little impact. Probably the greatest social effect was that, in the expanding econ- omy of the 1970s, employment opportunities kept up with the rap- idly growing labor force, which somewhat reduced poverty. One study (cited by Marshall-Silva 1984) has compared income distribu- tion for 1968 and 1975, as derived from national household surveys. Middle and lower-middle income groups improved their rel- ative positions at the expense of the upper 10 percent. At the same time, the position of the poorest 20 percent deteriorated slightly and that of the richest 0.5 percent improved slightly. According to the national accounts, between 1971 and 1982 the total real wage bill per worker grew at 3 percent a year, somewhat below real GDP per capita, which rose at 6 percent. Minimum real wages increased at 5.2 percent a year between 1970 and 1983, but the peak was reached in 1980; between that year and 1983, wages declined by 24 percent (see table 11-5). The share of compensation to workers and employees represented in the national accounts fluctuated between 31.1 and 34.8 percent during the entire period. Again, the peak came in 1980-which was also the year of highest oil prices. Social indicators show a distinct improvement, although social projects were undeniably directed less to rural areas then to the big urban centers, where the political pressures were most intense. Be- tween 1970 and 1981 the illiteracy rate fell from 31.6 percent to 14.5 percent. In rural areas the percentage decline was greater, but the base was higher; rural illiteracy was still 25.8 percent in 1980. Secon- dary school enrollment rose from 22 percent in 1970 to 56 percent in 1982, from below to above the norm for middle-income coun- tries. Despite concern over quality and wasteful duplication (by 1982 there were at least sixteen universities), this was a consider- able achievement. Those with access to drinking water increased from 28.4 percent to 43.5 percent of the population between 1970 and 1980; in rural areas the increase was from 10.3 percent in 1974 to 15.6 percent in 1980. Life expectancy, which in 1970 was 54.6 years for males and 57.6 for females, in 1980 was 61.3 and 63.6 years respectively. The number of people per doctor declined from 2,200 in 1972 to 760 in 1981, while the infant mortality index fell from 108 in 1970 to 76 in 1982. Public spending on education in rural areas increased from 23.8 Ecuador 195 percent of the total in 1975 to 37 percent in 1980. But in health it de- clined in the same period from 7 to 5.8 percent. Some idea of the changes during the oil era is given by a study com- paring social indicators for the province of Bolivar with those for the country as a whole. Bolivar is one of the poorest provinces of Ec- uador, with considerable out-migration. In 1974 it had 3.5 percent of the country's cultivated area and 2.2 percent of agricultural pro- duction; by 1980 those percentages had increased to 5.1 percent and 5.8 percent respectively. And whereas the rural population sup- plied with drinking water increased countrywide from 10.3 percent in 1974 to 15.6 percent in 1980, it increased in Bolivar from 5.7 to 16.5 percent. Expenditures on public health per capita increased in Bolivar from $4.10 to $12.50 between 1976 and 1981. To sum up, Ecuador presents a less focused, less dramatic picture than some of the other oil producers in this study. The oil windfalls relaxed resource constraints, and the government took ad- vantage of this to increase traditional public functions and to ex- pand subsidy and credit programs. Any radical departure from this pattern would have been politically difficult and bureaucratically impossible. On the positive side, Ecuador managed to avoid some serious investment errors, to raise growth, and to expand much- needed public services. However, Ecuador accumulated a substan- tial foreign debt that contributed to a foreign exchange crisis and low growth at the end of the second oil boom; moreover, the pat- tern of incentives funded by oil favored capital-intensive tech- niques. Ecuador did not take advantage of its oil wealth to establish strong export-based manufacturing industries. And since its funda- mental socioeconomic problems remain, the danger is that falling oil prices coupled with external debt will bring about a reversal of some of the gains bought with the oil windfalls. Notes 1. In 1972-78, 13 percent of oil revenues went to the armed forces, a per- centage which greatly increased in later years. 2. The largest state enterprise, CEPE increased its activities until it con- trolled about 70 percent of oil production. Hastily formed, subject to politi- cal pressures and continuous changes in its direction and management, CEPE inevitably developed inefficiencies. Between 1975 and 1982 the esti- mated real cost per barrel of the oil it produced doubled. Nevertheless, it did develop technical abilities in a new field and made Ecuadorians feel mas- ters of their main source of wealth. 3. Oil revenues accruing to the National Defense Board tripled in real terms from 1973 to 1974, then fell off slightly until 1979, when they tripled again. 196 Country Studies 4. See Marshall-Silva (1984). World Bank estimates suggested that the 16 percent of the population with the highest income received 60 percent of the oil subsidy. 5. Data do not permit the identification of the part of BIS liabilities that cor- responds to official reserve holdings. 6. The target for the public sector deficit was overfulfilled; the deficit practi- cally disappeared in 1983. 7. The caveat is due only to lack of information on the rate of return in the Algerian steel industry. 8. For a discussion of agricultural policies and prospects, see World Bank (1985), chap. 6. Chapter 12 Indonesia: Windfalls in a Poor Rural Economy with Bruce Glassburner With a per capita income of only $200 in 1974, Indonesia was the poorest country in the world to receive a substantial oil windfall. With 129 million inhabitants, it was also the most populous. Its wind- falls, at 16 percent of nonmining GDP in 1974-78 and 23 percent in 1979-81 (see tables 5-2 and 5-3), were comparable to those of the other exporters in this study. But on a per capita basis, they were far lower. Oil exports in 1979 were only $62 per person in Indone- sia, as opposed to an unweighted average of $778 in the other five comparators. Indonesia's oil industry is one of the worlds's oldest.' The quest for oil began in 1871, only twelve years after the earliest drillings in Pennsylvania. The first commercial oil was produced in 1885. Produc- tion was only 25,000 barrels a day before World War I but had grown to 160,000 barrels a day by the start of World War II. In the early 1950s Caltex began developing its giant fields in central Suma- tra (notably the Minas field), but it was not until the pioneering agree- ment on production-sharing in 1967 and the attainment of political and economic stability that Indonesia's oil boom really began. With weak administration of the non-oil tax system, oil provided the over- whelming bulk of tax revenues before 1973. Yet it accounted for only a third of exports because of the richness of the country's re- source base. The government of Indonesia therefore had less experi- ence than the governments of most other oil exporters in managing an economy strongly dependent on oil for foreign exchange. Despite this lack of experience, in a number of respects In- donesia's economic performance since 1973 has been unusually good. Especially noteworthy has been the strength of the non-oil trad- Bruce Glassburner is professor of economics emeritus at University of California, Davis. 197 198 Country Studies able sectors-agriculture and manufacturing-and of non-oil ex- ports. This is in distinct contrast to both the predictions of eco- nomic models centering on the "Dutch disease" and the record of countries such as Nigeria and Trinidad and Tobago. In Indonesia high non-oil growth rates were maintained during the two oil booms. And, although evidence is sketchy and sometimes contradic- tory, at least some benefits from spending the oil income seem to have spread to most segments of the population, both rural and urban. Excessive accumulation of debt was avoided, and after 1981 the country showed signs of being able to adjust to the slowdown in oil and other primary export markets with smaller cost than its comparators. How did Indonesia manage to avoid the serious problems that beset other oil exporters? Good luck (notably breakthroughs in rice technology and events which postponed overambitious plans to build resource-based industry) and an abundant supply of labor rela- tive to oil income played a part. But so did three distinctive fea- tures of Indonesia's economic policy: * The government accepted the need for a flexible exchange rate policy in large part because the idea that Indonesia's compara- tive advantage would be in oil for some time was never widely accepted. * Macroeconomic management was prudent and adaptive. The government was prepared to use key macro instruments to spread the absorption of windfalls over time and limit the im- pact of the oil booms on the non-oil tradable sectors. It was also prepared to undertake major structural reforms to help adjust to the oil glut. * The government was committed to a broad development strat- egy that emphasized raising rural incomes. It devised pro- grams to achieve this and succeeded in enlisting the participa- tion of millions of small farmers. As in other oil exporters, so in Indonesia institutions and policies in place before 1973 were important in determining the response to the shocks of 1974 and 1979. Perhaps they were even more impor- tant in Indonesia than elsewhere because of the exceptional continu- ity in policymaking. The same government has been in power since 1966, and key members of its economic team served from before the first oil boom through the oil slump of the mid-1980s. This un- usual stability made it possible for the government to fashion consis- tent, long-term economic policies in response to the oil shocks of the 1970s. This chapter seeks to explain Indonesia's relative success in manag- Indonesia 199 ing its oil windfalls. It begins by outlining the country's main politi- cal and economic features before 1973, first under the "Guided De- mocracy" of the Sukarno period and then under the "New Order" of the Suharto government. The following sections consider the macroeconomic responses to the first and second windfalls and sum- marize the set of measures-far more comprehensive and more speedily implemented than in the comparator countries-that were introduced to adjust to the downturn after 1981. Sectoral issues are next discussed, including the role of policy in agricultural perform- ance (particularly of rice), the effort to avoid undue dependence on oil, and the performance of non-oil exports. What is known, or at least is thought to be known, about how oil affected income distribu- tion is then summarized. An overall assessment of Indonesia's use of its oil windfalls concludes the chapter. Politics and the Economy before 1973 Indonesia proclaimed its independence from the Dutch after the end of the Japanese occupation in 1945. It achieved international rec- ognition as a sovereign state in 1949, after a four-year revolutionary struggle against the Dutch attempt to reestablish control. From Independence through the Sukarno Period, 1949-66 As a newly independent nation, Indonesia's economic prospects were bleak. Some twenty years of depression, occupation, and revo- lutionary struggle had greatly disrupted trade, damaged infrastruc- ture, and undermined the economic system. Moreover, there were few trained nationals with expertise in economics or experience in statecraft (Glassburner 1964; Kahin 1952). It is hardly surprising, therefore, that economic performance in the 1950s was very poor. The situation was made even bleaker by the unstable political cli- mate. The effort to emulate European parliamentary democracy led to a proliferation of political parties. Seven cabinets fell in the first six years of independence, and the first general election in 1955 failed to establish a national consensus. An atmosphere of deep disil- lusionment set in as accelerating inflation, capital flight, shortages of basic commodities, and increasingly open corruption grew ram- pant. In the late 1950s President Sukarno, who had previously been con- tent to play a largely symbolic role as a charismatic nation-builder, in- troduced his idea of Guided Democracy. He proposed to "bury the political parties" and return to the constitution of 1945, which cen- tered a great deal of power in the hands of the president himself. 200 Country Studies By 1959, with the support of the army, Sukarno had overcome the op- position to his plan. Presidential decrees were given the force of law, and the role of the legislature was diminished. The period 1959-65 was marked by continued economic decline. Sukarno elevated political goals over economic ones and con- centrated on building national pride and international prestige rather than productive capacity.2 Indeed, according to one ob- server, Guided Democracy and the accompanying "guided econ- omy" proved to be "in terms of disruption of the economy and capi- tal consumption, worse than most wars" (Higgins 1972). The withering of international trade reduced tax revenues. Unemploy- ment, both open and disguised, may have been as high as 25 per- cent of the labor force (Glassburner 1971, p. 431). With the end in 1957 of any effective restraint on creating money, inflation acceler- ated and was estimated at more than 600 percent in the early 1960s. Shortages of food, especially rice, contributed to political insta- bility. Foreign exchange reserves and domestic savings turned nega- tive. The extent to which output and income fell will never be known because of the demise of the statistical service. But even a small drop in GDP was critical for the large share of the population al- ready living at a subsistence level, especially in densely populated Java and Bali. The political crisis of 1965-66, precipitated by an end to the tenu- ous balance between the army and the Communist party, culmi- nated in a coup that stripped Sukarno of his power. Congress named General Suharto acting president in 1967; in 1968 the "act- ing" was dropped. Sukarno, confined to his residence, died in 1970. Stabilization and Rehabilitation under Suharto's "New Order," 1966-73 The condition of the economy when Suharto came to power could hardly have been worse. One of the new leader's first steps upon tak- ing office was to seek help in devising a strategy to meet the eco- nomic challenge. From the Faculty of Economics of the University of Indonesia (FEUI) a team of presidential economic advisors was formed which was to have an unusual degree of political influence and continuity.3 The central role of the FEUI in designing and imple- menting economic policy was strengthened when other members of the faculty were named to important government posts. In ef- fect, FEUI'S Institute for Economic and Social Research became, and has remained, the research arm of the government. This too con- tributed much to the continuity of Indonesia's economic policy. Indonesia 201 In 1966 the FEUI economists presented a major document on eco- nomic policy. Called the Tracee Baru ("New Line"), this document sig- naled a shift in priorities from international affairs to the domestic economy. The short-term objectives were stabilization and rehabilita- tion. Specific priorities included: * Strengthening agriculture, infrastructure, and industry, in that order * Reviving foreign trade and promoting exports * Reorienting import policy to facilitate domestic production and meet basic needs. Import policy was to be integrated into stabili- zation policy, with foreign exchange practices rationalized to avoid "wasteful" uses * Launching a wide-ranging program of military and civilian aus- terity, coupled with a strong effort to reduce the budget deficit rapidly by increasing internal revenues * Reforming the "confused" banking system and reestablishing central banking powers in the Bank of Indonesia. The details of the nation's economic priorities were to be specified each year through an integrated program of budgetary, fiscal, mone- tary, price, wage, and balance of payments policies. Remarkably, given the chaos in the monetary and fiscal systems, stabilization was soon accomplished. The rate of growth of the money supply was reduced substantially each year, and interest rates were raised successively until positive real rates were reached in May 1969 (Grenville 1981). Inflation plunged from more than 600 percent in 1966 to 84 percent in 1968, 9 percent in 1970, and only 4 percent in 1971. The exchange rate, persistently overvalued since 1955 (Gillis 1984, p. 11), was devalued. And, since widespread eva- sion had convinced the economic team of the futility of trying to re- strict foreign exchange transactions, controls were first modified and then abolished. The foreign exchange market has remained free to this day-an enduring legacy. The government recognized, moreover, that the large budgetary deficits had to be curbed. After 1967 the budget was, in a sense, "balanced"-although foreign aid and loans for development were counted as revenue rather than financing. This balanced budget prin- ciple was to become another enduring legacy of the Sukarno per- iod, even though it was sometimes applied quite flexibly.4 The economic turnaround in just a few years was impressive. Ex- ports recovered quickly, and net foreign reserves turned positive by 1972. Both industrial and agricultural production accelerated. The growth of trade and output helped to raise revenues; taxes col- 202 Country Studies lected by the central government climbed from a low of 4 percent of GDP in 1965 to 10 percent in 1970. ROLE OF THE MARKETPLACE. Despite the Suharto government's com- mitment to a free foreign exchange market and a balanced budget, and despite the support for the private sector expressed in the Tracee Baru, the economic regime of Indonesia was, and still is, far from laissez-faire.5 At the microeconomic level, regulation was pervasive-a legacy of Dutch colonial rule as well as of the guided economy of Sukarno. Both domestic and foreign investment was con- trolled by the planning ministry, Bappenas. Permission to invest might be denied for a variety of reasons, such as the judgment that a particular market was "saturated." To obtain clearance for new in- vestment, or even to add a new line of production, could take two years or more. Such measures insulated inefficient firms from domes- tic and foreign competition.6 In addition, thousands of ordinances af- fected every aspect of production. Many were unpublished and served only to generate opportunities for side payments to rent- seeking officials. Trade too was regulated. Effective protective rates, though not es- pecially high on average for broad sectors, were widely dispersed and strongly favored some activities at the expense of others. Stud- ies of Indonesia's system of trade intervention in the mid-1970s indi- cate that some sectors had negative value added at world prices. The effective rate for the tire and tube industry was estimated at 4,315 percent and for the motor industry 718 percent, whereas the rate for batik production, which was adversely affected, was a nega- tive 35 percent. At one point the capital intensity of protected, import-competing sectors was four times that of exporting sectors. Regulations, protection, and subsidies sometimes influenced techni- cal choices to the detriment of labor absorption.7 One objective of the plethora of regulations was to further the eco- nomic interests of the pribumi-autochthonous Indonesians-over the interests of those of Chinese extraction. The latter, who consti- tute only 3 percent of the population, had long wielded dispropor- tionate economic power. The Chinese proved adept, however, at cir- cumventing the regulations, sometimes by forming associations, called cukong relationships, with pribumi. In reality, the main effect of the regulations was to slow private investment. Whether mea- sured by the residual of total over public investment, by permit ap- provals from Bappenas, or by direct capital inflows, private invest- ment was disappointingly low in Indonesia in the 1970s and was to remain so into the 1980s. Indonesia 203 The regulation issue should not be left without some comment on economic planning. Under the Suharto regime, four five-year plans have been formulated. Repelita I, the plan for 1969-73, was es- sentially a "wish list." Nevertheless, its modest goals were gener- ally met or exceeded.8 The second plan was obsolete before it was printed because of the unanticipated oil bonanza. The third plan per- iod saw the favorable shock of a second windfall in 1979-80, and then the traumatic downward plunge in the price of oil and the vol- ume of exports in 1982-83. The fourth plan was therefore initiated in unfavorable circumstances. All these plans have been mainly exercises undertaken for public consumption. There has never been a real attempt to formulate any- thing comparable to the development plans of India and Pakistan, which set broad targets for a twenty-year period. The relatively ca- sual pattern of economic planning in Indonesia reflects the pragma- tism of the economic team and the government. Yet there is consider- able government intervention at all levels, and in that sense the economy might appropriately be called "planned" or "regulated" (McCawley 1978). NATIONAL PRIORITIES. Certain priorities identified by the Suharto re- gime well before the first oil shock were to exert a powerful influ- ence on the way in which oil revenues were used and the speed with which they were absorbed. The drive to stabilize the economy was closely linked to the need to expand and stabilize food sup- plies, particularly rice. Indonesia had become the world's largest im- porter in a thin and volatile world market. Food security was there- fore the main goal of the first development plan. A variety of programs to intensify rice cultivation were introduced, and they ex- panded rapidly. The period 1966-73 represented a learning phase in the design and replication of these rural development programs. Serious errors were made, and although agricultural output recov- ered from the low levels of the mid-1960s, growth through the early 1970s was slow. Rice output actually fell by 4 percent in 1972. This raised food prices sharply and increased dependence on im- ported rice, which totaled 3.6 million tons in 1972-74, or 8.3 per- cent of domestic production. Before the first oil shock, food self- sufficiency seemed a distant target. A second goal was to restore the conditions for sustained growth through rural reconstruction, especially by rehabilitating roads and irri- gation works neglected under Sukarno. To help achieve this goal, the central government funded labor-intensive public works proj- ects under Instruksi Presiden (INPREs) and associated programs. 204 Country Studies These programs, which were run at the county and village levels, were later to play an important role in distributing oil income widely. A third goal was the consolidation of national control. The govern- ment strengthened its administrative network, which consisted of a joint civilian-military system extending to the 383 counties (kabupaten) and towns and down to most of the 60,000 villages (desa).9 This system gave Suharto a firm power base in the bureauc- racy and the military, with important consequences for political sta- bility. It was also an essential factor in the government's success in implementing the rural programs and promoting agricultural growth. Two additional features of pre-1973 Indonesia were to influence use of the oil windfalls. First, the institutional power base of the Su- karno government was, in addition to the army, the Golongan Karya, or GOLKAR. GOLKAR is not a political party in the narrow sense but rather a coalition of diverse "functional groups"-in- cluding groups of workers, women, farmers, and youths-that dominate the parliament. GOLKAR represents a broader range of inter- ests than the usual political party. This reduced rivalry over how oil revenues were to be spent, both before and after the first shock. 10 Second, Indonesia's population was overwhelmingly rural. In the archetypal mineral economy, the rural sector is presumed to have lit- tle political power because the urban economy gets foreign ex- change from the mining sector and is therefore less dependent on do- mestic agriculture for food and export revenues. But in Indonesia the power of urban centers was limited; in 1970 only about 17 per- cent of the population lived in urban areas and most of these had close rural ties. Rural-urban migration caused by an urban bias in public spending could have swamped Indonesia's cities and towns. The rural predominance, however, helped to ensure the wide distri- bution of oil rents. Macroeconomic Responses to the Oil Windfalls, 1974-81 This section begins with the Pertamina crisis because of its impor- tant influence on the government's response to the oil windfalls. Pertamina, the state oil company, was by far the largest and most au- tonomous of the state enterprises. Like state enterprises (and espe- cially national oil companies) in some other countries, it had be- come an economic fiefdom of diversified operations.1" In 1972 a presidential decree required state enterprises to get official approval before negotiating long-term foreign loans. To get around this, Indonesia 205 Pertamina increased its short-term borrowing.12 In 1974 it failed to meet its obligations. Shock waves rebounded through the econ- omy, and the government was forced to allocate the equivalent of two-fifths of one year's oil windfall to repay Pertamina's debt. Although the economic system managed to avoid major disrup- tion, the Pertamina crisis was seen at the time as a national trag- edy. In retrospect, however, the debacle can be considered as a fortu- nate event. Because of its timing, just after the first oil price increase, it triggered an immediate and noninflationary use of oil in- come to repay foreign debt. The crisis reinforced the natural cau- tion of the government and its determination to avoid becoming too dependent on oil or excessively indebted. Thus it delayed the for- mulation of ambitious plans for capital-intensive, resource-based in- vestments until the second oil boom. This, in turn, allowed more time in Indonesia than in other oil countries for such plans to be re- viewed and scaled down with the onset of the oil glut. The First Windfall Indonesia's main response to the first oil shock was to increase spending on development. Table 12-1 records the dramatic change in the fiscal accounts. From 1973-74 to 1977-78 the government allo- cated 49 percent of the windfall in mining value added to in- vestment-virtually all of it public-as measured by the rise in development expenditures over the 1970-72 level (see table 5-2). Although "development expenditures" are commonly taken as an in- dicator of public investment in Indonesia, this category excludes investments by public enterprises not financed by the central govern- ment and includes some noninvestment expenditures."3 Another one- third of the windfall (again, using 1970-72 as the base) went to re- duce the trade and nonfactor service deficit, and the remaining 18 percent was spent on consumption (half public, half private). Inflows of project and program aid also rose, enabling total govern- ment expenditures to exceed revenues, as shown in table 12-1, with- out violating the balanced budget rule. This, together with favor- able trends in the non-oil terms of trade, enabled international reserves (less gold) to rise from $805 million in 1973 to $2.5 billion in 1978. Except that Indonesia spent a relatively high share on agricul- ture, its distribution of development spending by broad economic ac- tivity did not differ much from that of most other oil exporters in this study. During the first oil shock the agricultural sector received a substantial share of development spending (about 13 percent; see table 7-1). This estimate includes investments in irrigation works Table 12-1. Indonesia: Central Government Budget and Financing (billions of rupiah) Item 1973-74 1974-75 1975-76 1976-77 1977-78 1978-79 1979-80 1980-81 1981-82 1982-83 1983-84 1984-85 Revenue 968 1,754 2,242 2,906 3,534 4,266 6,697 10,227 12,213 12,418 14,433 16,194 Corporate tax on oil 345 973a 1,249 1,619 1,949 2,309 4,260 7,020 8,628 8,170 9,520 10,367 Current expenditures 713 1,016 1,333 1,630 2,149 2,744 4,062 5,800 6,978 6,996 8,412 10,101 Development expenditures 451 962 1,398 2,054 2,157 2,556 4,014 5,916 6,940 7,360 9,899 10,459 Balance (project plus program aid) -198 -224 -488 -788 -771 -1,033 -1,379 -1,489 -1,705 -1,938 -3,878 -4,411 Undisbursed development expenditures ... ... ... ... ... ... 833 1,908 572 -434 789 n.a. 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 Banking system claims on central governmentb n.a. n.a. -5 -339 -613 -878 1,703 -3,619 -4,179 -3,757 -5,062 -7,657 International reserves minus gold (millions of U.S. dollars)b 805 1,490 584 1,497 2,509 2,626 4,062 5,391 5,014 3,144 3,718 4,773 . . . Negligible. n.a. Not available. a. Exclusive underpayment of revenues, estimated at about Rp 340 billion, owed to the government by Pertamina. b. End of period. Source: Revenues and expenditures, Indonesia Ministry of Finance; undisbursed expenditures, World Bank; banking system claims, Bank Indone- sia; reserves, IMF. Indonesia 207 but not the heavy subsidies on fertilizer. These subsidies, shown in table 12-2, averaged 11 percent of all domestically financed develop- ment spending during the period. As usual in the sample of oil- exporting countries (see table 7-1), the largest share of development spending-in Indonesia's case, 43 percent-went for economic infra- structure. What was unique in Indonesia, however, was that such spending extended to rural as well as urban areas. Much of the spending on infrastructure was used to fund labor-intensive public works projects undertaken through INPRES and implemented at the county and village levels; see table 12-2 and discussion below. Al- though INPRES funding shifted in favor of sector-specific applica- tions in 1975-76, it changed little as a share of total development spending and thus increased rapidly after the first oil shock. Industry was not neglected in public investments. Twelve per- cent of domestically financed spending on development was allo- cated to capital participation, mostly in public industrial firms. The in- dustrial sector as a whole is estimated to have received almost 20 percent of development spending. Increased domestic spending out of the windfall contributed to a sharp appreciation of the real effective exchange rate, although ris- ing inflation was actually triggered by an adverse domestic supply shock-the poor harvest of 1972, which sent food prices skyrocket- ing. It is difficult to specify an equilibrium real exchange rate for the years just before the first oil shock because in 1967-72 the ru- piah was continuously devalued from 150 to 415 per dollar, where it remained for seven years. But relative to a 1970-72 base of 100, the average real effective exchange rate appreciation in 1974-78 was 33 percent, the highest in the sample (see table 6-1). Warr (1986) has constructed an index of the real exchange rate (the relative price of traded to nontraded goods) by comparing com- ponents of the price index that are representative of the traded and the nontraded sectors. His results point to a major shift of about 30 percent against non-oil tradables during the first oil boom. Table 12-3 reveals the dip in oil exports that occurred in 1978-79 -a decline of 13 percent in dollar terms from the preceding fiscal year. Both the government and analysts of the Indonesian econ- omy concluded that the boom was about over, and the outlook was generally thought to be cloudy."4 In November the rupiah was deval- ued from 415 to 625 per dollar. This was not done because of bal- ance of payments problems; reserves, equivalent to four months' im- ports, were considered adequate. Rather, the objective was to assist the non-oil traded sectors, many of which-such as rubber, coffee, and the emerging export sector of manufactured crafts-were rela- tively labor-intensive. Another reason for devaluation was to main- Table 12-2. Indonesia: Selected Components of Development Spending (percentage of total less project aid) Item 1973-74 1974-75 1975-76 1976-77 1977-78 1978-79 1979-80 1980-81 1981-82 1982-83 1983-84 INPRES program' General 13.6 13.2 13.9 11.2 11.8 10.6 8.1 7.5 8.5 9.8 8.9 Sectoralb 5.7 3.2 10.2 7.3 9.7 11.2 9.3 8.4 8.4 8.2 12.8 Fertilizer subsidy 9.8 29.6 14.5 8.4 2.2 5.2 4.6 6.3 8.0 7.7 7.6 Capital participation 12.1 11.9 11.7 13.0 11.8 8.2 9.4 10.6 9.1 10.8 9.8 Other 58.8 42.1 49.7 60.1 64.5 64.8 68.6 67.2 66.0 63.5 60.9 a. Labor-intensive public works projects. b. Primary schools (about 60 percent), health, roads, afforestation, markets. Source: Indonesia Ministry of Finance. Indonesia 209 tain the domestic purchasing power of falling dollar-denominated oil tax receipts; in fiscal 1977-78 these made up 65 percent of cen- tral government revenues. Plans were also initiated to contain pub- lic spending in line with the perceived tightening of resources. The expenditure-reducing measures introduced in Indonesia at the end of the first oil boom were not unusual. The governments of most oil-exporting countries began to rein in their spending as the first oil windfall subsided. But the Indonesian government was the only one to effect a major program of switching expenditures from foreign or traded goods to domestic or nontraded goods through ex- change rate adjustment."5 The Second Windfall In some ways the devaluation of November 1978 could not have been more poorly timed, coming as it did just before another oil price rise. This second shock raised Indonesia's mining sector wind- fall, as measured in table 5-3, from 15 percent of nonmining GDP in 1978 to 26 percent by 1980. And this time, unlike in the mid-1970s, a wide spectrum of opinion outside Indonesia expected oil prices to keep increasing in real terms. A less cautious government might have encouraged greater for- eign borrowing by public firms and increased its own foreign borrow- ing (as the government of Mexico did) or at least have maintained the balanced budget rule. But though the Indonesian government balanced its budget in an accounting sense, it ran substantial de facto surpluses in 1979-81 by slowing down the disbursement of funds allocated to various uses in the formal budget. Claims of the banking system on the central government fell sharply, and interna- tional reserves (less gold) soared from $2.6 billion in 1978 to more than $5 billion by 1980. For measures of the surplus, see table 12-1. Tables 5-2 and 5-3 confirm a tightened macroeconomic "absorp- tion effect" in this period. In 1979-81 (relative to the 1970-72 base pe- riod) 22 percent of the second windfall was consumed, and 36 per- cent was invested at home. But 42 percent of the windfall was used to reduce trade and nonfactor deficits, with the result that the cur- rent account swung from a deficit of 1.6 percent of nonmining GDP in 1974-78 to a surplus of 2.3 percent. Although spending did rise after the second oil price increase, the Indonesian government pur- sued a policy of restraint in that it kept absorption well below the level permitted by the availability of oil income and foreign aid. This had not been the case during the first oil boom. Holding down spending proved to be partially effective in reduc- ing the extent of real exchange rate appreciation and in maintaining Table 12-3. Indonesia: Oil and Non-Oil Export Performance Exports 1971-72 1972-73 1973-74 1974-75 1975-76 1976-77 1977-78 1978-79 1979-80 1980-81 1981-82 1982-83 1983-84 Value (millions of U.S. dollars) Net oil exports, n.a. n.a. 641 2,638 3,138 3,710 4,352 3,785 6,308 9,345 8,379 5,788 6,016 Net LNG exports, 0 0 0 0 0 0 93 225 667 1,256 1,382 1,378 1,355 Nonhydrocarbon exports 784 977 1,905 2,033 1,873 2,863 3,506 3,996 6,171 5,485 4,034 3,928 5,235 Timber 170 275 720 615 527 885 943 1,130 2,166 1,672 952 899 1,123 Rubber 215 211 483 425 381 577 608 774 1,101 1,078 770 614 962 Paim oil 45 42 89 184 142 147 202 221 257 178 79 103 83 Coffee 54 83 79 92 112 330 626 508 715 588 343 363 469 Tin 64 70 98 166 158 181 253 324 388 454 437 349 309 Miscellaneous (mainly manufactures) 56 76 77 114 144 196 245 361 636 570 613 730 1,167 Ratio of nonhydrocarbon to oil exports (percent) n.a. n.a. 297 77 60 77 81 105 98 60 50 68 89 Volume growth rates (percent) 1971-72 to 1980-81 1980-81 to 1983-84 Timber 3.1 -22.3 Rubber 1.8 4.2 Palm oil 6.6 -13.4 Coffee 13.9 4.9 Tin 4.6 -7.2 Miscellaneousb 15.8 27.9 Total nonhydrocarbon exportsb 11.1 -0.9 n.a. Not available. a. Exports are net of imports of goods and services. b. Dollar values are deflated by unit values of manufactured exports from developed to developing countries to derive volume index. Source: Bank Indonesia. Indonesia 211 competitiveness after the devaluation. Although price increases total- ing 42 percent in 1979 and 1980 eroded a substantial part of the real devaluation, the nominal rate was allowed to slide against the dol- lar, and the real effective exchange rate was maintained in 1979-81 at an average index of 104. This was considerably closer to the 1970-72 parity (100) than were the effective rates of exchange of the other five oil exporters in this study, which averaged 140.16 The profile of Indonesia's real effective exchange rate over the two oil shocks differs, therefore, from that of most of its compara- tors, as figure 12-1 makes clear. In Indonesia appreciation was ini- tially sharp but was substantially corrected rather early on. In con- trast, most of the other oil exporters continued to let their exchange rates appreciate-particularly Nigeria, which pegged its nominal ex- change rate throughout the period despite high domestic inflation and tightened import controls as export revenue fell off. The policies pursued by the Indonesian government during the sec- ond oil shock had a satisfactory effect on growth. Helped by good harvests, the economy rebounded after a slump in 1979 to grow at nearly 10 percent in 1980. In 1967-72 the nonmining economy had grown at 8.5 percent as it recovered from the stabilization period; it came close to maintaining this high rate during the windfall peri- ods, with a growth rate for 1972-81 of 8.2 percent. This was the sec- Figure 12-1. Indonesia and Comparators: Real Effective Exchange Rates 300 - Nigeria . 250 - g 200 - Z 4N 150 ; > S ; =Average 100 Indonesia 0: 50 50 1970 1972 1974 1976 1978 1980 1982 1984 Source: World Bank data. 212 Country Studies ond highest rate in the sample after Algeria. And because of its substantially lower population growth, Indonesia led all the com- parators in the growth rate of nonmining GDP per capita. Adjustment in the Downturn, 1982-85 Indonesia's net oil exports fell from $9.3 billion in 1980-81 to $8.3 bil- lion in 1981-82, and the next year they plunged to $5.7 billion."7 To make matters worse, the non-oil terms of trade deteriorated as impor- tant markets for primary products collapsed. The current account re- versed abruptly; a surplus of $2.8 billion in 1980 had become a defi- cit of $5.3 billion, or more than 8 percent of GDP, by 1982. At the same time, growth of total GDP slowed to 2.2 percent, while growth of nonmining GDP at 3.6 percent was far below the 8 percent level of the previous decade. What happened in Indonesia after the second plunge in oil prices again demonstrates the willingness of the government to act deci- sively and prudently to curb any emerging macroeconomic imbal- ance. The turnaround in the oil market was speedily diagnosed as not being merely short-term, and adjustments were made while re- serves were still high. In response to the adverse developments, in late 1982 and 1983 the Indonesian government introduced the most wide-ranging set of adjustment measures of any country in this study, including: * Exchange rate adjustment. After the rupiah had drifted down- ward against the dollar, from 625 in November 1978 to 703 in March 1983, it was devalued to 970 per dollar. * Tight fiscal policy. This involved, among other steps, further in- creases in the domestic prices of oil derivatives. These prices had traditionally been held far below world levels. The eco- nomic cost of the oil subsidy averaged some 5 percent of GDP dur- ing the two oil booms, while the budgetary subsidy peaked at 2.4 percent of GDP in 1981.18 * After the price increases, most subsidies were restricted to kero- sene and diesel, and in 1985 the budgetary fuel subsidy was only 0.6 percent of GDP. Moreover, with the recovery of rupiah revenues in 1983 (partly as a result of devaluation) the govern- ment switched from a temporary phase of overspending in 1982-83 back to underspending in keeping with its balanced bud- get rule (see table 12-1). * A drastic rephasing of the public investment program, particu- larly in heavy industry. By the end of 1981 industrial projects in the pipeline implied foreign exchange expenditures of $20 bil- Indonesia 213 lion. Postponements and cancellations saved an estimated half of this figure, with negligible long-run implications for em- ployment since the cancelled projects were highly capital- intensive. 19 * Between Repelita III (April 1979-March 1984) and the first two years of Repelita IV (1984-85 and 1985-86) the share of develop- ment spending going to industry and mining fell by 8 percent- age points, while the share to economic infrastructure rose by 2.5 points and the share to the social sectors rose by more than 6 points. In addition to reducing expenditures, this postpone- ment of industrial programs reallocated expenditures from in- dustry to infrastructure and social sectors and was an impor- tant expenditure-switching policy. Import coefficients were far lower, and employment coefficients higher, for the expanding than for the contracting sectors, and this cushioned the effect of falling demand. Estimates of import and employment coeffi- cients in table 12-4 take into account direct employment and im- port demands as well as demands from domestic sectors supply- ing intermediate products. The table indicates the potentially large impact on demand that accompanies such changes in the composition of public investment. * Financial reform and liberalization. The main objective of finan- cial reform was to raise interest rates on deposits and so pre- vent the capital flight that might otherwise be provoked when people realized that devaluation would follow the deterioration Table 12-4. Indonesia: Import and Employment Coefficients Import Employment Sector requirements' requirementsb Productive sectors 0.71 0.14 Agriculture 0.45 0.22 Industry 0.87 0.09 Mines 0.91 0.06 Economic infrastructure 0.69 0.23 Irrigation 0.50 0.31 Energy 0.79 0.16 Social sectors 0.44 0.45 All sectors 0.57 0.33 Note: Aggregations are based on the plan allocations for Repelita IV. a. Coefficients show the import requirements (in rupiah) per rupiah of develop- ment expenditures. b. Coefficients show the labor requirements (in person-years) per million rupiah of development expenditures (in 1980 prices). Source: World Bank data. 214 Country Studies in the oil market. Real interest rates on term loans from state banks rose from -6 percent to 10 percent; real rates on work- ing capital rose to 15-20 percent. These measures proved to be effective in stemming, and indeed reversing, the tendency to- ward capital flight that arose in 1982.20 Comprehensive tax reform and simplification. Changes ef- fected in 1983 were designed to broaden the non-oil tax base, in- crease its yield, and so compensate for an anticipated decline in oil taxes. Although it is too early for a full assessment of these adjustment measures, they appear to have been successful. In 1982-83 the real ef- fective exchange rate of 108.5 was near its previous level. But in 1984 it depreciated to 91.5 (see table 6-1), and growth, assisted by good harvests, recovered to 5 percent. Exports of manufactured goods expanded rapidly, and starting in 1979 exports of liquefied nat- ural gas (LNG) began to be significant. Even so, Indonesia did not escape all the adjustment costs associ- ated with the slowdown of demand and the supply shocks of more costly energy and imported intermediates. Capacity use fell sharply in the cement and other investment goods sectors, and, although data are scant, the rate of urban unemployment appears to have in- creased. Avoiding the Dutch Disease Notwithstanding its large oil revenues, between 1972 and 1981 Indo- nesia was actually able to achieve a combined share of agriculture and manufacturing in its nonmining GDP that was larger than the norm for the evolution of this share with rising per capita income (see table 6-3). And this was done even though Indonesia began the decade with a "normal" sectoral distribution of economic activ- ity, unlike Algeria and Venezuela, the other two countries that also raised their non-oil traded share. Despite a low level of private capital formation during this per- iod, the growth of manufacturing as a share of GDP in Indonesia out- stripped "normal" growth by a considerable margin, possibly be- cause of its continued low wages relative to other Southeast Asian countries. In contrast, agriculture lost ground quite rapidly. Its share of nonmining GDP fell by an average of 1.5 percent a year, not much less than the drop in the share of Nigerian agriculture of 1.9 percent a year. But Indonesia was a far poorer country than Ni- geria, with more rapidly growing income per capita. In Indonesia, therefore, the shift of economic activity out of agriculture was rela- Indonesia 215 tively "normal." In Nigeria agriculture's share fell at three times the "normal" rate. It is Indonesia's unusual success in agriculture and rural development, particularly as measured by the standards typical for mineral exporters, that this section seeks to explain. Agricultural Performance During the two oil booms and through 1984, Indonesia's agricul- tural performance was on the whole impressive, especially for food crops. In particular, the output of rice grew an average of 5.7 per- cent a year in 1972-84, accelerating from 3.7 percent in 1972-77 to 7.2 percent in 1977-82. Table 12-5 gives the figures for rice output. Even though the output of other crops rose less dramatically, Indonesia's performance in increasing food output per capita far outpaced that of the other oil exporters as well as that of develop- ing countries and of the world as a whole. Table 6-4 presents compar- ative data based on a set of uniform international price weights. By 1982-83 the index for Indonesia's food output per capita had reached 133 percent of its 1969-72 average; the corresponding ratio for the world was only 106 percent, and the unweighted average ratio for the rest of the sample was 91 percent. Sectoral Policies As noted above, improving the productivity of the rural economy in general and of rice cultivation in particular had been a major goal of the government since the mid-1960s. Much of its effort was di- rected at the most populated areas, Java and Bali. There, centuries of "agricultural involution" had resulted in tiny, highly fragmented landholdings that were cultivated with exceptionally labor-intensive techniques.21 In the early years of Suharto's rule, much was made of the need for Indonesia to become self-sufficient in rice. That the government was preoccupied with this goal is not surprising given the political impact of the price and availability of rice in Indonesia. Indeed, the price of rice was the focal point of anti-Sukarno sentiment in the street riots of 1965, and the availability of rice at a stable price has been the prime measure of the Suharto government's success in man- aging the economy. With the international market thin and volatile, increased domestic output seemed the only solution. BIMAS AND INMAS. The cornerstone of the government's effort to raise rice production was a succession of programs known as BIMAs-an acronym for bimbingan massal, or mass guidance. This ap- Table 12-5. Indonesia: Rice Output and Intensification Average Area Area under intensification program (thousands of hectares) Total Ratio of Rice output yield harvested Of Of Percentage fertilizer use rice price (thousands (tons per (thousands which which of area (thousands of to fertilizer Year of tons) hectare) of hectares) BIMAS INSUS INMAS INSUS Total covered metric tons) price 1969 12,249 2.28 8,014 1,309 - 821 - 2,130 27 119.8 0.58 1970 13,140 2.41 8,135 1,248 - 845 - 2,093 26 192.4 0.66 1971 13,724 2.46 8,324 1,396 - 1,393 - 2,789 34 197.0 0.68 1972 13,183 2.46 7,987 1,203 - 1,966 - 3,169 40 244.4 0.79 1973 14,607 2.59 8,403 1,832 - 2,156 - 3,988 47 308.1 0.79 1974 15,276 2.64 8,509 2,676 - 1,048 - 3,724 44 379.2 0.76 1975 15,185 2.63 8,495 2,683 - 1,957 - 3,640 43 393.3 1.05 1976 15,845 2,78 8,368 2,424 - 1,189 - 3,613 43 422.6 0.98 1977 15,876 2.79 8,360 2,059 - 2,181 - 4,240 51 415.6 0.86 1978 17,525 2.89 8,929 1,960 - 2,888 - 4,848 54 556.8 1.01 1979 17,872 2.97 8,850 1,571 - 3,452 - 5,023 57 617.6 1.07 1980 20,163 3.29 9,005 1,374 420 4,142 640 5,516 61 698.7 1.21 1981 22,286 3.49 9,382 1,384 587 4,802 1,119 6,186 66 1012.2 1.50 1982 22,837 3.74 8,988 1,296 832 5,047 2,113 6,343 71 1240.6 1.71 1983 24,006 3.85 9,162 1,401 955 5,222 2,484 6,623 72 1364.7 1.93 1984 25,825 3.94 9,636 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. - Not applicable. n.a. Not available. Sources: Output, yield, and area from the Indonesia Bureau of Statistics Intensification Supplement to the President's Report to Participants, August 1984. Fertilizer use from Timmer (1985), table 1. Indonesia 217 proach emerged more or less spontaneously in the early 1960s to spread the use of new high-yielding and pest-resistant varieties of rice and to intensify production. Five elements were essential to the BIMAS programs: improving water control, expanding the use of fertilizer, pesticides, and improved seeds, introducing better meth- ods of cultivation and weed control, and strengthening village coop- eratives (Mears and Moeljono 1981). The programs also granted credit on favorable terms. BIMAS programs have undergone many changes in response to ex- perience and circumstances since first being introduced, and they have expanded greatly. The early programs were marred by mis- takes and entailed a high degree of coercion; the later ones were more successful.2 In 1967 BIMAS was supplemented by INMAS- intensifikasi massal, or mass intensification-which provided credit for fertilizers, pesticides, and high-quality seeds to past participants in BIMAS. Another effort, introduced in 1968, relied heavily on Euro- pean firms to supply chemical inputs on a very large scale. But the cost of these programs was higher than expected, and repayment by farmers worse than expected. A revised program, Bimas Nasional, was launched in 1970; it re- lied more on the extension service, the national agricultural supply firm, the People's Bank, and the Logistics Board (Bulog).3 Coercion was steadily reduced in favor of incentives, and village organiza- tion was emphasized as the way to increase local participation. Unfor- tunately, the expansion of the village units was much too rapid and many of them did a poor job. By 1970 more than half of the 3,800 units that had been established were reported to be insolvent (Arndt 1977a). Nevertheless, the effort to utilize local units re- mained a cornerstone of Indonesia's agricultural policy. Despite their shortcomings, these programs laid the foundation for agricultural growth. At the start of the 1970s BIMAS and INMAS COV- ered 30 percent of the harvested rice area. This proportion in- creased steadily until coverage reached 72 percent in 1983. Table 12-5 records the development and shows how yields have risen, espe- cially since 1977. In that year new techniques and varieties began to defeat the wereng pest, which had been largely responsible for the slow growth in output during the mid-1970s.24 INSUS. In addition to the new wereng-resistant varieties, a coordi- nated planting program, called INSUS (intesifikasi khusus, or special in- tensification), was adopted. This program, initiated during the wet season of 1979-80, encourages farmers to schedule planting so that large areas of rice ripen at the same time. It had been observed that 218 Country Studies wereng damage was greater when the pest could feed on areas that ripened in succession. PRICING OF RICE AND FERTILIZER. With the advent of Bimas Nasional in 1970 came acceptance of the principle that farmers' price incentives must be strong enough to induce them to adopt new technologies and to raise production. As Timmer (1975) has pointed out, concern with incentives to farmers was a turnabout from well-established tradition in Indonesian agricultural policy. Until the 1970s the main concern had been with the protection of the interests of consumers-and, in particular, with maintaining the real value of salaries of public servants and the military.25 A modest step in the direction of incentive pricing was under- taken in 1968 with the adoption of Rumus Tani (farmer's formula), which sought to maintain a 1:1 ratio between the price of a kilo- gram of rice and a kilogram of chemical fertilizer. The general instabil- ity and geographic variation of prices made it very difficult to imple- ment this formula, however. In 1970 a policy of establishing floor and ceiling prices for rice was formally adopted. The Logistics Board, or Bulog, was to maintain a floor price that would guarantee reasonable returns to farmers in the BIMAS program. Bulog was also responsible for building a sufficient buffer stock from domestic pur- chases and imports to be able to prevent retail prices from rising above the farm price plus reasonable margins. The objective was to try to protect the interests of both producers and consumers by pro- viding adequate incentives to sustain output growth while shield- ing consumers from price hikes in bad crop years. By reducing sea- sonal price variations, the policy was also expected to make rice production and trade less risky. In the good crop years of 1970 and 1971, it was possible to meet the established price targets. But Bulog did not build up stocks large enough to meet the challenge of the next poor crop year, 1972. Rice prices doubled that year. As a result, a major program was undertaken to build storage facilities, and there was an intensi- fied effort to diversify procurement geographically and stimulate more local participation. By 1978 Bulog's storage capacity was sufficient to stock 1 million tons of rice in much improved facilities, with better geographical spread and rapidly improving transport-all of which contributed to Bulog's market leverage. By 1982 storage capacity had reached 2.5 million tons. These huge facilities were filled primarily by very large acquisitions of rice from international markets, both on commer- cial terms and through concessional food aid.26 It was, of course, the oil bonanza that provided the foreign exchange for both the stor- Indonesia 219 age facilities and the stock buildup. And yet until the production spurt of 1978-81, it appeared that Indonesia's long-sought self- sufficiency in rice would remain a distant target despite quite credit- able rates of growth. One objective of the intensification programs was to encourage fer- tilizer use; Indonesia's huge gas reserves and investments in ferti- lizer plants were thus a key component of the strategy. Fertilizer use rose from 120,000 metric tons in 1969 to 1,364,700 metric tons in 1982; this represents an annual growth rate of 16 percent. The price of fertilizer was held well below the floor price of gabah (rough rice). Table 12-5 indicates that the relative price of fertilizer to rice was reduced to only 30 percent of its initial value from 1969 to 1983. During 1981-83 the fertilizer subsidy averaged 7.8 percent of domestically funded development expenditures, or about 1 per- cent of GDP. The impact of the fertilizer subsidy on output has been assessed econometrically by Timmer (1985). He concludes that a substantial part of increased yield and output can be explained by the decrease in relative fertilizer prices. Moreover, the effect of low fertilizer prices is found to offset, by a large margin, the disincentive to pro- duction caused by holding the domestic gabah price at some 15 per- cent below its world level.27 The implication of these findings is that the fertilizer subsidy has a high benefit-cost ratio, particularly if it does not result in Indonesia's becoming a net rice exporter. Fertilizer pricing must be considered when assessing Bulog's rice pricing policy. Although at times Bulog has been criticized for at- tempting to suppress price increases to favor consumers, it is not ap- parent that Bulog has systematically depressed rice prices as op- posed to trying to stabilize them. Given the goal of self-sufficiency and the thinness of the international rice market, a cheap food pol- icy could not be sustained in the long term, at least not without offset- ting input subsidies. Subsidizing fertilizer use while containing rice prices was an attempt to raise productivity in a key non-oil tradable sector while distributing some of the windfall gains to rice consum- ers and avoiding a greater dependence on imports. The relative contribution of these measures remains uncertain. But it appears that the combination of the new planting patterns, the wereng-resistant varieties, improvement of infrastructure in the rural areas, sustained heavy investments in irrigation, continued sub- sidization of inputs (fertilizer in particular), price policy, and sev- eral years of good weather have combined to turn the rice situation around in a dramatic way. Clark (1972) showed that virtually half the caloric intake of Indone- sians came from food sources other than rice. So-called palawija 220 Country Studies crops (non-rice staples) were also promoted, but at a lower level, so that rice tonnage rose from 46 percent of total food crop production in 1971 to 51 percent by 1981. Yet palawija tonnage still outpaced pop- ulation, growing at 2.8 percent a year. INPRES. The rural public works programs-an important part of the strategy-also originated before the first oil shock. In 1973 INPRES ab- sorbed 19.3 percent of domestically funded development spending. INPRES funded labor-intensive, productive investments at the coun- ty level, paying wages similar to those available in agriculture. Similar programs at the village level provided only materials, which were used by unpaid volunteer labor.28 Although no accurate statis- tics are available, these programs appear to have created substan- tial employment. It is estimated that they generated a quarter of a million person-years of work in 1970 and about 1.5 million person-years in 1982-equivalent to 2.7 percent of the labor force. The programs also seem to have contributed positively to rural recon- struction and social developments; unlike public works programs in some other oil-exporting countries, they were not simply transfer pro- grams. The largest of the sectoral INPRES programs was the build- ing of primary schools. Between 1970 and 1978 the primary enroll- ment rate rose from 77 to 100 percent, closing the gap between Indonesia and the average middle-income country. Growth of Non-Oil Exports Indonesia turned in a strong non-oil export performance during the period of the oil booms as well as in the downturn after 1981. This performance, unusual for an oil exporter, indicates that between 1971 and 1981 non-oil exports grew across a wide front (see table 12-3). It also shows steady export diversification, with growth of "miscellaneous" (mainly manufactured) products at 16 percent dur- ing the boom periods and 28 percent during the slump. Several factors, such as the reconstruction and improvement of channels of trade and transport, contributed to export growth, but macroeconomic policy played an important role too. Export growth was especially rapid in the aftermath of the two large devaluations of 1978 and 1983 (see table 12-3). In the year after the first devalua- tion, manufactured exports (39 six-digit categories of Brussels Tariff Nomenclature) rose by 260 percent, with many firms entering ex- port markets for the first time. The increase in miscellaneous ex- ports after the second devaluation was similarly dramatic. During the 1970s agricultural and other primary exports bene- fited from buoyant markets. For this reason, the growth rate of the Indonesia 221 value of nonhydrocarbon exports, deflated by world manufacturers' prices, rose more than 11 percent, faster than most major non-oil ex- port volumes. Rising world prices also compensated exporting sectors to some extent for exchange appreciation. But after 1981 commodity markets weakened, and the country experienced a deterioration in many non-oil as well as oil sectors, as shown in table 12-3. By this time manufactured exports were so large and re- sponsive that they almost offset the decline in primary commodi- ties. Thus they were a key component of adjustment to the world re- cession. Another component in Indonesia's strategy of diversifying out of oil as a source of foreign exchange was to move into resource-based industrial (RBI) investments. The production of liquefied natural gas (LNG) began to make a positive contribution to the balance of pay- ments and to fiscal revenues, especially after 1979; in 1983 it repre- sented the equivalent of 23 percent of oil exports in net terms. Indo- nesia relied less heavily on RBI development than many other exporters, but investments in this area were still considerable. They included $4.3 billion in metal processing (steel and aluminum), $4.2 billion in refining, and $1 billion in methanol and ammonia plants. As in other oil-exporting countries, the RBI program was to suf- fer from adverse world market trends and domestic markets too small to absorb full capacity output. The steel industry was pro- tected through import restrictions; in the mid-1980s it was esti- mated to be selling at 30-40 percent above the cost of imported steel-a heavy burden for industrial users. Manufacturing in gen- eral was oriented toward the domestic market, frequently with high levels of protection, and this may have implications for the post- boom adjustment period. Yet Auty (1986a) considers the Indone- sian RBI experience to have been less risky and more successful than that of most other oil producers. It was relatively diversified, better phased over time, and less central to the country's overall de- velopment strategy. The financing of Indonesian RBI projects was also more cautious than in other countries. For example, foreign eq- uity in the Asahan aluminum smelter was 80 percent, compared with less than 20 percent in Venezuelan VENALUM plants. This paid off in terms of start-up and operating efficiency (Auty 1986a). Indonesia's LNG projects were also cautiously financed, and they were protected to some extent from downturns in the world market by a combination of foreign equity and long-term contracts with Japan. Unlike Algeria's larger LNG industry, which was operating at only 40 percent of capacity in the mid-1980s, Indonesia's plants oper- ated at full capacity. LNG exports rose from $162 million in 1977-78 to $2.40 billion by 1983-84, overtaking those of Algeria, which in- 222 Country Studies creased from $250 million to only $2.14 billion over the same per- iod. Distribution of Benefits Wage trends and the distribution of income in Indonesia are contro- versial issues. Some scholars, usually those with a more microeco- nomic focus, have tended to the view that income inequality and per- haps absolute poverty as well have worsened during the windfall periods. Others, often those with a more aggregated perspective, have pointed to rapid per capita growth in all categories of consump- tion and to increasing off-farm rural employment opportunities and have argued that gains have been widely distributed. To delve exten- sively into such complex issues is beyond the scope of this study, but some tentative observations are offered.29 Indonesia's great abundance of labor in comparison with other oil exporters held both advantages and dangers. Insofar as higher spending of oil income can shift labor from extremely low-pro- ductivity to higher-productivity work or even increase the participa- tion rate, it obviously has greater potential to stimulate growth than if the expansion of some (typically nontraded) sectors causes productive factors to move out of other (traded) sectors and so con- tracts their output. Unlike in Nigeria, for example, the increased demand for workers that arose from rapidly growing public pro- grams could be met without seriously affecting the non-oil tradable sectors, in particular agriculture. There is some evidence that the In- donesian labor supply grew more rapidly in the 1970s than in the pre- vious decade; the World Bank's (1980b) report on employment sug- gests that the growth rate of the labor force jumped from 1.5 percent in 1961-71 to 4.7 percent in 1971-76, although this has been disputed by other studies such as Jones (1981). But labor demand for seasonal crops, which in 1975 provided the equivalent of 21.6 million full-time jobs out of a total of 46.3 million in the economy (Downey and Keuning 1983) was particularly suscep- tible to changes in production techniques and social arrange- ments.0 When traditional methods are as labor-intensive as in Indo- nesia, any change is more likely than not to displace workers. Such changes might well be hastened by oil-funded modernization and commercialization. Earlier research on agricultural wages in Java found no upward trend in real wages during most of the 1970s despite rising agricul- tural productivity.3" Later years saw the introduction of wereng- resistant varieties of rice, the accelerated growth of rice output, and an increase in rural public spending. Some researchers, such as Col- Indonesia 223 lier and others (1982), began to report sizable increases in real agricul- tural wages and nonfarm rural employment. This trend was con- firmed by a panel study of households in six villages in West Java between 1977 and 1983. The results suggest a 15 percent rise in the av- erage real wage for hoeing, planting, and unskilled and skilled con- struction work in these villages (see table 6-2). The panel study also indicates that an important source of in- creases in household income was the diversification of economic ac- tivity: households shifted from low-productivity cottage industries, such as the weaving of bamboo mats,32 to higher-paying work in con- struction and transport. The study suggests further that the size dis- tribution of rural income has remained roughly unchanged. In both 1976 and 1983 the poorest 40 percent received 14 percent of the total income, while the richest 20 percent received 53 percent in 1976 and 52 percent in 1983. Other studies suggest a considerable increase in urban real wages during the windfall years.33 As for income distribution, the sit- uation may have been different in urban and rural areas. Certain studies (for example, Booth and Sundrum 1981) suggest a tendency for income differentials to widen, but data are far from adequate to paint a definitive picture. Conclusion It is sometimes suggested that Indonesia's success in managing its oil windfalls reflected the good health of its economy before 1973. But this is misleading. Other oil exporters that experienced a less fa- vorable outcome after 1973 had also been in good economic shape be- fore then; the non-oil growth of the sample countries in 1967-73, an unweighted average of 7.3 percent, was extremely high relative to other developing countries. A more accurate statement, there- fore, might be that Indonesia's good performance during the oil booms reflected the institutions developed earlier to nurse the econ- omy back to health, the approach to policy set in the Suharto govern- ment's formative years, and the unusual degree of continuity. All the main institutional components of the post-shock period-BIMAS, INPRES, the balanced budget policy, a free foreign exchange market- predated the oil shocks. So did the cautious approach to macroeco- nomic policy, which was reinforced by the Pertamina disaster, and the willingness to adjust the exchange rate. These positive policy and institutional features combined with Indonesia's abundance of labor and non-oil resources as well as with technical changes in the food crop sector, which oil revenues helped to accelerate by financing the diffusion of improved tech- 224 Country Studies niques. Together, they resulted in a distinctive pattern of growth for the former Dutch colony. To a surprising extent, the poorest coun- try in the sample has managed to avoid the most serious problems of the Dutch disease, and its economic performance stands out as being relatively successful. Notes 1. The industry is described in Hunter (1971). 2. The conflict over Irian Jaya is the prime example. The Dutch contin- ued to control this territory after 1949, and Sukarno resorted to diplomatic and military maneuvers to oust them. In 1963 Irian Jaya was finally trans- ferred to Indonesia. 3. Of the five academics on the team, three were fairly recent Ph.D.'s from the University of California, Berkeley, who came to be called "the Berkeley Mafia." In due course, all five became cabinet members. Pro- fessor Widjojo Nitisastro, formerly dean of the FEUI and subsequently minis- ter for economics as well as director of the Planning Commission, became- and remained in 1986-the leading figure in the nation's economic affairs. 4. Less successful were attempts to put the state enterprises on an unsubsi- dized basis (Glassburner 1984, p. 13). 5. The role of the government, as defined in the Tracee Baru, is not to con- trol economic activity "to the greatest extent possible" but to lead the nongovernment sector and to mobilize and develop the potential of the peo- ple. The tone of the document strongly favors allowing the private sector to expand much more than was allowed under Sukarno. At the same time, the document declares that the country's resources are to be controlled for the benefit of all the people. "Free-flight liberalism" (a phrase always ren- dered in English) is viewed as giving rise to the exploitation of humanity. Thus, although the economic advisers clearly wanted to revive Indonesia's market system, they did not advocate anything close to laissez-faire. 6. Pitt and Lee (1981) found that technical inefficiency in the Indonesian weaving industry in the early 1970s was as high as 38 percent. 7. Hill (1980) analyzes the impact of policy on the choice of technology in the weaving industry. Finding a potentially wide range of choices, he con- cludes that policies such as subsidized interest rates may have been impor- tant in encouraging the use of capital-intensive methods. 8. The notable exception was self-sufficiency in rice; the actual growth in rice output was 4.5 percent a year for 1968-73, only half the targeted amount. 9. Tinker and Walker (1973) describe the Indonesian administrative sys- tem. The development of administrative capability at local levels strength- ened the ability to implement and decentralize rural projects. 10. In the span of interests represented, GOLKAR resembled the PRI in Mex- ico. But there was no succession of presidents in Indonesia as there was in Mexico. Indonesia 225 11. Even before the first oil shock, Pertamina had taken advantage of its po- sition as a collector of oil taxes to build a conglomerate comprising several thousand filling stations, a fleet of tanks and trucks, several office build- ings, a data processing center, a tourism complex, a fertilizer factory, an air- line, and several rice plantations. 12. A similar situation in Venezuela led public enterprises to incur a heavy short-term debt burden; see chapter 15. 13. Total public investment is actually larger than development spend- ing. Recently available information on public investment during 1974-83 (but not earlier) suggests that the public sector was responsible for the bulk of capital formation recorded in the Indonesian national accounts. 14. Arndt (1978) provides a good indication of expectations at this time. 15. Algeria did effect a switch in spending after 1978 by redirecting the pub- lic investment program toward sectors making more intensive use of domes- tic factors of production (chapter 10). Indonesia was to follow this strategy after the second oil shock, as described below. 16. With regard to the sectoral impact of exchange rate shifts, Warr's (1986) index of the Indonesian real exchange rate suggests a smaller shift in relative prices between traded and nontraded sectors than the move- ments in the real effective exchange rate would be expected to produce. This is mainly because of the failure of the so-called law of one price in the tradables sectors, where domestic prices fell below those abroad after the de- valuation. This does not invalidate the resource allocation arguments for ex- change rate adjustment, but it does suggest that the differential effect on rela- tive prices will be felt more in specific trade-related activities than across broad sectors. 17. Net oil exports are gross exports less imports of goods and services at- tributed to the oil sector. This measure is conventionally used to express the contribution of the sector to foreign exchange earnings in Indonesia. 18. The budgetary cost of the oil subsidy was far less since it reflected mainly the explicit subsidization of refined imports from Singapore. 19. It was estimated that the capital cost per job of the more capital inten- sive investments would have been about $1 million. 20. With an open capital market the impact on private investment of such measures is probably less than that of controlling interest rates. The con- trasting Venezuelan case (see chapter 15) suggests that interest rate ceilings may have further depressed investment by encouraging domestic savers to transfer funds abroad. 21. Geertz (1966) analyzes the process of agricultural involution, which in- volves progressive fragmentation of landholdings and the raising of land pro- ductivity through the intensive application of labor. 22. Heavy pressure to participate was widely acknowledged both before and after 1970; see Glassburner (1978b). 23. Originally a procurement agency for the military, Bulog was later prom- inent in the implementation of rice pricing policies, as discussed below. 24. The wereng is a hopper that carries a virus. It was controlled tolerably well in the 1960s and early 1970s by widespread use of insecticides, but 226 Country Studies new breeds of the insect that could survive even heavy spraying evolved in the mid-1970s. Many areas of Java repeatedly suffered heavy crop losses. 25. Government employees' real incomes were hedged by partial pay- ments in kind, that is, by a rice allocation. 26. Rice imports totaled 9.1 million tons in 1972-78 and 6.0 million tons in 1978-79 to 1983-84. Bulog's "end stock" for crop year 1983-84 was 1.3 mil- lion tons. 27. The relation between world and domestic rice prices is estimated by Timmer (1985) for 1982 only. 28. Note the contrast between these Bantuan Desa programs and the DEWD programs of Trinidad and Tobago (chapter 14). DEWD paid effective wages several times higher than those available in agriculture, even though agricultural workers were substantially unionized. 29. For some contrasting views on these issues, see Booth and Sundrum (1981), Palmer (1978), and World Bank (1980b). 30. It has been estimated that harvest labor was reduced 54 percent after a switch from the traditional bawon system to the tebasan system, which in- volves the sale of standing crops to middlemen not bound by local custom to employ other villagers (World Bank 1984). 31. These studies are discussed in World Bank (1984). 32. Hart and Sisler (1978) found such activities to pay even less than those associated with food crops. 33. Warr (1986) analyzes wage levels in seven branches of manufacturing during 1975-82. He finds two periods of rapid increase: 1975-77 and 1979-82. Overall, there is a clear upward trend. Chapter 13 Nigeria: From Windfall Gains to Welfare Losses? with Henry Bienen Nigeria's oil revenues and its size-it is by far the most populous Sub- Saharan country-make it distinctive in Africa; at the same time, it shares many of the features common to African countries. Oil was discovered in Nigeria in 1956 and began to be exported in 1958. By the mid-1960s the economic and political significance of oil had be- come obvious, and in 1969 the state assumed a major role in the pe- troleum sector. By the early 1970s, with oil constituting more than half of Nigeria's commodity exports, the sectoral structure and trade patterns of the economy were beginning to show the effects of the oil income. But Nigeria was still far from being dependent on oil. Agriculture remained the dominant sector, accounting for some 40 percent of non-oil GDP and 42 percent of commodity ex- ports, and employing roughly 70 percent of the work force. This state of affairs, unfortunately, was to be short-lived. By the end of the decade, Nigeria was widely cited as suffering acutely from the "oil syndrome." Non-oil exports had collapsed, the rela- tive size of the agricultural sector was declining sharply, and the real effective exchange rate was appreciating rapidly. The mineral sec- tor's share of GDP went from about 1 percent in 1960 to more than 25 percent in the late 1970s. Petroleum exports rose to well over 90 percent of total exports in 1979, and they accounted for four-fifths of total government revenue. This extraordinary rise in oil income not only provided the govern- ment with the financial resources to undertake new programs; it also permeated the very institutions that formulated policy and pro- foundly influenced the decisionmaking processes. By the mid-1980s Nigeria was the only one of the six countries in this study in which Henry S. Bienen is James S. McDonnell Distinguished Professor and direc- tor of the Center of International Studies at Princeton University. 227 228 Country Studies the overall standard of living had probably sunk below its pre-shock level. Whether the oil windfalls have conferred a lasting value on Ni- gerian society is, therefore, a particularly pertinent question. This analysis of the impact of oil on Nigeria's economy stresses three themes. First, the specific uses to which oil income was put were conditioned by the country's distinctive social and political makeup. More so than in the other sample countries, the central gov- ernment presided over a federation with strong ethnic and regional loyalties, which gave rise to great rivalry over the delivery of public services. Class and sectoral interests, in contrast, were relatively weak. The state was under intense pressure to deliver public ser- vices rapidly to large numbers of people. This pressure, which per- sisted through several discontinuous changes of government, was re- sponsible for a heavy concentration of public investment in the nontraded sectors. That strategy, in turn, goes far to explain Nigeria's difficulty in accommodating to a weakening oil market after 1981. Second, there was little pressure to use oil revenues to improve the productivity of traditional agriculture. In this respect Nigeria dif- fered greatly from Indonesia, the other oil exporter with a large agri- cultural sector. In Indonesia agriculture prospered; in Nigeria it lan- guished. Although technical factors were somewhat more favorable in Indonesia than in Nigeria, another important factor was the politi- cal and institutional differences between the two countries, which manifested themselves in different priorities and capabilities. Third, Nigeria's macroeconomic adjustment to decelerating oil rev- enues differed considerably from that of most of the other export- ers. Indonesia took decisive action to devalue its currency in 1978 and again after 1981, so it can be regarded as a speedy exchange rate adjuster. Ecuador, Venezuela, and Trinidad and Tobago were more reluctant adjusters, but each effected at least one large devalua- tion some time after 1981 as the foreign balance came under pres- sure. The role of the exchange rate is not apparent in Algeria's tightly controlled economy-and its inflation rate was, in any event, relatively low. In Nigeria the government made no effective move, as inflation soared, to realign the real effective exchange rate by adjusting the nominal rate. Instead, it turned more and more to various types of quantitative import restrictions. In effect, this strat- egy redistributed part of the oil revenue from the government to fa- vored importers and to certain other agents and activities, and so se- verely aggravated the fiscal problem caused by falling oil revenues. Nigeria's prices moved more and more out of line with those of its trading partners. The result was a vicious cycle of rising distor- tions, declining efficiency, falling non-oil output, fiscal deficits, infla- Nigeria 229 tion, and disruptive cuts in public spending. Of all the countries in the sample, Nigeria experienced the most severe economic contrac- tion after 1981. This chapter investigates how political and institutional factors in Nigeria shaped decisions about using the oil revenues, and how these decisions affected the non-oil economy. The first section lays out the economic and political background of the pre-shock period. The second summarizes the three persistent concerns of the Niger- ian leadership both before and after the first oil price rise: growth and modernization, indigenization, and income distribution. The third section discusses the time profile and the composition of pub- lic spending after 1973 and the disastrous interaction of fiscal, trade, and exchange policies. Then the impact of the spending on the non-oil economy is assessed. In an attempt to account for the poli- cies pursued, the pressures on the government and on the adminis- trative structure of the country are analyzed. Conclusions are drawn in the final section. Nigeria before the Oil Boom In the five years before the first oil boom, Nigeria's economy grew rapidly as it recovered from two military coups in 1966 and a destruc- tive civil war in 1967-70, the causes of which were not unrelated to oil. Recovery was helped by rising oil income. In 1967-72 nonmining GDP grew at 9.2 percent, well above the population growth rate of about 2.6 percent. Economic Background Despite its oil, Nigeria in the early 1970s was still a poor country with an economy resembling those of many other African nations. It was heavily dependent on agriculture as a source of food, employ- ment, investable surplus, and exports. In 1960, 71 percent of the labor force was engaged in agriculture, and this share seems to have fallen only slightly by 1970. Production was mostly rainfed and dominated by smallholders using simple techniques, unim- proved varieties, and few modern inputs. Although arable land far exceeded cultivated land, extensive farming practices required that large areas lie fallow to regenerate soil fertility. It is unclear to what extent the land frontier could have been pushed back without adopt- ing improved techniques or risking long-term ecological damage. A high correlation has long existed in Nigeria between ethnic com- munity and export crop: cocoa is overwhelmingly grown by Yoruba in the west, groundnuts by Hausa in the north, palm products by Table 13-1. Nigeria: Export-Crop Producers' Potential Earnings Withheld by Marketing Board, 1947-62 Total withholdings Percentage of Crop Crop years (millions of £) potential earnings Cocoa 1947-48 to 1953-54 61.8 39.4 1954-55 to 1961-62 53.9 26.1 1947-48 to 1961-62 115.7 31.9 Groundnuts 1947-48 to 1953-54 39.5 40.0 1954-55 to 1960-61 22.3 14.9 1947-48 to 1960-61 61.9 24.9 Palm kernels 1947-54 37.0 29.2 1955-61 31.3 27.1 1947-61 68.3 28.1 Palm oil 1947-54 13.9 17.0 1955-61 18.5 25.6 1947-61 32.4 21.0 Cotton 1949-50 to 1953-54 9.7 42.3 1954-55 to 1960-61 4.8 11.2 1949-50 to 1960-61 14.5 22.1 Note: The mechanisms for withholding potential earnings include trading surpluses, export duties, and produce sales tax. Source: Helleiner (1966) as cited in Bienen (1985), table 15. Nigeria 231 Ibo in the east. This division along ethnic and regional lines-as well as geographic dispersion, repression, and a lack of resources- made it difficult for Nigeria's farmers to form cohesive producer orga- nizations.' As a result, small farmers were poorly organized com- pared with industrial, bureaucratic, and military personnel. It is no surprise, therefore, that urban groups exerted more political pres- sure on policymakers than did rural groups. The trade regime had long discriminated against agriculture. From the late 1940s until the 1960s, producers of agricultural ex- ports were subject to heavy and discriminatory taxation through a marketing board system, introduced by the colonial government, that withheld on average about one-fourth of potential earnings. Table 13-1 tells the story. For some, notably cocoa producers, this tax- ation was a shift from the practice of the early and middle years of the colonial occupation, when they had retained most of their earn- ings. For others, it was a continuation of previous policies of colonial and local administrators. Although one intention of the marketing boards had been to stabilize producers' earnings, an important goal was to create capital assets that could be used to foster development. Indeed, the large reserves that accumulated were the direct or indi- rect source of about 95 percent of the funds made available to devel- opment boards and corporations in 1946-62 (Rimmer 1981).2 As the country moved first toward regional self-rule in the 1950s and then to independence in 1960, the marketing boards and their as- sociated institutions became more and more politicized. Rural inter- ests suffered doubly. Producers, besides being heavily taxed, re- ceived below-market prices for their crops and were sometimes cheated by the licensed buying agents. Yet little of the revenue thus generated was plowed back into efforts to increase agricultural productivity. Instead, the funds were channeled to development cor- porations, whose activities were centered on the cities, and to other urban clients of political leaders (Bienen 1985, pp. 18, 20). This flow of surplus from rural to urban areas effected by the marketing boards increased after independence and continued into the early 1970s. Although the pattern varied somewhat among regions,3 the net re- sult of this policy was a relative neglect of agriculture, especially of smallholder production. Despite this, between 1960 and 1967 export crops grew in volume at an annual compound rate of 4-6 percent (Watts 1983, p. 469), while staple food production kept pace with pop- ulation growth and food imports grew only moderately. As a re- sult, little concern was voiced at the time either about food self- sufficiency or about the traditional export sector (Bienen 1985, p. 29). It was to manufacturing rather than to agriculture that politicians 232 Country Studies and planners looked for economic autonomy in the early 1960s. These first years of independence mark the beginning of a shift in the structure of manufacturing from processing traditional primary products for export to processing imported materials for the domes- tic market. Import-substituting industries grew in part because of the greater protection accorded the domestic market. Protective mea- sures included both quantitative restrictions and tariffs. Import du- ties represented almost 40 percent of all revenue collected by the fed- eral government in 1969-70, or 32 percent of import value. Because domestic industrial activities with low value added were favored, ef- fective protection rates were doubtless high in many cases, with wide sectoral variation. The impact of government spending on the economy was quite small until oil became important. In 1965-66, for example, total reve- nue collected by the federal and state governments amounted to only 12.0 percent of GDP, and their combined capital spending was only 5.5 percent of GDP. Almost half of this amount was allocated to economic services (with 12 percent of this category going to agri- culture). Only 14.7 percent of total capital spending, or 0.8 percent of GDP, was allocated to social and community services. Education, in particular, suffered-a situation made worse by the civil war. Pri- mary enrollment rates remained at a virtual standstill: 36 percent in 1960 and 37 percent in 1970 (in contrast to norms of 80 percent for low-income and 88 percent for middle-income economies). And the secondary enrollment rate in 1970 was a tiny 4 percent compared with 25 percent in middle-income countries (table 7-3). The fiscal situation began to change in the late 1960s as more and more oil was produced. Oil taxes collected by the government rose from 2.8 percent of nonmining GDP in 1970 to 8.5 percent in 1972. This permitted a rise in both current and capital spending as well as a slight fiscal surplus. Then, with the huge and unanticipated surge in oil revenues in 1974, federal government outlays skyrock- eted in an unprecedented manner. Political Background As a part of its colonial legacy, Nigeria-like many other African countries-found itself with a set of institutions that allowed a high degree of regional autonomy. Indeed, the political history of Ni- geria from 1960 until the present can be viewed as a continuous strug- gle between regional autonomy and central control. In the 1950s, under a constitution granting wide powers to the three regions of the country-Northern, Eastern, and Western- elections were held for regional parliaments as well as for the Fed- Nigeria 233 eral House of Representatives.4 A single party came to dominate each region. Ibo, Hausa-Fulani, and Yoruba leaders ruled in the east, north, and west, respectively, although only the Western Re- gion was ethnically homogeneous. Indeed, about half of Nigeria's population was neither Hausa-Fulani, Ibo, nor Yoruba. Even after in- dependence, political power and economic strategies continued to be defined in regional and ethnic rather than class terms. This ethnic- regional competition led to a continuous search for a constitutional formula to hold together the Nigerian Federation, and to an on- going battle over the regional allocation of public revenue, most of which derived from the sale of export crops. In none of the dominating parties were farming interests well repre- sented. Nor was there any active commitment-even among popu- list parties and leaders-to increasing equity. Pressures from poor farmers were blunted by ethnic cleavages and by the widespread de- sire to leave the agricultural sector, while pressures from the urban elites grew sharper. Not surprisingly, then, Nigeria's rulers had lit- tle inclination to invest in agriculture. Both their views on develop- ment and their political priorities led them to look elsewhere. CIVIL WAR. By 1965 ethnic tensions in the country had intensified greatly. Electoral politics broke down because of ethnic-regional con- flicts over appointments to civil service and military positions; over investment in infrastructure, plants, and social services in different parts of the country; over how revenue was allocated by the federal government to the regions; and over the interjection of northern power into the Western Region. Violence increased, and peaceful elections could not be carried out everywhere.' Nigeria's civilian government was toppled by a military coup in January 1966. A second coup followed six months later. During the period between the two coups, ethnic violence broke out against Ibos in northern and western cities. When negotiations with the country's military leaders collapsed, the Ibo area constituted itself as the Republic of Biafra and tried to secede. After a devastating civil war, during which an estimated 2 million people died of fam- ine, disease, and battle wounds, Biafra surrendered in January 1970. Oil had played a role in precipitating the civil war. Before 1960 the governments of the Northern and Western regions had favored applying the principle of derivation to the allocation of federal revenue-arguing, in effect, that revenue collected by the federal government should be allocated to the regions in accordance with how much revenue each region had generated. At that time, ground- nuts and cocoa, which were grown in the Northern and Western re- 234 Country Studies gions, were the main sources of export revenue. But when petro- leum was found in the Eastern Region and in the delta areas of what became the Midwestern Region, the Northern and Western re- gions reversed their former position, while the Eastern Region be- came the new champion of the principle of derivation.6 Although the changes in revenue allocation that became effective in 1966 returned a larger share of revenue to the regions than be- fore, they worked to the disadvantage of the Eastern Region. A rela- tively poor part of the country, it hoped that secession would make it financially secure. But Biafra lost control of its ports and oil facili- ties early in the war and was deprived of funds, blockaded, and even- tually starved into submission. OIL, WAR, AND CENTRALIZATION. Early in the civil war the military government replaced the four regions with twelve states. The pur- pose was to gain the support of minorities living in northern and southeastern Nigeria, who wanted to be free from the domination of large ethnic groups. Yet this step also set the stage for centraliz- ing the government since it diluted regional power. Rising oil revenues likewise contributed to this centralizing trend. The output of oil declined in 1967 and 1968 because of the civil war, but by 1969 it exceeded the 1966 figure of 152.4 million barrels a year, reaching 197.2 barrels in 1969 and 750.4 million barrels in 1973. The surge in revenues, which would increase far more dramati- cally in 1974, meant that as the power of the central government grew it had an easy way to keep the federal coffers filled. The civil war itself hastened centralization. The military had be- come extremely conscious during the war of Nigeria's fragility and its vulnerability to outside powers. This stood in stark contrast to the prevalent mood in the late 1950s, when Nigeria's leaders fully ex- pected their country to become the "giant of Africa" by virtue of its size and potential wealth. The sense of vulnerability in the late 1960s was heightened by the fact that Nigeria's oil was being devel- oped by foreigners. Concerned about sovereignty as well as na- tional security, the military officers were committed to a strong cen- tral government. In this they were joined by civil servants, many of whom came from minority ethnic communities and had risen in influ- ence when Ibos fled to Biafra. By the time Biafra was defeated, the Nigerian army had ex- panded from a small force of around 10,000 in 1966 to more than 200,000 in 1970. The ethnic composition of the army had altered too: Yoruba soldiers increased from relatively few to about 20 per- cent, whereas Ibo technical officers were gone. A significant num- Nigeria 235 ber of high-level officers, including General Yakubu Gowon, the mili- tary head of state, came from smaller ethnic groups in the Middle Belt between the north and south. Although military spending de- clined somewhat from its wartime peak, this large army was to be an important claimant for resources, and its leaders were to make crit- ical decisions about the use of the new oil revenues. Policy Goals since Independence Nigeria has been one of the most politically violent countries in Af- rica. The First Republic came to an end in 1966. Between that year and 1979 the country was led by four military leaders. General A. Ironsi inherited command after the first coup of 1966, but was assassi- nated in July 1966. General Y. Gowon served as head of state until July 1975, when he was replaced by General Murtala Mohammed. When General Mohammed was assassinated in an abortive coup in February 1976, his second in command, General 0. Obasanjo, suc- ceeded him. He remained in power until a new civilian regime, headed by an elected president, Shehu Shagari, took office in late 1979. But the Second Republic lasted only four years. Shagari was ousted by a coup in December 1983. And yet another military re- gime took over in August 1985. Despite these frequent changes in government, certain broad eco- nomic and political concerns have persisted from before the first oil price increase until the end of the second boom. The country's lead- ers have, of course, embraced many different objectives at various times, but three goals stand out as consistently important: * To grow and modernize * To indigenize the economy, in both the oil and non-oil sectors * To improve income distribution. But in Nigeria's case this goal must be interpreted as distribution among ethnic groups and re- gions rather than among socioeconomic classes or occupations. In the late 1970s the military government committed itself to two more concrete goals which were to have a considerable fiscal im- pact. The first was to establish a steel industry, which can be seen as a specific manifestation of the drive for modernization and indus- trial self-reliance. To appease regional rivalries, steel-producing ca- pacity was divided among several locations. The second objective, which again related to the regional factor, was to build a new capi- tal at Abuja. This was nearer to the center of the country than Lagos, away from the coast and the domination of Yorubaland, and closer to the politically powerful north. 236 Country Studies Growth The first major statement on economic policy by a Nigerian military regime was made just before the second coup of July 1966. In June of that year Guidelines for the Second National Development Plan was pub- lished. The main objectives set forth were a high overall rate of eco- nomic growth, rapid industrialization, increased production of food for domestic consumption without relaxing efforts in the export sec- tor, and a drastic reduction in unemployment. The Guidelines made clear that growth was the priority, should the objectives prove incon- sistent. This emphasis on growth as the prime goal was to continue through the 1970s. The Second National Development Plan for 1970-74, having been delayed by the civil war, was prepared in a context of economic opti- mism. The plan called for avoiding uncertainty and instability, for building national unity, and for exploiting the economic base pro- vided by oil. The goals articulated included "a just and egalitarian so- ciety" and "a land of bright and full opportunities for all citizens"- but the real priority was ambitious economic growth. Nationalism was also stressed, although in practice there were many compro- mises with foreign interests in this period. When the Third National Development Plan for 1975-80 was pre- pared, the spirit at the top of the Nigerian government was expan- sive. The Gowon regime felt that it had the resources both to increase welfare and to lay the groundwork for a modern and powerful Nigeria. The third plan noted that oil was a wasting asset and that the productive capacity of the non-oil economy had to be de- veloped. At the same time, a more even distribution of income and control of inflation were high priorities. It is not clear from the plan document whether growth was intended to take precedence over all other objectives. By the time the fourth plan began, in January 1981, concern over the wasting of the smallholder agricultural sector had mounted. This sector was therefore singled out as vital for growth. Indigenization Perhaps the most striking goal of the Nigerian military regimes of the early and mid-1970s was the indigenization of the economy. Heated debate took place both within and outside government about the meaning of such a policy and how it was to be carried out. The indigenization decrees put forward by the Gowon and Obasanjo regimes in 1972 and 1977 reflected pressures that pre- dated the first military coup and even Nigeria's independence.' Ni- Nigeria 237 gerians had long been conscious of the role of foreigners in their econ- omy. The Tax Relief Act of 1958 required transnational corporations seeking "pioneer" status to increase Nigerian ownership and person- nel and to use Nigerian materials. And immigration laws estab- lished quotas for expatriates to increase the opportunities for Nigeri- ans to participate in management. In the petroleum sector specifically, Nigeria's actions paralleled those of Venezuela and the Arab oil exporters. In 1959, through the Petroleum Profits Tax Ordinance, the government instituted a fifty- fifty profit-sharing arrangement with foreign concerns. In 1966 it reduced by one-half the rate at which companies were allowed to depreciate capitalized investment. In 1967 it applied OPEC terms in Ni- geria. In 1969 a Petroleum Decree established the state's option to own shares in commercial oil operations. In 1971 the Nigerian Na- tional Oil Company (NNOC) was created, and between 1972 and 1975, in accordance with other indigenization measures, the state in- creased its share of ownership and altered the terms of compensa- tion and buy-back arrangements in its favor.8 NNOC'S participation in foreign companies up to 1974 is estimated to have cost Nigeria $1.3 billion, but receipts to the government from NNOC'S shares, along with revenues from royalties, were more than ten times that amount (von Lazar and Duerstein 1976, p. 11). In terms of direct reve- nue use, indigenization was not costly.9 Both the first and second plans stressed increased national con- trol over the economy. But it was the Nigerian Enterprise Promo- tion Decree of 1972 that set out the broad objectives and the means to increase that control. During the late 1960s and early 1970s, the na- tions' planners were committed to acquiring and controlling produc- tive assets, especially strategic national resources. This was to be done by the government acting either alone or in partnership with private concerns. Although the planners, along with others, were well aware of the inefficiency and corruption of Nigerian statutory corporations and state-owned companies, they nevertheless pushed for control of the oil and other mining sectors. From the viewpoint of military and civilian policymakers as well, indigenization meant increasing Nigeria's control over its own re- sources. Since it was also seen as a way to expand public involve- ment, the intention was not merely to hand over foreign-owned eq- uity to the private sector. Yet policymakers were not completely unresponsive to pressures from Nigerian businessmen either (Hoogvelt 1979; Biersteker 1980). Moreover, although some military officers and civil servants wanted to move ahead and nationalize large sectors of the economy, most of them-including those in the highest positions-were committed to maintaining the private sec- 238 Country Studies tor as an important actor. Businessmen, for their part, were much more interested in participation and in windfall profits than in con- trol over foreign enterprises. As for indigenization in the rural sector, the colonial companies had pulled out of agricultural production and trade in the 1950s and 1960s. In the 1970s some Nigerian businessmen did push into ag- riculture, but not at all to the extent that they moved into services and manufacturing. Thus the indigenization programs did not stem the decline of the agricultural export sector. Overall, indigenization has proceeded rather slowly in Nigeria, and the government has adjusted its programs from time to time. Recognizing the country's acute shortage of indigenous technical and managerial skills, those in both the public and private sectors have acted cautiously so as not to jeopardize the prime goal of eco- nomic growth. Distribution In Nigeria formulas for allocating revenues between the federal and state governments have an influence on public spending patterns and their distributional effects. This is because various states as well as the federal government have different responsibilities and pri- orities, which often reflect their different constituencies. Under the Nigerian federal system, some responsibilities are vested in the cen- tral government only, some in the states only, and some in both. Even before the constitutional changes made at the end of 1979, the relationship of the states to the federal government had been shifting. Specifically, as the states' power to raise revenue through marketing boards, taxation, and other means was reduced, the funds allocated to them by the federal government became an ever larger share of their spending. This centralizing trend has been rein- forced by oil. Almost all observers of Nigeria agree that interregional income dif- ferences have been politically more important than interpersonal ones, and that income differences per se have been less of a prob- lem than the more visible uneven distribution of schools, roads, health services, and the like. In fact, regional disparities in per ca- pita income have been less than most other regional disparities that have been measured. Although the Second National Development Plan increased the emphasis on the welfare of the common man and on a more equita- ble distribution of income, no direct assault was launched on the size distribution of income or on sectoral or rural-urban differences. Indeed, such an assault would have been difficult to mount given Nigeria 239 the conditions of Nigerian economic and political life. There was no landed aristocracy with lands that could be expropriated, and the government, as noted above, was unwilling to risk a sharp decline in growth by taking over private businesses. Nigeria's leaders and planners in the 1970s therefore preferred to guide investment through the use of incentives, to create more entrepreneurial know- how, and to rely on indigenization to expand opportunities. The pri- vate sector, for its part, was expected to develop better habits of sav- ing and thereby raise economic levels in the country (Onage 1975, p. 61). At the same time, through public sector programs-including elec- trification, water supply, health services, cooperatives, community development in the rural areas, and housing in the urban areas- the government would seek to raise the standard of living of the poor. Universal free primary education was regarded as the main hope for bringing about equal opportunities because of the widely perceived link between education and access to the modern econ- omy. There is no doubt that equity became a more salient issue. Gen- eral Mohammed came to power in 1975, but the Third National Devel- opment Plan, launched in April 1975, carried on the distributional strategy of the second plan. Again, no direct assault on inequity was proposed. Oil revenues were to be used to create an infrastruc- ture for self-sustaining growth and to expand greatly the delivery of welfare benefits. True, the third plan emphasized not only an in- crease in income but a more even distribution, largely to be ef- fected by delivery of public services. Yet the Guidelines to the Third Plan, published in September 1973, had argued that basic amenities should not be provided to communities of less than 20,000. This pol- icy is not easy to reconcile with the desire to use social services and public utilities to relieve poverty. There could be no rapid and fron- tal assault on the size distribution of income or on rural-urban imbal- ances with such a criterion for service delivery. Use of the Oil Windfalls, 1974-84 The price that Nigeria received for its oil increased steadily between 1971 and 1973; in nominal dollar terms the rise was about 22 per- cent. Then, in nine months from late 1973 to mid-1974, the govern- ment's oil revenues almost quintupled because of much higher prices, greater production, and an increase in its share of the oil reve- nues through greater public ownership and higher taxes and royal- ties. Thus the Nigerian government found itself in mid-1974 with much more revenue than it had anticipated. Its budget surplus for Table 13-2. Nigeria: Federal Budget, 1970-84 (percentage of nonmining GDP) Item 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982d 1983d 1984d Total revenuea 10.7 16.0 16.2 24.0 40.6 34.3 33.9 33.0 24.9 42.2 47.2 37.2 29.7 25.7 28.2 Non-oil revenue 7.9 8.2 7.7 7.5 7.7 7.2 7.5 8.6 8.1 7.7 8.7 8.7 7.0 7.0 5.4 Current expenditureb 8.5 8.6 9.6 10.9 11.6 15.9 13.2 14.7 13.3 18.1 21.1 21.1 | 42.7 39.6 37.0 Capital expenditureb 3.6 4.9 5.7 7.5 17.8 27.5 29.5 28.0 20.3 26.7 26.5 29.6 Balancec -1.5 2.6 0.8 5.6 11.1 -9.1 -8.9 -9.7 -8.8 -2.6 -0.5 -13.5 -13.0 -13.9 -8.8 Claims on government by the banking system n.a. n.a. n.a. n.a. -13.5 n.a. 2.7 n.a. 12.3 n.a. 11.0 18.0 23.4 34.2 n.a. Loss of real income due to exchange appreciation n.a. n.a. n.a. n.a. n.a. 3.3 5.0 3.5 4.6 7.2 10.1 5.2 3.9 3.9 8.8 n.a. Not available. Note: To compute these estimates, the cumulative surplus of 1973-74 was equally distributed over spending out of oil income in 1975-78. For 1979-82 more detailed estimates of the domestic deficit (the excess of spending on domestic goods over domestic revenues) are available. This pres- ent method yields a slightly lower estimate of the domestic deficit; the estimates of real income losses due to real effective appreciation are correspond- ingly more conservative. a. Includes all federally collected revenues. b. Apportions statutory and nonstatutory transfers to states to current and capital expenditure according to the limited data available. c. Federal government only. d. Provisional. Source: World Bank data. Nigeria 241 1974-75 was very large, and the balance of payments surplus shot up about twenty times from the relatively small sum of 153 million naira in 1973. Nigeria's leaders suddenly found themselves with vast new resources that could be used to accomplish their eco- nomic, political, and social objectives. The Expenditure Response When oil prices surged the Nigerian government had just pub- lished the Guidelines to the Third Plan, which put forward a rela- tively expansionary scenario for public spending. The response of the government to the unexpected windfall, and to the expectations of further increases in oil income, is shown in table 13-2. This table in- cludes all federally collected revenues, an increasing share of which was transferred to the states after 1974. The transferred funds have been apportioned to recurrent and investment spending based on the limited information available on the states' spending patterns. After an initial lag in 1973 and 1974, when large surpluses were ac- cumulated abroad, public capital spending accelerated rapidly from only 3.6 percent of nonmining GDP in 1970 to 29.5 percent by 1976. This acceleration was so strong that it alone absorbed more than the entire increase in oil income between 1970 and 1976. The excess, to- gether with a considerable rise in current spending, resulted in a sub- stantial deficit that was financed by drawing down reserves accumu- lated in 1973 and 1974 and by expanding the money supply. WAGE INCREASES. Current spending rose sharply in 1975, largely be- cause the Gowon government moved to implement the recommenda- tions of the Public Service Review Commission, better known as the Udoji Commission. In response to intense pressure from public employees, the average pay for civil servants was doubled (retroac- tively to April 1974), with rises of up to 130 percent for employees in the lower grades. Other organized workers clamored to keep up, and this fueled inflation. It has been estimated that the Udoji wage ac- cords increased the public sector wage bill by 50-60 percent. Commentators have interpreted the Udoji accords as an attempt by General Gowon to stay in power (despite an earlier promise to withdraw in 1975) by giving a pivotal sector of the population a siz- able share of the oil wealth (Oyediran and Ajibola 1976; Joseph 1978). For just as the Udoji awards were announced, General Gowon also stated that the country was not ready to return to civil- ian rule.10 In addition to wanting to build support among its immediate constituency-officials in the civilian public sector and the military- 242 Country Studies the government was in a hurry to use its oil revenues. It wanted to avoid being accused of not spending when Nigeria was so poor. In- creasing wages was one way to spend money rapidly. Investing heav- ily in education, transportation, and construction was another. Offi- cials involved in formulating plans for the public sector have related how orders came down to multiply spending on education and transportation. By 1975-76 new programs were under way, and expenditures increased sixfold over the level of only three years before (Schatz 1981). Not surprisingly, committing funds hast- ily to new domestic investment led to waste and corruption. As inflation accelerated after 1974 (raising domestic costs relative to oil prices) and the oil market weakened, the ratio of total reve- nue to nonmining GDP fell from 41 to 25 percent. Although spend- ing pressures were moderated somewhat, the budget swung from a surplus of 11 percent of nonmining GDP in 1974 to a deficit of 9 per- cent in 1975-78. THE DOWNTURN. The second oil price rise, despite its considerable ef- fect, was not sufficient to bring the budget back into surplus even in the peak year of 1980. Expecting a quick reversal of the deteriorat- ing oil market, the government failed to adjust adequately to falling revenues, although it cut many programs sharply. The result was the emergence of large fiscal deficits, which averaged 12.3 percent of nonmining GDP in 1981-84 and generated further inflationary pres- sure. As shown in tables 5-7 and 5-8, Nigeria's foreign borrowing prior to 1981 had been relatively cautious compared with that of the other sample countries. At the end of 1980, gross external debt was only $9.0 billion (of which $3.5 billion was short-term) while re- serves were $10.6 billion. But cumulative current account deficits of $17.2 billion in 1981-83 were mirrored in a rise in debt to $19.7 bil- lion (including trade arrears of about $5 billion) and a fall in re- serves to $1.3 billion. The result was a sharp deterioration of Nigeria's credibility on international capital markets. Discussions with the IMF for an Extended Fund Facility and with the World Bank for a structural adjustment loan stalled over issues of devalua- tion, trade liberalization, and domestic energy pricing.' Patterns of Public Spending It is not possible to present a comprehensive picture of the composi- tion of Nigerian public spending funded by oil revenues because in- formation is lacking on the expenditure patterns of the states. Statu- tory and nonstatutory allocations to the states represented about Nigeria 243 one-fourth of all federal outlays after the first oil shock.12 Although the states employ about twice as many civil servants as the federal government, state administration is often very weak. Detailed data are scarce, but there is good reason to suspect that while the states have been at least as responsive to popular wishes as the federal gov- ernment in their spending patterns, they have been less financially responsible in their overall budget and revenue processes. A reasona- bly good indication of the use of oil revenues, however, is provided by the available information on capital spending by the federal gov- ernment (see table 13-3). For a comparison with the other countries in this study, see table 7-1. More than in the other five countries studied, public investment in Nigeria emphasized the nontraded sectors-even though the share of general services (including defense) fell after 1976. Table 13-3 indicates a large rise in the allocation to transport and communi- cations. Much of this went into a program of constructing trunk roads. Since rural feeder roads were neglected, however, only farm- ers in the most densely populated areas were within even five miles of a road that could accommodate motor vehicles by 1979. Ac- cess costs thus remained very high. This policy, like the one of pro- viding amenities only to large communities, suggests the extent to which public spending on infrastructure was directed to the urban rather than the rural population. The contrast to Indonesia is nota- ble. Capital spending on education also increased sharply, from 3.9 per- cent of the total in 1973-74 to 18.2 percent in 1975-76. The massive universal primary education system was launched in 1975-77. It was estimated that the program would cost the government 17 mil- lion naira by 1982, but by 1978 it was already costing four times that amount. The demand was far larger than projected; primary en- rollment shot up from 3.5 million in 1970 to 9.5 million in 1976 and to 13 million by 1980. The drop in federal allocations to education after 1975-76 in table 13-3 is misleading. As a result of constitu- tional changes, authority for primary education was shifted from fed- eral to local government. Nonstatutory transfers to states (997 mil- lion naira by 1979, or 3 percent of non-oil GDP) went mainly to fund educational programs. Accelerated investment in building roads as well as schools caused an exceptional boom in the construction sec- tor. In the non-oil tradable sectors-manufacturing and agriculture- it is obvious from table 13-3 how investment was concentrated in manufacturing. The share of industrial investment (mostly in heavy manufacturing) rose especially rapidly after the decision to build up a steel industry in the late 1970s. Agriculture's share, quite minor be- Table 13-3. Nigeria: Capital Expenditures of the Federal Government, 1973-80 (percentage of total) Sector 1973-74 1974-75 1975-76 1976-77 1977-78 1978-79 1979-80 General services 26.0 16.6 26.3 19.9 19.0 17.8 17.4 Defense 11.9 10.4 9.4 8.5 9.5 11.3 10.1 Other 14.1 6.2 16.9 11.4 9.5 6.5 7.2 Community services 1.3 13.2 5.4 9.5 9.5 10.0 5.2 Housing 1.3 10.5 4.0 4.9 3.4 2.8 1.1 Water resources . . . 2.6 0.7 4.2 4.6 3.9 3.1 Other . . . 0.1 0.7 0.4 1.5 3.3 1.0 Social services 5.2 10.4 19.2 11.6 5.9 5.6 10.2 Education 3.9 9.7 18.2 10.6 5.0 4.8 8.2 Health 1.3 0.7 1.0 1.0 0.9 0.8 2.0 Economic services 36.0 36.3 37.6 50.1 58.4 56.2 56.6 Agriculture 5.3 5.6 4.2 1.7 2.4 2.1 3.2 Power 2.7 5.0 2.7 3.0 3.5 11.6 8.8 Manufacturing, mining, and quarrying 8.0 7.2 7.8 11.6 23.9 16.8 20.1 Transportation and communication 20.0 18.5 22.9 33.8 28.6 25.7 24.4 Transfers and loans on-lent to states 31.5 23.4 11.6 8.9 7.2 10.4 10.0 . .. Negligible. Sources: Nigerian Federal Ministry of Finance and World Bank data. Nigeria 245 fore the oil price rise, declined further, so that in 1975-78 the sector received only 2.6 percent of total federal capital outlays. Agricul- ture had been budgeted for 6.5 percent of capital spending in the third plan (1975-79), about the same level as in the second plan (1970-74). But severe problems in administering and coordinating in- stitutions and programs limited disbursements. The military rulers' preference was to centralize control over agri- culture. In 1973 the power to set producer prices was transferred from the marketing boards, which had long been criticized for their negative effect on agricultural production, to the federal govern- ment by creating a national board for each commodity. Also in 1973 export duties and state producer sales taxes were replaced by a 10 percent federal tax, which itself was abolished the following year. The declared purpose of these reforms was to benefit produc- ers, but an equally strong motive may well have been the desire of military leaders and civil servants to centralize their power (see Rimmer 1981). All through the 1970s government policies affecting agriculture were contradictory. But there was consistency in that the strategy for improving production relied primarily on the public sector (Olayide 1980, pp. 159-70). Instead of funds going to existing smallholder activities or extension services, they were used largely to promote the growth of large-scale, mechanized, parastatal farms for food production. The government intervened heavily in such agri- businesses, and capital allocations to government food companies (the National Livestock Production Company, the National Grains Production Company, and the National Rootcrop Production Com- pany) increased sharply during 1975-79. But success stories were few. 13 In 1976 the government took over fertilizer distribution from the states and placed it in the hands of the newly created Ministry of Ag- riculture. Fertilizer imports by individual states had been growing rapidly since the early 1970s. The programs, politically popular with bureaucrats, provided an important vehicle for getting money into agriculture by subsidizing inputs (Idachaba 1982). Increasing the use of fertilizer became a central element in World Bank- financed agricultural development projects, although the Bank tried and eventually succeeded in reducing the subsidy given to ferti- lizer. 14 Other important initiatives in the agricultural sector were the River Basin Development Authorities (RBDAs) and the integrated agri- cultural development projects (ADPS). Started in 1973, the RBDAS were originally seen as an attractive centralizing mechanism since irri- gation works would follow river flows and thus cut across state 246 Country Studies lines. But largely for geographical reasons, many of the RBDAS were organized on a statewide basis. They became relatively strong claim- ants for funds; the importance of water resources in capital spend- ing is shown in figure 13-1. At the end of the 1970s, however, the Figure 13-1. Nigeria: Federal Government Capital Budget Allocations to Agriculture and Water Resources, 1976-77 to 1981 16 / All agriculture 14 _+ water _ ~~~~~~~~~~~~// 12 _' O / 10 / v _ / - Water resources ~3 8 0~~~~~ 6 ]° 6 _ N/>' t \x /- / All agriculture 4 ~ Crops 2 _ Fishery Livestock I __ = ore Forestry 1976-77 1977-78 1978-79 1979-80 Average 1980 1981 1976-77 to 1979-80 Note: 1980 and 1981 figures for agriculture include crops and rural development; 1980 figures refer to nine months (April-December) only; 1981 figures refer to the calen- dar year. Source: Idachaba (1982), figure 14 as cited in Bienen (1985). Nigeria 247 RBDAS were not providing significant increases in food output, and in 1980-82 only 40,000 tons of grain were being produced on irri- gated land.'5 Begun in 1974 to increase grain production, the purpose of the ADPS was to supply an improved package of services to the small farmer. ADPS expanded rapidly, so that by 1979 they covered 7.5 per- cent of the population in nine states. The impact of the ADPS is un- clear, however. They have been criticized by some for poor manage- ment and for benefiting large farmers more than small ones."6 It might be hypothesized that decisionmaking units closer to the rural areas-that is, on the state rather than the federal level- would be more sensitive to farmers' interests. But this is not obvi- ously the case in Nigeria. Throughout the 1970s and into the 1980s the states are believed to have had shortfalls from planned expendi- tures in agriculture at least as great as those of the federal govern- ment. And a 1980 government report on revenue allocation re- vealed that, between 1970 and 1975, the states had diverted significant funds away from agriculture even when the federal gov- ernment had made specific, not matching, grants to agriculture. Deficits, Inflation, and Exchange Rates: A Very Vicious Cycle Like most of the other oil exporters, Nigeria experienced a sharp burst of inflation in the mid-1970s as public spending rose. The gov- ernment may have thought that large wage increases would result in a similarly large increase in real private consumption. But since im- ports were restricted through tariffs and direct controls, and domes- tic food supply was inelastic, prices could be expected to rise to choke off excess demand. Efforts to reduce inflation, such as by hold- ing down utility prices, contributed to growing public sector defi- cits."' With an exchange rate pegged to a currency basket of trading partners and inflation averaging 20 percent in 1973-78, the real naira appreciated from its 1970-72 purchasing power parity (100) to an average of 129 in 1974-78. Except for Indonesia, this was the sharpest appreciation in the sample. Unlike most of the other exporters, however, Nigeria followed a policy of tightening import restrictions and raising tariffs, particu- larly when oil revenues started to decline. As early as the mid-1970s the focus of tariff policy had begun to shift from raising revenue to protecting domestic industry; table 13-4 indicates the pro- gressive escalation in effective rates of protection. On average, the net effective rate for the industrial groups listed rose by 177 per- cent, assuming a constant shadow-priced exchange rate. The effect of increasing protection was to appreciate the real effective ex- 248 Country Studies Table 13-4. Nigeria: Net Effective Rates of Protection for Selected Industrial Groups (percent) Industrial group 1977a 1979-80a Consumer goods industries 55.8 124.9 Processing mainly domestic raw materials 12.7 81.7 Processing mainly imported raw materials 74.6 146.9 Intermediate and capital goods industries 1.5 37.7 Processing mainly domestic raw materials -9.9 16.4 Processing mainly imported raw materials 24.4 65.6 Assembly industries 79.4 215.8 a. The exchange rate is shadow-priced at 65 percent. Source: World Bank data. change rate even more by shifting the basis of adjustment onto a smaller set of sectors (see chapter 3). At the same time, the public deficits were financed internally to a large extent, mostly by the Central Bank. The banking system's claims on the government rose from -13.5 to 34.2 percent of nonmining GDP in 1974-83 (see table 13-2). The combination of tighter trade restrictions and money creation caused the real effective ex- change rate to appreciate still further, to 163 in 1979-81, 194 in 1982-83, and 287 in 1984. By this point it had appreciated almost twice as much as the next most appreciated currency in the sample, that of Trinidad and Tobago. As a result of this appreciation, real income should have been redis- tributed from the government, which suffered a sharp reduction in the purchasing power of its dollar-denominated income from oil ex- ports, to private agents and economic activities. Let us leave aside for the moment the question of who appropriated this transfer, how- ever, and consider only its magnitude. An accurate estimate would require knowledge of the movement in the price index of public ex- penditures relative to public revenues caused by exchange apprecia- tion, and hence knowledge of the shares of oil taxes spent on domes- tic and foreign goods. For simplicity, suppose that all imports of plant and transport equipment are public purchases, that all other public expenditures are for domestic goods, and that the price of these goods moves in line with the general price level. Table 13-2 shows the loss in the government's real income im- plied by the appreciation of the real effective exchange rate from its 1970-72 base. For the decade 1975-84, the ratio of this loss to Nigeria 249 nonmining GDP averaged 5.5 percent. For comparison, the ratio of the federal deficit to nonmining GDP averaged 8.9 percent. Although the former estimate is rough, it indicates the magni- tude of the vicious circle induced by responding to inflation and fall- ing oil revenues by tightening import restrictions rather than by ad- justing the nominal exchange rate. First, the purchasing power of oil revenues shrinks more rapidly because of rising domestic costs. Next, domestic financing of the resulting increase in the fiscal defi- cit places greater upward pressure on the price level. This in turn is met by tighter import restrictions and a more appreciated real effec- tive exchange rate. This erodes the value of oil income still further, in- creasing the pressure on domestic financing, and so on. As a complement to the adverse macroeconomic effect, cutbacks in public demand leave projects uncompleted and thus lower the re- turns to public investment. And import restrictions reduce the effi- ciency of the non-oil economy, as discussed below. Some Effects of Oil Spending Because oil wealth has had so profound an impact on Nigeria's politi- cal economy, it is possible to analyze its effects from many dif- ferent perspectives. One of the main political effects, as noted above, has been the impetus given to the centralization of power. From an economic perspective, crucial to any assess- ment of how Nigeria managed its windfalls are the effects on growth, sectoral structure (the Dutch disease), and consump- tion. Growth Had there been no oil windfalls, in all likelihood Nigerian growth rates would have slowed substantially once the recovery from the civil war was complete. It is difficult, therefore, to project growth in the absence of the windfalls. Still, it would be very hard to argue that the overall impact of the windfalls was to stimulate growth. Nonmining GDP grew by only 5.3 percent a year during 1972-81, and it contracted by 5.6 percent a year in 1982-84. This gives a growth rate for the years 1972-84 of only 2.5 percent a year, slightly below the population growth rate of around 2.6 percent- despite by far the largest investment boom in Nigeria's history. Poor growth appears to have been caused by a number of fac- tors. The first is the high proportion of investment that went for phys- ical and social infrastructure, and the supply-side failure of that in- vestment to stimulate other economic activities. The second factor Table 13-5. Nigeria: Oil and Non-Oil Export Performance, 1973-84 Non-oil exports 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982a 1983, 1984, Volumeb Cocoa 214 194 175 219 168 192 218 187 108 125 232 125 Palm kernels 138 186 171 272 186 57 51 50 92 64 80 14 Rubber 49 61 61 34 28 31 34 31 24 27 29 29 Groundnuts 199 30 5 2 0 0 0 0 0 0 0 0 Total non-oil exports (millions of 1975 naira) 549 578 350 343 366 375 360 270 141 76 133 63 Petroleum/total exports (percent) 84.7 92.9 93.2 93.3 93.4 90.5 93.4 95.9 96.9 97.5 94.3 96.8 a. Estimated. b. A constant-price measure. Source: World Bank data. Nigeria 251 appears to be the lowering of investment quality, which was re- lated to the relaxation in approval procedures that accompanied the acceleration of public programs. A third factor was the disruptive ef- fects of cutbacks in investment after the second oil boom. In 1985 it was estimated that about 14 billion naira would be needed to com- plete public industrial projects alone. But only about 10 billion naira was projected to be available for all public investments through 1990. Both the acceleration and the deceleration of invest- ments therefore appear to have effectively eliminated the potential payoffs to a large part of domestic capital formation.'8 A fourth factor, especially relevant to the decline in output after 1981 (and Nigeria's decline in nonmining output was the largest in the sample), seems to have been the the way the allocation of for- eign exchange affected industrial activity. The inefficiency of the li- censing system caused shortfalls in imported raw materials and spare parts that severely affected industry. According to one esti- mate, in 1984 the manufacturing sector operated at only 30-40 per- cent of capacity. One indication that the proximate cause of this was on the supply side, rather than on the demand side, is the fact that the prices of manufactured goods soared, apparently reflecting scarcity premiums. The importance of inefficiencies and supply con- straints in explaining observed wage and consumption trends is fur- ther highlighted below. The Performance of Non-Oil Tradables As table 6-3 indicates, during 1972-81 the composition of Nigeria's nonmining economy swung substantially more to the nontradable sector relative to its "norms." The average annual decline in the share of agriculture in nonmining GDP was the largest in the sam- ple, and it took place despite the relatively low growth of income per capita. Nigeria's non-oil trade performance is summarized in table 13-5 and figure 13-2. Although export taxes on agricultural exports were reduced when oil income soared and were largely eliminated after 1976, this was insufficient to compensate for the initial demand-led appreciation of the exchange rate and the later antiexport bias caused by the intensification of import restrictions. By the mid- 1970s non-oil exports had become insignificant, falling to one- seventh of their 1973 value (in 1975 naira) by 1982. Nigeria became a net importer of agricultural products in 1975. Imports soared, espe- cially in 1975-78 and 1980-82 (see figure 13-2). Estimates of agricultural production levels and trends are avail- able from at least five different sources, but they diverge a great 252 Country Studies deal."9 It is therefore difficult to obtain a clear picture of agricultural performance and consumption of agricultural products during the boom years and thereafter. The various time series agree in only one respect: total calorie consumption per capita declined from 1970 to 1982, even when imports are taken into account. Inasmuch as imports rose, domestic production would have fallen even fur- ther behind. For example, in terms of a bundle of commodities val- ued at common international prices, the U.S. Department of Agricul- ture (USDA) estimated that per capita food output (not availability) fell from an index of 100 in 1969-72 to 85 in 1982-83 (see table 6-4). Despite this extremely gloomy supply picture, however, real pri- vate consumption per capita is estimated to have risen considerably Figure 13-2. Nigeria: Agricultural Imports and Exports 2,200- 2,000- 1,800 2 Surplus 1,600 [] Deficit 1,400 Y 1,200 Imports . 1,000 800 600 400 2,00 Exot 1955 1960 1965 1970 1975 1980 1984 Source: Agribusiness Associates, "Agricultural Investment: The Opportunities and Realities for U.S. Agribusiness Companies," prepared for OICD/USDA, p. 51, as cited in Bienen (1983), and World Bank data. Nigeria 253 during the boom period. Between 1970 and 1978, for example, real private consumption a head rose by about 4 percent a year accord- ing to national accounts data (a population growth rate of 2.6 per- cent is assumed), although by the mid-1980s real private consump- tion levels had declined back to about their level of 1970. Even allowing for possibly growing inequality of income distribution and for substitution effects caused by a rise in the relative price of food, such growth in private consumption would not have been possible without at least some growth in food availability per capita. The supply data are therefore not fully consistent with likely demand trends. It seems probable that agricultural performance, at least in the food sector, may have been somewhat better than the production data indicate. It is still likely that agricultural production, at best, barely kept up with population growth, and it may have fallen some- what behind. Nigeria's poor performance in food production was not the sim- ple result of a drop in relative food prices caused by its massive im- ports. On the contrary, if 1970-72 is taken as the base, the price of food to consumers was higher in 1974-83 than the price of other com- ponents of the consumer price index (see table 6-5). It is possible that this reflected higher distribution margins rather than higher farmgate prices-the data are insufficient to judge. A more proba- ble explanation, however, is that agriculture, the largest employer of unskilled labor, lost workers to the rapidly growing construction sector and was also adversely affected by the spread of primary edu- cation, which reduced the use of child labor on farms.20 With stag- nant levels of productivity, agriculture would not have been able to raise per capita income enough to retain sufficient workers to sat- isfy food demand at constant rural-urban terms of trade.2' One factor sometimes cited in accounting for the far better perform- ance of Indonesia's food crop sector is the Green Revolution in rice production. As described in chapter 12, this was brought about by the development and dissemination of improved high-yielding and disease-resistant varieties and the adoption of improved farming techniques. No such dramatic advances were made in the rainfed ag- riculture of Nigeria, but even so, yields were low.22 The gap was con- siderable between actual practice and potential improvements. The obstacles were partly technical but also infrastructural and institu- tional ones. The inefficiency of input distribution services (notably for fertilizer) and the lack of a good extension service were serious problems.23 These, in turn, reflected the lack of urgency that succes- sive governments assigned to the development of agriculture, espe- cially among smallholders. Unlike Indonesia, Nigeria had very lim- ited success in using oil income to raise rural productivity. 254 Country Studies Consumption and Wages So far the Nigerian story is relatively clear. But now we come to three related and puzzling issues. First, who appropriated the consid- erable transfers from the government implied by the sharp real appre- ciation of the naira? Second, have there been real wage gains in the Nigerian economy? Third, if real wages in urban activities did in- deed fail to rise when public spending rose, how was it possible to continue to attract large numbers of workers from an agricultural sec- tor that supposedly was not in labor surplus? On the first issue, the expected counterpart to the transfer from the fiscal budget would be a gain in real absorption as prices of goods rose less rapidly than prices of domestic non-oil output be- cause of the cheapening of imports. But table 5-2 indicates no such gain in 1974-78-in fact, the effect goes the other way. There is a sub- stantial gain in consumption in 1979-81, equivalent to a transfer of 4.1 percent of nonmining GDP (see table 5-3). This is offset, how- ever, by increases in the relative price of investment goods, so that the overall effect is eliminated. Where, then, did the transfers from the real exchange rate movement go? Several explanations are possible. One, of course, is that the rele- vant data are flawed. The situation is not improved, however, if the consumer price index is substituted for the consumption deflator. A second possibility is that the increased resource cost of transport and distribution systems, strained by greatly increased import lev- els, could have offset the transfer. This might have been a factor, but only in the initial stages of the oil boom. A third explanation could be capital flight, effected by overinvoicing imports. Had this at- tained macroeconomic proportions, the transfer would have been off- set by apparent non-oil terms of trade losses. The fourth possibility, and probably the only one able to account for a transfer as large as that computed above, is the absorption of the transfer in rent-seeking activities induced by the increased re- sort to tariffs and quantitative restrictions. Such a loss would in- clude not only the real resources expended by importers on obtain- ing licenses, but the general decrease in allocative and X-efficiency of an economy increasingly operating behind quantitative restric- tions that inhibited competition. If this interpretation is correct, a sub- stantial part of Nigeria's windfall gains were eaten up in the re- duced efficiency of its non-oil economy, which resulted largely from the protectionist policies adopted in response to the falling pur- chasing power of oil receipts. The question of whether real wages rose has been investigated (see, for example, Collier 1985), but here, too, are problems of inter- Nigeria 255 pretation. The available wage data, though incomplete and confus- ing (especially when juxtaposed with estimates of the evolution of in- come distribution and private consumption) appear to argue against a substantial real wage gain in urban activities. If anything, they argue for a real wage loss in construction and more so in the public sector. Such an evolution is quite unusual; the pattern for the other countries suggests considerable real wage gains in the windfall period (see table 6-2). Stagnant or falling real wages are not necessarily inconsistent with a rise in private consumption per capita. Much depends on changes in the distribution of private income, and if-as in Nigeria- labor is rapidly reallocated from low-wage activities (such as farm- ing) to higher-wage urban activities (such as construction), private in- come and consumption per capita can rise even with static or falling real wages for specific activities. Diejomaoh and Anusionwu (1981) concluded that there was a sharp trend toward inequality within the modern sector from 1970 to the mid-1970s, with the Gini coefficient rising from 0.55 to 0.70. For the urban sector as a whole, the bottom layer of which is the self-employed in the informal sec- tor, the studies reported in Bienen (1983) also suggest a rise in ine- quality from 1970 through 1976.24 Income inequality in the rural sec- tor has been estimated to be low, with a Gini coefficient of about 0.3 (Bienen 1983, p. 37). According to some estimates, the trend of rural wages may have been upward. In contrast to the data of table 6-2, Collier (1985) suggests that in 1973-81 real wages in agriculture may have risen by about 50 percent, although such an estimate seems very high given the macroeconomic context. Whatever the precise pattern of real wages, the main reasons for their failure to rise faster appear to lie in (a) the impact on the food sector of labor market pressures created by oil-financed spending, and (b) an insufficiency of permitted imports to make up the supply- demand gap. Food import policy was, indeed, erratic during the per- iod, so that the sector is best classed as semitraded, although in prin- ciple many staple items are tradable. Increasing nominal urban incomes combined with the economy's shift toward producer goods to bid up the price of the main consumer goods; this, together with low growth and falling efficiency, reduced or eliminated real wage gains. If urban wages were not rising, how was labor drawn out of the farm sector? Collier (1985) provides evidence of a substantial wage gap between urban and rural-or between modern and traditional- activities in the early 1970s. He estimates that unskilled wages in non- agricultural activities were about 80 percent higher than those in agri- culture in 1973, essentially because of the political clout of public 256 Country Studies sector employees and the impact of public wage levels on the urban private sector through "fair wage" clauses in employment con- tracts. But by the early 1980s, Collier finds, the difference had all but disappeared. According to this view, the large initial gap made possible the expansion of labor use in the modern sector, and real urban wages fell because of the scarcity of key wage goods. Accounting for the Policy Response Nigeria's strategy for managing its extraordinary windfalls cannot be regarded as a carefully articulated one in the sense that, say, Algeria's can. Nigerian policymakers do not seem to have reso- lutely set out to pursue a strategy heavily biased toward infrastruc- ture and education and away from agriculture, but they were in- clined to believe that achieving productivity gains in traditional smallholder agriculture would be very difficult (see Bienen 1985, p. 11). What, then, explains the policy response? Did the choices made regarding the use of oil revenues reflect a representative bal- ance of individual preferences? Or did they result from a combina- tion of the political process and administrative constraints? A useful point of departure is the observation that there has been no clear development of class-based politics in Nigeria. Nei- ther peasant organizations nor trade unions have exerted strong pres- sure for income redistribution. Nor have informal urban workers con- stituted an effective force for radical change. The very meanings of wealth and equity differ within and among communal and occupa- tional groups. Relationships between workers in the formal and infor- mal sectors are complex, and they vary by city, region, and ethnic group. This weakens pressures from the urban poor. Neither in the 1950s, nor in the 1960s, nor in the 1979 elections did political parties organize on the basis of class, income group, or occupation. Surveys help to explain why this pattern has not devel- oped. Morrison (1981) analyzes a 1974 national survey. He argues that, although in reality social and occupational mobility may have decreased, respondents still assumed the possibility of social mobil- ity. Views differed markedly, however, among states (and thus among ethnic groups) about whether people received a "fair share" in life and whether ordinary people were better or worse off since in- dependence. Earlier surveys had yielded many of the same results, with communal cleavages stronger than occupational ones.15 Furthermore, there were strong positive associations between sta- tus and participation, between fathers' and children's levels of educa- tion, and between occupation and education. Mobility and opportun- Nigeria 257 ity were thus related to an individual's start in life and in particular to educational attainment. Many rural Nigerians wanted to acquire the skills needed to compete in urban labor markets, so spending on education was popular, despite widespread concern about the quality of the education provided. The personal preferences of many rural inhabitants coincided with the political preferences of Nigeria's leaders as well as with their recognition of the administra- tive problems inherent in any attempt to reform smallholder agricul- ture. Both individuals and politicians have pushed, therefore, for more funds to be spent on education at all levels.26 The push to spend funds on the agricultural sector was far weaker. As already noted, agricultural interests in Nigeria had never managed to surmount regional differences to form a signifi- cant pressure group, and the widespread desire of rural residents to move out of the sector further undermined any efforts in this direc- tion. The political power of those in the agricultural sector was, as in most African countries, dispersed and unorganized. When mili- tary regimes were in power, elections did not count, so the rulers could not be called to account for the relative neglect of agriculture in government spending. Even in the 1950s and 1960s, when govern- ments were elected and when groundnuts and cocoa were impor- tant sources of export revenue, farmers wielded neither political nor economic power. Cocoa farmers in the Yoruba areas, for exam- ple, never formed a well-organized pressure group as did planters in the C6te d'Ivoire. The one major exception to the weakness of agricultural interests was the Agbekoya revolts of 1968-69 in the Western Region. Agbekoya, which means in Yoruba "farmers who reject injustice," was an agrarian populist movement and a revolt against high taxes and poor services. It centered in areas where cocoa production had been declining.27 The revolts were put down with some loss of life. After initial relief, not much was done to benefit these areas. Mili- tary officers tended to see the movement as a plot hatched by for- mer politicians and were insensitive to farmers' grievances (Adebayo 1969). Those in revolt could not sustain the pressure on the government. The pressures were not echoed throughout Ni- geria, nor did they recur on a similar scale in the 1970s. Policymakers were thus relatively free to focus on urban rather than rural problems and to follow their own predilections. They be- lieved that urban inequality and urban poverty could be attacked through service delivery, whereas investments that would benefit small and medium-size farmers were difficult to make, especially in- vestments in agricultural extension, new seeds and other inputs, 258 Country Studies marketing facilities, and so on. This abandonment of the smallholder in the 1970s, before making any major commitment to ag- riculture, has not been unique to Nigeria in Africa. Nigeria also had many bottlenecks in administration. Public sec- tor investment programs suffered because of constraints in the imple- mentation process. This was especially true for agriculture, where expenditures increasingly fell short of projections as the 1970s progressed. The government had difficulty getting funds out to the rural areas, in part because it was hard to find good proj- ects to fund and in part because the administrative networks to de- velop and implement rural projects were very weak. It was much easier to contract for the building of roads and ports and to invest in plants. In addition, the elites of the state, both civilian and military, had in- terests of their own, which they were in a position to further. It has been suggested that the Udoji awards were one attempt to distrib- ute oil income to civil servants as patronage. As noted earlier, the elites were in a hurry to spend oil income because of the patronage available from large projects and because they did not wish to be charged with not spending "the people's money" in worthwhile, usually highly visible, ways.28 Agriculture suffered from this empha- sis on rapid spending. Without lengthy institution-building efforts (such as the improvement of extension services) there was no effec- tive way to reach the smallholder or to improve inefficient institu- tions. Spending priorities were therefore shaped partly by policy- makers' (correct) interpretation of widespread desires, partly by the distinctive structure of government, and partly by administrative practicality. A final question concerns the possible reasons for Nigeria's deci- sion to adjust to falling oil income through trade restriction rather than devaluation, despite the disastrous macroeconomic implica- tions of such a choice. Two explanations suggest themselves. First, it is clear that by the early 1980s the exchange rate issue had come to be regarded as one of national sovereignty rather than simple eco- nomics. It was almost surely the case that Nigeria's non-oil compara- tive advantage still lay in the export of agricultural products. A large devaluation accompanied by a substantial liberalization of trade could therefore have seemed threatening to the modern indus- trial sector, thought to be so essential to making Nigeria the "giant of Africa." Second, it may be that the scarcity rents generated by re- strictions had given rise to a pressure group sufficiently powerful to influence policy. Nigeria 259 Summing Up: Have the Windfalls Benefited Nigeria? As is apparent from this analysis, it is not easy to judge how much the oil windfalls have benefited the Nigerian economy-or, indeed, whether they have benefited the Nigerian people at all. Yet some con- clusions are obvious. Unlike Ecuador, Venezuela, and Trinidad and Tobago, the other countries in this study that registered substantial declines in their nonmining economies after 1981, Nigeria appar- ently did not translate its oil wealth into substantial gains in real per capita private consumption (relative to the hypothetical scenarios with- out windfalls in chapter 5) during the boom years. Unlike Indonesia, it did not emerge from the windfall era with a strengthened agricul- tural sector. Unlike Algeria, it did not invest in industrial sectors with the potential (even if not fully realized) of generating appreciable non- oil exports. Nigeria's non-oil tradable sectors almost certainly weak- ened during the decade of windfalls. And certain of its import substitu- tion ventures, notably in steel, will probably involve a net economic loss through debt service and high production costs. In the context of the present sample of countries, the growth stimu- lus of Nigeria's oil income took mainly the indirect route of strength- ening key infrastructure and social sectors. Nigeria did improve its physical infrastructure significantly, even though it failed to extend key services to rural areas. But perhaps its major achievement has been the use of oil income to fund almost universal primary educa- tion. Without this, it can be argued, a long-run transformation of agri- culture would be far more difficult. Enrollment in secondary and higher education also increased. By 1980 Nigeria had virtually closed the education gap between itself and the other developing oil exporters, at least in quantitative terms. How Nigeria can trans- late such improvements into future growth will largely determine the long run value of the oil windfalls. Notes 1. For a discussion of the difficulty of forming rural interest groups, see Bates (1981). 2. The literature on the marketing boards is very large. See Helleiner (1966), Wells (1974), and Rimmer (1981). 3. See Bienen (1985, pp. 20-23) for a discussion of regional variations in public spending on agriculture. 4. A fourth, Midwestern, region was later created. 5. The history of this period is treated in Coleman (1964), Mackintosh (1966), Sklar (1966), Panter-Brick (1970), Dudley (1973), and Oyediran (1979). 260 Country Studies 6. For a discussion of revenue-sharing in Nigeria, see Rupley (1981) and Rimmer (1981). 7. For a discussion of Nigeria's indigenization policy, see Biersteker (1978, 1980) and Akinsanya (1981). 8. For the early years of the Nigerian oil industry, see Schatzl (1969) and Pearson (1970). For accounts of the early 1970s, see Meyer and Pearson (1974) and von Lazar and Duerstein (1976). 9. It was not Nigeria's imposition of OPEC terms on foreign companies that affected their willingness to invest in Nigeria's oil potential in the late 1970s and early 1980s, but the fact that Nigeria was consistently a price hawk in OPEC. Indigenization may well have slowed capital flows, how- ever, especially after 1977. And there may have been indirect costs associ- ated with restrictions on foreign managers and technical personnel. Some companies whose shares were bought by the states-particularly textile firms-probably were poor purchases and costly to public purses. 10. The Udoji wage increases and the burden of inflation triggered a back- lash against General Gowon, who was already unpopular because of his an- nounced intention to stay in power beyond 1975. In July 1975 he was re- placed by General Mohammed, who advocated "clean government" and used the rhetoric of populism to build support. 11. As in other countries in the sample, domestic oil prices had not been raised to world levels, although the average divergence and the implicit sub- sidy seem to have been lower in Nigeria than in some of the other export- ers. 12. Loans and nonstatutory allocations became important to the states after 1976. Between 1969 and 1977 federal lending to the states rose well over 1,000 percent; federal development in the states also rose markedly, from 91.6 million naira to more than 5 billion naira. 13. For further discussion, see Bienen (1985, p. 63). 14. The use of fertilizers in the late 1970s was subsidized up to a rate of 85 percent (Norton 1983, p. 41). 15. For a summary of the results achieved by the RBDAS, see Bienen (1985, pp. 67-69). 16. For a summary of (sometimes divergent) views on the ADPS, see Bienen (1985, pp. 69-75). 17. Utility prices rose by only 5 percent between 1973-74 and 1979-80. In this period the corporate consumer price index for rural and urban centers in- creased by 176 percent; see Bienen (1983, tables 13 and 18). 18. A critical view of spending on the new federal capital at Abuja illus- trates the losses from both acceleration and deceleration of investment. It points out the waste involved both in unfinished construction and in its poor quality (Washington Post, 20 January 1984, pp. A23-28). 19. The five sources are Nigerian Federal Office of Statistics (FOS), Cen- tral Bank of Nigeria, Nigerian Federal Department of Agriculture, the Food and Agriculture Organization (FAO), and the U.S. Department of Agricul- ture (USDA). To give some idea of the larger divergences, in 1981 the FOS esti- mated cassava production at 580,000 tons and the USDA estimated it at 11.8 million tons. Nigeria 261 20. The share of Nigeria's work force actively engaged in agriculture fell from about 70 percent throughout the 1960s to 55 percent in 1980. 21. If real agricultural wages did rise as suggested by Collier (1985), the sec- tor would have been squeezed with labor costs rising faster than prices. 22. Cotton yields in Nigeria, for example, were just over half those in Mali. 23. The ratio of extension workers to farmers was estimated to be 1 to 2,500 in Nigeria in the late 1970s, compared with 1 to 200 in India and 1 to 250 in Kenya (Guidelines for the Fourth National Development Plan, 1981-85, p. 26). Some observers believe that the extension service was in worse shape in the late 1970s than a decade earlier. 24. As noted by Bienen (1983, p. 36), there are severe problems with Niger- ian data on income distribution, especially after 1975. Bienen's study sur- veys research on income distribution covering the period to 1976. 25. See Verba, Nie, and Kim (1978, p. 103, table 6-1). Fieldwork for this study was carried out in 1966. 26. By the time of the 1981 budget, the planned capital expenditure for uni- versities amounted to more than half the total capital allocated to educa- tion. 27. The most important accounts of the Agbekoya revolts are in Beer (1976) and Beer and Williams (1976). 28. In countries with an extensive history of public corruption, it may be difficult for governments to save abroad openly since they are vulnerable to accusations that money is being siphoned into officials' pockets. This point has been made several times in discussion with officials from countries included in this study. Chapter 14 Trinidad and Tobago: Windfalls in a Small Parliamentary Democracy with Richard Auty The smallest of the oil-exporting developing countries in this study, Trinidad and Tobago has a slow-growing population of a little more than one million and an area only one-fiftieth that of Ecuador, the next smallest country. It is also the most densely populated, with three times as many people per square mile as Indonesia or Ni- geria. Although its 1974 per capita income was slightly less than Venezuela's, its 1982 per capita income of US$6,840 placed it at the top of the group of all middle-income developing countries. Trinidad and Tobago is a democracy. Unlike Venezuela-the only other comparator with an elected government throughout the wind- fall period-its government is modeled on the British rather than the American system and is therefore less prone to disagreements be- tween the executive and the legislature over how to manage oil reve- nues. Also unlike Venezuela, the same political party dominated the government during and after the entire windfall period. Another distinctive feature of the country is that before the first windfall the gradual decline in oil output from established fields had precipitated a period of austerity. By 1974 it was widely under- stood that the new offshore finds developed in the early 1970s were limited. The public was therefore aware of the need to pro- ceed with caution in using the income from the country's finite oil resources. Except for its size-which, as this study confirms, reduced the ef- fectiveness of certain strategies for absorbing oil rents-these fea- tures enhanced the chances for Trinidad and Tobago to put its wind- falls to good use. In relation to the comparator countries, it began with some important advantages. Other features, however, made it more difficult for the govern- Richard M. Auty is lecturer in geography at the University of Lancaster, United Kingdom. 262 Trinidad and Tobago 263 ment to use oil income to increase the productivity of the non-oil economy. These included a society strongly divided along ethnic, po- litical, and economic lines and an economy that has been described as the most regulated in the Caribbean except for Cuba's (Black and others 1976, p. 209). Faced with the demands of powerful inter- est groups, the government has traditionally accepted a great deal of the responsibility for maintaining employment and income. These factors created strong pressures to distribute windfall gains widely. Political pressure to raise and sustain consumption was to be one serious weakness in the development strategy. At the same time, the above features, combined with windfall gains, relaxed efforts to raise productivity in the non-oil economy, to arrest agriculture's decline, and to improve manufacturing's com- petitiveness. Trinidad and Tobago lacked effective means to make productive investments in agriculture and competitive light indus- try, particularly as the windfalls pushed up wages. Investments ei- ther were not effective or suffered from lack of demand, and in its ea- gerness to use oil revenues for large projects the government downplayed the risks and decided to go ahead on its own. Conse- quently, the country ended the second boom far more dependent on oil than before, even though it had accumulated a cushion of for- eign assets. This chapter assesses the impact of the development strategy im- plemented in Trinidad and Tobago during the decade of the oil wind- falls. It first outlines the conditions of the political system and of the economy before the first oil shock. The response to the boom of 1974-78 is then described. In absorbing its windfall gains, the coun- try did, in fact, proceed much more cautiously than the other five oil exporters. Nevertheless, many of the policies adopted during these years-notably to extend state ownership of enterprises, to in- crease subsidies to consumers and failing firms, and to promote gas- based industrial development-adversely affected the economy later on. The discussion of the second boom, 1979-81, therefore em- phasizes state ownership, subsidies, and the performance of the non- hydrocarbon-based tradable sectors. As in Algeria, gas-based indus- trial development played a central part in the Trinidadian strategy of "sowing the oil." This strategy is the subject of a separate sec- tion because of its crucial importance in the 1980s and beyond. Con- clusions follow in the final section. Before the First Boom Trinidad and Tobago evolved into a parliamentary democracy after World War II, and by 1956 its two-party structure was set. The politi- 264 Country Studies cal process, however, had to contend with ethnic cleavage. In 1970 blacks constituted 43 percent of the population, East Indians 40 per- cent, and people of mixed race 14 percent. The island's black popula- tion was strongly represented in government and predominant in in- dustry. The East Indian population, descendants of indentured laborers imported to work on the plantations after slavery was abol- ished in the 1830s, was generally poor, largely rural, and heavily de- pendent on sugar production for a livelihood (Black and others 1976). Political Structure The People's National Movement (PNM), which held power from 1956 to 1986, has a primarily black constituency. The party was led by the increasingly autocratic Dr. Eric Williams until his death in 1981, and then by George Chambers. The Democratic Labour Party (DPL), the leading opposition, has drawn its support more from the East Indians.' Influenced by the views of Lewis (1950), the PNM, upon coming to power, made industrialization its main goal, initially with heavy re- liance on foreign capital, technology, and management. Unlike Lewis (1950, 1972), however, it did not emphasize competitive, labor- intensive exports as an essential part of its strategy. The DPL at- tacked the government for depending too much on "exploitative for- eign capital." But because it represented an unwieldy coalition of rural East Indians, urban poor, and a small European elite, the party never cohered, despite receiving almost half the votes in the 1956 election. The DPL'S weakness allowed the PNM to emerge as a strong, unified government largely free of competition within its ranks. In the mid-1970s this lack of opposition enabled the govern- ment to keep public spending more disciplined in Trinidad and To- bago than in most of the other oil exporters. The lack of an effective opposition, however, also encouraged extra-parliamentary attempts to influence government policy. Early in 1970 street demonstrations and a mutiny in the army gave voice to widespread dissatisfaction over growing economic problems caused, indirectly, by declining oil revenues. Despite a sweeping elec- toral victory in 1971 (the DPL boycotted the elections and the PNM won every seat in Parliament), the PNM recognized that there was enough discontent in the country to challenge its power. So it reluc- tantly shifted to a populist program with three main goals: * To extend public ownership in both the oil and non-oil sectors * To redistribute income more equitably * To accelerate industrial diversification out of oil. Trinidad and Tobago 265 Economic Structure OIL SECTOR. Trinidad and Tobago is a long-established oil producer. Oil was discovered in 1857, and Shell became the first producer in 1913. By the 1950s British Petroleum, Royal Dutch Shell, and Texaco operated oil refineries and augmented domestic crude with imports from Venezuela. By the early 1970s petroleum extraction and refin- ing accounted for almost three-quarters of exports, one-fifth of gov- ernment revenue, and one-fifth of GDP. The fortunes of the oil industry had long determined the rhythm of the overall economy. In the late 1960s both oil production and re- fined exports began to fall as traditional fields were exhausted. Pub- lic spending dropped from 23 to 20 percent of GDP in 1971-73, and a modest budget deficit of 4.7 percent of GDP was financed half from foreign and half from domestic sources. Between 1970 and 1973, GDP growth slowed from an annual average of almost 5.5 percent over the preceding fifteen years to 3.6 percent a year (4.5 per- cent for nonmining GDP). Inflation rose from an average of 2.5 per- cent a year in the 1960s to more than 10 percent in the early 1970s. Marine exploration between 1969 and 1971 revealed substantial new sources of recoverable oil and gas. Oil production rose from a low of 129,000 barrels a day in 1971 to 191,000 in 1973, exceeding the previous peak in 1969. Although oil reserves were expected to last only about ten years, the new finds reduced the immediate pres- sure for politically difficult structural change. Their anticipated earn- ings permitted the economy to tide itself over the interim period by borrowing abroad. Even before the disturbances of 1970, the government had taken a small step toward nationalizing the oil industry when it bought Brit- ish Petroleum's refinery and field facilities (Sandoval 1983). Then the discovery of natural gas spurred a growing debate on how best to use it. Amoco, which had made the critical discoveries on the east coast, proposed in 1971 to ship gas to the United States as lique- fied natural gas (LNG). This drew criticism from a younger genera- tion of economists, who were influenced by dependency theory. They argued for using gas as a fuel or a feedstock for local industry under national (rather than foreign) control instead of continuing to export raw materials (Girvan 1970; Parsan 1981). NON-OIL SECTORS. The non-oil sectors of Trinidad and Tobago's econ- omy had long relied on protection from foreign competition (Seers 1964; Brewster 1972). In the early 1970s manufacturing-though heavily protected by a "negative list" of prohibited imports and by other measures, which in many cases doubled prices relative to 266 Country Studies those of imports-accounted for only 22 percent of nonmining GDP, with half of this from refining. Agriculture, dominated by sugar, ac- counted for only about 6 percent. In 1971 the government had moved into non-oil entrepreneurial activity by buying a majority holding in Caroni, the largest sugar company. Food imports repre- sented 226 percent of value added in all nonsugar, nonexport agricul- ture. Unlike many Caribbean islands, Trinidad and Tobago enjoyed little tourist trade. The economy, therefore, was heavily dependent on oil earnings on the eve of the first oil shock. Non-oil tradables-manufacturing and agriculture-accounted for just under 28 percent of nonmining GDP, in contrast to the Chenery-Sryquin "norm" of 42 percent for countries with a comparable per capita income.2 This was despite the fact that economic diversification had been a major goal of the third five-year plan (1969-73). Income was unevenly distributed, the result of a pronounced dual- ism in the economy between high-technology oil and manufactur- ing sectors on the one hand and a low-skill agricultural sector on the other. In 1973 agriculture generated one-sixteenth of the value added per worker in oil refining and one-fifth of that in chemicals. Agriculture itself was dualistic: value added per worker in sugar was almost four times value added in the rest of the sector. The elite oil workers (less than 4 percent of the country's labor force) earned US$5,000 a year, compared with US$750 for all other work- ers and US$325 for agricultural workers (Black and others 1976). The effects of this uneven income distribution were accentuated by high and rising unemployment, which topped 14 percent in the early 1970s. These disparities provided labor with a powerful impe- tus to organize; 70 percent of all workers were unionized, a figure ex- ceeded in only a few small European countries. The First Oil Shock, 1974-78 As shown in figure 14-1 and table 5-2, the 1974-78 windfall of Trini- dad and Tobago was unusually large compared with that of other capital-importing oil exporters-equivalent to 39 percent of non- mining GDP.3 This was because new oil fields began to produce in sub- stantial quantities just as the embargo of 1973 caused the world- wide price of oil to quadruple. Corporate and other taxes (Trinidad and Tobago does not follow the common OPEC system) were rapidly adjusted to siphon off five-sixths of increased revenues. Oil taxes, charted in figure 14-2, jumped from 20 percent of government reve- nues before the boom to 60 percent in 1974-78. Trinidad and Tobago 267 Figure 14-1. Trinidad and Tobago: The Oil Windfall and Its Use, 1973-81 Unweighted Average of Algeria, Ecuador, Indonesia, Nigeria, Trinidad and Tobago, and Venezuela 50 :c 40 .° 30 Trade and nonfactor services surplus .E30- 20 / ~_ ~Oil windfall =:20-i52 I~nv stment Consumption 1970-72 1974 1976 1978 1980 Trinidad and Tobago E 50 ~-40 gL 40 t ~ / \ Oil windfall 'X 30 - ........ Trade and * -------- Constant-price E 20 - . nonfactor services consumption 0>: / .-- surplus /\ T Conupto lo° L 4Q